ednmag.comment By Brian Dipert, Technical Editor
Dont forget about the end user
hile researching my recent series of video articles, I had the opportunity to evaluate three DTV (digital-television) receivers: Mitsubishis
(www.mitsubishi-tv.com) SR-HD5, Panasonics (www. panasonic.com) TU-HDS20, and Princeton Graphics (www.princetonhdtv.com) HDT-2000. I used them with a
roof-mounted (20-ft-pole) Terk (www.terk.com) TV35 antenna, Princeton Graphics AF3.0HD wide-screen display, Panasonics Technics SA-DX1050 audio receiver, and a multivendor surroundplus-subwoofer speaker set. First, the good news: all three receivers solidly tuned in all NTSC (National Television Systems Committee) and ATSC (Advanced Television Systems Committee) channels in my broadcast area, a ringing endorsement of 8-VSB (eight-level vestigial-sideband) demodulation and particularly impressive given my neighborhoods poor reception characteristics. Go to www.antennaweb.org and enter my office address (1864 52nd St, Sacramento, CA, 95819) to get a visual explanation of the challenge the receivers faced. And, when I watched a film-captured program or a video-captured program that has been carefully converted from interlaced to progressive scan, with an accompanying 5.1-channel audio track, the multimedia quality blew me away. Enough of this type of programming and my wife and I might actually watch more than one or two hours of television per week. Alas, enough programming is only one of several serious challenges facing DTV as the Federal Communications Commissions 2006 deadline for analog-to-digital conversion looms. I could spend

www.ednmag.com

an entire editorial ranting about the chicken-or-the-egg-type finger pointing going on now among broadcasters, cable operators, and equipment suppliers. The fact is, though, that even if all the broadcasters immediately switch over to creating their new television material in a wide-screen, progressive-
DELAYING DTV: A GOOD IDEA.
scan-captured digital format, most of what DTV will broadcast for decades to come will be digitally converted analog-video archives. Think, for example, about all those Happy Days and Gilligans Island reruns! How does DTV handle this analog- and interlace-sourced material? Horribly, if my experience highlighted by my high-resolution progressive-scan display, which uncovers quality sins that older interlaced sets obscureis any indication. In my survey of morning talk shows, mid-day soap operas, evening sitcoms, and weekend sports programming, I saw lots of blocky MPEG compression artifacts, temporal jaggies, colorfringe moire patterns, fuzzy edges, and overall soft images. Was this mediocre quality a function of poor encoding at the broadcast end or of poor decoding at the reception end? I dont know for sure (al-
though I strongly suspect the former). But I dont care, and the average consumer wont either. Its sad but true that digital 480i generally looks worse than analog NTSC, either on my old TV or after I ran it through a Focus Enhancements (www.focusinfo.com) QuadScan Elite or Silicon Image (www.siimage.com) iScan Pro video processor before displaying it on the AF3.0HD. Some broadcasters insistwhen displaying 4-to-3 aspect ratio material in a 16-to-9 formaton horizontally stretching the image, thereby creating disproportionately short-and-fat objects. If you watch only one television channel, your eyes and brain eventually compensate, but channel surfers dont have that luxury; why the broadcasters dont keep the 4-to-3 aspect ratio and simply put gray vertical bars on either side is beyond me. Video quality, unfortunately, isnt DTVs only shortcoming. The Princeton Graphics HDT-2000 provides only a sixchannel analog output and no S/PDIF (Sony/Philips Digital Interface)a baffling omission considering that digital-audio signals run around inside the unit. Newer audio receivers provide six-channel analog inputs, but they target DVD-Audio or SACD (Super Audio CD) players. And hardly any receiver more than a year old has them. The Mitsubishi SR-HD5 almost got audio right; the unit outputs Dolby Digital or PCM (pulse-code modulation) over S/PDIF when it receives ATSC broadcasts, and PCM for NTSC. I say almost because the bit streams that both the SR-HD5 and the Panasonic unit generate occasionally confuse the digital-decoding circuitry in my audio receiver, necessitating one or sometimes several transitions away

