Beolab 90


TheBeoLab90isaloudspeaker conceptfromBang&Olufsenthat givesthecustomeranunprecedented levelofcontrolofitsacoustic behaviourandperformance. Unlike almostallotherconventional loudspeakers,theBeoLab90canbe alteredbytheusertobehaveasifit werecompletelydifferentloudspeakers fordifferentlisteningsituations. ImaginethatyouhaveapairofBeoLab 90loudspeakers,perfectlypositioned inyourlisteningroom,withasingle chairinthecorrectlocation,asis showninFigure1.1. Yousitinthischair tolistentoarecording–tohear sparklinghighfrequenciesandatight, punchybassthatextendstothelowest audiblefrequencybandsalongwith theaccurateandpreciseplacementof theinstrumentswithinthespacein frontofyou(betterknownas“stereo imaging”).
Speaker Distance
Perceived location of left loudspeaker
Figure 1.1: The sound stage presented byBeoLab90’sforactivelisteningwhen sitting in the sweet spot with properly placedloudspeakers.
Now,imaginethatyouhavethesame loudspeakersinthesamepositionsin thesameroom,butyou’vemovedto thesofaasinFigure1.2(orperhaps you’restillinthesamechairasbefore asinFigure1.3)andyouprefertohear musicinthebackgroundwhileyou readabook. Inthiscase,thebass precisionandtheimagingofthe recordingisnotimportant–youjust wantacloudofsoundthatdoesnot distractyouwhileyouread. Usingyour controlleryousimplyswitchthe
behaviouroftheBeoLab90’stodeliver thisexperienceinstead.
Speaker Distance Compensation
Perceived location of left loudspeaker
Figure 1.2: The sound stage presented by BeoLab 90’s for passive listening whennotsittinginthesweetspot.
Speaker Distance Compensation
Perceived location of left loudspeaker
Figure 1.3: The sound stage presented by BeoLab 90’s for passive listening whensittinginthesweetspot.
Finally,imaginethatyouinviteyour friendsforapartyoryou’rejust walkingaroundtheroom. Imagingisof nointeresttoanyone–youwanta loudspeakerthatcandeliverthesame experiencetotheentireroomatthe sametimebysendingsoundinall directionssimultaneously. Again,with yourcontroller,youchangetheBeoLab 90’sacousticalbehaviourtosuitthe occasion.
Speaker Distance Compensation
Perceived location of left loudspeaker
Figure 1.4: The sound stage presented by BeoLab 90’s for background music when the listeners are moving around theroom.
Thesethreescenariosillustratethe primarylisteningmodesthatthe BeoLab90candeliver. We’llcallthe firstoneactivelistening–sinceyour primaryactivityistolistentothe recording. We’llcallthesecondone passivelistening,since,inthiscase, listeningtomusicissecondaryto anotheractivity(inourexample, reading)inastationarylistening positionorarea. We’llcallthethird casebackgroundmusic,whichis similartopassivelistening,however, thereisnodeterminedlistening position(eitherbecausethereare manylisteners,orlistenersaremoving aroundthelisteningarea,orboth). In ordertobeabletodothis,thereare differentparametersinsidetheBeoLab 90thatareadjustable. Youcanchoose toaltereachoftheseparameters individuallyaccordingtoyour preferencesandlisteningpositions– andthensavethesettingstoapreset forfutureuse.
1.1 Features
BeoLab90givesyouthepowerto makethesechangesusingalarge numberof“handles”–controllersthat letyouchangetheacoustical behaviouroftheloudspeaker. Among thesefeatures,therearethreethat standout: • BeamWidthControl • BeamDirectionControl • ActiveRoomCompensation Inadditiontothese,theBeoLab90has amanyotherparametersthatgiveyou awiderangeofcustomisation possibilitiessuchas: • SpeakerDistance(for time-alignmentatthelistening position) • SpeakerLevel • BasicToneControls (Bass,Treble,FrequencyTilt, SoundEnhance)
7
• SelectableSoundDesigns • 10-bandParametricEqualiser
1.1.1 BeamWidthControl
Whenarecordingengineermakesa recordinginawell-designedstudio,he orsheissittingnotonlyina carefully-designedacousticalspace, butaveryspecialareawithinthat space. Inmanyrecordingstudios, thereisanareabehindthemixing consolewheretherearenoreflections fromthesidewallsarrivingjustafter thedirectsoundfromthe loudspeakers. Thisisaccomplished eitherbyputtingacoustically absorptivematerialsonthewallsto soakupthesoundsoitcannotreflect (asshowninFigure1.5),ortoangle thewallssothatthereflectionsare directedawayfromthelistening position(asshowninFigure1.6).
"Reflection-free Zone"
Figure1.5:Onewaytoreducetheproblemofsidewallreflectionsistoabsorb thematthewallssothatthereisnoreflection. Thisisasolutionoftenusedin recording studios, however, it also results in an unnatural-sounding “dead” room.
"Reflection-free Zone"
Figure1.6: Analternativemethodtoreducetheproblemofsidewallreflections is to re-direct them away from the listeningposition,producinga“reflectionfree zone”. This method is often used inrecordingstudiosthatareinitiallydesigned with the help of an experienced acousticalconsultant.
Thisisdifferentfromyourlivingroom whichhasnotbedesignedprimarilyas anacousticalspace. Ithassidewalls thatreflecttheenergyfromyour loudspeakersandsendthatsoundto youatthelisteningposition–a situationthatismorelikethatwhichis showninFigure1.7.
"Reflection-free Zone"
Figure 1.7: The direct sound (in red) from the loudspeakers is influenced by the reflections off the side walls (in grey).
Inordertogetthesameacoustical behaviourinyourlivingroomthatthe recordingengineerhad,wehaveto reducetheamountofenergythatis reflectedoffthesidewalls. Ifwedonot wanttochangetheroom,onewayto dothisistochangethebehaviourof theloudspeakerbyfocusingthebeam ofsoundsothatitstaysdirectedatthe listeningposition,butitsendsless soundtothesides,towardsthewalls. ThisisoneoftheoptionsthatBeoLab 90givesyou–tomakethebeamof sounddirectedoutthefrontofthe loudspeakernarrowertoreducethe levelofsidewallreflections,sothatyou
getamoreaccuraterepresentationof thesoundtherecordingengineer heardwhentherecordingwasmade.
"Reflection-free Zone"
Figure 1.8: BeoLab 90 solves the problem of side wall reflections by reducingtheamountofacousticenergythat is radiated towards the side walls – so thereislessenergytoreflect.
However,ifyou’resharingyourmusic withfriendsorfamily,dependingon wherepeoplearesitting,thebeam maybetoonarrowtoensurethat everyonehasthesameexperience. In thiscase,itmaybedesirabletomake BeoLab90’ssoundbeamwider. Ofcourse,thiscanbeextendedtoits extremewheretheBeoLab90’sbeam widthisextendedtoradiatesoundin alldirectionsequally. Thismaybea goodsettingforcaseswhereyouhave manypeoplemovingaroundthe listeningspace,asmaybethecaseat aparty,forexample. Thisoptiontochangethepatternof theradiationofsoundfromtheBeoLab 90iscalledBeam Width Control.
1.1.2 BeamDirectionControl
Almostallloudspeakersaredesigned toradiatesoundforwards–so,inorder togetthebestexperiencefromyour loudspeakers,youhavetobelocated directlyinfrontofthem. However, BeoLab90givesyouthefreedomto changethedirectionofthesound beamdirectedfromtheloudspeaker. Youcanselectoneoffivedirectionsas beingthe“acousticalfront”ofthe loudspeaker. Ifyou’resittingtothe sideoftheloudspeakerasisshownin Figure1.2youcanchoosetorotatethe soundbeamsothatitisdirectedmore towardsyourlisteningpositioninstead ofinfrontoftheloudspeakers.
8
Front
Front Right
Back Right
Front Left
Back Left
Figure1.9: BeoLab90hasfivebeamdirectionsavailable,allowingyoutooptimisefordifferentlisteningpositions.
1.1.3 ARC:ActiveRoom Compensation
In2002,Bang&Olufsenintroduced theBeoLab5whichincludedABC – Automatic Bass Calibration. Thiswasa systemthatusedamicrophoneto measuretheeffectsofthelistening room’sacousticalbehaviouronthe soundoftheloudspeaker,andthen createdafilterthatcompensatedfor thoseeffectsinthelowfrequency band. Asasimpleexample,ifyour roomtendedtoincreasetheapparent basslevel,thentheBeoLab5’swould reducetheirbasslevelbythesame amount. BeoLab90takesthisconcepttoanew levelwithitsActive Room CompensationorARC.Usingan externalmicrophone(availablefrom yourBang&Olufsendealer),youcan measuretheeffectsofyourroom’s acousticalbehaviourindifferentzones intheroomandsubsequentlyselect optimisedcompensationfiltersfor differentsituations. Forexample,you cancustomiseafilterforthesofa,and anotherforyourdiningarea. Incases whereyouaremovingbetweenthese locations,youcansimplyselecta combinationofbothfilterstocreatea singlecompensationfilterthat improvesthesoundexperienceinboth locations. TheBeoLab90alsooffersanother developmentinacousticalroom compensation: multichannel processing. Thismeansthatthe
loudspeakersnotonly“see”eachother ashavinganeffectontheroom–but theyhelpeachothertocontrolthe room’sacousticalinfluence.
1.1.4 Performance
BeoLab90hasbeendesignedfromthe outsettodeliveranunparalleledaudio performance. Measureddirectlyin frontoftheloudspeaker,ithasa frequencyrangethatexceedsthe limitsofhumanhearingatnormal listeninglevelsascanbeseeninthe comparisonplotinFigure1.10. (see theTechnicalSpecificationsformore details).
9
16
32
64
125
250
500
Frequency (Hz)
1,000
2,000
4,000
8,000
16,000
32,000
64,000
Piano Human Hearing BeoLab 90 Pipe Organ
Double Bass
Piccolo
Male Voice Female Voice
16
32
64
125
250
500
Frequency (Hz)
Sound Presssure Level (dB SPL)
20
-20
0
60
40
100
80
140
120
1,000
2,000
4,000
8,000
16,000
32,000
64,000
Threshold of Pain
BeoLab 90 Maximum SPL
Threshold of Hearing
Cymbals
Figure1.10: Theapproximatefrequencyrangesofexamplesoundsources. Thedarkergreybarsshowthe frequencyrangesofthefundamentalfrequencies. Thegradientbarsshowtheharmoniccontent. Thewhite lineinthePianorangeshows“MiddleC”.BeoLab90’sfrequencyrangeisshownforcomparison.
10
ApplicationInformation
2.1 MenuNavigation
Figure2.1:Presstheleft-pointingarrowheadatthetopleftofthescreentoreturntothepreviousmenu.
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure2.2:Pressthe“...” iconinthetop righttoswitchtotheeditmodeforthe currentscreen.
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure2.3:Circularselectionbuttonsallowforoneitemfromthelisttobechosen.
Room Compensation
Sweet Spot
Sofas
Entire Room
...
Dining Table
Room Compensation
Sweet Spot
Sofas
Entire Room
...
Dining Table
Room Compensation
Sweet Spot
Sofas
Entire Room
Dining Table
+
Figure 2.4: Rounded square selection buttonsaretogglesthatallowformore thanitemfromthelisttobechosen.
11
2.2 MenuMap
Beolab 90 Presets ...
Volume / Mute PRESETNAME PRESETNAME Beam Width
Latency Mode Parametric EQ Loudness Frequency Tilt Sound Enhance Sound Designs Preset Number
ZONENAME ZONENAME +
Beam Direction Speaker Role Speaker Distance Speaker Level
Room Compensation Beam Control PRESETNAME Advanced RenameTone Controls +
Wireless PowerLink Automatic Power Link XLR RCA S/P-DIF Optical USB Audio Inputs ...
Bass Treble Tone Controls
Beam Control
Room Compensation ...
Gain Offset
Listening Preset Detect Threshold
Max Input Voltage Control Volume Input Impedance Time-Out
INPUTNAME
About
Guide Network
Reset
Power Enhance Max Volume Default Volume
System
Automatic Update UTC Time Search for Update Submit Log Product information About
Speaker
Preset Input Systemetc.
Advanced
Figure2.5: SimplifiedandgeneralisednavigationalmapfortheBeoLab90interface. Someitemsshownaboveareonlyvisiblewhen the menu is in edit mode which is entered by pressing the “...” icon at the top right of some menus. Also, not all parameters are availableforallitems(e.g. differentinputshavedifferentoptions).
12
QuickSetupGuide
Itistypicallysimplesttoeditoneofthe factorydefaultPresetswhendoinga first-timesetup. Entertheeditmodebypressingthe “...” icononthetoprightofthescreen andselectthePresetyouwishto modify,orpressthe“+”icontocreate anewpreset. EntertheBeamControlmenu
• SelectthedesiredBeamWidth (andBeamDirection,if applicable) • AssigntheappropriateSpeaker Role(Left/Right)totheMaster loudspeaker • SettheDistancesfromeach loudspeakertolisteningposition • SettheSpeakerLevelsat listeningposition
Oncethisiscomplete,itisadvisedto createanActiveRoomCompensation (ARC)filter. Tostartthisprocess,
connectthemicrophonetothepanel onthebackoftheMasterloudspeaker andentertheARCmenu.
• Entertheeditmodebypressing the“...” icononthetoprightof thescreen • Pressthe“+”icontocreatea newARCfilter • Followtheinstructionsfor microphoneplacementand performthethree measurements. • Oncethemeasurement procedureiscomplete,selectthe ARCfilteryouhavemadeto applyittothelisteningPreset.
Enterthe“Advanced”menu
• VerifythattheLatencyModeis setto“Auto” • IfyoursourceisaBang& Olufsentelevision,takenoteof thePresetNumberinthismenu
RenamethePresetifdesired. Theloudspeakersarenowoptimised forathird-partysourceandreadyfor use. IftheBeoLab90’sareconnectedtoa BeoVisiontelevision:
• CreateaSpeakerGroupinthe television • IncludetheBeoLab90’sinthe “SpeakerConnections”menu • AssigntheappropriateSpeaker Rolestotheloudspeakers • SettheDistancesfromeach loudspeakertolisteningposition • SettheSpeakerLevelsat listeningposition • AssigntheSpeakerPresetinthe BeoVisionmenutomatchthe appropriatePresetnumberinthe BeoLab90’s
13
Presets
4.1 WhatisaPreset?
Almostallparametersthatcanaffect theaudiocharacteristicsoftheBeoLab 90canbepre-programmedandsaved asapreset thatiseasilyandquickly selectablebytheenduser. Apreset containsawiderangeofcontrolsthat canbecustomisedtosuitboththe listener’spersonalpreferencesandhis orherlocationinthelisteningroom. Presetscaneitherbeselected manuallyusingtheBeoLab90 interfaceortheycanbeselected automaticallyasisexplainedin AutomatingPresetSelection.
4.2 Presetmanagement
4.2.1 SelectingaPreset
Thelistofcurrently-availablepresets areshowninthePresetSelectmenu, anexampleofwhichisshowninFigure 4.1. Fromthismenu,youcanmanually selectapresetbyclickingonitsicon asshown,oryoucanmovedeeperinto theEditPresetmenuasshownin Figure4.2.
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
+
Presets
Easy Chair
Party
Sofa Left
...
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure 4.1: Select a Preset by pressing its icon on the right of the screen as shown. Thecurrently-selectedPresetis indicatedwithacheckmark.
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure 4.2: Press the three dots at the toprightofthescreentoentertheedit mode.
4.2.2 CreatingaPreset
Inordertocreateanewpreset,enter theeditmode(asshowninFigure4.2) andpressthe“+”iconinthePreset menu. Thiswillstartaprocesswhere youcannamethepresetandeditits parameters.
Easy Chair
Party
Sofa Left
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
+
Presets
Easy Chair
Party
Sofa Left
...
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure4.3: Pressthe“+”icontocreate a new preset. Note the check mark on the top right of the screen which indicates that we have entered the “edit” mode.
4.2.3 EditingaPreset
Toedittheparametersofanexisting preset,pressitsassociatediconafter youhaveenteredthepresetmenu’s editmode(bypressingthe‘...’ iconon thetoprightofthescreen).
Easy Chair
Party
Sofa Left
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
+
Presets
Easy Chair
Party
Sofa Left
...
Presets
Easy Chair
Party
Sofa Left
...Presets
Easy Chair
Party
Sofa Left
...
Figure4.4:Pressanywhereonapreset’s linetobegintoedititsparameters.
4.2.4 DeletingaPreset
Inordertodeleteapreset,enterthe presetmenu’seditmodeandswipeto theleftatanypositionintherow. This willrevealan“x”ontherightsideof thescreen. Pressingthe“x”willdelete thepreset.
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
Presets
Easy Chair
Party
Sofa Left
+
Presets ...
Presets
Easy Chair
Party
Sofa Left
...Presets ... Figure 4.5: An example of deleting a preset. To delete the “Party” preset, swipetotheleftonitsrow. Thiswillreveal the “x” on the right of the screen. Pressthe“x”todeletethepreset.
14
ControlParameters
5.1 BeamWidthControl
Thebeamofsoundthatisradiated fromtheBeoLab90canbeadjustedby selectingfromthreeoptions:
• Narrow • Wide • Omni
Beam Control
Master
Distance
Level
Left
3.5 m
5.0 dB
Master
Distance
Level
Left
3.5 m
5.0 dB
Master Left
3.5 m
5.0 dB
Wide sound
... Beam Control
Wide sound
... Beam Control
Level
Wide sound
Figure 5.1: The Beam Control menu for the Master loudspeaker. To switch between the Master and the Slave loudspeaker, swipe the upper area of the screenleftandright.
Notethat,ifanActiveRoom Compensationzonehasbeencreated, andisappliedtothepresetyouare currentlyediting,thenanappropriate filtersettingwillbecalculatedeach timeyouchangeBeamWidth. This calculationtakesapproximately20 secondstoperform,andisindicatedby aprogresswheel. Ifyouwishto auditiondifferencesinBeamWidth morequickly,thiscanbedoneby disablingtheARCfilterforthepresetin theRoomCompensationmenu.
5.1.1 Narrow
Sitinthe“sweetspot”–alocationin yourlisteningroomthatisexactlythe samedistancefromeachofyour BeoLab90’s,andwherethetwo loudspeakersarefacing(shownin Figure5.2). UsingyourBeoLab90 interface,settheBeamWidthControl to“Narrow”.
Figure5.2:A“perfect”loudspeakerconfigurationwithBeoLab90’s. Bothloudspeakersareaimedatthelisteningposition. The distance from the listening position to each loudspeaker is the same as the distance between the two loudspeakers. For more detailed information on loudspeaker configuration, see Appendix 1: Recommendations for CriticalListening
Whilefacingapointlocatedatthe centrebetweenthetwoloudspeakers, play“Tom’sDiner”(recordedby SuzanneVegain1987forheralbum “SolitudeStanding”). Vega’svoice shouldappeartofloatataposition betweenthetwoloudspeakers. Ifher voicedoesnotappeartobelocated exactlymid-waybetweenthetwo loudspeakers,itislikelythatyouare sittingslightlyclosertoone loudspeakerthantheother–inother words,toonesideofthesweetspot. Trymovingslightlyside-to-sideand payattentiontothelateralmovement ofVega’svoiceinspace. Thisabilityforapairofloudspeakersto delivertheillusionofasoundcoming fromalocationinspacebetweenthem iscalledphantom imagingorstereo imagingorsimplyimaging. Nowpayattentiontotheapparent distancetothevoice. IftheBeam WidthControloftheloudspeakersis
setto“Narrow”mode,thevoicewill appeartofloatingroughlyhalf-way betweenyouandtheloudspeakers. ThisisshowninFigure5.3.
tri
tri
tom
tom
tom
hh
voice
bongos
bk
bk bass
bass
gtr
synth pad at end
sax
snare
"choir""choir"
cow
synth fx around 2:20
shake
tri
tri
Figure 5.3: A map of the phantom imagelocationofthevoice(showninred) in Suzanne Vega’s recording of Tom’s Diner. BeamWidth: Narrow
ChangethetracktoJenniferWarnes’s recordingof“BirdonaWire”fromthe Album“FamousBlueRaincoat: The SongsofLeonardCohen”. Inthis recording,therearemanymore instrumentsandvoices,however,it shouldbeveryeasytolocatethe positionofeachofthosesourcesas comingfromsomewherebetweenthe twoloudspeakers. Apartialmapof theselocationsisshowninFigure5.4.
tri
tri
tom
tom
tom
hh
voice
bongos
bk
bk bass
bass
gtr
synth pad at end
sax
snare
"choir""choir"
cow
synth fx around 2:20
shake
tri
tri
Figure 5.4: A map of the phantom imagelocationsofinstrumentsandvoices inJenniferWarnes’srecordingofBirdon aWire. BeamWidth: Narrow.
BeoLab90isabletodeliversucha precisestereoimagingforactive listeningbecauseitisabletoreduce theamountofenergyinthereflections offthesidewallsofyourlistening
15
room. Thisgivesthesameresultatthe listeningpositionasifyouused acousticallyabsorptivematerialson yourwalls,orchangedthegeometryof yourlisteningroomtoavoidhaving firstreflectionsinthelisteningposition.
Front Right
Back Right
Figure5.5:Conceptualdrawingshowing thebeamwidthoftheNarrowBeam.
Youshouldnote,however,thatthere areside-effectstousinganarrow beamwidth. Themostobviousmaybe inthelowfrequencybehaviourofyour BeoLab90’s. Generally,theoverall impressionwillbethatthebass contentis“tighter”orhasmore “punch”whentheBeoLab90isin narrowmode. However,thiseffectis alsodependentonthesettingof anotherparameterdescribedin LatencyMode.
Figure 5.6: Press the sector (or “pizza slice”) on the BeoLab 90 interface to changetheBeamWidthtoNarrow.
Asecondpotentialsideeffectisthe sensitivityofthesystemtoanincorrect listeningposition. Youmaynoticethat, innarrowmode,itiscriticalthatyou areseatedatexactlythecorrect listeningpositioninordertoachieve bothpreciseandaccuratestereo imaging. Smalldeviationsinlistening positionmayresultinnoticeable detrimentsinthespatial representationofyourrecordings.
−15
−10
−5
0
5
30
210
60
240
90
270
120
300
150
330
180 0
Beam Width: Narrow
 
 
  64 125 250 500 1000 2000 4000 8000
Figure5.7:Polarplotofthedirectivityof theNarrowBeam. Latency: Long
Frequency [Hz]
Horizontal Angle [°]
Beam Width: Narrow
 
 
100 1000 10,000 −150
−100
−50
0
50
100
150
−12
−10
−8
−6
−4
−2
0
Figure 5.8: Full frequency range directivityplotoftheNarrowBeam. Latency: Long. Contours in steps of 3 dB, normalisedtotheon-axisresponse.