November 8, 2001 | edn 33

ednmag.comment

from and then back to the desired channel before sound comes out the speakers. Do the television receivers cause this problem, or does the audio receiver cause it? Again, I suspect the former, because the Technics handles my DVD players Dolby Digital and PCM outputs with aplomb. But again, it doesnt matter. The S/PDIF, Dolby Digital, and PCM formats have been around long enough that less-than-100% interoperability is inexcusable. My final gripe concerns channel tuning and identification. All three units contain automatic-channel-scanning and programming features, yet the SRHD5 stubbornly refuses to find the ATSC version of channel 58. When I manually punch 35-2 into my remote control, however, the SR-HD5 reports a strong signal and dutifully adds the channel to its list. Channel designations represent baffling idiosyncrasies. One DTV receiver might report the digital version of Fox network channel 40 as channel 61; another might call it 401. The same unit might even switch from one designation to the other after autoreprogramming. After using these supposed secondand third-generation DTV receivers, I find it hard to believe that they went through much focus-group testing before their release. Theyre just too clum-
THE LAST THING ANYONE WANTS IS ANOTHER WIDGET WHOSE REALITY FALLS FAR SHORT OF THE PROMISES MADE TO CONSUMERS.
sy and nonintuitive to operateeven for a technoid like myself. Quite a few pundits are predicting that the FCC will extend its self-imposed 2006 deadline to give everyone more time to make the analog-to-digital switch. This idea is great. The technology is ready for prime time, but the implementations are not. And the last thing anyone wants is another widget whose reality falls far short of the promises made to consumers. You can reach Technical Editor Our engineering Brian Dipert at 1-916-454-5242, mentalities are both fax 1-916-454-5101, our biggest blessings e-mail bdipert@ pacbell.net. and our greatest curses. To paraphrase a common expression, they result in us getting so close to the forest that we lose sight of the trees. We sometimes kid around about how we cant explain what we do to our spouses, relatives, and friends. Thats a problem, not something to brag about. Its a bug, not a feature.
34 edn | November 8, 2001

techtrends By Brian Dipert, Technical Editor
YOU MAY PLAN TO INCLUDE ONLY TWO SPEAKERS, JUST A SET OF HEADPHONES, OR SUPPORT FOR ONE- AND TWO-CHANNEL SOUND SOURCES IN YOUR NEXT DESIGN. NOT TO WORRY. PSYCHOACOUSTICS, PROCESSORS, AND PLENTY OF MEMORY CAN STILL GIVE YOUR CUSTOMERS AN IMMERSIVE AUDIO EXPERIENCE. HEAR IS HOW.
Decoding and virtualization
bring surround sound to the masses
At a glance..64 Havent heard enough?.64 For more information.72
eople commonly use the terms stereo,two channel, and two speaker to mean the same thing, but theyre committing
an etymological error. Stereo derives from the Greek word stereos, meaning solid. In other words, a stereo recording-and-listening environment is one that reproduces the 3-D atmosphere of the
original performance. Bob Dylans Live 1966 ideally should sound just like it did when first performed in Manchester, England, and Miles Davis Kind of Blue should transport you to 1959 and Columbias 30th Street Studio, New York, even if youre listening to them in your living room in Sacramento, CA, in 2001. And, when the Tyrannosaurus Rex in the movie Jurassic Park growls on-screen, you should hear it over your shoulder and 20 feet above you, and the hair on the nape of your neck should rise. As early as the 1930s, Bell Labs researchers JC Steinberg and WB Snow determined that you need at least three transducers to realistically reproduce an audio source (Figure 1a and b). Their research didnt comprehend additional speakers necessary to replicate the acoustics of the listening environment. Monophonic radios, phonographs, and tape players, eventually supplanted by their two-channel variants and by audio CDs, existed for reasons of the economics and technology limitations of the time, not because they delivered realism. Some of you may be familiar with the illfated quadraphonic (four-channel) systems that briefly appeared in the early 1970s. The theory behind quadraphonic audio was fairly solid, but the implementations werent. Manufacturers proprietary systems and corresponding media were incompatible with each other, and compelling content was lacking. Deluded audiophiles like to point to the failure of quadraphonic sound as justificaOctober 25, 2001 | edn 63

www.ednmag.com

techtrends Surround sound
tion for two-channel-only audio (Reference 1). The fact is surround sound in movie theaters has thrilled consumers for decades. Leopold Stokowski and the Philadelphia Orchestra performed the classical music in Disneys (www.disney.com) 1940 film Fantasia in Fantasound, and it is but one early example of surround sound. Now-ubiquitous Dolby Surround first appeared in the mid-1970s, Dolby Digital debuted with the movie Batman Returns in June 1992, and Jurassic Park followed in 1993 in DTS (Digital Theater Systems) surround. Dolby Surround-encoded television broadcasts and videotapes are now commonplace and, along with Dolby Digital- and DTS-aware DVD and audio CD players, have brought surround music, movies, and other programs into living rooms and automobiles. DVD-Audio and SACD (Super Audio Compact Disc), assuming they succeed in the market (which is not a foregone conclusion), will accelerate this awareness, as will immersive gaming and other computerbased audio environments (references 2 and 3). Teleconferencing, another poAT A GLANCE

Armed with an auditory system understanding, a powerful-enough processor, and sufficient RAM and ROM, you can give your customers an immersive audio experience no matter what the sound source is. Artificially created and artfully extracted reflections and reverberations can both approximate and alter the ambiance of the original recording environment. Low- and high-frequency enhancement compensate for bass-deficient speakers and lossy compression artifacts. Interaural time and intensity adjustments and head-related transfer-curve data enable you to place phantom sound sources all around a listener. Speakers and headphones require different audio-processing schemes.
tential surround-sound application, enables listeners to differentiate individual speakers, talking at the same time at the
other end of the line, within a group. Another common argument of the audiophile Luddites against audio with more than two channels is the baffling statement that they need only two speakers because they have only two ears. In reality, the human auditory system, working with other sensory faculties, such as vision, can accurately locate a sound source in 3-D space using both absolute and differential time-of-arrival, intensity, and frequency cues along with training in how the head, shoulders, and ears modify incoming sounds. Even if the sound source is directly in front of you, the echoes and reverberations of the recording environment will alter it unless the recording occurs in an anechoic chamber. Because the end listening environment doesnt have the same acoustics, you need audio processing to re-create a semblance of the original. Now that listeners have enjoyed surround sound with portions of their television programming, movies, and music collections, theyd like to extend immersiveness to all of their multimedia experiences, regardless of the sources characteristicsthat is, monophonic or

HAVENT HEARD ENOUGH?

If this article has whetted your appetite to learn more about surround sound, I encourage you to check out references 1 to 10 in the main article. Surround Sound Professional magazine (www.surroundpro.com), which audio pioneer Tomlinson Holman heads, is also an excellent source of news and information and has particular appeal to recording engineers. The magazine sponsors a conference in Beverly Hills, CA, each December, as well as seminars at the Consumer Electronics Show (www.cesweb.org) and other forums. Speaking of conferences, the biannual Audio Engineering Society (www.aes.org) convention is an impressive group of some of the finest minds in audio, and the AES also sponsors periodic conferences on audio topics. An AES membership is worthwhile if only for the 10-issue-per-year Journal of the Audio Engineering Society. Most of the vendors listed in the For more information. sidebar have Web sites flush with white papers, application notes, other documentation, presentations, and sound samples. Each thinks its own approach is the best, but, by surveying a number of them, you gain a breadth of knowledge on technology and product alternatives. And speaking of Web sites, I list some that have been particularly useful to me in my research. Fire up the Google (www.google.com) search engine using keywords surround sound to find hundreds of others. Floyd Toole, long-time and well-known audio researcher, has several detailed white papers on the Harman Web site (www.harman.com). David Griesinger, principal scientist at Lexicon, a division of Harman, provides a large number of his AES and other papers and presentations for free downloading at www.world.std. com/~griesngr. You can even see a picture of him sketching Figure 3! Well-known audio engineer Bobby Owsinksi runs Surround Associates, and the companys Web site at www.surroundassociates.com contains several helpful articles. The Ambisonics FAQ is at www.stanford.edu/~mleese/Am bisonic/faq_latest.html. Also, check out www.ambisonics.net for more information on this set of techniques, developed in the 1970s and intended for the recording, studio processing, and reproduction of the complete sound field that occurs during an audio performance. Bob Stuart from Meridian Audio, the developer of the MLP compression scheme used in DVD-Audio and one of the key definers of the DVD-Audio standard, maintains the Acoustic Renaissance for Audio Web site at www.meridian-audio.com/ara, which contains a number of interesting papers. The Ambiophonics Web site at www.ambiophonics.org discusses how to optimize the acoustics of typical home listening room environments. The 3D Audio Immersion site at www.3dai.net keeps up to date on the latest news in the world of surround sound. Finally, watch for my Jan 10, 2002 cover story, which will describe and compare product alternatives, including complete encoding/decoding systems.