5.1.2 Wide
Asmentionedabove,whentheBeoLab 90’saresettoanarrowbeamwidth, theyaresomewhatunforgivingofa mis-placementofthelistening position. Thisisparticularlynoticeable whenyouarelisteningtorecordingsor movieswithfriendsandfamily. Consequently,inmoresocialor passivelisteningsituations,itislikely preferablethattheBeoLab90’shavea widerbeamwidth,moresimilarto BeoLab5loudspeakers. Althoughthis willlikelyresultinmoreenergyinthe sidewallreflections,italsoensuresthat thereisamoreequaldistributionof thedirectsoundacrossawider listeningareaintheroom.
Front
Front Right
Front Left
Figure5.9:Conceptualdrawingshowing thebeamwidthoftheWideBeam.
Figure5.10:Pressthecurvedlineshown above to change the Beam Width to Wide(Front).
−15
−10
−5
0
5
30
210
60
240
90
270
120
300
150
330
180 0
Beam Width: Wide
 
 
  64 125 250 500 1000 2000 4000 8000
Figure5.11: Polarplotofthedirectivity oftheWideBeam. Latency: Long
Frequency [Hz]
Horizontal Angle [°]
Beam Width: Wide, Front
 
 
100 1000 10,000 −150
−100
−50
0
50
100
150
−12
−10
−8
−6
−4
−2
0
Figure5.12: Fullfrequencyrangedirectivity plot of the Wide Beam. Latency: Long. Contours in steps of 3 dB, normalisedtotheon-axisresponse.
16
ThesideeffectsoftheWidebeam widtharedependentonthestrengthof thesidewallreflections,however,in manysituations,fourdifferenteffects maybeaudible. Thefirstisthattheapparentdistance tothevarioussourcesinthestereo mixwillcollapseslightly,resultingin theperceptionthatthesourcesinthe recordingareroughlythesame distancefromthelisteningpositionas theloudspeakersthemselves. This meansthat(relativetothenarrow mode)veryclosesourceswillmove furtherawayandveryfarsourceswill moveclosertothelisteningposition. Secondly,theapparentwidthofthe sourceswillbecomeslightlylargerwith lesspreciseleft-rightlocations. Youwill nothavepinpointlocationsasin narrowmode–imagingbecomes slightlymore“cloudy”or“fuzzy”. This isduetotheextraenergyreflectedoff thesidewalls.
bongos
"choir"
tri
Figure5.13: Amapofthephantomimagelocationofthevoice(showninred) in Suzanne Vega’s recording of Tom’s Diner. Beam Width: Wide. Compare to Figure5.3
tri
tri
tom
tom
tom
hh
voice
bongos
bk
bk bass
bass
gtr
synth pad at end
sax
snare
"choir""choir"
cow
synth fx around 2:20
shake
tri
tri
Figure5.14: Amapofthephantomimagelocationsofinstrumentsandvoices inJenniferWarnes’srecordingofBirdon aWire. BeamWidth: Wide. Compareto Figure5.4
Thirdly,theoveralltimbreortone colourofthesoundmaychangeasa resultofincreaseinfluenceofthe sidewallreflectionsatthelistening position. Finally,asmentionedabove,the overall“punch”ofthebasswillchange whencomparedtothenarrowmode.
5.1.3 Omni
Insomesituations,itmaybe preferablethattheBeoLab90’sradiate soundinalldirectionsequally. One exampleofthisarewhenyouare throwingapartyandhavemany guestslisteningtomusic simultaneouslyfrommanydifferent locationsintheroom. Anotherexample iswhenyouhavefewerpersonsinthe room,buttheyaremovingaroundto differentlocationsandsimplywant backgroundmusicwhiletheydoso. Insuchsituations,youcansetthe BeoLab90’stodeliveran“Omni”beam widthwheresoundisradiatedequally inalldirectionsinthehorizontalplane.
Figure 5.15: Conceptual drawing showing the beam width of the Omnidirectional(Omni)Beam.
Figure 5.16: Press the outside circle in the Beam Control menu to change the BeamWidthtoOmni.
−15
−10
−5
0
5
30
210
60
240
90
270
120
300
150
330
180 0
Beam Width: Omni
 
 
  64 125 250 500 1000 2000 4000 8000
Figure5.17: Polarplotofthedirectivity oftheOmniBeam. Latency: Long
17
Frequency [Hz]
Horizontal Angle [°]
Beam Width: Omni
 
 
100 1000 10,000 −150
−100
−50
0
50
100
150
−12
−10
−8
−6
−4
−2
0
Figure5.18: Fullfrequencyrangedirectivity plot of the Omni Beam. Latency: Long. Contours in steps of 3 dB, normalisedtotheon-axisresponse.
ThesideeffectsoftheOmnibeam widtharesimilartothoseoftheWide beamwidth–theonlydifferenceis thattheyaremorenoticeable. Distancestosoundsourcesbecome evenmoresimilartothedistancethe loudspeakers,left-rightimaging becomeslessprecise(butmore forgivingofincorrectlistener placement),andtheinfluenceofwall reflectionsbecomesmoreaudible.
Figure5.19: Amapofthephantomimagelocationofthevoice(showninred) in Suzanne Vega’s recording of Tom’s Diner. Beam Width: Omni. Compare to Figure5.3
hh
bongos
"choir"
tri
tri
Figure5.20: Amapofthephantomimagelocationsofinstrumentsandvoices inJenniferWarnes’srecordingofBirdon aWire. BeamWidth: Omni. Compareto Figure5.4
Inaddition,incaseswherethe loudspeakersarelocatedneararear wall,thetimbraleffectsofreflections frombehindtheloudspeakersmayalso becomemoreaudible. Experiencedreaderswillnoticethat, althoughinthelowfrequencybands, the“omni”settingresultsinan omnidirectionaldirectivity,thereare measurable“lobing”effectsinthe higherfrequencybands. Thisis primarilycausedbythedistances betweenthemidrangeandtweeters whichhavebeenoptimisedforthe narrowbeamwidth,however,ina passivelisteningorbackgroundmusic situation,thiswillnotdetractfromthe overallperformanceofthe loudspeaker.
5.1.4 Comment
Notethattheaboveillustrations connectingBeamWidthstolistener positionaremerelythat–illustrations. Itshouldalsobesaidthatchangingthe BeamWidthoftheBeoLab90has non-intuitiveconsequencesonthe perceivedsoundoftheloudspeakers. Forexample,theoverallsensationof “punch”inthebassmaybedifferent forthethreeBeamWidths,regardless ofyourlocationinthelisteningroom. Consequently,itmaybethatyou prefertheoverallsoundofaparticular BeamWidth,evenifyouarenotsitting “inthebeam”.
5.2 BeamDirectionControl
Theremaybecaseswhereyouare sittingoff-axistotheloudspeakers,far awayfromtheso-called“sweetspot” inthelisteningroom. Dependingon theplacementofyourloudspeakers, thismayevenincludelistening positionsthatarebehindthe loudspeakers. Inthesesituations,it maybedesirabletochangethe principaldirectionofradiationofthe soundfromtheBeoLab90’s,rotating thebeamsothatitisbetterdirected towardsthelisteningposition. Thisis possibleusingtheBeamDirection ControlfeatureoftheBeoLab90. WhentheBeamWidthissetto“Wide”, itispossibletochangethedirectionof thebeambyselectingfromfive options:
• Front • FrontLeft • FrontRight • BackLeft • BackRight
Thesefivedirectionsareillustratedin Figure1.9aswellasFigures5.9,5.22, and5.23. ItshouldbenotedthattheBeam Directioncontrolisonlyavailablewhen theBeamWidthcontrolissetto “Wide”. Thisisbecausethenarrow beamwidthisonlypossibleduetothe clusterofthreetweetersandthree midrangedriversonthefrontofthe loudspeaker. Also,sincetheomni beamwidthiscircular,itsrotation wouldberedundant.
Figure5.21:Pressthecurvedlineshown above to change the Beam Width to WidewithaLeftFrontdirection.
18
Notethatthebeamdirectionsofthe twoBeoLab90’sinapairare independent,however,ifyouare adjustingtheMasterloudspeaker’s direction,theSlave’sdirectionwillbe automaticallyadjustedtomatch. (e.g. IfyousettheMastertoLeftBack,then theSlavewillalsobesettoLeftBack.) Ifyouwouldliketheloudspeakersto bedirectedintwodifferentdirections (e.g. LeftBackandRightFront),you shouldadjusttheMasterloudspeaker first,andthenadjusttheSlave.
Front
Front Right
Figure 5.22: Conceptual drawing showingthebeamwidthoftheMediumBeam intheLeftFrontdirection.
Figure 5.23: Conceptual drawing showingthebeamwidthoftheMediumBeam intheRightBackdirection.
Notethat,ifanActiveRoom Compensationzonehasbeencreated, andisappliedtothePresetyouare currentlyediting,thenaappropriate filtersettingwillbecalculatedeach timeyouchangeBeamDirection. This calculationtakesapproximately20 secondstoperform,andisindicatedby
aprogresswheel. Ifyouwishto auditiondifferencesinBeamDirection morequickly,thiscanbedoneby disablingtheARCfilterforthePresetin theRoomCompensationmenu.
5.3 SpeakerDistance
TheSpeakerDistancecontrolisusedto ensurethatthetimesofarrivalofthe loudspeakers’signalsatthelistening positionarematched,despitetheir beingplacedatdifferentdistances fromthelisteningposition. Thevalue displayedonthemenushouldbethe distancefromthelisteningpositionto eachloudspeaker. Theresultofthis alignmentisthatthecloser loudspeaker’ssignalisdelayedto matchthetimeofarrivalofthesound fromthemoredistantloudspeaker.
Notethat,sincetheListeningPosition canbedifferentfordifferentPresets, thesedistancesmaynotnecessarilybe thesamefromPresettoPreset. Units Metres,Feet Range(m) 0.0–10.0 Range(ft) 0.0–30.0 Resolution 0.1 FactoryDefault 1.0m
5.3.1 AdjustingSpeaker Distanceswitha rotatedBeamDirection
IfyouaremeasuringtheSpeaker Distancesmanually,thenthe measurementshouldbemadefrom thelisteningpositiontothetweeter associatedwiththeBeamDirectionas illustratedinFigure5.25.
Figure 5.25: The Speaker Distance for each loudspeaker should be measured from the listening position to the relevanttweeterforthegivenBeamDirections.
5.3.2 AdjustingSpeaker Distancesformorethan onelisteningposition
Incaseswherethereismorethanone listenerpresent,theSpeakerDistances canbeoptimisedbymeasuringeach loudspeaker’spositionrelativetothe closestlisteningposition,asisshown inFigure5.26.
Figure 5.26: The Speaker Distance for each loudspeaker should be measured fromtherelevanttweeterforthegiven BeamDirectionstotheclosestlistening position.
AutomatedMeasurementof SpeakerDistance Thedistancefromthelisteningposition toeachBeoLab90canbemeasured automaticallyusingthemicrophone. Thisisdonebypressingthe microphoneiconatthebottomofthe BeamControlmenu(showninFigure 5.27)andfollowingtheinstructions.
19
Master
Distance
Level
Left 3.5m
dB5.0
180  directional sound
Master
Distance
Level
Left 3.5m
dB5.0
180  directional sound
Master
Level
Left 3.5m
dB5.0
180  directional sound
Figure 5.27: Pressing the microphone icon in the Beam Control menu starts the Speaker Distance measurement procedure.
5.4 SpeakerLevel
ThesensitivityofanytwoBeoLab90’s hasbeencalibratedduringtheir creationtobewithin0.2dBofeach otheratanythird-octavefrequency bandwithintheirfrequencyrange.1 However,therearecaseswhere,due toplacementinthelisteningroom, roomacoustics,orthelistening positionrelativetotheloudspeakers, youmaywishtofine-tunetherelative levelsofthetwoloudspeakers. This canbedonewiththeSpeakerLevel adjustment. ItisrecommendedthattheSpeaker Levelsshouldbeadjustedatthe listeningposition. Notethatthiscanbe performedeitherbeforeorafteran ActiveRoomCompensationprofilehas beencreated–theARCcompensates foranyadjustmentsautomatically. TheSpeakerLevelforeachBeoLab90 inthepairisadjustedfromtheBeam Controlmenu,showninFigure5.1.
5.5 SpeakerRole
TheBeoLab90iscreatedasapairof loudspeakers–one“master” loudspeakerwhichhastheconnection panelfortheinputsignalsandone “slave”loudspeaker. Sinceboththeleftandrightaudio channelsareinputtoyourmaster loudspeaker,thereisnophysicalway
ofknowingwhichloudspeakerisonthe leftandwhichisontheright(compare Figures5.28and5.29asanexample). Asaresult,theinterfaceallowsyouto swaptheSpeakerRole,toensurethat thecorrectaudiochannelis reproducedbythecorrectloudspeaker. TheselectionofLeftorRightforthe MasterandSlaveloudspeakersisdone intheBeamControlmenu,shownin Figure5.1.
Master (Left)
Slave (Right)
Slave (Left)
Master (Right)
Master (Right)
Slave (Left)
Slave (Right)
Master (Left)
Figure 5.28: An example of a loudspeaker configuration where the Masterloudspeakershouldbeassignedthe SpeakerRoleof“right”.
Master (Left)
Slave (Right)
Slave (Left)
Master (Right)
Master (Right)
Slave (Left)
Slave (Right)
Master (Left)
Figure 5.29: An example of a loudspeaker configuration where the Masterloudspeakershouldbeassignedthe SpeakerRoleof“left”.
Inadditiontothis,sincethebeam directioncanberotatedtothebackof theloudspeakers,itispossiblethat,for somepresets,youwillwishtoswapthe leftandrightSpeakerRoles(compare Figures5.29and5.30asanexample). NotethataMaster/Slavepairof BeoLab90’scannotsharethesame SpeakerRole. Ifyouwishtosendthe sameaudiosignaloutofboth loudspeakers,thiswillhavetobe arrangedusingthesourcedevice.
Master (Right)
Slave (Left)
Figure 5.30: An example of a loudspeaker configuration where the Masterloudspeakershouldbeassignedthe SpeakerRoleof“right”.
5.6 ActiveRoom Compensation
Forageneralintroductiontothe effectsofroomacousticsonthesound ofaloudspeaker,pleasereadAppendix 3: TheInfluenceofListeningRoom AcousticsonLoudspeakers Itshouldbenotedthattheacoustical behaviourofaroomcanchange considerablywhenwindowsordoors areopenedandclosed. Consequently, foroptimaltuning,itisrecommended thatARCprofilesbemadeforthese cases,particularlyifthischangeis madeoften(e.g. patiodoors).
5.6.1 CreatinganewARC Zone
AnewActiveRoomCompensation zonecanbecreatedbypressingthe “+”iconintheRoomCompensation Editmenu. (EntertheRoom CompensationEditmenubypressing thethreedotsatthetoprightofthe RoomCompensationSelectmenu.)
1Preliminaryspecification. Subjecttochange.
20
Room Compensation ...
Room Compensation ...
Room Compensation
Sweet Spot
Sofas
Entire Room
Dining Table
+
Figure 5.31: Press the “+” icon in the Room Compensation Edit menu to createanewARCzone.
Thiswillstartaprocedurewhereyou willbeguidedthroughtheprocessof positioningthemicrophoneindifferent locationstooptimisedtheARCfilters. EachARCiscreatedusing measurementsmadeinthree microphonelocations,andamaximum of5differentARCZonescanbe created. Foradditionalguidance,pleasesee Appendix5: Microphoneplacement strategywhencreatingARCZones.
5.6.2 SelectinganARCZone
Itispossibletocreateupto5different ActiveRoomCompensationZonesthat canberecalledeithermanually,or automaticallyaspartofaPreset. Inordertodisabletheactiveroom compensationfilters,simplyde-select theminthemenu.
Figure 5.32: An example of a listening space showing four different overlappingARCzonesinred,blue,green,and yellow.
Room Compensation
Sweet Spot
Sofas
Entire Room
...
Dining Table
Room Compensation
Sweet Spot
Sofas
Entire Room
...
Dining Table
Room Compensation
Sweet Spot
Sofas
Entire Room
Dining Table
+
Figure 5.33: Selecting one or more ActiveRoomCompensationzonesaccordingtoyourlisteningarea(s)intheroom. Note that it is possible to select more thanonezonesimultaneously.
5.6.3 CombiningARCZones
UptothreeARCZonescanbeselected simultaneouslytocreateafilterthat incorporatesthemeasurementsfrom theapplicableareasofthelistening room. Note,however,thataddinganextra zonetoacurrentonemaycompromise thequalityoftheaudiosignalinthe originalzone. Forexample,ifyouhave twoARCZones,oneforthe“Sweet Spot”andtheotherforthe“Dining Table”,addingtheDiningTablezoneto theSweetSpotzonewillreducethe qualityoftheARCfilteringinthesweet spotlocation. Thisisduetothefact thatsomeofthefilteringrequiredto
compensatefortheroom’sacoustical effectsinthediningareamaynotbe requiredinthesweetspot. AlsonotethatchangingRoom Compensationzoneswillcausean approximately20-secondbreakinthe audiosignalastheBeoLab90 calculatesandupdatestheappropriate filters. Thisisnormal.
5.7 Volume
ThevolumeoftheBeoLab90is controllablefrom0to90instepsof1 dB.NotethatVolumeStep0isafull mute. Initsdefaultsettings,BeoLab90has beencalibratedtomatchthelevelof otherBang&Olufsenloudspeakersfor itsPowerLinkandWirelessPowerLink inputs. Tables8.1and8.3showthe outputleveloftheloudspeakerfor variousinputsandparameters. Notethat,althoughtheVolumecontrol oftheBeoLab90isdisabledforPower LinkandWirelessPowerLinksources,2 thevolumeofthesourceisduplicated ontheBeoLab90app. Thisisto ensurethatchangestoadifferent sourcearematched. Forexample,sayyouhaveaBeoVision AvantconnectedtoPowerLinkinput andaCDplayerconnectedtothe S/PDIFinputoftheBeoLab90. You startbylisteningtoaCDatahigh volumelevel,thenswitchtowatching thetelevisionnewsatalowlevel(set onthetelevision). Whenyouswitch backtolisteningtoCD,thevolumeof theBeoLab90willautomaticallyhave beenchangedtothelowsettingofthe television. Youmayalsonoticethatthe volumesettingoftheBeoVision televisionisalsodisplayedonthelight ringsontopoftheBeoLab90’s.
5.8 Mute
Pressingthemutebuttoninthecentre ofthevolumewheelreducesthe volumetoafixedvalueof0.
2Thisrestrictionismadetopreventincorrectcalibrationoflevelsinsurroundsoundconfigurations.
21
Figure5.34:TheVolumecontrol(theexterior circle) and the Mute control (the iconinthecentreofthecircle).
Inordertounmutethesound,either pressthemutebuttonagain,oradjust thevolume. Notethat,ifthevolumesettingofthe BeoLab90washigherthanthestartup volumewhenmuted,thenthevolume settingafterunmutingwillbethesame asthestartupvolume.
5.9 ToneControls
TheToneControlsontheBeoLab90 consistoftraditionalTrebleandBass controls. Theseareglobaladjustments thatareappliedtoallPresetsandto bothloudspeakerssimultaneously.
5.9.1 Treble
TheTrebleadjustmentallowsyouto changetherelativeamountof high-frequencysoundgloballyusinga highshelvingfilterwithafixed turnoverfrequencyof8kHzandaQof 0.707. Notethatthegainatthe turnoverfrequencyisonehalfthe maximumgainappliedbythefilterin decibels. Forexample,whenthegain ofthecontrolleris-4dB,thegainat8 kHzis-2dB. TheTreblecontrolisappliedtoaglobal filterandthereforeisappliedtoall Presets. Itisalsoisindependentofthe settingsofotherequalisation controllersinthesystemsuchasthe FrequencyTilt,SoundEnhanceand ParametricEqualisercontrols. The rangeofthecontrollerisfrom-6.0dB to+6.0dBinstepsof0.5dB.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure5.35: MagnitudeResponses,Treblecontroller. Notethatthisfilterisapplied to both loudspeakers simultaneously.
5.9.2 Bass
TheBassadjustmentallowsyouto changetherelativeamountof low-frequencysoundgloballyusinga lowshelvingfilterwithafixedturnover frequencyof120HzandaQof0.707. Thegainattheturnoverfrequencyis onehalfthemaximumchangeingain appliedbythefilterindecibels. For example,whenthegainofthe controlleris+6dB,thegainat120Hz is+3dB. TheBasscontrolisaglobalfilterand thereforeisappliedtoallPresets. Itis alsoisindependentofthesettingsof otherequalisationcontrollersinthe systemsuchastheFrequencyTilt, SoundEnhanceandParametric Equalisercontrols. Therangeofthe controllerisfrom-6.0dBto+6.0dBin stepsof0.5dB.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure 5.36: Magnitude Responses: Bass controller. Note that this filter is appliedtobothloudspeakerssimultaneously.
5.10 AdvancedControls
TheAdvancedControlssectiongives theuseranalmost-surgicalcontrol
overthetimbralcharacteristicsofthe BeoLab90usingacombinationof legacyBang&Olufsenaudio processing,standardequalisationtools foundinprofessionalstudio equipment,andproprietaryprocessing availableonlyinthisloudspeaker. TheAdvancedControlsoftheBeoLab 90are
• LatencyMode • Loudness • FrequencyTilt • SoundEnhance • SoundDesign • ParametricEqualiser
5.10.1 LatencyMode
InordertocontroltheBeamWidthof thesoundradiatingfromtheBeoLab 90,acustomisedFiniteImpulse Response(FIR)audiofilterisselected foreachwoofer,midrangeandtweeter. Thesefiltersareappliedtoeachofthe DSP’s18audiooutputchannels. However,inordertocontrolthevery lowfrequencybands,itisnecessary forthewoofers’FIRfilterstobevery long. Oneimplicationofthisisthatit takessometimebetweenthemoment anaudiosignalenterstheinputofthe loudspeakerandthemomentitexits theloudspeakerassound. Thelowerin frequencytheBeamWidthControlis extended,thelongerthelatency(or delay)oftheloudspeaker. Thisultimatelymeansthatthereisa directrelationshipbetweentheoverall latencyoftheloudspeakerandits soundcharacteristics–especiallyin thelowfrequencybands. Oneexample ofthiseffectis: thelongerthelatency, the“tighter”thebass. However,thismaymeanthat,for somesourcesandprogrammaterials, thereisalossofsynchronisation. For example,initslongestlatencysetting, theloudspeakermaybetoolateto maintainlipsynchwithsome televisionsorsomemultiroom systems. Thisiswhythelatencyofthe
22
loudspeakerisuser-selectablebetween twodifferentsettings.
Auto IfyouareusingBeoLab90’switha currentBeoVisiontelevision3,thenthe LatencyModeshouldbesetto“Auto”. Thiswillallowthetelevisiontomanage thelatencymodeoftheloudspeakers automatically. Notethat,iftheLatencyModeissetto “Auto”andtheinputisneitherPower LinknorWirelessPowerLink,thenthe BeoLab90willdefaulttoaHigh LatencyModeof100ms.
High Toachievethehighestpossiblelevelof audioqualityfromtheBeoLab90,the internaldigitalprocessingmusttake 1/10thofasecondinordertocontrol thelow-frequencybehaviourofthe system. Thisisselectedbysettingthe LatencyModeto“High”,thus deliveringtheultimatepossiblesound qualityfromtheloudspeaker. However,therearecaseswheresucha longdelayintheloudspeakerwill resultinlossofsynchronisationwith otherdevicesinthesystemsuchas thevideo(lipsynch)orother loudspeakersinasurroundsystem. If youareexperiencingsuchproblems, thenthelowerlatencymodeshouldbe selected. ThelatencyoftheBeoLab90in“High” latencymodemeasuredusingan analogueinputis100ms.
Low Insomecases,aBeoLab90is connectedtoasystemthatrequiresa lowerlatency. Oneexampleofthisisa casewheretheloudspeakeris connectedtoanon-B&Otelevisionor surroundprocessor. Asecondexample isanon-B&Omultiroomsystemthat lackstheabilitytoadapttodifferent loudspeakerlatenciesthroughoutthe
network. Anotherexamplewouldbea multichannelloudspeakersetupwitha non-B&Osurroundprocessoranda mixtureofdifferentloudspeakersinthe configuration. Inthiscase,theoveralldelayofthe BeoLab90shouldbesetto“Low”to ensuresynchronisationwithother loudspeakersinthesystem. ThelatencyoftheBeoLab90in“Low” latencymodemeasuredusingan analogueinputis25ms.
EffectsofLatencyModeon BeamWidth
Frequency
Beam Width
Front SideSideBack Back
Figure 5.37: Conceptual plot showing therelationshipbetweenLatencyMode and a Narrow Beam Width over frequency. The black curve shows a High latency mode. The red curve shows a low latency mode. Note that the high frequency beam width is the same for both latency modes. Only the beam widthofthelowfrequencybandswiden forlowerlatencies.
5.10.2 Loudness
Sadly,humanhearingisimperfect. Oneoftheissuesthatweallsuffer fromisthatourperceptionofthe timbreor“tonecolour”ofasoundis notconstantwithlisteninglevel. We arelesssensitivetolowfrequencies whentheyareplayedatlowlistening levels. Inotherwords,ifyouare listeningtomusicatahighleveland youturndownthevolume,youwill noticethat,thelowerthevolume,the lessbassyoucanhear. Thisisalsotrue ofhighfrequencies,albeittoalesser extent.
TheLoudnesssettinginyourBeoLab 90’scounteractsthiseffect. Asyou reducethevolume,thebassandtreble levelsareautomaticallyincreasedto compensateforyourreduced perceptionintheouterfrequency bands. Ifyoudonotwishthissettingenabled, LoudnessshouldbesettoOFF. NotethattheLoudnesstoggle (whetheritisonoroff)isstoredwith thePreset,sodifferentmodescan havedifferentsettings. Options On/Off Default On
10 100 1,000 10,000
0
2
4
6
8
10
12
Frequency (Hz)
Gain (dB)
Figure5.38:Magnituderesponsesofthe loudnessfunctionatvarioussettingsof thevolumecontrol.
Notethat,whenconnectedtomost Bang&Olufsensources,theLoudness functionintheBeoLab90willbe disabledforthePowerLinkand WirelessPowerLinkinputs. Thisis becauseinthesecases,theLoudness functionisperformedbythesource ratherthantheloudspeaker.
5.10.3 FrequencyTilt
FrequencyTiltcanbeconsideredtobe acombinationofBassandTreble settingsinasingleparameter. When FrequencyTiltissettoalowvalue,the lowfrequencycontentofyouraudio signalisincreasedandthelevelofthe highfrequencycontentisreduced. IftheFrequencyTiltissettoahigh value,thentheoppositewillbetrue. TheFrequencyTiltfunctionwillhave noeffectontheaudiosignalatits middlesetting.
3BeoPlayV1,BeoVision11,Avant,AvantNG,14,Horizon,Eclipse–orlater
23
NotethatFrequencyTiltcanhave differentsettingsfordifferentPresets. Therangeofthecontrollerisfrom-6.0 dBto+6.0dBinstepsof0.5dB.As canbeseeninFigure5.39,acontroller settingof+6.0willresultina peak-to-peakmagnituderesponse deviationofapproximately6dB, howeverthemaximumdeviationfrom aflatresponseisonly3dB.
10 100 1,000 10,000 −4
−3
−2
−1
0
1
2
3
4
Frequency (Hz)
Gain (dB)
Figure5.39: MagnitudeResponses,FrequencyTiltcontroller. Notethatthisfilter is applied to both loudspeakers simultaneously.Blackcurvesshowtheresultforpositiveslidervalues,redcurves shownegativeslidervalues.
5.10.4 SoundEnhance
TheSoundEnhancesettingissimilarto theFrequencyTiltsettinginthatit affectsthelowandhighfrequency bandswithasingleslider. Increasing theSoundEnhancevaluewillincrease thelevelofthebassandtreblebands whilereducingthemidrange. DecreasingtheSoundEnhancevalue willhavetheoppositeeffectandwill enhancethemidrange. TheSoundEnhancesettingwillhave noeffectonthisaudiosignalatits middlesetting. NotethatSoundEnhancecanhave differentsettingsfordifferentPresets. Therangeofthecontrollerisfrom-6.0 dBto+6.0dBinstepsof0.5dB.As canbeseeninFigure5.40,acontroller settingof+6.0willresultina peak-to-peakmagnituderesponse deviationofapproximately6dB, howeverthemaximumdeviationfrom aflatresponseisonly3dB.
10 100 1,000 10,000 −4
−3
−2
−1
0
1
2
3
4
Frequency (Hz)
Gain (dB)
Figure 5.40: Magnitude Responses, Sound Enhance controller. Note that this filter is applied to both loudspeakers simultaneously. Black curves show theresultforpositiveslidervalues,red curvesshownegativeslidervalues.
5.10.5 SoundDesign
Attheendofthedevelopmentprocess, allBang&Olufsenloudspeakersgo throughafinaltuningprocesswhere theloudspeaker’stimbreisevaluated indifferentlisteningenvironments. In ordertoachieveanoptimisedbalance betweentheon-axisfrequency responseandthethree-dimensional “powerresponse”,filtersareincluded inthesignalpathtogivethe loudspeakerafinalsound design. TheBeoLab90isnoexceptiontothis– asaresult,ithasacustom-tuned, factory-defaultsounddesignforevery combinationofbeamwidths,beam directions,andlatencymodes. However,theremaybesomespecific caseswherethistuningisnot applicable. Oneexampleofthisisa casewheretheBeoLab90isusedina listeningroomsuchasarecording studiowhereacousticalabsorptionhas beenappliedtothevarioussurfaces. Inthiscase,itmaybepreferableto usetheBeoLab90asa“studio monitor”styleofloudspeaker,where theoveralltuningisdesignedto deliveraflatmagnituderesponse whenmeasuredon-axistothe loudspeakerinafreefield. TheSoundDesigncontrolallowsyouto switchbetweenthesetwotunings. Itis currentlyplannedthatadditionalsound designswillbemadeavailablein futuresoftwarereleases.
5.10.6 ParametricEqualiser
For a general introduction to equalisation, please see Appendix 2: Introduction to Parametric Equalisers. Incaseswhereamoredetailedcontrol ofthefrequencyresponseofthe loudspeakerisneeded,a10-band parametricequaliserisavailable. This allowsyoutosculptthetimbral balanceoftheloudspeakerwithahigh degreeofprecision. Whenthegainsofalltenfiltersinthe ParametricEqualiseraresetto0dB, theprocessingblockisautomatically disabled. Figure5.41isgivenasarough“map” offrequencyasreferencewhenusing theParametricEqualiser. Additional guidanceisgiveninTable8.5onpage 33
440.0 Hz
660.0 Hz
880.0 Hz
1320.0 Hz
220.0 Hz 261.6 Hz 330.0 Hz
110.0 Hz
165.0 Hz
55.0 Hz
55.0 Hz
Figure 5.41: Pitch vs. Fundamental frequencyforreferencepurposeswhen equalising.
TheParametricEQconsistsofone low-shelvingfilter,onehigh-shelving filter,and8reciprocalpeak-dip(or peaking)filters,eachwithavariable Frequency,Gain,andQ. Thefilters’frequencyrangesarelisted inTable8.4onpage33andarelimited toISO1/6thoctavecentres. Formore information,pleaserefertoTable8.7 onpage34. TheavailableQ’softhefiltersare limitedtothevalueslistedinTable8.6 onpage33. Notethatallfiltersareimplementedin series,andthatfrequenciesmaybe overlappedincaseswhereadditional gainisdesired.
24
AllfiltersintheParametricEQsection areimplementedasminimumphase filters. Inordertoensurephasematchingof thetwoloudspeakersandthereforeto maintainphantomimaging characteristics,identicalParametric Equaliserparametersareappliedto bothloudspeakerssimultaneously.
MagnitudeResponsePlots Low-ShelvingFilter TheParametricEqualiserhasone low-shelvingfilteravailablewitha frequencyrangeof32.0Hzto500.0Hz andaQrangeof0.4to1. Thegain rangesfrom-6.0dBto+6.0dBin stepsof0.5dB.
Figure5.42:MagnitudeResponses,lowshelvingfilter:Fc=100Hz,Gainvaried from-6.0to+6.0dB,Q=0.7.
Figure5.43:MagnitudeResponses,lowshelving filter: Fc varied from 32 Hz to 500Hz.,Gain=±6dB,Q=0.7.
10 100 1,000 10,000 Frequency (Hz)
-6
-4
-2
0
2
4
6
Gain (dB)
Figure5.44:MagnitudeResponses,lowshelvingfilter: Fc=100Hz,Gain=±6 dB,Qvariedfrom0.4to1.Notethat,for Q values greater than 0.7, there is an overshoot in the magnitude response. For a monotonically increasing (or decreasing)response,theQshouldnotbe settoavaluegreaterthan0.7.
PeakingFilters TheParametricEqualiserhaseight reciprocalpeak-diporpeakingfilters available. AllpeakingfiltershaveaQ valuethatrangesfrom0.5to8.0 wheretheQisbasedonabandwidth definedbythehalf-gainpoints4. The gainrangesfrom-6.0dBto+6.0dBin stepsof0.5dB.Thepeakingfilters havearangeof5octaveswith differinglimitsasfollows:
• Fourlow-frequencyfilterswitha rangeof25.0Hzto400.0Hz. • Threemid-frequencyfilterswith arangeof250.0Hzto4.0kHz. • Onehigh-frequencyfilterwitha rangeof2.5kHzto40.0kHz.
Figure 5.45: Magnitude Responses, Peakingfilter: Fc=100Hz,Gainvaried from-6.0to+6.0dB,Q=1.
Figure 5.46: Magnitude Responses, Peaking filters: Examples of Fc varied from25Hzto25kHzonone-octavecentres,Gain=±6dB,Q=1.
10 100 1,000 10,000 Frequency (Hz)
-6
-4
-2
0
2
4
6
Gain (dB)
Figure 5.47: Magnitude Responses, Peakingfilters: Fc=100Hz,Gain=±6 dB,Qvariedfrom0.3to8.
High-shelvingFilter TheParametricEqualiserhasone high-shelvingfilteravailablewitha frequencyrangeof1.0kHzto16.0kHz andaQrangeof0.4to1. Thegain rangesfrom-6.0dBto+6.0dBin stepsof0.5dB.
Figure 5.48: Magnitude Responses, high-shelving filter: Fc = 10 kHz, Gain varied from -6.0 to +6.0 dB, Q = 0.7. Note that this filter is applied to both loudspeakerssimultaneously.
4For more information on this, please see “The Equivalence of Various Methods of Computing Biquad Coefficients for Audio Parametric Equalizers” Robert Bristow-Johnson, Preprint3906,97thInternationalConventionoftheAudioEngineeringSociety,November1994
25
Figure 5.49: Magnitude Responses, high-shelving filter: Fc varied from 1 kHzto16kHz,Gain=±6dB,Q=0.7.
10 100 1,000 10,000 Frequency (Hz)
-6
-4
-2
0
2
4
6
Gain (dB)
Figure 5.50: Magnitude Responses, high-shelvingfilter:Fc=100Hz,Gain= ±6dB,Qvariedfrom0.4to1.Notethat, forQvaluesgreaterthan0.7,thereisan overshoot in the magnitude response. For a monotonically increasing (or decreasing)response,theQshouldnotbe settoavaluegreaterthan0.7.
Generalcommentregarding equalisation Althoughitispossibletoapplyagain tothesignalusingtheparametric equaliser,thiswillhaveimplicationson theheadroomofthetotalsystem. For example,ifyousetthegainofan equalisertobe+6dBat30Hz,and playasignalthatcontainsa30Hztone atmaximumlevel(e.g. 0dBFSona digitalinput)thenyouwillreachthe maximumpossibleoutputofthe loudspeaker(andthereforetheABL algorithmwillstarttoprotectthe loudspeaker)atavolumestepthatis6 dBlowerthanitwouldbewithoutthe filter. Thisisthereasonthata“ruleof thumb”forprofessionalaudio engineersitthat,whenadjusting parametersinaparametricequaliser,
wheneverpossible,a“cut”is preferabletoa“boost”. Forexample,if youwishtohavemorebass,itis smartertoreducethetrebleandturn uptheoverallvolumethantosimply increasethebass. Theresultwillbe thesameatlowervolumesettings,but therecanbearemarkabledifference athigherlisteninglevels.
5.11 AutomatingPreset Selection
Itisnotnecessarytomanuallyselect PresetsusingtheBeoLab90app. Itis possible,instead,tohavePresets triggeredtobeselectedautomatically usingoneoftwopossibleexternal controls: BySpeakerGroup(ifyou haveaBang&Olufsentelevisionsuch asaBeoVision11orBeoVisionAvant) orBySource.
BySpeakerGroup IfyouhaveapairofBeoLab90’s connectedtoaBang&Olufsen televisionsuchasaBeoVision11or BeoVisionAvantasshowninFigure 5.51,thenitispossibleto automaticallytriggerpresetsin tandemwiththetelevision’sSpeaker Group. Thisselectionisdoneinthe SpeakerGroupmenusonthe television,whereyoucanselectthe “SpeakerPreset”numberforthe BeoLab90asoneoftheparametersin theSpeakerGroup. SeetheBeoVision TechnicalAudioGuideformore informationaboutthis. Notethat,incaseswherea multichannelloudspeaker configurationincludesmorethanone pairofBeoLab90’sconnectedtoa BeoVisiontelevision,itwillbe necessarytoensurethatthePreset numbersarethesameforallpairsof BeoLab90’sinthesystem,sincethe televisionsendsoutonlyoneSpeaker Presetnumberforallloudspeakers
connectedtoit.
BySource ImagineyouhaveapairofBeoLab 90’sconnectedtotwonon-B&O sourcesasshowninFigure5.52.
• anAVSurroundProcessor connectedtotheXLRLineinputs. Thedeviceisalsoconnectedto otherloudspeakerstoforma multichannel(surround) configurationforwatching movies. • ahigh-resolutionaudioplayer connectedtotheS/PDIFinput.
Inaddition,youhaveconfiguredtwo PresetsinyourBeoLab90’s:
1. Optimisedformultichannel listeningwithalisteningzone thatencompassesmorethanone listeningposition(e.g. thewhole sofa). 2. Optimisedfor2.0Stereolistening withonlyone“sweetspot”inthe centreofthesofa.
Inthissituation,youwanttheAV Surroundprocessortoautomatically selectPreset1andthehigh-resolution audioplayertoautomaticallyselect Preset2. Inthisway,thereisnoneed tomanuallychangeBeoLab90presets.
Mixedsystems Notethatispossibletotriggerbothby sourceandbySpeakerGroupinmixed systemssuchasthatshowninFigure 5.53. Inthiscase,theBeoVision televisioniscontrollingtheBeoLab90 presetwithinitsSpeakerGroup parameters. However,theBeoLab90 canalsohaveapresetthatis automaticallytriggeredbytheaudio playerconnectedviaS/PDIF.
26
 Audio Player
AV Surround Processor
Power Amplifier
XLR
S/P-DIF
Audio Player S/P-DIF
Digital Power Link
Digital Power Link
Power Link
Digital Power Link
Power Link
Figure5.51: AnexampleofapairofBeoLab90’sconnectedtoaBeoVision11usingPowerLinkbl.
 Audio Player
AV Surround Processor
Power Amplifier
XLR
S/P-DIF
Audio Player S/P-DIF
Digital Power Link
Digital Power Link
Power Link
Digital Power Link
Power Link
Figure 5.52: An example of a pair of BeoLab 90’s connected to two third-party sources: an AV Surround Processor using XLR and a separateaudioplayerusingS/PDIF.Notethat,inthiscase,thelatencyoftheBeoLab90’smustbecarefullymanagedinthesetupof theloudspeakersandtheAVSurroundProcessorinordertoensurethatthemultichannelsystemisbehavingcorrectly.
 Audio Player
AV Surround Processor
Power Amplifier
XLR
S/P-DIF
Audio Player S/P-DIF
Digital Power Link
Digital Power Link
Power Link
Digital Power Link
Power Link
Figure5.53: AnexampleofapairofBeoLab90’sconnectedtooneB&OsourceusingPowerLinkandathird-partyaudioplayerusing S/PDIF.
27
Inputs
Ascanbeseenontheconnectorpanel showninFigure6.5,theBeoLab90has atotalofeightdifferentaudioinputsas follows: Bang&OlufsenProprietary
• PowerLink(analogue) • WirelessPowerLink(digital) Digitalinputs
• S/PDIF(or“coaxial”) • Optical • USBAudio Analogueinputs
• XLR(or“balancedline”) • RCAPhono(or“unbalancedline”) Wirelessinputs
• WiSA Thetechnicalspecificationsforthese canbefoundinInputs. Itispossibletoenableanaudiosource connectedtoaninputeithermanually (viatheBeoLab90interface)or automatically,asdescribedbelow.
6.1 InputsSelection
6.1.1 AutomaticSelection
Inputs
Power Link XLR
S/P-DIF
RCA
Optical
USB-Audio
WPL / WiSA
Automatic
AUTOMATIC SENSE
MANUAL SENSE
...Inputs
Power Link XLR
S/P-DIF
RCA
Optical
USB-Audio
WPL / WiSA
Automatic
AUTOMATIC SENSE
MANUAL SENSE
... RCA
Gain Offset
-76 dBV
50 kΩ
2.0 V
Detection threshold
Max input voltage
Input impedance
Time-out
Figure6.1: TheInputSelectmenu.
SelectionPriority IftheBeoLab90issettoautomatically detectaninputsignal,thenitmaybe necessarytocustomisethe prioritisationofthesources. For example,ifyouhaveaCDplayer connectedtotheS/PDIFinputanda turntableconnectedtotheXLRinput, andbothsourcesareplaying,this parameterallowsyoutodetermine whichsourceshould“win”andbe playedbytheBeoLab90. Thisprioritisationcanbepersonalised bychangingtheverticalorderofthe inputsonBeoLab90interfaceinthe InputSelectmenu(pressthe“...” icon atthetoprighttoentertheeditmode).
Inputs
Power Link XLR
S/P-DIF
RCA
Optical
USB-Audio
WPL / WiSA
Automatic
AUTOMATIC SENSE
MANUAL SENSE
...Inputs
Power Link XLR
S/P-DIF
RCA
Optical
USB-Audio
WPL / WiSA
Automatic
AUTOMATIC SENSE
MANUAL SENSE
...
Gain Offset
Detection threshold
Max input voltage
Input impedance
Time-out
Figure6.2:Thepriorityofautomaticallyselectedsourcescanbechangedbyrearrangingtheirorder.
6.1.2 ManualSelection
Theremaybecaseswhereyouprefer tomanuallyselectaninput. Inthis case,youcandragtheinputintothe “ManualSense”listatthebottomof theInputsmenuscreen. Inthiscase,a signalononeoftheseinputswillnot beautomaticallydetectedbythe BeoLab90andmustthereforebe switchedonandoffmanually. Notethatitispossibletomanually selectinputswithoutthesmartphone interfaceusingtheBeoRemote1 remotecontrol. Foradescriptionof howtodothis,includinginstructions onsettinguptheBeoRemote1,please seeSection21.
6.2 IndividualInput Parameters
Notethatnotallcontrolsareavailable forallinputs.
6.2.1 Re-naming
Itispossibletore-nametheinputs labelsintheBeoLab90interfaceby enteringtheeditmodeoftheInputs menu,selectinganinput,andthen press-and-holdthenameoftheinput atthetopofthescreen. Thispersonalisedname(i.e. “CD Player”or“Turntable”,forexample)
28
willbedisplayedthroughoutthe BeoLab90interface.
... RCA
Gain Offset 5 dB
-76 dBV
5 min.
50 kΩ
2.0 V
Detection threshold
Max input voltage
Input impedance
Time-out
Figure6.3:Renameaninputbypressing andholdingitsnameintheInputmenu
6.2.2 GainOffset
Youcanchangetherelativelevelsof theindividualinputsusingtheGain Offset parameter. Forexample,ifyou haveaparticularsourcethathasa loweroutputlevelthantheothers,its GainOffsetcanbeincreaseto compensate,makingitappeartohave thesamelevelasyourotheraudio sources. Range -12to12dB Resolution 1dB FactoryDefault 0dB NotethattheGainOffsetparameteris notavailableforthePowerLinkand WirelessPowerLinkinputs.
6.2.3 DetectionThreshold
TheBeoLab90canbesetto automaticallyturnitselfonby detectingthepresenceofasignalon theXLRandRCAlineinputs. However, dependingonyoursourceand/orthe styleofmusicyoutypicallylistento,it maybenecessarytomakethe detectionmoreorlesssensitive. This canbedoneusingtheDetection Thresholdcontrol. Forexample,ifyoulistentomusicwith alargedynamicrange,itmaybe necessarytolowertheDetection ThresholdtomaketheBeoLab90more sensitivetothepresenceofquiet
signals. Conversely,ifyouhavean audiosourcethathasahighernoise floor,itmaybenecessarytoincrease theDetectionThresholdinorderto maketheBeoLab90lesssensitive. Range -76to-46dBV Resolution 3dB FactoryDefault -76dBV SeeFigure6.4foragraphic representationofthethedetection thresholdrelativetothesignal strength. NotethattheDetectionThreshold parameterisnotavailableforthe PowerLinkinput,sincethe loudspeakerisautomaticallyturnedon andoffbythePowerLinksource. Notethat,fortheUSBAudio,S/PDIF andOpticaldigitalinputs,the Auto-detectioncontrolisusedinstead oftheDetection Threshold.
6.2.4 Auto-detection
TheBeoLab90canbesetto automaticallyturnitselfonby detectingthepresenceofasignalon itsdigitalinputsbysettingthe Auto-detectiontoON.However,thisis slightlydifferentfromthedetectionof analoguesignals,sincethedigital inputsaretriggeredbythepresenceof aanynon-zerosignalonthedigital audiostreamratherthanasignal aboveauser-definedlevel. Options On/Off FactoryDefault On NotethattheAuto-detection parameterisonlyavailablefortheUSB Audio,S/PDIFandOpticalinputs. Notethat,fortheXLRandRCA analogueinputs,theDetection Thresholdcontrolisusedinsteadofthe Auto-detection.