64 edn | October 25, 2001
two-channelor the listeners settings, such as in an automobile, an airplane, an office, a conference room, a living room, or on the jogging track. But, because of financial and aesthetic resistance to buying and installing center, rear, and subwoofer speakersthe so-called SAF (spousal-acceptance factor)many of them would like to gain an approximation of the true surround-sound experience using their current two-speaker or headphone configurations. Fortunately, just as with lossy audio compression, an in-depth understanding of both the strengths and the shortcomings of the auditory system, along with some DSP horsepower and memory, can credibly accomplish these seemingly divergent objectives (references 4 and 5). If youre starting with a single-channelthat is, monophonicaudio
STAGE SOURCE DIRECT-SOUND PULSE THREE MICROPHONES THREE ELECTRICAL CHANNELS
LOUDSPEAKER-BASE WIDTH B=2 TO 4m CENTER SCREEN 1 LEFT SCREEN 30 RIGHT
THREE LOUDSPEAKERS THREE DISTINCTSOUND PULSES
2 AUDITORIUM OBSERVER THREE PULSES TO EACH EAR (a)
STAGE SOURCE DIRECT-SOUND PULSE SCREEN OF MICROPHONES ELECTRICAL CHANNELS VIRTUAL SOURCE SCREEN OF LOUDSPEAKERS INDIVIDUAL POINT-SOURCE SOUND PULSES SINGLE RESULTANT SOUND PULSE O2 O1 OBSERVERS ONE PULSE TO EACH EAR AUDITORIUM

120 LEFT SPEAKER

120 RIGHT SPEAKER
NOTES: 1. SCREEN 1'S LISTENING DISTANCE IS 3H (21=33), WHERE H IS HEIGHT. THIS DISTANCE IS POSSIBLY MORE SUITABLE FOR A TV SCREEN. 2. SCREEN 2'S LISTENING DISTANCE IS 2H (22=48), WHERE H IS HEIGHT. THIS DISTANCE IS MORE SUITABLE FOR A PROJECTION SCREEN. H:SCREEN HEIGHT 3. SUBWOOFER IS NOT SHOWN; ITS LOCATION IS OPTIONAL.
CENTER LEFT 0 LEFT HEIGHT RIGHT HEIGHT LEFT WIDE SUBWOOFER ON FLOOR SUBWOOFER 90 ON FLOOR 60 LEFT HEIGHT HIGH 45 HIGH 45 RIGHT HEIGHT LEFT WIDE 60 RIGHT

Figure 1

LEFT SIDE 120 120
RIGHT SIDE RIGHT-SIDEDIFFERENCE DIPOLE
LEFT-SIDEDIFFERENCE DIPOLE (d)
Bell Labs research in the 1930s suggested that you need at least three speakers (a, courtesy Focal Press) to credibly reproduce a sound source (b, courtesy Focal Press). The 5.1-speaker configuration is common in todays home-theater settings (c, courtesy Focal Press), whereas Tomlinson Holmans 10.2-speaker proposal suggests one possible future arrangement (d, courtesy Tomlinson Holman Labs).