6.2.5 MaximumInputVoltage
Differentaudiosourceshavedifferent maximumanalogueoutputlevels. Typically,amaximumlevelfroma line-levelRCAoutputis2.0VRMS, however,differentmanufacturers
occasionallychoosetodeliverahigher outputlevelonsomemodels. Inordertomaximisethe signal-to-noiseratioofyouraudio system,theBeoLab90givesyouthe optiontochangetheMaximumInput VoltageforthetheXLRandRCAline inputs. Thedatasheetforyouraudio sourceshouldindicateitsmaximum outputlevel. ThevalueintheBeoLab 90interfaceshouldbesettomatch thisvalue. Ifthesourcehasahighermaximum outputlevelthanthatwhichissetin theBeoLab90interface,thismay causedistortionduetoclippingofthe signalattheloudspeaker’sinputs. Ifthesourcehasalowermaximum outputlevelthanthatwhichissetin theBeoLab90interface,thiswillcause yourmaximumoutputofthe loudspeakertobelower,andthe outputnoisefloortobeincreased. Options 2.0,4.0,6.5VRMS FactoryDefault 2.0VRMS NotethattheMaximumInputVoltage parameterisonlyavailablefortheXLR andRCAlineinputs.
6.2.6 Time-out
Incaseswheretheautomaticsignal detectionisusedtoturntheonBeoLab 90,theTime Out controlcanbeused todeterminethelengthoftimethe loudspeakercontinuestobepowered upaftertheaudiosignalhasstopped. Itmaybenecessarytoincreasethe lengthofthistimeifyoulistento musicwithanextremedynamicrange. Forexample,aquietpassageina pieceofmusicmaybebelowthe detectionthreshold. Ifthedurationof thatpassageislongerthantheTime Out value,thentheloudspeakerwillgo intostandbymodewhilethepieceis playing. Options 0-840seconds NotethattheTime-outfunctionisnot availableforthePowerLink,Wireless PowerLinkandWiSAinputs.
Figure6.4: The“DetectionThreshold”and“TimeOut”parameters
6.2.7 InputImpedance
IftheBeoLab90’sRCALineinputis connectedtoadevice’sheadphone outputthatusesaClass-Damplifier, theremaybeinstanceswherethis causesthenoisefloortoriseaudibly. Thisiscausedbytheinputimpedance oftheBeoLab90beingmuchhigher thanthatwhichisexpectedbythe headphoneamplifier’sdesigner. In ordertocorrectthisproblem,theinput impedanceoftheRCAinputcanbeset toalowvalueof50 Ω. However,iftheinputimpedanceofthe RCAinputissetto50 Ω anditis connectedtoadevice’sstandard low-impedancelineoutput,thismay haveadetrimentaleffectonthesignal. Forexample,themaximumpossible outputlevelwillbereduced. Insome cases,incorrectlysettingtheinput impedanceto50 Ω mayalsocause distortionoftheaudiosignal. Options 50 Ω,50kΩ FactoryDefault 50kΩ NotethattheInputImpedancecontrol isonlyavailablefortheRCAlineinput.
6.2.8 ControlVolumeof S/PDIF(orOptical)input usingPowerLink
Bang&Olufsenaudioproductsthatare abletosendtheaudiosignalonan S/PDIFoutputadditionallysendthe volumesettingonthedataconnection includedinthePowerLinkcable. Thisis usedforvariousreasons. Oneprimary exampleofthisiscurrentcustomers whoconnectaBeoSound9000toa
pairofloudspeakerssuchasBeoLab 5’sorBeoLab90’sviaaS/PDIFdigital connection. SincetheBeoSound9000 doesnotapplyvolumeregulationto theS/PDIFoutput,thevolumesetting mustbesentseparatelyonthePower Linkcableandappliedtotheaudio signalinsidetheloudspeakerinstead. ThisparameterontheBeoLab90 allowscustomerstousethevolume controlofaBang&Olufsensource (sentviaaPowerLinkconnection)and applyittoanaudiosignalcominginto theBeoLab90viaitsS/PDIFinput. Notethat,inorderforthisoptionto functionproperly,theS/PDIFinput mustbeassignedahigherprioritythan thePowerLinkinputintheSelection Priority. Thisfunctionisalsoindependently availableforsignalsontheOptical input. WhenthevolumeoftheBeoLab90is controlledbyanexternalPowerLink source,thevolumewheelinthe BeoLab90interfaceisgreyedoutand willnotrespondtotouchcommands. It does,however,displaythevolume settingassignedtotheBeoLab90by thePowerLinkdatasignal. Options Enabled/Disabled FactoryDefault Disabled
6.2.9 USBVolumeenabled
WhentheBeoLab90isconnected usingUSBAudiotoanaudiosource, youhavetheoptionofusingthe source’svolumeasanexternalcontrol forthegainoftheloudspeaker. This
alsomeansthatthevolumeofthe BeoLab90(setbyitsremotecontrol) wouldbereflectedontheuser interfaceoftheaudiosourceor softwareplayer. However,thisexternalcontrolofthe BeoLab90maynotbedesirableinall situations. Forexample,itisveryeasy toinstantlychangethevolumeofa softwareaudioplayertomaximum, whichwillbesurprisinglyloudwitha BeoLab90ifthechangewas accidental. Italsomaybepreferableto settheBeoLab90toastatic(e.g. low) volumesettingandtohavean independentadjustmentonthesource device. Inthesecases,theUSBVolume EnabledshouldbesettoDisable. Options Enabled/Disabled FactoryDefault Disabled NotethattheUSBVolumecontrolis onlyavailablefortheUSBAudioinput.
6.3 ConnectionPanels
TheconnectionpanelsontheMaster andSlaveBeoLab90’sareslightly differentinthataudiosignalscanonly beconnectedtotheMaster loudspeaker. Theaudiosignal connectionsfromyoursourcedevices shouldbeconnectedtotheMaster loudspeaker. Theonlyaudioinputon theSlaveloudspeakeristheDPLor DigitalPowerLinkinput.
30
OPTICAL
S/P-DIF
LEFT XLR RIGHTMIC / IR POWER LINK
LEFT RCA RIGHT
USB 5V    0.5 A
USB AUDIO
DPL DPL DPL / ETHERNET
USB 5V    0.5 A
DPL / ETHERNET
DPL DPL
Figure 6.5: Audio connection panel – Master loudspeaker. Analogue inputs are shown in blue. Digital audio connections are shown in red. Utility connectionsareshowninblack.
OPTICALLEFT XLR RIGHTMIC / IR POWER LINK
USB 5V    0.5 A
USB AUDIO
DPL DPL DPL / ETHERNET
USB 5V    0.5 A
DPL / ETHERNET
DPL DPL
Figure 6.6: Audio connection panel – Slave loudspeaker. Note that the only audio connectors on this loudspeaker arefortheDigitalPowerLinkconnection totheMasterloudspeaker.
IfyouarefamiliarwithXLRconnectors, youwillnoticethatthepush-button lockismissingontheXLRinput. Thisis intentionalandhasbeendonetohelp tominimiserattlingartefactswhen playingathighersoundpressure levels. Forspecificinformationregardingthe variousinputs,pleaseseeInputs.
31
System
7.0.1 About
The“About”menuallowsaccessto informationregardingtheloudspeaker aswellastheaudiosignalitis currentlyplaying. Selectthe“Speaker Info”,andtheneitherthe“Input Signal”todisplaydetailedinformation abouttheinputaudiosignalor “Temperatures”todisplaythecurrent temperaturesoftheloudspeaker drivers.
7.0.2 MaxVolume
TheMax Volumecontrolallowsyouto determinethelimitofthevolume control. Range 0–90 Resolution 1dB FactoryDefault 90 NotethattheMaxVolumeparameteris notavailableforthePowerLinkand WirelessPowerLinkinputs.
7.0.3 StartupVolume
TheStartup Volumecontrolallowsyou todeterminethevolumelevelwhen
theBeoLab90wakesasaresultofa detectedsignal,orismanuallyturned on. Range 0–90 Resolution 1dB FactoryDefault 42 NotethattheStartup Volume parameterisnotavailableforthe PowerLinkandWirelessPowerLink inputs. AlsonotethattheStartup Volumemaybeoverriddenbythe volumecontrolfromPowerLink(if enabledforS/PDIForOptical)oraUSB Audiovolumecontrol(ifenabled).
32
Tables
8.1 LoudspeakerSensitivity
Input dBSPL “MaxInput”LevelSetting PowerLink 88.0 6.5Vrms(Fixed) XLR,RCA 88.0 6.5Vrms XLR,RCA 92.2 4.0Vrms XLR,RCA 98.2 2.0Vrms Table8.1:UnweightedSoundPressureLevel(SPL)oftheBeoLab 90at1minafreefield(200Hz–2kHz). Inputsignalstrength: 125mVrms. VolumeStep: 90. AllotherparameterssettoFactory Defaults. Note that the input signal strength on the XLR inputismeasuredbetweenpins2and3.
Input Output USBAudio,S/PDIF,Optical 92.3dBSPL WirelessPowerLink 92.3dBSPL WiSA TBD Table 8.2: Unweighted Sound Pressure Level (SPL) of the BeoLab90at1minafreefield(200Hz–2kHz). Inputsignal:-30.0 dB FS. Volume Step: 90. All other parameters set to Factory Defaults.
8.2 VolumeControl
VolumeStep Output 90 92.3dBSPL 89 91.3dBSPL 88 90.3dBSPL . . 51 53.3dBSPL 50 52.3dBSPL 49 51.3dBSPL . . 2 4.3dBSPL 1 3.3dBSPL 0 -∞dBSPL Table8.3: UnweightedSoundPressureLevel(SPL)oftheaudio signalfromaBeoLab90at1minafreefield(200Hz–2kHz). Inputsignal: -30.0dBFS.Notethatthesevaluesconsideronly the output level of the audio signal and assume that thermal protectionhasnotbeenengaged.
8.3 ParametricEqualiser
Type Range(Hz) Filters Low-shelving 32.0–500.0 1 Peaking(LF) 25.0–400.0 4 Peaking(MF) 250.0–4.0k 3 Peaking(HF) 2.5k–40.0k 1 High-shelving 1.0k–16.0k 1 Table 8.4: Frequency ranges of Parametric EQ filters. The filter frequenciesareISO1/6-octavespacing.
Band Frequency(Hz) Ultrasonic 22.4kHzto40.0kHz Treble 4kHzto20kHz Midrange 125Hzto3.55k Bass 25Hzto112Hz Table8.5: FrequencyBandsforapproximateinformationonly.
FilterType Qvalues Low-shelving 0.4,0.5,0.6,0.7,0.8,0.9,1.0 Peaking 0.5,0.7,1.0,1.4,2.0,4.0,8.0 High-shelving 0.4,0.5,0.6,0.7,0.8,0.9,1.0 Table8.6: AvailableQvaluesofParametricEQfilters.
33
Frequency OctaveDivisions multiplier 1/12 1/6 1/3 1/2 1 1.0 x x x x x 1.06 x 1.12 x x 1.18 x 1.25 x x x 1.32 x 1.4 x x x 1.5 x 1.6 x x x 1.7 x 1.8 x x 1.9 x 2.0 x x x x x 2.12 x 2.24 x x 2.36 x 2.5 x x x 2.65 x 2.8 x x x 3.0 x 3.15 x x x 3.35 x 3.55 x x 3.75 x 4.0 x x x x x 4.25 x 4.5 x x 4.75 x 5.0 x x x 5.3 x 5.6 x x x 6.0 x 6.3 x x x 6.7 x 7.1 x x 7.5 x 8.0 x x x x x 8.5 x 9.0 x x 9.5 x Table8.7: ISOstandardfrequencycentresforfractionaloctave spacing.
34
Features
9.1 Resonance-basedSound Design
Averylargepartofthesoundtuningof theBeoLab90,likemanyotherBang& Olufsenloudspeakers,isbasedon acousticalmeasurementsperformedat manylocationsaround,aboveand belowtheloudspeaker.
Figure 9.1: The first BeoLab 90 prototype undergoing acoustical measurementsinTheCube.
Oneoftheimportantaspectsofthese measurementsistofindthebehaviour oftheloudspeakerintime. For example,ifasoundissenttothe loudspeaker,andthenstopped suddenly,doestheloudspeakeralso stop,ordoesit“ring”atsome frequencies(inexactlythesameway thatabellringswhenstruck). Ringing inthetimeresponseofaloudspeaker isanindicationthatithasaresonance –afrequencyatwhichit“wants”to move. Thisresonancehasa detrimentaleffectontheoverallsound oftheloudspeaker,sinceit“smears” soundsintime. Forexample,ifyouhavealoudspeaker thathasanaturalresonanceat110Hz (twooctavesbelowa“ConcertA”,to musicians)thenitwillnaturallyringat
thatnotewhenitis“hit”withan impulsivesignalsuchasakickdrum. If thesongthatthebandisplayingisnot inA(majororminor),butinB-Flat instead,thentherewillbea dissonancebetweenthenotesplayed mostofteninthesong,andthenote thatis“singingalong”withthekick drum. Thiscancontributetothe loudspeakersounding“muddy”(touse onlyoneword...). Thisiswhythemeasurement-based portionofthefilteringofallcurrent Bang&Olufsenloudspeakersis primarilydesignedtocounteractthe naturalresonancesinthesystem. So, forexample,ifoneofthewoofersin theBeoLab90hasanaturalresonance at110Hz,thenthatresonanceis mirroredwithanequal,butopposite phasebehaviourintheDigitalSignal Processingengine. Thetotalresultof thefilterintheDSPandthebehaviour ofthewooferisthatthereisno unwantedringingintheentiresystem. This,inturn,meansthatthe loudspeaker’sresponseiscontrolled notonlyinthefrequencydomainbutin thetimedomainaswell. Thisisonlypossiblewithanextensive setofmeasurementsofeach loudspeakerdriver’smechanicaland acousticalbehaviouranda custom-createdsetoffiltersforit.
9.2 Phase-OptimisedFiltering
Likeallaudiodevices,inorderforthe BeoLab90todeliveritslevelofsound performance,filtersareusedinthe DigitalSignalProcessing(DSP). Generally,anaudiofilterisadevice thatchangestheoverallresponseof thetheaudiosignal. Inthecaseof BeoLab90,theseareusedforvarious reasonssuchascontrollingthe relationshipbetweenthedifferent loudspeakerdrivers,actingas crossoverstodistributethecorrect signalstothetweeters,midrangesand woofers,andoptimisingtheoverall
magnituderesponseofthetotal system. Anaudiofilterhasaneffectonthe behaviourofthesignal’smagnitude (howlouditisatagivenfrequency) and/oritsphase(atypeofmeasureof theamountoftimeittakesagiven frequencytogetthroughthefilter). SincetheBeoLab90usesdigital insteadofanaloguefilters,weareable tochoosethecharacteristicsofeach filter’sphaseresponseindependently ofitsmagnituderesponse. For example,afiltercanbeimplemented tohavea“minimumphase”ora “linearphase”(thetwomostcommon responses)characteristic,regardlessof themagnituderesponseitisrequired todeliver. Thephaseresponseofeachfilterin BeoLab90’sprocessingchainhave beenindividuallytailoredaccordingto itsparticularfunction. Forexample, someofthecrossoverfiltershavebeen implementedaslinearphasefilters. MostfiltersintheActiveRoom Compensationalgorithmare implementedasminimumphasefilters (sinceroomresonanceshavea minimumphasecharacteristic). The BeamWidthControlfiltershave customisedphaseresponsesthatare dependentontheparticular frequency-dependendent characteristicsoftheindividual loudspeakerdriversthattheycontrol andarethereforeneitherminimum phasenorlinearphase.
9.3 AutomaticBass Linearisation(ABL)and ThermalProtection
AlmostallloudspeakersintheBang& Olufsenportfolio(includingBeoLab90) featureAutomaticBassLinearisationor “ABL”.Thisisanalgorithmthatwas patentedbyB&Oin1991andis custom-tunedforeachofourproducts. Itspurposeistoensurethat,whenthe physicallimitsofacomponentofthe
35
loudspeakerarereached(forexample, awooferisapproachingitsmaximum excursion,orapoweramplifierisclose toclipping)theloudspeakereither preventsthatlimitfrombeingreached, orthetransitiontothatlimitis “softened”(dependingonthe componentinquestion). Inaddition,BeoLab90’sprocessing continuallymonitorstheindividual temperaturesofmanyinternal componentsincluding:
• Individualloudspeakerdriver magnets • PowerAmplifiermodules • DSPcircuitboards • PowerSupplycircuitboards
Usingthisinformation,combinedwith thepowerthattheamplifiersdeliverto theloudspeakerdrivers,the temperaturesofmanymore componentswithinBeoLab90are calculatedusingcustomisedthermal modelsoftheloudspeaker. Ifthetemperatureofacomponent insidetheloudspeakerapproachesits “thermallimit”(thetemperatureat whichitstopsworkingdueto overheating)thesignalprocessingof theBeoLab90adjuststhesignalsto
protectthecomponent. Theexacttype ofadjustmentdependsonthe particularcomponentthatis approachingitslimits. Asasimple example,ifatweetervoicecoilis calculatedtobeapproachingitslimit, thenitsgainisreducedtoattemptto protectitfromdestruction. Itisimportanttostatethatthisdoes not meanthattheBeoLab90is indestructible–butitdoesmakeit verydifficulttodestroy. Moreinformationcanbefoundin Appendix6: ABL-AdaptiveBass Linearisation.
9.4 ThermalCompression Compensation
BeoLab90’sprocessingincludes automaticcompensationforchangesin loudspeakerdriverresponseasaresult ofinternalchangesintemperature. Formoreanin-depthdiscussionofthis feature,pleasereadAppendix7: ThermalCompressionCompensation.
9.5 Production“Cloning”
EveryBeoLab90thatleavesthe productionlineismeasuredina custom-builtanechoicchamberto ensurethatitsperformancematches themasterreferenceloudspeaker. This automatedmeasurementisperformed using18microphones(oneforeach loudspeakerdriver)wheresmall differencesintheresponsesarefound andcustomcorrectionfiltersare createdandloadedintotheDigital SignalProcessing. Thisensuresthat eachloudspeaker’sthird-octave smoothedresponsematchesthatof themasterreferenceloudspeaker within0.2dBbetween20Hzand20 kHz.1
Figure 9.2: An early prototype in the anechoic chamber at the end of the BeoLab 90 production line where every loudspeaker is measured and calibrated.
1Notethatthesevalueshavenotyetbeenfinalised.
36
TechnicalSpecifications
10.1 TotalSystem
Note: Total System measurements performed with Sound Design set to “Flat on-axis” and Active Room Compensation disabled. FrequencyResponse 25.2Hzto28.1kHz(±1dB,1/3octavesmoothed) FrequencyRange <12Hzto >43kHz(-10dB,ref. 200Hz-2kHz,unsmoothed) Sensitivity seeSection8.1 MaximumSPL XXdBSPL(C)@1m,on-axis SelfNoise(Digitalinput) XXdBSPL(C)@1m,on-axis SelfNoise(Analogueinput) XXdBSPL(C)@1m,on-axis
10.2 PreamplifierandProcessorSection
InordertosimplifycomparisonofBeoLab90’stechnicaldatatootherproducts,theinformationinthischapterhasbeendividedinto threesections:
• Preamplifier and Processor,equivalenttoasurroundprocessor,preamporreceiver • Power Amplifiers • Loudspeaker Drivers
Stereo Preamp
BeoLab 90
Power Amps Loudspeakers
Figure10.1: AblockdiagramoftheBeoLab90showingthecomparativesectionsintermsofcompetingdevices.
10.2.1 OverallSpecifications
Note: Hardware-only measurement. All filters and equalisation bypassed or removed from signal processing for measurements.
DigitalinputtoDACoutputs FrequencyResponse 0Hzto40kHz(+0dB,-1dB) FrequencyRange 0Hzto75kHz(+0dB,-3dB) THD+N 0.004%(997Hz,-1dBFS,22Hz–20kHz) DynamicRange 122dB(A)(997Hz,-60dBFS,20Hz–20kHz,AES17)
AnalogueinputtoDACoutputs FrequencyResponse TBD(+0dB,-1dB) FrequencyRange TBD(+0dB,-3dB) THD+N TBD(997Hz,-1dBFS,22Hz–20kHz) DynamicRange TBD(A)(997Hz,-60dBFS,20Hz–20kHz,AES17)
37
10.2.2 Inputs AnalogueInputs Analogue-to-DigitalConverter NotethatthesameADCmodelisusedforallanalogueinputs. Model TexasInstrumentsPCM4220 SamplingRate 192kHz(fixed) Resolution 24bits FrequencyResponse 10Hz–80kHz(+0dB,-0.2dB) FrequencyRange < 2Hz–85kHz(-3dB) DynamicRange(Typical) 122dB(A)(997Hz,-60dBFS) DynamicRange(Worst-case) 117dB(A)(997Hz,-60dBFS) THD+N 0.001%(997Hz,-1dBFS,22Hz–20kHz) ChannelSeparation 100dB(20Hzto20kHz) PassbandRipple ±0.001dB PowerLink Connector RJ45 InputImpedance 100kΩ AudioChannels 2 MaximumInputVoltage 6.5VRMS Features 5VcontrolvoltageforOn/Standby PowerLinkDatasupport Sensitivity 125mVRMSproduces88dBSPL(1m,on-axis,free-field) XLRLine InputImpedance(Single-ended) 50kΩ (Fixed) InputImpedance(Balanced) 100kΩ (Fixed) MaximumInputVoltage 2.0,4.0,6.5VRMS(Selectable) Features Differentialandimpedancebalanced Pinconfiguration Pin1 Audioground Pin2 Positivesignalinput(“hot”) Pin3 Negativesignalinput(“cold”) NotethattheXLRconnectorcasing(or“shell”)isconnectedtothechassisgroundoftheBeoLab90forshielding. RCALine InputImpedance 50 Ω ,50kΩ (Selectable) MaximumInputVoltage 2.0,4.0,6.5VRMS(Selectable)
DigitalInputs SamplingRateConverter NotethattheSRCisappliedtoalldigitalinputs. Model TexasInstrumentsSRC4392 Outputsamplingrate 192kHz(fixed) Outputwordlength 24bits THD+N 0.000014%(f=997Hz,0dBFS,22Hz–40kHz,unweighted) DynamicRange 138dB(f=997Hz,-60dBFS,22Hz–40kHz,unweighted) PassbandRipple ±0.008dB
38
S/PDIF SupportedFormat LinearPCM Samplingrate Standardsamplingratesupto192kHz Wordlength 24bits OpticalConnector SupportedFormat LinearPCM Samplingrate Standardsamplingratesuptoandincluding96kHz Wordlength 24bits USBAudioConnector SupportedFormat LinearPCM Samplingrate Standardsamplingratesuptoandincluding192kHz Wordlength 24bits WirelessPowerLink SupportedFormat LinearPCM Samplingrate StandardWiSAsamplingratesuptoandincluding96kHz(48kHzstandard) Wordlength 24bits WiSA SupportedFormat LinearPCM Samplingrate StandardWiSAsamplingratesuptoandincluding96kHz Wordlength 24bits
10.2.3 DigitalSignalProcessor
Model AnalogDevicesADSP-21489 Number 2 InstructionRate 400MHz Samplingrate 192kHz(fixed) Notes 32-bitfloatingpoint
10.2.4 DigitaltoAnalogueConverters
NotethatthesespecificationsincludetheanaloguestagesthatfollowtheDACoutputs. Model TexasInstruments/Burr-BrownPCM1798 AudioChannels 18 Samplingrate 192kHz(fixed) Wordlength 24bits FrequencyResponse 0Hzto40kHz(+0dB,-1dB) FrequencyRange 0Hzto75kHz(+0dB,-3dB) THD+N 0.004%(997Hz,-1dBFS,22Hz–20kHz) DynamicRange 122dB(A)(997Hz,-60dBFS,20Hz–20kHz,AES17) ChannelSeparation 110dB(20Hz–20kHz,AES17) LevelLinearity ±1dB(at-120dBFS)
39
10.3 PowerAmplifiers
10.3.1 TweetersandMidrangeSections
Oneamplifierperloudspeakerdriver Model Bang&OlufsenICEpowerAM300-X PeakVoltage 50V PeakCurrent 20A PeakPower 780W(into3.2 Ω) FrequencyRange < 2Hz– > 100kHz(+0dB,-3dB) THD+N 0.02%(20Hz–20kHz,100mW-300W,4 Ω,AES17) Features DualLoop3-ICEpower’sthird-generationClass-Dtopology Custom-modifiedforBeoLab90
10.3.2 WooferSection
Oneamplifierperloudspeakerdriver Model HelioxAM1000-1 PeakVoltage 100V(Software-limited,Hardwarecapableof150V) PeakCurrent 40A PeakPower 3125W(into3.2 Ω) FrequencyRange < 1Hz– > 20kHz(+0dB,-3dB) THD+N 0.05%(20Hz–20kHz,100mW-1000W,4 Ω,AES17) Features UnifiedClass-D(UCD) Custom-modifiedforBeoLab90
10.4 LoudspeakerDrivers
10.4.1 Tweeters
Model Scan-SpeakIlluminatorD3004/602000 Number 7 NominalImpedance 4 Ω EffectiveDiameter 30mm Features Textiledomediaphragm Symmetricaldrive,SD-2motor Non-resonantaluminiumrearchamber
10.4.2 Midranges
Model Scan-SpeakIlluminator12MU/4731T00 Number 7 NominalImpedance 4 Ω EffectiveDiameter 86mm Features Under-hungneodymiummotordesign
40
10.4.3 FrontWoofer
Model Scan-SpeakRevelator32W/4878T00 Number 1 NominalImpedance 4 Ω EffectiveDiameter 260mm Features Papersandwichconewithpatentedfoamfilling Symmetricaldrivemotor
10.4.4 Woofers
Model Scan-SpeakDiscovery26W/4558T00 Number 3 NominalImpedance 4 Ω EffectiveDiameter 212mm Features Anodisedaluminiumcone Fibreglassdustcap
10.5 PowerSupply
PowerConsumption Low-PowerStandby < 0.5W NetworkStandby < 2W Low-levelaudio/Idle Approximately150Wcontinuous SustainedMaxAverage 250W Peak > 18,000W(Duration < 1ms)
10.6 DigitalPowerLink
Technology AudioVideoBridge(AVB) SamplingRate 192kHz(fixed) Bitdepth 24 Features IncludesproprietaryB&Odatachannelsforinter-loudspeakercommunication
41
FAQ
11.1 Multichannelsystem setup
Incaseswhereyouusemorethanone pairofBeoLab90’sinaconfiguration, therearesomerecommendationsthat shouldbefollowedinordertofacilitate dailyuse.
11.1.1 Bang&Olufsen televisionassource
AsdescribedinSection5.11.0.1,a currentBang&Olufsentelevisioncan automaticallyswitchBeoLab90 PresetsaspartoftheSpeakerGroup function. However,itshouldbenoted thatagivenSpeakerGroupinthe televisionsendsonlyoneSpeaker PresetvalueonitsPowerLinkoutputs toallloudspeakersconnectedtothe television. Thismeansthatthepreset identificationnumbersinallBeoLab 90’smustmatchforagiven configurationcorrespondingtoa SpeakerGroupinthetelevision.
11.1.2 Third-partydeviceas source
Whenusingathird-partymultichannel deviceasasourceformorethanone pairofBeoLab90’s,eachMaster-Slave pairofloudspeakersshouldbe configuredcorrectlyforagivensource. Theresultingparametersshouldbe savedtoaPresetthatisthentriggered bytheappropriateinput. SeeSection AutomatingPresetSelectionformore information.
11.2 DoesBeoLab90support DSD?
DSDandDSDoverPCM(DoP)arenot currentlysupportedbyBeoLab90. In ordertoplayDSDaudiofiles,itis thereforenecessarytoconverttoPCM intheaudioplayerbeforesendingthe signaltotheloudspeakers.
Notethat,sincetheBeoLab90audio signalpathcontainsasignificant amountofdigitalsignalprocessing (DSP)whichisperformedonlinearPCM signals,aconversionofDSDtoPCMis requiredsomewhereintheaudio chain. Placingthisconversionprocess aheadoftheloudspeakers’inputs givestheusertheoptiontochoosehis orherpreferredfilterfortheprocess.
11.3 DoesBeoLab90support DXD?
DXDisnotcurrentlysupportedby BeoLab90,sinceitsdigitalinputswill notoperateatsamplingratesabove 216kHz. InordertoplayDXDfilesonthe BeoLab90,theaudiosignalwilleither havetobedownsampledto192kHz (maximum)orconvertedtoanalogue inadvanceofsendingthesignalsto theloudspeakers’inputs.
11.4 WhydoestheBeoLab90 sound“different”whenI switchtowatching television?
SomefeaturesoftheBeoLab90are disabledwhentheyareconnectedto currentBang&Olufsensources. Thisis toensurethatsimilaraudioprocessing isnotperformedtwice. Thereare cases,however,wherealthoughtwo processesaresimilar,theyarenot identical. Forexample,itmaybethe casethatthebassortreble adjustmentsintheBeoLab90donot havethesamefrequencyresponsesas thoseintheaudiosource. Formore informationaboutthis,pleasesee Section13.1. Itmayalsobethecasethatthe adjustmentofsomeofthese processorsaredifferentinthe loudspeakersandthesource. For example,ifthebassisincreasedinthe loudspeakers,andthendisabled
becausethePowerLinkinputis chosen,therewillbearesultant changeintimbreoftheloudspeakers. Theremayalsobeinstanceswherea Bang&Olufsensourceautomatically changesthelatencymodeofthe BeoLab90’sinordertopreservelip syncorsynchronisationwithmultiroom systems. Thiswillalsohavea potentiallyaudibleeffectontheaudio qualityoftheloudspeakers.
42
Setup“TipsandTricks”
SettingSpeakerLevels AlthoughitispossibletosetSpeaker LevelsusinganSPLmeter,thiscan,in somecases,beimprovedbymaking smallfinaladjustments“byear”while listeningtomusicinstead. Whilesitting inthepreferredlocation,playatrack withasolidcentrephantomimage (“Tom’sDiner”bySuzanneVegaisa goodchoice)andadjustoneofthe SpeakerLevelvaluestoplacethe centreimageinthecorrectlocation. Thisshouldbedoneafter theSpeaker Distanceshavebeenadjusted. ARCvs. BeamWidthandBeam Direction TheActiveRoomCompensation algorithmcalculatesacustomfilterfor theBeamWidth,BeamDirection,and LatencyModeofthepreset. When editingapreset,itiswisetoturnoff theARCfilterswhileswitchingbetween differentBeamWidths,Beam Directions,andLatencyModesinorder toavoidwaitingforthiscalculationto beperformed. Whentheappropriate settingshavebeenchosen,thengo backtoyourARCsetupandenablethe filters. iOSsettingsduringsetup Insomecases,particularlywhen makingARCmeasurements,itmaybe helpfultochangetheSettingsofyour iOSdevicetoensurethatitdoesnot “sleep”,sincethiscancauseittolose communicationwiththeBeoLab90’s SpeakerDistancesforlarger listeningareas WhenmeasuringtheSpeaker Distancesforlargerlisteningareas (withmorethanoneperson),itis sometimesbettertomeasurethe distancefromagivenloudspeakerto theclosestlistener. Thisisparticularly trueinmultichannelsystems. Fora“PartyMode”wherethereisno singlelisteningposition(eitherdueto thesizeofthecrowd,orthefactthat thelistenersaremovingthroughout
thespace),itisusuallybettertoset theSpeakerDistancesandSpeaker Levelstothematchingvalue(e.g. 1.0 mand0.0dB). Multichannelsystemswithmore than2BeoLab90’s Incaseswheretwoormorepairsof BeoLab90’sareusedinamultichannel system,thencareshouldbetakento ensurethatthePresetnumbersforthe differentpairsofloudspeakersmatch appropriately. Thisisbecausethe BeoVisiontelevisionsendsonlyone SpeakerPresetvaluetoallPowerLink andWirelessPowerLinkoutputs. BackgroundnoiseduringARC measurements Itisimportanttoensurethatthereis aslittleextraneousnoiseaspossible duringtheARCmeasurement procedure. Thisincludesturningoffair conditioningsystemsorperformingthe measurementsduringlow-traffichours, wherepossible. ARCfiltercalculationtime WhencreatinganARCfilterthatuses themeasurementsfrommorethan ARCZone,itisadvisabletowaituntil thefirstcalculationisdonebefore includingthesecondZone.
43
Troubleshooting
13.1 Somefeaturesinthe BeoLab90controlsare disabled
WhenconnectedtomanyBang& OlufsensourcesviaPowerLinkor WirelessPowerLink,somefeaturesin theBeoLab90maybedisabled. Thisis toavoiderrorssuchasmis-calibration ofthevolumesettingwithother loudspeakersinasurround configurationorduplicationof processing(e.g. turninguptheBass controllertwice: onceinthesource andonceintheloudspeaker).
13.2 LipSyncproblems
WhenusedwithanolderBang& Olufsentelevision(BeoSystem3-based orearlier)orathird-partytelevision, theLatencyModeoftheBeoLab90is notautomaticallycontrolledbythe source. Consequently,theLatency Modeshouldbesetto“Low”toensure synchronisationwiththevideosignal. Thiscanbedonemanuallyusingthe controlinterface(seesection5.10.1), orsetasthedefaultforthepreset triggeredbytheaudioinputconnected tothetelevision.
13.3 Echoproblems
13.3.1 Multiroomaudio systems
WhenaBeoLab90isusedwitha third-partymultiroomsystem,the loudspeaker’sLatencyModeshouldbe setto“Low”inordertoreducethe delaytimeoftheBeoLab90toa minimum. IftheLatencyModeissetto “High”andifitisimpossibletoadjust theexpectedloudspeakerlatencyin themultiroomsystem,thenthe BeoLab90’slatencywillbehigh enoughthattheyappeartoproducean audibleechorelativetoother loudspeakersinthesystem.
13.3.2 SurroundProcessors
WhenaBeoLab90isusedwithan olderBang&Olufsensurround processor(suchastheBeoSystem3or earlierdevices)orathird-party surroundprocessor,theloudspeaker’s LatencyModeshouldbesetto“Low”in ordertoreducethedelaytimeofthe BeoLab90toaminimum. Ifthe LatencyModeissetto“High”andifit isimpossibletoadjusttheexpected loudspeakerlatencyinthesurround processor,thentheBeoLab90’s latencywillbehighenoughthatthey appeartoproduceanaudibleecho relativetootherloudspeakersinthe system. Itmaybepossibleto“trick”some surroundprocessorsintocompensating forBeoLab90’slatencyinHigh LatencyModebyadding34.3m(112.5 feet)totheiractualdistancefromthe listeningposition. Thisvalue correspondstoa100mslatency. WhenusedinLowLatencyMode,8.6 m(28.1feet)shouldbeaddedtothe actualdistancefromthelistening position. Thisvaluecorrespondstoan 25mslatency.
13.4 Loudspeakersdon’tturn onautomatically
IftheBeoLab90’saresettorecognise theWirelessPowerLink/WiSAinput, thenallcabledinputsaredisabled. Ifacabledsourceisnotinthelistof signalssetforauto-detectionas describedinAuto-detection,thenitwill notautomaticallyturnonthe loudspeakers. ItispossiblethattheDetection Threshold,describedinDetection Threshold,issettotoohighavalueto detectthesignal.
13.5 Loudspeakersnever shutoff
13.5.1 Analoguesources
AdjusttheDetectionThresholdhigher asdescribedinDetectionThresholdto ahighervaluetopreventitfrom detectingnoiseontheinputcable.
13.5.2 Digitalsources S/PDIFandOptical Ensurethatthesignalonthedigital connectioneithershutsdown,or transmitsa“digitalblack”signal. The BeoLab90detectsanynon-zerosignal onthesedigitalinputsandwillturnon automaticallyasaresult.
13.6 Theapplicationorweb interfacedonotwork
EnsurethattheDigitalPowerLink cablebetweentheMasterandSlave loudspeakersisconnected. Ensurethattheloudspeakersandthe deviceareconnectedtothesame network.
13.7 Loudspeakersare distortingatlowlevels
TheMaximumInputLevelasdescribed inMaximumInputVoltagemaybeset totoolowavaluetobecompatible withtheaudiosource. Thiscancause theinputoftheBeoLab90toclip.
13.8 Loudspeakersarenoisy/ tooquiet
Ifthesourcehasavariableoutput level,thenthebeststrategyforgain managementistoincreasethe source’soutputleveltomaximumand usethevolumecontroloftheBeoLab 90’s. Thiswillensurethelowest possiblenoiseflooroftheoverall system.
44
Itisalsoimportanttoensurethatthe MaximumInputLevelasdescribedin MaximumInputVoltageissettothe correctvalueforthesourcedevice. Usingtoohighasettingwillresultinan elevatednoisefloor. SeeInputImpedanceonInput Impedance
13.9 USBAudionotworking
TheUSBAudioinputwillonlyaccept PCMsignalsupto192kHz. IfyoursourceisoutputtingDoP(also knownasDSDoverPCM)orPCM signalsathighersamplingrates(e.g. DXDat384kHz)therewillbenoaudio outputfromtheloudspeaker.
13.10 S/PDIFinputnot working
TheS/PDIFinputwillonlyacceptPCM signalsupto192kHz. IfyoursourceisoutputtingDoP(also knownasDSDoverPCM)orPCM signalsathighersamplingrates(e.g. DXDat384kHz)therewillbenoaudio outputfromtheloudspeaker. IfthevolumecontroloftheS/PDIF
inputusingPowerLinkhasbeen enabled,thenitisimportanttosetthe priorityoftheS/PDIFinputhigherthan thatofthePowerLink. Ifthisisnot done,thenthePowerLinksignalwill overridetheS/PDIFinputandthelatter willnotbeheard.
13.11 Opticalinputnot working
NotethattheOpticalinputwillnot acceptsamplingratesabove96kHz duetounreliabilityofanopticaldigital audioconnectionathighersampling rates. IfthevolumecontroloftheOptical inputusingPowerLinkhasbeen enabled,thenitisimportanttosetthe priorityoftheOpticalinputhigherthan thatofthePowerLink. Ifthisisnot done,thenthePowerLinksignalwill overridetheOpticalinputandthe latterwillnotbeheard.
13.12 Automaticswitchingof inputsnotbehavingas expected
IfyouareusingtheAutomaticinput selection,theremaybecaseswhere
theloudspeakerdoesnotbehaveas youwouldintuitivelyexpectduetothe Timeoutparameterofthecurrently selectedsource. Thisisbestexplained bygivingexamples. Takethecasewhereyouareplaying audiofromtwosources,aCDplayer connectedtotheS/PDIFinputanda turntableconnectedtotheXLRinput (viaanRIAApreamplifier),andletus assumethattheS/PDIFinputhasa higherselectionprioritythantheXLR input. Inthiscase,theloudspeakers willplaytheCDsignal. Ifyouthen pressSTOPontheCDplayer,the loudspeakerswillnotswitchtothe signalontheXLRinput(theturntable) untiltheS/PDIFinput’stimeout durationhaspassed. (SeeSection 6.2.6foradetaileddescriptionofthe Time-outparameter.) Thisbehaviourwouldalsobetrueif youwerefirstplayingasignalonthe CDplayer,youpressSTOP,andthen youstartplayingasignalonthe turntable. Again,untiltheS/PDIF input’stime-outdurationhaspassed, thesignalontheXLRinput(fromthe turntable)willnotbeautomatically selectedbytheBeoLab90.
45
Appendix1: RecommendationsforCriticalListening
14.1 Loudspeaker Configuration
TheBeoLab90providesyouwithan extremelywiderangeofparameters thatcanbeusedtoadjustthetimbral andspatialpresentationofyour recordingsforvariouslisteningrooms, loudspeakerplacementsandlistening positions. However,itisalwaysbestto startwithanoptimalconfigurationin yourlisteningroom. Firstconsidertherelationshipbetween theloudspeakersandthelistening positionitself. Thetwoloudspeakers andthelisteningpositionshouldbethe threecornersofanequilateraltriangle. Thismeansthatthedistancefromeach loudspeakertothelisteningposition shouldbethesameasthedistance betweeneachloudspeaker. (Seethe exactpositionformeasurementsin Figure14.1.) Thisalsomeansthatthe loudspeakerswillbe30◦ awayfrom thefront,centrelocation,directlyin frontofthelisteningposition.
Beam Origin Centre for Front
Figure 14.1: The centre of the sound beamintheverticalplaneislevelwitha positionbetweenthebottomtwotweeters and the top midrange driver as shown(ataheightof108.6cmfromthe floor).
Secondly,thetwoloudspeakersshould be“toed-in”by30◦. Thismeansthat theyshouldbeslightlyrotatedsothat theyarebothfacingthelistening position. ThisisshowninFigure14.2.
Figure 14.2: A “perfect” loudspeaker configuration with BeoLab 90’s. Both loudspeakersareaimedatthelistening position. The distance from the listening position to each loudspeaker is the same as the distance between the two loudspeakers.
Ifpossible,theheightofthelistening chairshouldbesetsothatthe listener’searsarelevelwiththecentre oftheverticalbeam,showninFigure 14.1. Thispointis108.6cmabovethe floor. Thenextconsiderationissymmetry withinyourlisteningroom. Itis commonlyrecognisedthatthebest stereoimagingwillbeachievedifthe listeningconfiguration(thetriangle formedwiththelistenerandthetwo loudspeakers)isplacedinleft–right centreoftheroom. Therefore,theside wallswillbothbethesamedistance fromthelisteningposition,andthe loudspeakerswillhavethesame distancetoitsadjacentwalls. Thisisto saythatthedistancefromtheleft loudspeakertotheleftwallisthesame asthedistancebetweentheright loudspeakerandtherightwall. The distancetothefrontwall(behindthe loudspeakers)shouldbethesamefor bothloudspeakers,butcertainlydoes nothavetobethesameasthe distancetothesidewalls.
Figure 14.3: An optimal placement for the loudspeakers with respect to adjacent walls. Note that the distance between each loudspeaker and its closestsidewallareidentical, andthatthe distances from the loudspeakers to the frontwallarealsomatched.
Figure14.4: Anless-optimalplacement fortheloudspeakerswithrespecttoadjacent walls. Note that the distances from the loudspeakers to the front wall arematched,however,thedistancebetweeneachloudspeakeranditsclosest sidewallarenotidentical.
Figure14.5: Anotherexampleofalessoptimal placement for the loudspeakers with respect to adjacent walls. The distances from the loudspeakers to the front wall are matched, however, the right loudspeaker and lacks a side wall nearby.
46
Figure14.6:Anexampleofaworst-case placementofloudspeakerswithrespect tothelisteningroom. Notwodistances betweenaloudspeakerandanadjacent wallmatcheachother.
Itshouldbenotedthattheprimary casualtyofpoorloudspeaker placementinalisteningroomwillbe thespatialrepresentationofyour recordings. Theprecisionandaccuracy ofthestereoimaging,aswellasthe sensationofenvelopmentfromthe recordingwillbeadverselyaffectedby earlyreflectionpatternsthatarenot matchedforthetwoloudspeakers. Thisproblemisminimisedbyusing BeoLab90’snarrowbeamwidth, however,eventhismodecanbenefit fromcorrectloudspeakerplacementin theroom. Finally,itisrecommendable(but certainlynotrequired)thatthe loudspeakersbepositionedaminimum of1mfromtheclosestwalls. The ActiveRoomCompensationalgorithm willcompensateforchangesinthe BeoLab90’stimbralresponsecaused byadjacentboundaries. However, placingtheloudspeakersslightly distantfromreflectivesurfaceswill reducetheseboundaryeffects,and thereforealsoreducetheamountof compensationthatisrequiredbythe ARCfilters.
14.2 ListeningRoom Acoustics
TheBeoLab90hastwofeaturesthat canovercomesomedetrimental effectsofthelisteningroom’s acousticalbehaviour(BeamWidth ControlandActiveRoom
Compensation). However,thesoundof anyloudspeakercanbeoptimisedby improvingtheroom’sacoustics. Oneofthemainacousticalproblemsin listeningroomsisthatofroommodes orresonances. Theseoccurbecause theroombehavesverymuchlikean organpipe,naturally“singing”at specificfrequenciesthatare determinedbythedimensionsofthe room. Withoutcorrectacoustical treatment,theseresonancesare almostunavoidable. Itispreferredto ensurethattheresonancesinthe room’sthreedimensions(length, width,andheight)donotoverlapeach other. Thismeansthatthebetter listeningroomshavecomplex relationshipsbetweenthesethree dimensions. Forexample,a“worst case”foralisteningroomwouldbea cube,whereallthreedimensionsare identical,thusallresonanceshavethe samefrequencies. Anext-worstcaseis onewhereadimensionisamultipleof another,forexample,aroomthatis 9mx6mx3m. Inabestcase,the ratiosoftheroom’sdimensionswould havenon-simplevalues(e.g. 1: 2.16: 2.96–so,asanexample,3mx6.48m x8.88m).1 Asecondissueinmanylisteningrooms isthatofhard,reflectivesurfaces– particularlyinlocationswherethe soundfromtheloudspeakerisdirectly reflectedtothelisteningposition. Therearetwowaystoalleviatethis problem: absorptionanddiffusion. In ordertoabsorbasoundwavesothatit doesnotreflectoffasurface,an absorptivematerialsuchasfibreglass insulationoracousticalfoammustbe placedonthesurface,orinthepath takenbythereflection. Areflectioncan bediffusedbymakingthereflective surfaceirregular. Forexample,placing abookcaseatthepointofreflection willhelpasadiffusorifthebooksare arrangedinrandomheightsand depths. Finally,itiswisetoabsorbthesound wavesthatwouldbereflectedoffthe floor(e.g. withcarpetorarug)and ceiling(usingabsorptiveceilingtiles).
Thiswillalsohelptoreducetheoverall reverberationtimeoftheroom.
14.3 Loudspeakers
For“critical”or“serious”listening sessions,itisrecommendedthatthe upperfabricframeberemovedfrom theloudspeaker’shighspeakersection.