66 edn | October 25, 2001
APPROXIMATELY A source, it might seem impossible at first + 100-Hz glance to create a two-channel variant HIGHPASS FILTER + never mind a full-blown sur+ Figure 2 LEFT OUTPUT + round-sound representationof it (Reference 6). Recall, though, that peoAPPROXIMATELY B ple generally regard low-frequency 2- TO 10-mSEC + DELAY MONOPHONIC sounds as nondirectional; that is, the hu+ INPUT man auditory system can perceive them RIGHT OUTPUT + no matter what location in the listening APPROXIMATELY 100-Hz area they come from. Therefore, you can LOWPASS FILTER apply a lowpass filter to the source and direct frequencies of less than 100 Hz or so only to those channels whose transducers will likely be able to reproduce the Frequency subdivision, followed by time delay and mixing, is one means of converting monophonic frequencies, such as a distinct 0.1 sub- audio into a two-channel representation that retains monophonic-speaker backward compatibility woofer channel. (courtesy DV Magazine). Next, delay the highpass-filter output by an adjustable amount of time if you seat rows or in which the listeners are one of the two channels; the stereowant to allow for user customization and milling about instead of remaining in wide button in low-end consumer-electhen add it to the nondelayed signal to one location, requires a wide, deep sweet tronics gear frequently activates this norcreate one output channel and subtract spot. A similar situation exists in auto- mally undesirable technique. This techit from the nondelayed signal to create a mobiles, in which neither the driver nor nique obliterates center imaging of second channel (Figure 2). The results any of the passengers is in an ideal lis- shared-channel content. A more elaboare true and complementary comb fil- tening location. Bigger sweet spots result rate technique that less destructively scatters, and, when you present the two chan- in a more immersive surround-sound ters the sound-source-directional cues nels through two speakers, the frequen- experience at the possible expense of re- involves first calculating the channel cies split evenly between them. Running duced sound-source-positioning preci- (AB) and (BA) information and then both channels through the same speaker sion. employing frequency-dependent time, cancels out the delay and re-creates the The sweet-spot characteristics also de- spectrum, and overall intensity alteration original monophonic signal. Advanced pend on the anticipated placement of the to create the perception that the sound technologies from some of the compa- two front speakers and on the geometri- is originating beyond the physical nies that the sidebar For more infor- cal relationship between their spacing boundary of each speaker. mation. lists employ more elaborate and the listeners location. If the speakers Next, how do you create listeningfiltering techniques to transform mono- are several yards apart, as with most au- room acoustic effects for the rear surphonic audio. Patent searches and con- dio/video receiver setups, the sweet spot round speakers? First, you must differenference papers reveal some clues on these is naturally wider; however, the center- tiate among early reflections, echo, and advanced techniques, but you must sign perceived location of audio material that late reverberations (Figure 3). These a nondisclosure agreement to get all the they share is indistinct. Adjacent trans- three phenomena result from sound redetails (see sidebar Havent heard ducer placement, such as with speakers flecting off objects before entering listenenough?). In contrast, some simple and on either side of a television tube or com- ers ears, but the auditory system percheap pseudo-two-channel algorithms puter display, creates a narrow sweet spot ceives them in different ways. Early simply subdivide the audio into multi- but a well-defined center location for di- reflectionsthose lagging behind the frequency bins, allocating the bins alogue and other common material. To original sound by as much as 30 msec among the various channels. enhance the center-channel characteris- enable the ear and brain both to locate the Now, youve got a two-channel audio tics, you determine what material the sound source and to perceive the room clip, either as an original recording, front left and right channels share and dimensions. Their amplitude depends on which, lets assume, contains no matrix- then emphasize it using HRTF (head-re- the reflectivity of objects the sound waves encoded center and mono-surround in- lated-transfer-function) transforma- bounce off before entering your ear, formation, or one resulting from the ear- tions. Ideally, center-channel informa- whereas their delay is a function of room lier mono-to-two-channel conversion. tion transmits through a dedicated width, depth, and height and of the presNext, you might want to first alter the speaker, because a phantom center, ence of reflective objects within a room. sweet spot, or region within which a which you create by coupling the left and We perceive direct reflectionsthose person listening to the audio can hear right front speakers, exhibits timbre that that bounce off only one or only a few both channels. For a computer user sit- differs from the real thing. objects before entering the earas ting before a display, the sweet spot can In addition to emphasizing the shared echoes beyond 30 msec; hence, these rebe narrow and shallow, which often re- information, you might also want to flections tend to degrade the sound. Consults in more accurately perceived sound broaden the audio image created by data versely, sounds that have been reflected positioning and which is particularly ap- in one channel and not in the other. One many times with attenuation at each repealing with 3-D games. A home-theater quick and dirty means of accomplishing flection point come at a listener from all setting, in which listeners may be in 10- this goal involves inverting the phase of directions. Many of these low-amplitude,