14.4 SourceDevices
Whenconnectinganaudiosourceto theBeoLab90,therearesomebasic, generalrulesthatshouldbefollowedin ordertogettheoptimalperformance fromyoursystem. Notethattheseare generalrules–sothereareexceptions.
• Ifpossible,thesourceshouldbe connectedtotheBeoLab90 usingadigitalaudioconnection. • Ifthesourcedevicehasavolume controlitshouldbedisabledand theBeoLab90’svolumecontrol shouldbeusedinstead • Ifthesourcehastwoanalogue outputs: onevolume-regulated andtheotheratafixedlevel,the fixed-leveloutputshouldbeused • Ifyouareconnectingasource usingaline-levelanalogueinput (RCAorXLR),checkthesource device’sdatasheettofindits maximumoutputlevelandset thevalueappropriatelyonthe BeoLab90(SeeMaximumInput Voltage). Ifthemaximumoutput ofyourdeviceisgreaterthanthe BeoLab90’smaximumpossible setting(6.5VRMS)thenitis recommendablethatthesource device’soutputlevelisreducedif possible,eitherwithinitsown settingsorusinganexternal attenuator. Table14.1andFigure 14.7showthenecessary attenuationtoreducevarious voltagelevelsto6.5VRMS.
1See“Roomdimensionsforsmalllisteningrooms”byDr. TrevorCoxforagoodintroductiontothistopic.
47
MaxOutput Attenuation 7.0VRMS -0.64dB 9.0VRMS -2.83dB 11.0VRMS -4.57dB 13.0VRMS -6.02dB Table 14.1: Examples of minimum attenuation required to externally converttheMaximumOutputLevelsfroma sourcedeviceto6.5VRMSattheinput oftheBeoLab90.
6 8 10 12 14 16 18 20 22 24 −12
−11
−10
−9
−8
−7
−6
−5
−4
−3
−2
−1
0
Maximum Output Voltage (V rms)
Attenuation (dB)
Figure 14.7: The minimum attenuation required to apply to a source with a given Maximum Output level to reduce it to 6.5 V rms in order to prevent clippingtheBeoLab90analogueinputs.
14.5 Cablerecommendations
Thereareinnumerablebeliefsand opinions,bothfoundedandunfounded, regardingcablesusedforconnecting audiodevices. Thefollowingisasmall setofrecommendationsthatarebased oncommonpracticesforwiring professionalaudiosystemssuchasare foundinrecordingandmastering studios. Decisionsregardingthe specificthebrandorconstructionof thecablesusedforconnectingBeoLab 90arelefttothereader’spreferences.
14.5.1 Analoguecables
Inordertoensurethatthenoisefloor ofanaloguesourcesisaslowas possible,thefollowingguidelinesare recommended: Usecableswithgoodshielding(or screening)toreduceRF(Radio Frequency)interferenceontheaudio signalsfromexternalsources. WhenconnectingtheXLRinputtothe XLRoutputofasourcedevice,use twistedpair(preferablybonded-pair)
or“starquad”cablestoensurethebest possiblematchingoflow-frequency magneticinterferencenoiseatthe differentialinput. Thiswillensurethe highestpossiblecommonmode rejectionandlowestnoisefloor. Note, however,thattheuseofstarquad cableconstructiongenerallyhasa higherinherentcapacitancethana twistedpaircableofthesamelength andwillthereforehaveahigherlossof high-frequencysignalsoverlonger cableruns. Avoidgroundloopswhenconnecting audiodevicestoeachother. Inordertoreducemagneticinductance ofinterference(typically50Hzor60 Hz“hum”)frompowercablesonthe audioinputs,itisalsogoodpracticeto physicallyseparatesignalcablesand mainscablesasmuchaspossible. In caseswherethesecablesmustcross eachother,itisrecommendedthat theycrossata90◦ angle. Forathoroughguidetoinstallationof high-endaudioequipment,“Audio SystemsDesignandInstallation”by PhilipGiddingsishighlyrecommended. Althoughthisisbookintendedfor installationofaudiodevicesin recordingandmasteringstudios,the practicesandrecommendations detailedthereinarealsoapplicableto consumer-levelaudioequipment.
14.5.2 Opticalcables
Itisrecommendedthathigh-quality opticalcablesareusedfortheBeoLab 90,particularlyforlongercableruns. Thisisduetothefactthatthereis attenuation(dimmingorlossoflight intensity)oftheopticalsignalonthe plasticorglassfibreinthecable. This attenuationisproportional(indB)to thelengthofthecable. Therefore,in ordertoensurethattheoptical receiverontheBeoLab90hasan adequatesignalatitsinput,thelight attenuationonthecableshouldbe minimisedeitherbyusingshortcables orhigh-qualityopticalfibre. Traditionally,manypeoplehave
claimedthatopticaldigitalsignalsare lessreliablethanelectricalconnections (suchastheAES/EBUandS/PDIF protocols)duetohigherlevelsofjitter causedbythelimitationsoftherise andfalltimeoftheLEDinthe transmitter. TheBeoLab90usesa very-high-qualitysamplingrate converteratitsinputforalldigital signalswhichattenuatesthejitterof incomingsources,therebyreducing thisconcernconsiderably.
14.5.3 S/PDIFcables
WhenconnectingasourcetoBeoLab 90’sS/PDIFinput,itisrecommended thatacablewitha75 Ω impedanceis used. Thiswillensurethatthereareno reflectionsofthesignalonthecable whichmayincreasethelevelofjitterat theinputoftheBeoLab90. Notethat thisrecommendationisparticularly trueforlongercableruns. Itshould, however,bestatedthatthesampling rateconverteratthedigitalinputsof theBeoLab90isveryeffectiveat attenuatingjitterartefactscaused eitherbythesignalsourceorproblems inthecabling.
14.6 ACmainscables
Itishighlyrecommendedthatan additionaldeviceusedtofiltertheAC powerfromthemains(sometimes calledan“audiophilemainsfilter”or “powerpurifier”,forexample)not be usedwiththeBeoLab90. Thisis becausetheinternalpowersupplyof theBeoLab90hasacustom-designed filterthatreducesnoiseonitsAC mainsinput. Thisfilterhasbeen optimisedforthetime-variantcurrent demandsoftheBeoLab90,makinga genericexternalfilterredundant(at best)ordetrimental(atworst)tothe performanceoftheloudspeaker. Similarly,itisunnecessarytousea so-called“exotic”or“audiophile” mainscablefortheBeoLab90.
48
Appendix2: IntroductiontoParametricEqualisers
Almostallsoundsystemsofferbass andtrebleadjustmentsforthesound– thesearebasicallycoarseversionsofa moregeneraltoolcalledanequaliser thatisoftenusedinrecordingstudios. Onceuponatime,ifyoumadea long-distancephonecall,therewasan actualphysicalconnectionmade betweenthewirerunningoutofyour telephoneandthetelephoneatthe otherendoftheline. Thiscausedabig probleminsignalqualitybecausealot ofhigh-frequencycomponentsofthe signalwouldgetattenuatedalongthe wayduetolossesinthewiring. Consequently,boostercircuitswere madetohelpmaketherelativelevels ofthevariousfrequenciesmoreequal. Asaresult,thesecircuitsbecame knownasequalisers. Nowadays,of course,wedon’tneedtouse equaliserstofixthequalityof long-distancephonecalls(mostly becausethecommunicationpathsuse digitalencodinginsteadofanalogue transmission),butwedousethemto customisetherelativebalanceof variousfrequenciesinanaudiosignal. Thishappensmostofteninarecording studio,butequaliserscanbeagreat personalisationtoolinaplayback systeminthehome. Thetwomainreasonsforusing equalisationinaplaybacksystemsuch astheBeoLab90’sarepersonal preferenceandcompensationforthe effectsofthelisteningroom’s acousticalbehaviour. Equalisersaretypicallycomprisedofa collectionoffilters,eachofwhichhas upto4“handles”or“parameters”that canbemanipulatedbytheuser. These parametersare
• FilterType • Gain • CentreFrequency • Q
15.1 FilterType
TheFilter Typewillletyoudecidethe relativelevelsofsignalsatfrequencies withinthebandthatyou’reaffecting. Althoughthereareupto7different typesoffiltersthatcanbefoundin professionalparametricequalisers,the BeoLab90containsthethree most-usedofthese:
• Low-shelvingFilter • High-shelvingFilter • PeakingFilter
15.1.1 Low-shelvingFilter
Intheory,alow-shelving Filter affects gainofallfrequenciesbelowthecentre frequencybythesameamount. In reality,thereisabandaroundthe centrefrequencywherethefilter transitionsbetweenagainof0dB(no changeinthesignal)andthegainof theaffectedfrequencyband.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure 15.1: Example of a low-shelving filter with a positive gain. Frequencies below approximately 80 Hz have been affected.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure 15.2: Example of a low-shelving filterwithanegativegain. Frequencies below approximately 80 Hz have been affected.
Notethatthelow-shelvingfiltersused intheBeoLab90definethecentre frequencyasbeingthefrequency wherethegainisonehalfthe maximum(orminimum)gainofthe filter. Forexample,inFigure15.1,the gainofthefilteris6dB.Thecentre frequencyisthefrequencywherethe gainisone-halfthisvalueor3dB, whichcanbefoundat80Hz. Somecareshouldbetakenwhenusing low-shelvingfilterssincetheiraffected frequencybandsextendto0HzorDC. Thiscancauseasystemtobepushed beyonditslimitsinextremelylow frequencybandsthatareoflittle-to-no consequencetotheaudiosignal. Note, however,thatthisislessofaconcern fortheBeoLab90,sinceitisprotected againstsuchabuse.
15.1.2 High-shelvingFilter
Intheory,ahigh-shelving Filter affects gainofallfrequenciesabovethe centrefrequencybythesameamount. Inreality,thereisabandaroundthe centrefrequencywherethefilter transitionsbetweenagainof0dB(no changeinthesignal)andthegainof theaffectedfrequencyband.
49
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure15.3: Exampleofahigh-shelving filter with a positive gain. Frequencies above approximately 8 kHz have been affected.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure15.4: Exampleofahigh-shelving filterwithanegativegain. Frequencies above approximately 8 kHz have been affected.
Notethatthehigh-shelvingfiltersused intheBeoLab90definethecentre frequencyasbeingthefrequency wherethegainisonehalfthe maximum(orminimum)gainofthe filter. Forexample,inFigure15.4,the gainofthefilteris-6dB.Thecentre frequencyisthefrequencywherethe gainisone-halfthisvalueor-3dB, whichcanbefoundat8kHz. Somecareshouldbetakenwhenusing high-shelvingfilterssincetheir affectedfrequencybandscanextend beyondtheaudiblefrequencyrange. Thiscancauseasystemtobepushed beyonditslimitsinextremelyhigh frequencybandsthatareoflittle-to-no consequencetotheaudiosignal.
15.1.3 PeakingFilter
Apeaking filter isusedforamorelocal adjustmentofafrequencyband. Inthis case,thecentrefrequencyofthefilter
isaffectedmost(itwillhavetheGain ofthefilterappliedtoit)andadjacent frequenciesoneithersideareaffected lessandlessasyoumovefurther away. Forexample,Figure15.5shows theresponseofapeakingfilterwitha centrefrequencyof1kHzandgainsof 6dB(theblackcurve)and-6dB(the redcurve). Ascanbeseenthere,the maximumeffecthappensat1kHzand frequencybandstoeithersideare affectedless.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure15.5: Exampleoftwopeakingfilters. Theblackcurveshowsafilterwith apositivegain,theredcurveshowsthe reciprocal with a negative gain. The centrefrequencyofthisfilteris1kHz.
YoumaynoticeinFigure15.5thatthe blackandredcurvesaresymmetrical– inotherwords,theyareidentical exceptinpolarityofthegain. Thisisa particulartypeofpeakingfiltercalleda reciprocal peak/dip filter –so-called becausethesetwofilters,placedin series,canbeusedtocanceleach other’seffectsonthesignal. NotethatBeoLab90usesreciprocal peak/dipfilters.
15.2 Gain
Ifyouneedtomakeallfrequenciesin youraudiosignallouder,thenyoujust needtoincreasethevolume. However, ifyouwanttobealittlemoreselective andmakesomefrequencybands louder(orquieter)andleaveother bandsunchanged,thenyou’llneedan equaliser. So,oneoftheimportant questionstoaskis“howmuchlouder?” or“howmuchquieter?” Theanswerto
thisquestionisthegainofthefilter– thisistheamountbywhichissignalis increasedordecreasedinlevel. Thegainofanequaliserfilterisalmost alwaysgivenindecibelsordB1.Thisis ascalebasedonlogarithmicchanges inlevel. Luckily,it’snotnecessaryto understandlogarithmsinordertohave anintuitivefeelfordecibels. Thereare reallyjustthreethingstoremember:
• againof0dBisthesameas saying“nochange” • positivedecibelvaluesare louder,negativedecibelvalues arequieter • Addingapproximately6dBtothe gainisthesameassaying“two timesthelevel”. (Therefore, subtracting6dBishalfthelevel.)
15.3 CentreFrequency
So,thenextquestiontoansweris “whichfrequencybandsdoyouwant toaffect?” Thisispartiallydefinedby thecentre frequencyorFcofthefilter. Thisisavaluethatismeasuredinthe numberofcyclespersecond2,labelled Hertz orHz. Generally,ifyouwanttoincrease(or reduce)thelevelofthebass,thenyou shouldsetthecentrefrequencytoa lowvalue(roughlyspeaking,below 125Hz). Ifyouwanttochangethe levelofthehighfrequencies,thenyou shouldsetthecentrefrequencytoa highvalue(say,above8kHz).
15.4 Q
Inalloftheabovefiltertypes,there aretransitionbands–frequencyareas wherethefilter’sgainischangingfrom 0dBtothedesiredgain. Changingthe filter’sQ3 allowsyoutoaltertheshape ofthistransition. ThelowertheQ,the smootherthetransition. Inboththe caseoftheshelvingfiltersandthe peakingfilter,thismeansthatawider
1The“B”isacapitalbecauseit’snamedafterAlexanderGrahamBell. 2Thisisliterallythenumberoftimesaloudspeakerdriverwillmoveinandoutoftheloudspeakercabinetpersecond. 3Notethat,althoughtheterm“Q”isusedthroughoutthismanualandtheBeoLab90interfaceforbothpeakingandshelvingfilters,thisisincorrect. Tobetechnicallycorrect, theterm“S”(orshelfslope)shouldbeusedforshelvingfilters.
50
bandoffrequencieswillbeaffected. Thiscanbeseenintheexamplesin Figures15.6and15.7.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure 15.6: Example of two lowshelving filters. The black curve shows a filter with a Q of 0.4, the red curve showstheafilterwithaQof1. Forboth filters, the centre frequency is 100 Hz andthegainis+6dB.
10 100 1,000 10,000
−6
−4
−2
0
2
4
6
Frequency (Hz)
Gain (dB)
Figure15.7: Exampleoftwopeakingfilters. Theblackcurveshowsafilterwith a Q of 0.5, the red curve shows the a filter with a Q of 8. For both filters, the centre frequency is 1 kHz and the gain is+6dB.
ItshouldbeexplainedthattheQ parametercancauseashelvingfilter tobehaveslightlystrangely. Whenthe Qofashelvingfilterexceedsavalueof 0.707,thegainofthefilterwill “overshoot”itslimits. Forexample,as canbeseeninFigure15.8,afilterwith againof6dBandaQof4willactually haveagainofalmost13dBandwill attenuatebyalmost7dB. Thisover-andundershootingofthe filter’smagnituderesponseisthe reasontheQofthehigh-shelvingand low-shelvingfiltersintheBeoLab90’s parametricequaliserhavebeenlimited toamaximumvalueof1. Notethat,whentheQissettoavalue of1,thentheresultingovershootof thefilter,indecibels,isapproximately 7.4%higherthanthestatedgainofthe
filter. Forexample,ifGain=6dBand Q=1,thenthemaximumactualgain ofthefilterwillbe6*1.074=6.44dB. IfGain=3dBandQ=1,thenthe maximumactualgainofthefilterwill be3*1.074=3.22dB.
10 100 1,000 10,000
−10
−5
0
5
10
Frequency (Hz)
Gain (dB)
Figure15.8:Exampleoflow-shelvingfilterswithaQofmorethan1. Theblack curve shows a filter with a Q of 0.7 for reference, the red curves shows filters with Q’s of 1, 2, and 4. The centre frequency of this filter is 100 Hz and the gain is +6 dB. Note that some of these values are not possible in the parametric equaliser in BeoLab 90.
51
Appendix3: TheInfluenceofListeningRoomAcousticsonLoudspeakers
Aroomcomprisedoflargeflat reflectivesurfaceswithlittleacoustical absorptionhasaverydifferent acousticalbehaviourfromarecording ormasteringstudiowherethefinal decisionsaboutvariousaspectsofa recordingaremade. Consequently, thismusthaveaneffectonalistener’s perceptionofarecordingplayed throughapair(assumingstereo reproduction)ofloudspeakersinthat room. Theinitialquestiontobeasked is“what,exactly,aretheexpected effectsoftheroom’sacoustical behaviourinsuchacase?” Thesecond is“iftheroomhastoomuchofan effect,howcanIimprovethesituation (e.g. byaddingabsorptionorchanging thephysicalconfigurationofthe systemintheroom)?” Thethird,and possiblyfinalquestionis“howcana loudspeakercompensate(oratleast account)fortheseeffects?” Theeffectaroom’sacoustical behaviourhasonaloudspeaker’s soundcan,atasimplelevel,be consideredunderthreegeneral headings:
• EarlyReflections • RoomModes • Reverberation
16.1 EarlyReflections
Earlyreflections,fromsidewallsand thefloorandceiling,haveaninfluence onboththetimbre(tonecolour)and thespatialcharacteristicsofastereo reproductionsystem. Wewillonly discussthetimbraleffectsinthis article.
Figure16.1: Thesoundarrivingatalistenerfromaloudspeakerinaroomwith only one wall. Note that the sound arrives from two directions – the first is directly from the loudspeaker (in red). Thesecondisa“firstreflection”offthe wall(inblue).
Let’sstartbyassumingthatyouhave aloudspeakerthathasamagnitude responsethatisperfectlyflat–atleast from20Hzto20kHz. Wewillalso assumethatithasthatresponse regardlessofwhichdirectionyou measureitin–inotherwords,it’sa perfectlyomnidirectionalloudspeaker. Thequestionis,“whateffectdoesthe wallreflectionhaveonthemeasured responseoftheloudspeaker?” Verygenerallyspeaking,theansweris thatyouwillgetahigherlevelatsome frequencies(becausethedirectsound andthereflectionaddconstructively andreinforceeachother)andyouwill getalowerlevelatotherfrequencies (becausethedirectsoundandthe reflectionworkagainsteachotherand “canceleachotherout”). Whatis potentiallyinterestingisthatthe frequenciesthataddandthe frequenciesthatcancelalternateas yougoupthefrequencyrange. Sothe totalresultlookslikeacomb(asina combthatyouusetocombyourhair, if,unlikeme,youhavehairtocomb). Forexample,takealookatFigure16.2.
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.2: Distancetoloudspeaker= 2m. Distancetowall=1m. Wallisperfectly reflective and the loudspeaker is perfectly omnidirectional. The red line isthemagnituderesponseofthedirect sound. The blue line is the magnitude response of the reflected sound. The blacklineisthemagnituderesponseof thecombination.
Youcanseethat,attheverylowend, thereflectionbooststhelevelofthe loudspeakerbyaapproximately5dB (oralmosttwotimesthelevel)atthe listeningposition. However,asyougo upinfrequency,thetotalleveldropsto about15dBlessbeforeitstartsrising again. Asyougoupinfrequency,the levelgoesupanddown. This alternationactuallyhappensata regularfrequencyspacing(e.g. anotch atmultiplesof200Hz)butitdoesn’t lookregularbecausetheX-axisofthe plotislogarithmic(whichbetter representshowweheardifferencesin frequency). Whathappensifwemovethewall furtheraway? Well,twothingswill happen. Thefirstisthatthereflection willbequieter,sothepeaksand notcheswon’tbeaspronounced. The secondisthatthespacingofthepeaks andnotchesinfrequencywillget closertogether. Inotherwords,the effectstartsatalowerfrequency. Forexample,takealookatFigure16.3.
52
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.3: Distancetoloudspeaker= 2m. Distancetowall=3m. Wallisperfectly reflective and the loudspeaker is perfectly omnidirectional. The red line isthemagnituderesponseofthedirect sound. The blue line is the magnitude response of the reflected sound. The blacklineisthemagnituderesponseof thecombination.
Conversely,ifwemovethewallcloser, wedotheopposite(theproblemgets worse,butatahigherfrequency),as canbeseeninFigure16.4.
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.4: Distancetoloudspeaker= 2m. Distancetowall=0.25m. Wallis perfectlyreflectiveandtheloudspeaker is perfectly omnidirectional. The red line is the magnitude response of the direct sound. The blue line is the magnituderesponseofthereflectedsound. The black line is the magnitude responseofthecombination.
So,ifyouhavearoomwithonlyone wallwhichisperfectlyreflective,and youhaveaperfectlyomnidirectional loudspeaker,thenyoucanseethat yourbestoptionistoeitherputthe loudspeaker(andyourself)veryfaror veryclosetothewall. Thatwaythe artefactscausedbythereflectionare eithertooquiettodoanydamage,or haveaneffectthatstartsattoohigha frequencyforyoutocare. Thenagain, mostroomhavemorethanonewall, thewallsarenotperfectlyreflective, andtheloudspeakerisnotperfectly omnidirectional.
So,whathappensinthecasewhere theloudspeakerismoredirectionalor youhavesomeabsorption(better knownas“fuzzystuff”)onyourwalls? Well,eitherofthesecaseswillhave basicallythesameeffectinmostcases sinceloudspeakersaretypicallymore directionalathighfrequencies–soyou getlesshighenddirectedtowardsthe wall. Alternatively,fuzzystufftendsto soakuphighfrequencies. So,ineither ofthesetwocases,you’llgetlesshigh endinthereflection. Let’ssimulate thisbyputtingalowpassfilteronthe reflection,asshowninFigure16.5, 16.6and16.7whichhaveidentical distancesasthesimulationsinFigures 16.2,16.3,and16.4–forcomparison.
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.5: Distancetoloudspeaker= 2m. Distancetowall=1m. Wallisabsorptiveand/ortheloudspeakerisdirectional at high frequencies. The red line isthemagnituderesponseofthedirect sound. The blue line is the magnitude response of the reflected sound. The blacklineisthemagnituderesponseof thecombination.
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.6: Distancetoloudspeaker= 2 m. Distance to wall = 3 m. Wall is absorptive and/or the loudspeakers is directional at high frequencies. The red line is the magnitude response of the direct sound. The blue line is the magnitude response of the reflected sound. The black line is the magnitude responseofthecombination.
10 100 1,000 10,000 −20
−15
−10
−5
0
Frequency (Hz)
Gain (dB)
Figure16.7: Distancetoloudspeaker= 2 m. Distance to wall = 0.25 m. Wall is absorptive and/or the loudspeaker is directionalathighfrequencies. Thered lineisthemagnituderesponseofthedirect sound. The blue line is the magnituderesponseofthereflectedsound. The black line is the magnitude responseofthecombination.
Whatyoucanseeinallthreeofthe previousplotsisthat,asthehigh frequencycontentofthereflection disappears,thereislessandlesseffect onthetotal. Thebottomplotis basicallyaproofoftheage-oldruleof thumbthatsaysthat,ifyouputa loudspeakernexttoawall,you’llget morebassthanifit’sfartherfromthe wall. Sincethereisnotmuchhigh frequencyenergyradiatedfromthe rearofmostloudspeakers,Figure16.7 isaprettygoodgeneralrepresentation ofwhathappenswhenaloudspeakeris placedclosetoawall. Ofcourse,the exactbehaviourofthedirectivityofthe loudspeakerwillbedifferent–butthe generalshapeofthetotalcurvewillbe prettysimilartowhatyouseethere. So,theendconclusionofallofthisis that,inordertoreduceundesirable artefactscausedbyawallreflection, youcandoanycombinationofthe following:
• movetheloudspeakerveryclose tothewall • movetheloudspeakerfarther frontthewall • sitveryclosetothewall • sitfartherawayfromthewall • putabsorptiononthewall
However,thereisoneinteresting effectthatsitsontopofallofthis– thatisthefactthatwhatyou’llseeina
53
measurementwithamicrophoneisnot necessarilyrepresentativeofwhat you’llhear. Thisisbecausea microphonedoesnothavetwoears. Also,thedirectionthereflectioncomes fromwillchangehowyouperceiveit. A sidewallreflectionsoundsdifferent fromafloorreflection. Thisisbecause youhavetwoears–oneoneachside ofyourhead. Yourbrainusesthe sidewallreflections(or,moreprecisely, howtheyrelatetothedirectsound)to determine,inpart,howfarawaya soundsourceis. Also,since,inthe caseofsidewallreflections,yourtwo earsgettwodifferentdelaytimeson thereflection(usually),yougettwo differentcomb-filterpatterns,where thepeaksinoneearcanbeusedtofill inthenotchesintheotherearandvice versa. Whenthereflectioncomesfrom thefloororceiling,yourtwoearsget thesameartefacts(sinceyourtwo earsarethesamedistancetothefloor, probably). Consequently,it’seasily noticeable(andit’sbeenprovenusing science!) thatafloororceiling reflectionhasabiggertimbraleffect onaloudspeakerthanalateral(or sideways)reflection.
16.2 RoomModes
Roommodesareacompletelydifferent beast–althoughtheyexistbecauseof reflections. Ifyoupluckaguitarstring, youmakeadeflectioninthestringthat movesoutwardsuntilithitstheendsof thestring. Itthenbouncesbackdown thestring,bouncesagain,etc. etc. As thewavebouncesbackandforth,it settlesintoatotalresultwhereitlooks likethestringisjustbouncingupand downlikeaskippingrope. Thelonger thestring,thelowerthenote,because ittakeslongerforthewavetobounce backandforthonthestring. Youcan alsolowerthenotebyloweringthe tensionofthestring,sincethiswillslow downthespeedofthewavemoving backandforthonit. Thelastwayto lowerthenoteistomakethestring heavier(e.g. bymakingitthicker)– sinceaheavierstringisharderto
move,thewavemovessloweronit. Theairinapipebehavesexactlythe sameway. Ifyou“pluck”theairinthe middleofapipe(say,byclappingour hands,orcoughing,ormakingany noiseatall)thenthesoundwave travelsalongthepipeuntilithitsthe end. Whethertheendofthepipeis cappedornot,thewavewillbounce backandtravelbackthroughthepipe intheoppositedirectionfromwhence itcame.1 Asthewavebouncesback andforthoffhetwoendsofthepipe,it alsosettlesdown(justliketheguitar string)intosomethingcalleda “standingwave”. Thisisthepipe’s equivalentoftheskippingrope behaviourinthestring. Theresultis thatthepipewill“resonate”orringat anote. Thelongerthepipe,thelower thenotebecausethespeedofthe soundwavemovinginairinthepipe staysthesame,butthelongerthe pipe,thelongerittakesforthewaveto bouncebackandforth. Thisisbasically howallwoodwindinstrumentswork. What’sinterestingisthat,intermsof resonance,aroomisbasicallyabig pipe. Ifyou“pluck”theairintheroom (say,bymakingsoundwitha loudspeaker)thesoundwavewillmove downtheroom,bounceoffthewall,go backthroughtheroom,bounceofthe oppositewall,etc. etc. (Ofcourse, otherthingsarehappening,butwe’ll ignorethose.) Thiseffectismost obviousonagraphbyputtingsome soundinaroomandstopping suddenly. Insteadofactuallystopping, youcanseetheroom“ringing” (exactlyinthesamewaythatabell ringswhenit’sbeenhit)atafrequency thatgraduallydecaysastimegoesby. However,it’simportanttoremember thatthisringingisalwayshappening– evenwhilethesoundisplaying. So,for example,akickdrum“thump”comes outofthespeakerwhich“plucks”the roommodeanditrings,whilethe musiccontinueson.
Time (sec)
Figure 16.8: The concept of the effect ofaroommodeandActiveRoomCompensation. See the associated text for anexplanation.
Figure16.8showstheconceptofthe effectofaroommodeandhowit’s dealtwithbyActiveRoom Compensation. Thesoundcomingout oftheloudspeakerisshownonthetop plot,inblack. Theresponseofthe loudspeakerandasingleroommodeis shownbelow,inred. Youcanseethere thattheroommodekeeps“ringing”at onefrequencyafterthesoundfromthe loudspeakerstops. Therearetwoaudibleeffectsofthis. Thefirstisthat,ifyourmusiccontains thefrequencythattheroomwantsto resonateat,thenthatnotewillsound louder. Whenyouhearpeopletalkof “unevenbass”ora“one-note-bass” effect,oneofthefirstsuspectsto blameisaprominentroommode. Thesecondisthat,sincethemodeis ringingalongwiththemusic,the overalleffectwillbemuddiness. Thisis particularlytruewhenonebassnote causestheroommodetostartringing, andthiscontinueswhenthenextbass noteisplaying. Forexample,ifyour roomroomringsonaC#,andthebass playsaC#followedbyaD–thenthe roomwillcontinuetoatC#,conflicting withtheDandresultingin“mud”. This isalsotrueifthekickdrumtriggersthe roommode,soyouhaveakickdrum “plucking”theroomringingonaC#all throughthetrack. Ifthetuneisinthe keyofF,thenthiswillnotbepretty.2 Inorderfortheloudspeakerto compensatefortheeffectoftheroom mode,ithastonotonlyproducethe 1Whetherthepipeisclosed(capped)oropenonlydeterminesthecharacteristicofthereflection–therewillbeareflectioneitherway. 2Doasearchfor“tritone”or“diabolusinmusica”.
54
signalitshould(showninblack)butit mustalsoproduceasignalthat counter-actstheringingintheroom mode. Thisisshowninthelowerplot inblue. Ascanbeseenthere(most easilyintheringingafterthesignal hasstopped),theloudspeaker’s compensationsignal(thebluecurve)is themirrorimageoftheroom’s “misbehaviour”(inred). Ifyouadd thesetwocurvestogether,theresultis thattheycanceleachotherout,and theresultistheblackcurve. Ifyouwouldliketocalculatea predictionofwhereyou’llhavea problemwitharoommode,youcan usethefollowingequation:
• metricversion: frequency=172/(lengthinm) • imperialversion: frequency=558/(lengthinfeet)
Thiscalculationwillproducethe fundamentalfrequencyoftheroom modeinHzforthedimensionofthe roomrepresentedby“length”. Your mostaudiblemodalproblemswillbeat thefrequenciescalculatedusingeither oftheequationsabove,andmultiples ofthem(e.g. 2timestheresult,3 timestheresult,andsoon). So,forexample,ifyourroomis5m wide,yourworst-casemodes(forthe room’swidth)willbeat172/5=34.4 Hz,aswellas68.8Hz,103.2Hzandso on. Rememberthatthesearejust predictions–butthey’llcomepretty close. Youshouldalsorememberthat thisassumesthatyouhavecompletely immovablewallsandnoabsorption–if thisisnottrue,thentheseverityofthe actualproblemwillvaryaccordingly. Sadly,thereisnotmuchyoucando aboutroommodes. Therearewaysto managethem,including,butnot exclusivetothefollowingstrategies:
• makesurethatthethree dimensionsofyourlistening roomarenotrelatedtoeach otherwithsimpleratios • putupmembraneabsorbersor slotabsorbersthataretunedto
themodalfrequencies • placeyourloudspeakerinanode –alocationinaroomwhereit doesnotcoupletoaproblematic mode(however,notethatone mode’snodeisanothermode’s antinode) • sitinanode–alocationina roomwhereyoudonotcoupleto aproblematicmode(seewarning above) • useroomcorrectionDSPsoftware suchasARCintheBeoLab90
16.3 Reverberation
Reverberationiswhatyouhearwhen youclapyourhandsinabigcathedral. It’sthecollectionofalotofreflections bouncingfromeverywhereasyougo throughtime. Whenyoufirstclapyour hands,yougetacoupleofreflections thatcomeinseparatedenoughintime thattheygettheirownlabel–“early reflections”. Afterthat,thereareso manyreflectionscomingfromsomany directions,andsodenselypacked togetherintime,thatwecan’t separatethem,sowejustcallthem “reverberation”or“reverb”(although you’lloftenhearpeoplecallit“echo” whichisthewrongwordtouseforthis. Reverbiswhatyougetwhenyouhave alotofreflectivesurfacesinyourroom –butsinceit’ssoirregularintimeand space,itjustmakesawashofsound ratherthanaweirdcomb-filtereffect likewesawwithasinglereflection. So, althoughitmakesthings“cloudy”–it’s morelikehavingafogonyourglasses insteadofascratch,orasoft-focus effectonakitschyphotographofa fieldofflowers.
16.4 Solutions
Aswe’veseen,ifyourlisteningroomis normal,youhaveatleastthesethree basicacousticproblemstodealwith. Eachproblemhasadifferentsolution... Thefirstsolutionhasalreadybeen startedforyou. Asisexplainedinthe
sectiononSoundDesign,thefinal tuningofeveryBang&Olufsen loudspeaker(includingtheBeoLab90) isvoicedinatleastfourroomswith verydifferentacousticalbehaviours rangingfromavery“dead”livingroom withlotsofabsorptiveanddiffusive surfacestoalargerandvery“live” spacewithaminimalisticdecorating, andlargeflatsurfaces. Oncewehave asinglesounddesignthatisbasedon thecommonelementsthoserooms, wetesttheloudspeakersinmore roomstoensurethatthey’llbehave wellunderallconditions. ThesecondsolutionisBeoLab90’s ActiveRoomCompensationwhichwill correcttheeffectsofboundaries (walls)androommodesonthetimbre oftheloudspeakeratthelistening position(s). Usingmeasurementsofthe characteristicsoftheloudspeakerat thelisteningpositions,theARC algorithmthencreatesafilterthatis usedto“undo”theseeffects. For example,iftheloudspeakeriscloseto awall(whichwillgenerallyresultina boostedbass)thenthefilterwill reducethebasssymmetrically. Similarly,ringingcausedbyroom modeswillbeactivelycancelledby bothBeoLab90’s. Thatway,thelossin thefilterandthegainduetotheroom willcanceleachother. Thethirdsolutionisuniquetothe BeoLab90–BeamWidthControl. This allowsyoutocustomisetherelative levelsofthedirectsoundandthe reflectedsoundatthelistening position. Theresultofthisisthat,even ifyouhaveacousticallyreflectiveside walls,theBeoLab90canstilldeliver anaccurateandpreciserepresentation ofthespatialpresentationofyour stereorecordings.
16.5 Conclusions
Ofcourse,thissectiondoesnotcover everythingthereistoknowaboutroom acoustics. And,ofcourse,youcan’t expectaloudspeakertosoundexactly thesameineveryroom. Ifthatwere true,therewouldbenosuchthingasa
55
“good”concerthall. Aroom’s acousticalbehaviouraffectsthesound ofallsoundsourcesintheroom. On theotherhand,humansalsohavean amazingabilitytoadapt–inother wordsyou“getusedto”the characteristicsofyourlisteningroom. However,thereisnodebatethat,due tomanyissues(thefirsttwothatcome tomindarefrequencyrangeand directivity)twodifferentloudspeakers willbehavedifferentlyfromeachother intwodifferentrooms. Inotherwords, ifyoulistentoloudspeaker“A”and loudspeaker“B”inashowroomofa shop,youmightpreferloudspeaker“A” –butifyoutookthemhome,youmight preferloudspeaker“B”.Thiswouldnot besurprising,sincewhatyouhearis notonlytheloudspeakerbutthe loudspeaker“filtered”bythelistening room. Thisisexactlywhy,evenwith automatedroomcompensation algorithms,somefinetuningmaybe necessarytoachieveasoundthatbest suitsyourroomandyourtastes.
56
Appendix4: LoudspeakerDirectivityandDistancePerceptioninStereoImaging
17.1 DistancePerceptionin RealLife
Gotothemiddleofasnow-covered frozenlakewithaloudspeakeranda friend. Sitthereandcloseyoureyes andgetyourfriendtoplacethe loudspeakersomedistancefromyou. Keepyoureyesclosed,playsome soundsoutoftheloudspeakerandtry toestimatehowfarawayitis. Youwill bewrong(unlessyou’reVERYlucky). Why? It’sbecause,inreallifewithreal sourcesinrealspaces,distance information(inotherwords,the informationthattellsyouhowfaraway asoundsourceis)comesmainlyfrom therelationshipbetweenthedirect soundandtheearlyreflectionsfrom wallsinyourlisteningroom. Ifyou don’thaveanyearlyreflections,then youdon’thaveanydistance information. Addtheearlyreflections andyoucanveryeasilytellhowfar awayitis.1
17.2 DistancePerceptionina StereoRecording
Recordingengineershaveabasictrick forcontrollingtheapparentdistanceto asoundsourceinastereorecording usingtheso-called“dry-to-wet”ratio —inotherwords,therelativelevelsof thedirectsoundandthereverberation. Tobehonest,thisisabitofan over-simplification,butit’satthelevel ofknowledgeonewouldtypicallyhave ifonewerejuststartingoutrecordinga buddingrockbandinagarage. Manyclassicalrecordingsaremade withapairofmicrophones. An instrumentthatisontheleftsideof thepairwillproduceasoundthatis slightlylouderorslightlyearlierinthe leftmicrophonethanintheright microphone. Thismeansthat,when yousitinthesweetspotandlistento thestereorecording,youwillhearthat sourceontheleftsideofthestereo
image. Thiseffectistruenotonlyfor thedirectsoundoftheinstruments arrivingatthemicrophonepair,but alsofortheacousticreflectionsoffthe varioussurfacesintherecording space. So,iftherecordingengineer hasbeenpayingattention,the distanceinformation(therelationship betweenthedirectsoundandthe reflections)hasbeencapturedinthe recording. Thismeansthatwhenyou listentotherecording,younotonly cantellwheretheinstrumentsare fromlefttoright,butalsotheirrelative distances.
17.3 CombiningtheTwo
So,weknowthatearlyreflectionstell yourbrainhowfarawaythesound sourceis. Nowthinktoaloudspeaker inalisteningroom: Case1: Ifyouhavealisteningroom thathasnosidewalls,thenthereare noearlyreflections,and,regardlessof howfarawaytheloudspeakersare,a soundsourceintherecordingwithout earlyreflections(e.g. aclose-mic’ed vocal)willsoundclosertoyouthanthe loudspeakers. Case2: Ifyouhavealisteningroom withearlyreflections,andthe loudspeakersarelessdirectionalsuch asBeoLab90’swiththeirBeamWidth settoWideorOmni,thentheearly reflectionsfromthesidewallstellyou howfarawaytheloudspeakersare. Therefore,theclose-mic’edvocaltrack fromCase1cannotsoundanycloser thantheloudspeakers–yourbrainis toosmarttobetoldotherwise. Case3: Ifyouhavealisteningroom withsidewallsandthereforeearly reflections,buttheloudspeakersare directionalsuchthatthereisnoenergy beingdeliveredtothesidewalls,then theresultisthesameasinCase1. Thistimetherearenoearlyreflections becauseofloudspeakerdirectivity
insteadofwallabsorption,butthe effectatthelisteningpositionisthe same. ThisisthecasewithBeoLab90 whenitsBeamWidthissettoNarrow. Theconclusionisthat,inordertoget anaccurateandpreciserepresentation ofthespatialpropertiesinastereo recording,youshouldtrytominimise thelevelsoftheearlyreflectionsfrom thesidewallsinyourlisteningroom. However,thismeansthatyouare optimisingthesoundforthesweet spot–on-axistobothloudspeakers. Whenlisteningwithfriends,itmaybe necessarytowidentheloudspeakers’ BeamWidths.
1Thishasbeenproveninvariouslisteningtests. Forexample,gocheckout“PsychoacousticEvaluationofSyntheticImpulseResponses”byPerRubak&LarsG.Johansenasa startingpoint.
57
Appendix5: MicrophoneplacementstrategywhencreatingARCZones
AsisdiscussedinActiveRoom Compensation,itispossibletocreate settingsfordifferentARC Zones(or listeningareas). Thisisdoneby placingamicrophoneinthreedifferent locationswithinthezoneand performinganARCmeasurementat eachposition. Thissectiongivessome recommendationsregardingwhereto placethemicrophoneforthe measurements.
18.1 Generalinformation
18.1.1 Backgroundnoise
Itisimportanttoensurethat extraneousbackgroundnoiseiskeptto aminimumduringthemeasurement procedure. Thisdoesnotonlyinclude mid-rangefrequencies(e.g. speech) butalsolow-frequencynoise. Therefore,forexample,itis recommendedthatairconditioning systemsbeturnedoff,andthe measurementsareperformedduring low-traffichours. Thisisbecause,in somecases,themeasurementmay interpretbackgroundnoiseasartefacts oftheroom’sacousticalbehaviour. (Forexample,theprocessmayresult inreducedbassintheloudspeakersif atruckwasidlingoutsidetheroom duringthemeasurementprocedure.)
18.1.2 Microphone OrientationandHolder
Themicrophoneshouldbesecurely held(e.g. onacameratripodor microphonestand),pointingupwards asisshowninFigure18.1.
Figure 18.1: Recommended microphoneorientation.
Itisnotrecommendedthatthe microphonebehand-heldduetothe lengthofthemeasurementprocedure andthefactthatthemicrophone shouldnotmoveduringthe measurement. Extraneousnoise causedbyholdingthemicrophonemay alsoaffectthemeasurementaccuracy.
18.1.3 Height
IftheARCmeasurementisforonlyone listenerwhoneverchangesposition (e.g. never“slouches”inthelistening chair),thentheheightofthe microphoneshouldberoughlythe sameastheheightofthatperson’s ears,typically100–120cmabovethe floor.
Figure 18.2: Recommended microphoneplacementheightforonelistener showningray.
IftheARCmeasurementzoneis intendedformorethanonelistening position,orforlistenersofdifferent heights,thenitmaybebeneficialto changetheverticalpositionofthe microphoneforthethree measurements. Forexample,setthe centremicrophonepositionat ear-height,thefront-leftposition slightlylower(10-20cm),andthe back-rightpositionslightlyhigher (10-20cm). Thiswillprovidethe calculationwithadditionalinformation regardingtheeffectsofroommodesin theverticaldimensionthatmay benefitlistenersofdifferentheights.
18.1.4 DoorsandWindows
Doorsandwindowsinthelistening roomshouldhavethesameposition duringthemeasurementaswhenthe roomwillbeusedforlistening. So,if
younormallylistentomusicwiththe doorsclosed,thentheyshouldalsobe closedduringthemeasurement procedure. Thisisbecauseopeninga doororawindowcanhavea significanteffectontheacoustical behaviourofalisteningroom. Ifdoorsmaybeopenedorclosedfor differentlisteningsituations(e.g. patio doorsleadingfromthelivingroomto theoutdoors)thentwodifferentARC Zonesshouldbecreatedseparatelyfor thetwodifferentscenarios.
18.2 Onelisteningposition
IfanARCZoneconsistsofonlyone listeningposition,itisrecommended thatthethreemicrophonepositions are:
• thelocationofthelistener’s head,asshowninFigure18.3 • oneachsideofthelistening position(approximately30cmto eithersideofthelistening position). Oneplacementshould beslightlyforward (approximately20cm)andthe othershouldbeslightlybehind (approximately20cm).
Asmentionedabove,themicrophone shouldbeplacedroughlyatear-height.
Figure18.3: RecommendedARCmicrophone placements (in red) for IfanARCZoneconsistsofmorethan onelisteningposition(e.g. asofa)then themeasurementshouldbeperformed onceforeachposition. Figures18.4 and18.5showaexamplesofzones consistingofmanypossiblelistening positions. Themicrophoneshould placedatthreepositions(at ear-height)distributedroughlyevenly throughouteachzonetocreatethe ARCfilterforeachsituation.
Figure18.4: RecommendedARCmicrophoneplacements(inred)foramultiple listeningpositions.
Figure 18.5: Example ARC microphone placements(inred)forapassivelisteningandbackgroundmusicsituationsat adiningtable.
Notethat,incaseswherethereis overlapbetweendifferentARCZones, themeasurementscanbecombined bycombiningARCProfilesinthe BeoLab90interfaceinsteadof duplicatingmeasurements. An exampleofthisisshowninFigure18.6.
Figure 18.6: An example of avoiding duplicatemicrophoneplacementswhen ARC Zones overlap. The “sweet spot” zoneismeasuredusingthemicrophone placementsshowninred. Theotherlistening positions on the sofas are measured for a second ARC Zone using the microphone placements shown in blue. These two are combined by selecting both in the interface to completelycovertheentirearea.
59
Appendix6: ABL-AdaptiveBassLinearisation
19.1 AGeneralIntroduction toABL
AlmostallloudspeakersmadebyBang &OlufsenincludeAdaptiveBass LinearisationorABL.Thisincludesnot onlyour“standalone”loudspeakers (theBeoLabseries)butalsoour smaller“BeoSound”loudspeakersand televisions. Theonlyexceptionsinthe currentportfolioareourpassive loudspeakers,headphones,andthe BeoLab5. Thereisnoonetechnicaldefinitionfor ABL,sinceitisincontinualevolution– infactitmaychangefromproductto productaswelearnmoreandas differentproductsrequiredifferent algorithms. Speakingverybroadly, however,wecouldsaythatitreduces thelowfrequencycontentsenttothe loudspeakerdriver(s)(e.g. thewoofer) whentheloudspeakerisaskedtoplay loudly–buteventhisispartially inaccurate. Itisimportanttonotethatitisnot the casethatthisreplacesa“loudness function”whichmay(ormaynot)be equalisingforEqualLoudnessContours (sometimescalled“Fletcher-Munson Curves”). However,since(generally) thebassispulledbackwhenthingsget loud,itiseasytoassumethistobe true. Whenwearedoingthesounddesign foraloudspeaker(whichisbasedboth onmeasurementsandlistening),we ensurethatweareoperatingata listeninglevelthatiswellwithinthe “linear”behaviouroftheloudspeaker anditscomponents. (Typically,the sounddesignisdoneatastandard playbacklevelwherea-20dBFS full-bandpinknoiseproduces70dB(C) atthelisteningposition.) Thismeansthat
• thedrivers(usuallythewoofers) aren’tbeingaskedtomovetoo far(inandout) • theamplifierisoperatingwithin
itslimits • thepowersupplyisoperating withinitslimits,and • nothing(notthepowersupply, theamplifiers,northevoice coils)isgettingsohotthatthe loudspeaker’sbehaviouris altered.