68 edn | October 25, 2001
DIRECT diffuse late reflections, or reverberations, SOUND simultaneously arrive at the listener. A EARLY certain amount of reverberation is genSOURCE REFLECTION EARLY erally desirable. Think, for example, how REFLECTIONS LATE much richer your voice sounds when you REVERBERATIONS LEVEL sing in the shower. Conversely, in rooms DIRECT LISTENER with little reverberation, the sound you hear is often unpleasant. EARLY You create the perception of reREFLECTION Figure 3 flections and reverberations by TIME employing RAM-based delay lines, along with signal processing, to modify the au- The auditory system perceives early reflections, echo, and late reverberations in different ways dio as a real-life reflection would. Differ- (courtesy Focal Press). ent delays, intensities, and spectral transformations can create the illusion of a operations. Mix and play back 100- and to create the illusion of more speakers cavernous concert hall or an intimate jazz 150-Hz tones in an audio-editing pro- than actually exist. Recall that, in the ITU club. Artificial ambiance seems appealing gram, and youll also hear what sounds configuration, each ear of each listener in theory, but reality is often under- like a 50-Hz tone. At the high end of the perceives sounds that originate from all whelming, especially if a listener exag- frequency spectrum, Kenwoods (www. six speakers. Interaural time differences gerates the effect. An acoustical model kenwood.com) Supreme technology in- play a key role in locating sound sources that sounds good with a symphony or- terpolates high-order fundamental tones of frequencies of 1 kHz and lower (Figchestra, for example, might sound horri- from lower-order harmonics that have ure 4). Conversely, interaural intensity ble with a solo pianist or vocalist. Short survived lossy encoding. Thomson Mul- differences are the primary means by cuts in memory and processing power to timedias (www.thomsonmultimedia. which the auditory system locates the reduce system cost and power consump- com) MP3pro compression scheme takes source of sounds higher than 1 kHz. tion leave the resulting reverberation similar advantage of harmonics to shift When a sound source is close to a listensounding artificial. Artificial ambiance high frequencies beneath harms way er, the spherical outward radiation of also clashes with other ambiance already during encoding, subsequently restoring sound emanating from an off-center in the audio. them during decoding. source and the resulting level difference An alternative approach is more proAt this point, youve created front left between the ears of the listener are addicessing-intensive but more authentic in and right channels, a center channel, one tional factors in determining location. its results. It involves analyzing and ex- or more rear channels, and maybe even The unique shape of each persons head tracting this existing ambiance and send- a subwoofer channel. Todays most com- and shoulders is an appreciable barrier to ing the reflections and reverberations to mon ideal reproduction setting is a six- and spectral modifier of sound waves, the rear channels. Audio engineers cap- speaker setup like the one that the and the spacing between ears, shape of ture this ambiance during the original ITU (International Telecommunication each ear, and shape of the auditory canal recording by using binaural, hypercar- Union) defines (Figure 1c). Aesthetic leading to the eardrum are also key idendioid, or omnidirectional microphones, considerations can drive subwoofer tifiers of a sounds direction. or they can add the ambiance to the au- placement because the channels sound HRTF curves summarize these listendio during mixinga high-tech version isnt directional. However, if you place er-specific phenomena. By transforming of singing in the shower. Directing re- the subwoofer against a wall, especially in a sounds timing and frequency spectrum flections and reverberations to the rear the corner of a room, you will perceive its using the HRTF data, you can theoretispeakers can be an effective arrangement sound as the loudest. For music repro- cally place a sound anywhere in space in an automobile. The traditional auto duction, all other speakers should have around a listeners head using only two audio configuration replicates the right full frequency response, and the sur- speakers. Some algorithm developers and left audio channels in both the front round-sound speakers should be direct- claim that an audio engineer can employ and the rear speakers. This so-called ly radiating. Conversely, for movies, the basic amplitude-based rather than the dual-stereo configuration significantly center channel often carries only dia- more elaborate HRTF-based alternatives diminishes the listeners appreciation: logue, and the surround-sound speakers to give the illusion of a sound moving The speakers not only are in poor loca- find most use in reproducing special ef- from front to back. Some algorithm detions but also bombard your ears with fects. Such home theaters often employ velopers claim that if an audio engineer destructive crosstalk from both the front bipole or dipole surround-sound speak- has employed basic amplitude-based efand the rear. ers for immersiveness. Unfortunately, fects rather than more elaborate HRTFAfter creating additional audio chan- they also often trade off speaker fre- based techniques to give the illusion of a nels, you might also want to enhance the quency response and other characteris- sound moving from front to back, an perceived low frequencies of the audio to tics to reduce cost. HRTF-aware virtual-surround-sound compensate for anticipated bass-defisystem can deliver a more realistic result cient speakers or the high frequencies to TWO-CHANNEL VIRTUALIZATION than a front- and back-speaker combicounterbalance the effects of lossy comWhat if having more than two speak- nation. Because our ears are on the sides pression. Harmonics play a part in both ers isnt an option? In this case, you need of our heads, were particularly sensitive

70 edn | October 25, 2001
to shortcomings in phantom midlaterFRONT al speakers, which HRTF transCENTER Figure 4 forms account for, particularly if SOURCE those transforms also adjust for interaural crosstalk effects. PATHS TO LISTENER'S Conversely, when a front-placed TWO EARS rsin speaker pair creates virtual surround, its difficult to spin a psychoacoustically alr tered sound completely around to the r back of your head and equally challenging to communicate sound source height LEFT RIGHT EAR EAR cues. If youre playing a video game, your head is stable and centered, and you also FALSE have visual suggestions in the form of SOURCE on-screen objects to aid in locating sound. However, in the absence of visual assistance, you may experience frontREAR CENTER versus-back location confusion. In real life, youd turn your head to help find the Frequency-dependent interaural time and intensound. But, because the sounds location sity differences enable the ears and brain to is virtual and the HRTF-transformation determine a sound sources location, but lackeffect depends on your heads orientation ing collaborating visual or head reorientation to the real speakers, this reflex reaction cues, these differences may result in front-vermakes the problem worse. sus-back confusion (courtesy Focal Press). Ideally, HRTFs should be customizable for each listener, because we all have by, it does not serve as a point source.) different head shapes, ear spacings and When audio engineers mix music and contours, and concha openness and movie soundtracks, they assume that depth. Multilistener environments ob- users will play the end results on tradiviate the effectiveness of such cus- tionally placed speakers. As noted, even tomization, however. Some HRTF algo- in a simple two-speaker configuration, rithms also adapt their responses to each ear senses not only the output of its plane sound sources, to nearby sources corresponding front speaker, but also a where spherical effects are important, spectrally modified and time-delayed and to listeners who are not in the sweet version of the opposite speaker. Howevspot. (For example, when a train speeds er, this acoustic crosstalk doesnt occur with headphones, in which each channel is isolated to only a single ear. Uncompensated audio, intended for speakers but played back through headphones, produces an annoying in-the-head effect, compared with a compensated alternative, which mixes each channel with a time-delayed and spectral-modified version of the opposite. HRTF transformations for virtualization with more than two speakers over headphones take a different form from their two-speaker equivalents. Because the speakers are on the ears and follow the head regardless of its orientation, sweet-spot issues disappear, more realistically achieving full 360 sound placement and convincing vertical positioning. However, the direct coupling of the transducer to the ear effectively removes outer-ear HRTF effects, which can make it difficult to achieve consistent results from user to user. Headphone-virtualization algorithms are also sometimes complicated because they cant use the acoustics in the listening room to help recreate the original environment. On the other hand, if the listening-room acoustics are detrimental (echo-conducive, for example), you should probably begin with a clean slate. To achieve more realistic reproduction of sound-source height (think of a rocket taking off) and to improve the realism of front-to-back and back-side-toside sound source movement, audio

FOR MORE INFORMATION.