Thisiswhatismeantby“linear”—it’s fancywordfor“predictable”. In addition,itshouldbestatedthatifwe werelisteningtoloudspeakersathigh levelsdaily,wewouldgetincreasingly badatourjobsduetohearingloss. So,wedothetuningatalisteninglevel whereweknowthingsarebehaving– rememberthatwealwaysdoitatthe samecalibratedleveleverytimefor everyloudspeakersothatwedon’t changesounddesignbalancedueto shiftsassociatedwithequalloudness contours. (Ifyoutunealoudspeaker whenit’splayingloudly,you’llwindup withaloudspeakerwithlessbassthan ifyoutuneditquietly. Thisisbecause you’reautomaticallycompensatingfor differencesinyourownhearingat differentlisteninglevels.) Afterthattuningisdone,thenwego backtothemeasurementstosee wherethingswillmisbehave. For example,inordertocompensatefor therelativelysmallcabinetbehindthe woofer(s)intheBeoSound8/BeoPlay A8,weincreasetheamountofbass thatwesendtotheamplifiersforthe woofersaspartofthesounddesign. If weleftthatbassboostinthetuning whenyouturnupthevolume,the loudspeakerwouldgoupinsmoke–or atleastsoundverybad. Thiscouldbe because:
• thewooferisbeingpushedor pulledbeyonditslimits,or • becausetheamplifierclipsor • thepowersupplycannotsupply morecurrentor • somethingelse.
So,afterthetuningprocessis complete,weputtheloudspeakerina smalltorturechamberroughlythesize ofaclothescloset,putonsomedance music(ormodifiedsynthetictest signals)andturnupthevolume. While that’splaying,we’recontinually monitoringthesignalthatwe’re sendingtotheloudspeaker,thedriver excursion,thedemandsonthe electronics(e.g. theamp’s,DAC’s, powersupply,etc.) andthe temperatureofvariouscomponentsin theloudspeaker,alongwithanumber ofotherparameters. Armedwiththatinformation,weare ableto“know”howthoseparameters behavewithrespecttothe characteristicsofthemusicthatis beingplayed(e.g. howlouditis,in variousfrequencybands,forhowlong, inboththeshorttermandthelong term). Thismeansthat,whenyouplay musicontheloudspeaker,it“knows”:
• howhotitisatvariouslocations inside, • theloudspeakerdrivers’ excursions, • amplifierdemands, • powersupplydemands, • andsoon. (Theactuallistvaries accordingtoproduct–theseare justsometypicalexamples...)
So,whensomeparametergetsclose toamaximum(e.g. theamplifierstarts togettoohot,orthewooferisnearing maximumallowableexcursion)then somethingwillbepulledback. What ispulledback? Itdependsonthe productandtheconditionsatthetime you’replayingthemusic. Itcouldbea bandoffrequenciesinthebassregion, itcouldbethelevelofthewoofer. Ina worst-case-last-ditchsituation,the loudspeakermightevenberequiredto shutitselfdowntoprotectitselffrom you(ortheguestsattendingyour party). Ofcourse,thereisno guaranteethatyoucannotdestroythe
60
loudspeakersomehow–butwedoour besttobuildinenoughprotectionto coverasmanyconditionsaswecan. Howisitpulledback(i.e. howquickly andbyhowmuch)? Thatalsodepends ontheproductandsomedecisionswe madeduringthesounddesignprocess, aswellaswhatkindof state-of-emergencyyourloudspeaker isin(somepeopleareverymeanto loudspeakers...). Notethatallthisisdonebasedonthe signalsthattheloudspeakerisbeing askedtoproduce. Soitdoesn’tknow whetheryou’veturnedupthebassor thevolume–itjustknowsyou’re askingittoplaythissignalrightnow andwhattheimplicationsofthat
demandareonthecurrentconditions (voicecoiltemperature,forexample) Thisissimilartothefactthattheseat beltsinmycardon’tknowwhythecar isstoppingquickly–maybeit’s becauseIhitthebrakes,maybeit’s becauseIhitaconcretewall–theseat beltsjustlockupwhenthey’reasked tomovetooquickly. Yourwoofer’s voicecoildoesn’tknowthedifference betweenEminemandStravinskywitha bassboost–itjustknowsit’shotandit doesn’twanttogethotter.
19.2 ABLandBeoLab90
InspiteofBeoLab90’smassivepower reservesandfourcapablewoofers,it
stillbenefitsfromtheinclusionofABL initsprocessing. Thisisduetothefact thattheBeoLab90’ssounddesign resultedinafrequencyrangethat extendstoapproximately10Hz. Playingathighlisteninglevels,sucha lowfrequencyextensionwouldresult inover-excursionofthewoofersifABL werenotincludedintheloudspeaker’s processing. However,itshouldbesaid thatwhereasatypicalBang&Olufsen loudspeakerwillhaveanABLoperating atfrequencybandsfromapproximately 100Hzanddown,theBeoLab90’sABL onlyoperatesbelowapproximately20 Hz.
61
Appendix7: ThermalCompressionCompensation
20.1 Introduction
Takeawooferandputitinan appropriately-sizedcabinet,connectit toanamplifier. Settheroom temperatureto20◦C.Wheneverything intheroomisthesametemperature, measurethewoofer’son-axis magnituderesponse. Turnuptheroomtemperatureto 100◦C.Wheneverythingintheroomis thesametemperatureagain,measure thewoofer’smagnituderesponseonce more. Youwillnoticethatthesetwo measurementslookverydifferent–but why? Whenyoureadamagazinereviewofa loudspeaker,itwillincludea measurementofits“frequency response”(moreaccuratelycalledits “magnituderesponse”)whichshows (ignoringmanythings)howloud differentfrequenciesarewhenthey comeoutoftheloudspeakerassuming thattheyallcameinatthesamelevel. Unfortunately,thisisonlyasmallpart ofthetruth. Wecanexplainaloudspeakerdriver’s electromechanicalcharacteristicsby breakingitdownintodifferent components(bothactualand analogical). Forexample,the suspension(whichiscomprisedofthe surroundandthespider)canbe thoughtofasaspring. Theelectrical analogyforthisisacapacitor. Ifyou takeallofthemovingpartsinthe loudspeakerdriver,theyalladduptoa massthathastobemoved–the electricalanalogyforthatmassisan inductor(sinceaninductorhassome electrical“inertia”). Someofthe componentsarenotanelectrical analogy–theyarerealelectrical components. Forexample,thevoice coil,sinceit’sacoil,actsasan inductor. Sinceitisathinbitofwire,it alsohassomeresistancetotheflowof electricalcurrentthroughit,soit’salso aresistor. Asimpleversionofthis
breakdownisshowninFigure20.1.
Amplifier
Voice Coil Resistance
Electrica Components Mechanical Components
Voice Coil Inductance
Energy Losses in Suspension
Inertia of Mass of Moving Parts
Springiness of Suspension
Figure 20.1: A simplified version of the actualelectricalandelectricalanalogies of mechanical components in a loudspeakerdriver.
Thisshowsthecomponentsofa movingcoildynamicloudspeakerasa verysimplified“circuit”. Ifthese componentsdon’tlookfamiliartoyou, don’tworry,it’snotthatimportantfor now. Somecomponentsinthecircuit areactualelectricalthingsandothers areanalogies–electrical representationsforamechanical componentinthesystem. Ifyouknowhoweachofthese componentsbehaves,andyouknow thecorrectvaluestoputinforagiven loudspeaker,andyouknowhowtodo therightmath,thenyoucancome closetogettingapredictionofthe responseoftheloudspeakerthat you’remodellingwiththecircuit. However,ifyoujustputinonevalue foreachcomponent,thenyou’re assumingthattheyneverchange–in otherwordsthatyou’redealingwitha “linear”system. Theproblemisthatthisassumptionis incorrect. Forexample,thevoicecoil resistance–theamountthatthewire inthevoicecoilresiststheflowof currentthroughitwhenthe loudspeakerdriverisnotmoving– changeswithtemperature. Thehotter thewiregets,thehighertheresistance goes. (Thisisanormalbehaviourfor mostresistors.) Ifthevoicecoil resistancechanges,thenthewhole systembehavesdifferently,sinceit isn’ttheonlycomponentinthecircuit. So,aswechangethetemperatureof thevoicecoil,thetotalresponseofthe loudspeakerchanges. Sadly,thetemperatureofthevoicecoil isn’tonlydependentontheroom
temperatureasitseemedtobeinthe beginningofthisdiscussion. Assoon asyoustartplayingsoundusingthe loudspeaker,itstartsheatingup. The louderthesignal,thehotteritgets. So asyouplaymusic,itheatsandcools. Thespeedwithwhichitheatsupand coolsdownisdependentonits “thermaltimeconstant”–abigwoofer withalargevoicecoilandmagnetwill takelongertoheatupandcooldown (andthereforehavealongerthermal timeconstant)thanasmalltweeter. Thisraisesatleastfourquestions:
• Howmuchdoesthetemperature varywhenIplaymusic? • Howdoestheresponseofthe loudspeakerchangewith temperature? • Howmuchdoestheresponseof theloudspeakerchangewith temperature? • Whatcanwedoaboutit?
20.2 Voicecoiltemperature
Atypicalloudspeakerdriveris,giveor take,about1%efficient. Thismeans thatapproximately1%ofthepower youpushintotheloudspeakerfromthe amplifierisconvertedintosound. The remaining99%islostasheat–almost allofitatthevoicecoilofthe loudspeaker. So,thelouderyour music,thehotteryourvoicecoilgets. Ofcourse,ifyouhaveawayofcooling it(forexample,byusingotherpartsof theloudspeakerasaradiatortoyour listeningroom)thenitwon’tgetas hot,anditwillcooldownfaster. Forexample,playpopmusicthathas beenmasteredatahighlevelandplay itatmaximumvolumeonaBeoLab90 whilstmonitoringthetemperatureof thevoicecoils. Whatyou’llseeifyou dothisissomethingliketheFigure 20.2.
62
Figure 20.2: The temperatures (in ◦C) ofthevoicecoilsofthedriversinaBeoLab5asaresultofplayingpopmusic atfullvolume. TheX-axisisthetimein minutes.
AsyoucanseeinFigure20.2,while playingmusic,thewoofervariedfrom amaximumtemperatureofabout 200◦Cdowntoabout110◦C. Thismeansthattheworst-case variationintemperatureofthewoofer wasabout90◦Cwhilstplayingmusic, andpeakedatabout180◦Cabove roomtemperature(whichwe’llassume is20◦C). Unfortunately,thistemperaturecannot bemeasureddirectly,sincewecannot putthermalsensorsdirectlyonthe drivers’voicecoils. Instead,we measurethetemperatureofthe loudspeakerdrivermagnets,anduse thatreal-timedatainputinadditionto thesignalthatwe’resendingtothe driverstocalculatethetemperatures ofthevoicecoilsbasedonthermal modelsofeachofthem. Asyoucan seeinFigure20.3,themagnet temperaturereactsmuchmoreslowly. Thesemeasurementsweretakenat exactlythesametimeastheones showninFigure20.2.
Figure 20.3: The temperatures of the magnetsoftheloudspeakerdriversina BeoLab5asaresultofplayingpopmusicatfullvolume. TheX-axisisthetime inminutes.
20.3 Loudspeakerresponse changes
So,nowthequestionis“whatdoesthis changeintemperaturedotothe responseoftheloudspeakerdriver?”. AsImentionedabove,thethingthat changesmostinthemodelshownin Figure20.1istheloudspeakerdriver’s voicecoilresistance. Forthoseofyou withabackgroundinreadingelectrical circuits,youmaynoticethattheone showninFigure20.1hassomereactive componentsinitwhichwillresultina resonanceatsomefrequency. For thoseofyouwithoutabackgroundin readingelectricalcircuits,whatthis meansisthataloudspeakerdriver(like awoofer)hassomefrequencyatwhich it“wants”toring–ifyouthumpitwith yourthumb,that’sthenotethatyou willhearringing–alittlelikeabellwith alowpitch. Asthevoicecoilresistancegoesup,its resistanceincreases,andwegenerally losesensitivity(i.e. levelorloudness) fromthewoofer. Inotherwords,the hotteritgets,thequieteritgets. However,thisonlyhappensatthe frequencieswheretheresistorisnot “overridden”byanothercomponent– saythemechanicalresonanceofthe wooferortheinductanceofthevoice coil. Thetotalresultisshownforvarious temperaturedifferencesinFigure20.4. Noticethattheseplotsshowthe changeinmagnituderesponseofthe
driverwithchangesintemperature. So,thecurveatthetopisthechange inthewoofer’smagnituderesponse (whichis0dBatallfrequencies–in otherwordsnochange)whenthe loudspeakerisplayingatthesame temperatureatwhichitwasmeasured (let’ssay,20◦Corroomtemperature). Asthetemperatureofthevoicecoil increasesabovethattemperature,you canseethatyouloseoutputintwo frequencybandsoneithersideofa “bump”intheresponse–thatbumpis attheresonantfrequencyofthe loudspeakerdriver. So,thelouderyouplay,themorelow endyoulose,apartfromapeakinthe response(whichalsoringsintime)at theresonantfrequencyofthedriver.
10 100 1,000− 4.5 −4 −3.5 −3 −2.5 −2 −1.5 −1 −0.5 0 0.5 Frequency (Hz) Sensitivity (dB) + 0º C + 20º C + 40º C + 60º C + 80º C + 100º C + 120º C + 140º C + 160º C + 180º C Figure 20.4: Sensitivity of BeoLab 90’s front woofer vs. the change in temperatureofitsvoicecoil.
20.4 Thesolution
Interestingly,everythingIsaidaboveis trueforeverymovingcoilloudspeaker. So,ifyou’rethekindofpersonwho believesthattheonly“proper” loudspeakerisonewhereyouhave nothingbutaloudspeakerdriver(ina cabinetofanykind,ornot)andan amplifier–andnoactivefiltering,then you’llhavetolivewiththekindof unpredictablebehaviourthatyousee above. However,sinceaBeoLab90 “knows”thetemperatureofthevoice coilofitsloudspeakerdrivers,and sinceithasbeenprogrammedwiththe curvesliketheoneshowninFigure 20.4,wecanactivelylineariseits response,makingitmuchmore predictable. Inessenceallweneedtodoistotake Figure20.4,flipitupsidedownand
63
makeafilterthat“undoes”theeffect oftemperatureontheloudspeaker’s response. Forexample,ifthewoofer’s voicecoilgets160◦Caboveroom temperature(whereweoriginally measuredit),itdrops3.2dBat20Hz, theBeoLab90knowsthisandadds3.2 dBat20Hz. Inordertodothis,the processingoftheBeoLab90includesa setoffilters(oneforeachdriver) whoseresponsevariesintimewiththe temperaturesofthethedrivers. The temperature-dependentfiltersforthe frontwooferareshowninFigure20.5.
10 100 1,000− 0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 Frequency (Hz) Gain (dB) + 0º C + 20º C + 40º C + 60º C + 80º C + 100º C + 120º C + 140º C + 160º C + 180º C Figure 20.5: Magnitude responses of thecompensatingfilterforBeoLab90’s front woofer vs. the temperature of its voicecoil.
It’simportanttonotethreethingshere.
• Thiscanonlybedonebecause weknowhowtheresponseofthe wooferchangesatdifferent temperatures(thisbehaviour wasfoundaspartofthe developmentprocess). • Thiscanbedonebecausethe loudspeaker“brain”(theDSP) knowsthetemperatureofthe voicecoilinrealtimeasyou’re playingmusic • Thisparticularfiltershownin Figure20.5shouldonlybe
appliedtotheappropriate loudspeakerdriver. Theother woofersandtheotherdrivers havedifferentbehavioursand shouldbeprocessedwiththeir owncorrectioncurves. Inother words,thisfilteringcanonlybe donebecausetheBeoLab90is anactiveloudspeakerwith independentfilteringforeachof the18loudspeakerdrivers.
20.5 Someextrainformation
Youshouldbeleftwithatleastone question. Isaidabovethat,asthe musicgetsloud,thewooferheatsup, soyouloseoutput,soweaddafilter thatcompensatesbyputtingmore signalintothedriver. However,this meansthattheproblemiscausedby thesignalbeingtooloud,andthe resultisthatwemakethesignal louder. However,thereisonemoretrickup oursleeve. Appendix6: ABL-Adaptive BassLinearisationdescribesBeoLab 90’sThermalProtectionalgorithm. This meansthattheDSPbrainknowsthe temperatureofthedriversand,ina worst-casesituation,turnsthelevels downtoprotectthingsfromburning up. So,ifwegobacktotheexampleof aBeoLab90playingatfullvolume, let’sseewhat’shappeningtothe signallevels.
Figure20.6:Thegains(indB)appliedto thesignalssenttothedriversinaBeoLab 5 as a result of playing pop music atfullvolume. TheX-axisisthetimein minutes.
ThesecurvesinFigure20.6showthe gainsappliedtothefrontwooferina BeoLab90atthesametimeasthe measurementsshowninFigures20.2 and20.3werebeingmade. Infact,if youlookcarefullyatFigure20.2 aroundthe5minutemark,you’llsee thatthetemperaturedropped–which iswhythegaininFigure20.6increases (becauseitcan!) inresponse. Now,don’tpanic. TheBeoLab90isn’t messingaboutwiththegainsofthe driversallthetime. Rememberthat thisexamplewasdoneatfull volume– which,foraBeoLab90isextremely loud. ThegainsshowninFigure20.6 area“last-ditch”effortofthe loudspeakertoprotectitselffroma verymeancustomer(ortheverymean childrenofacustomerwhoisawayfor theweekend). Thisistheequivalentof theairbagsdeployinginyourcar. You canguessthat,iftheairbagisoutside thesteeringwheelsomething significanthasoccurred. Many thanks to Gert Munch for his help in writing this section.
64
Appendix8: ControlofBeoLab90usingtheBeoRemote1
21.1 Introduction
TheBeoRemote1canbeusedinstead ofthesmartphoneinterfacetocontrol anumberofparametersontheBeoLab 90. Theusageandsetupoftheseare describedbelow Notethat,inordertocontrolthe BeoLab90’swiththeBeoRemote1, twothingsmustinitiallybeconfigured. Firstly,theinfra-redreceiver(also knownasan“IREye”)mustbe correctlyconnectedtotheMaster BeoLab90. Secondly,theBeoRemote 1mustbesettothe“BeoSound” productonitsdisplay. Inaddition,the BeoLab90mustbe“paired”tothe BeoSoundproduct(sincethereare multipleBeoSoundoptions).
21.1.1 Customisingthe ProductName
Itispossibletocustomisetheproduct namedisplayedontheBeoRemote1 (forexample,toread“BeoLab90” insteadof“BeoSound”). Thiscanbe doneasfollows:
1. PressLIST 2. ScrolldowntoSETTINGSand presstheroundGObutton 3. ScrolldowntoADVANCEDand pressGO 4. SelectPRODUCTSandpressGO 5. SelectRENAMEandpressGO 6. Selecttheitemyouwishto renameandpressGO 7. “Type”inthenewnameforthe item 8. Whenyouaredone,scrollallthe waytotherightofthelistof lettersandselecttheCHECK MARKtostore.
21.2 InputSelection
Itispossibletomanuallychangethe inputselectionofaudiosourcesusing theBeoRemote1. Inaddition,the displayoftheremotecontrolcanbe customisedtoreflectyourparticular setup.
21.2.1 SourceSelection
IfyoupresstheMUSICbuttononthe remote,youwillseeanumberof sourceslistedtherethatarescrollable usingtheUP/DOWNbuttons,and selectableusingtheGObuttononthe remotecontrol. Firstly,youwillneedto customisethislistasfollows:
1. pressLIST 2. ScrolldowntoSETTINGSand pressGO 3. ScrolldowntoMUSICsources andpressGO 4. SelectSHOWandpressGO 5. UsetheGObuttontoplacea checkmarknexttothefollowing itemsinthefollowinglist. (Itmay alsobehelpfultoremovethe checkmarksnexttotheitems thatarenotinthelist.) -NETRADIO -MUSIC -LINEIN -CD -A.MEM
NowpresstheMUSICbuttonandyou shouldseethese5itemsinascrollable listonthescreenoftheremote. These 5aremappedtotheinputsofthe BeoLab90asfollows:
BL90Input BR1Display Automatic Music RCA Line-In XLR A.Mem S/PDIF CD Optical NETRadio So,forexample,iftheBeoLab90isin Automaticmodeandyouplaying signalsonboththeS/PDIFandtheXLR inputssimultaneously,youwillhear theS/PDIFsignalifithasahigher selectionpriority. PressMUSICand selectA.MEMtoswitchmanuallytothe XLRinput. PressMUSICandselectCD toswitchtotheS/PDIFinput.
21.2.2 CustomisingtheInput Names
Itispossibletore-nametheseinputs ontheBeoRemote1’sdisplaysothat youdonothavetorememberthe mapping. Thiscanbedoneasfollows:
1. PressLIST 2. ScrolldowntoSETTINGSand pressGO 3. ScrolldowntoMUSICsources andpressGO 4. SelectRENAMEandpressGO 5. Selecttheitemyouwishto renameandpressGO 6. “Type”inthenewnameforthe item 7. Whenyouaredone,scrollallthe waytotherightofthelistof lettersandselecttheCHECK MARKtostore.
Now,whenyoupresstheMUSIC button,youwillseethelabelsyou haveenteredinsteadofthedefault oneslistedabove.
65
Index
ABL,35,60 about,32 activelistening,7 activeroomcompensation,9,20 AdaptiveBassLinearisation,60 ADC,38 AnalogueInputs,38 analogue-to-digitalconverter,38 ARC,9,20,58 audioformats DoP,42,45 DSD,42 DSDoverPCM,45 DXD,45 automaticbasslinearisation,35
backgroundmusic,7 bass,22 beamdirection,8 beamdirectioncontrol,18 beamwidth control,8,15 narrow,15 omni,17 wide,16,18 BeoRemote1,65 BeoVisiontelevision,connectionto,26
centrefrequency,50 cloninginproduction,36 coaxialdigital(S/PDIF),39 converter,digitaltoanalogue,39
DAC,39 dB,50 decibel,50 DigitalInputs,38 digitalsignalprocessor,39 digitaltoanalogueconverter,39 distance,19 DSP,39
earlyreflections,52 echo,44 editmode,menu,14 equaliser,parametric,24,33
Fc,50 features,35 featuresdisabled,44 filter high-shelving,25,49 low-shelving,25,49 peaking,25,50 phaseresponse,35
Q,50 reciprocalpeak/dip,50