For more information on products such as those discussed in this article, go to www.ednmag.com and click on the Reader Service link under the Tools & Services section. When you contact any of the following manufacturers directly, please let them know you read about their products in EDN.
Ambiophonics www.ambiophonics.org Enter No. 341 Bell Labs www.bell-labs.com Enter No. 342 Creative Labs www.creative.com Enter No. 343 DFX Power Technology www.fxsound.com Enter No. 344 Digital Theater Systems www.dtsonline.com Enter No. 345 Dolby Labs www.dolby.com Enter No. 346 Fosgate Audionics www.fosgateaudionics.com Enter No. 347 4Front Technologies www.oss3d.com Enter No. 348 Harman International www.harman.com Enter No. 349 Kenwood www.kenwoodusa.com Enter No. 350 Lake Technology www.laketechnology.com Enter No. 351 Lexicon www.lexicon.com Enter No. 352 Lucasfilm THX www.thx.com Enter No. 353 Meridian Audio www.meridian-audio.com Enter No. 354 QSound Labs www.qsound.com Enter No. 355 Sensaura www.sensaura.com Enter No. 356 Sony www.sony.com Enter No. 357 Sorient www.softamp.com Enter No. 358 Spatializer Audio Laboratories www.spatializer.com Enter No. 359 SRS Labs www.srslabs.com Enter No. 360 Surround Associates www.surroundassociates.com Enter No. 361 Thomson Multimedia www.thomsonmultimedia.com Enter No. 362 Tomlinson Holman Labs www.tmhlabs.com Enter No. 363 SUPER INFO NUMBER For more information on the products available from all of the vendors listed in this box, go to www.ednmag.com, click on the Reader Service link, and enter no. 366 3D Audio Immersion www.3dai.net Enter No. 364 WaveArts www.wavearts.com Enter No. 365
72 edn | October 25, 2001
pioneer Tomlinson Holman advocates a 10.2-channel system (Figure 1d). His approach builds on the 5.1-channel ITU standard with the addition of dual side that is, midlateralchannels, a rear center channel, dual subwoofers, and two height channels. The height channels target reproduction over speakers 45 horizontally and 45 vertically away from the listener. As an interim step to 10.2 channels, several 6.1-channel approaches are gaining popularity. Dolby Digital Surround EX, which Dolby Labs developed with Holman, matrix-encodes a rear center channel, as does DTS-ES Matrix. DTS-ES Discrete employs a distinct rear center channel; Digital Theater Systems designed its bit-stream format to be backward-compatible with respect to additional audio channels; higher precision sample sizes; and higher sampling frequencies, such as the upcoming 24-bit, 96-kHz DTS. And THX Ultra2 technology extracts seven full-range channels and a subwoofer channel from 5.1-channel source material. References 1. Guttenberg, Steve, and Brent Butterworth, Stereo vs 5.1: Is more more.or less?,Stereophile, August 2001, pg 49. 2. Dipert, Brian, Bassless buzz impairs advanced audios image, EDN, May 24, 2001, pg 20. 3. Dipert, Brian, Security scheme doesnt hold water(marking), EDN, Dec 21, 2000, pg 35. 4. Dipert, Brian,Digital audio gets an audition: part two, lossy compression, EDN, Jan 18, 2001, pg 87. 5. Dipert, Brian, Digital audio breaks the sound barrier, EDN, July 20, 2000, pg 71. 6. Rose, Jay, Stereotypes, DV Magazine, September 2001, pg 98. 7. Layton, Leonard, Surround sound takes to the air, IEEE Spectrum, August 2001, pg 56. 8. Kraemer, Alan, Two speakers are better than 5.1, IEEE Spectrum, May 2001. 9. Rumsey, Francis, Spatial Audio, ISBN 0-240-51623-0, Focal Press, Woburn, MA, 2001. 10. Coulter, Doug, Digital Audio Processing, ISBN 0-879-30-566-5, CMP Media, Lawrence, KS, 2000. Acknowled gments Thanks to Andrew Reilly from Lake Technology, Scott Willing from QSound Labs, Randy Roscoe from Spatializer Audio Laboratories, and Alan Kraemer from SRS Labs, for their assistance and feedback. Authors bio graphy Technical Editor Brian Dipert experiences surround sound every time the four dogs lying around him in his office sense that the mailmans at the front door. You can reach Brian at 1-916-454-5242, fax 1-916-4545101, bdipert@pacbell.net, and see his four-legged childrens pictures at www. bdipert.com.

74 edn | October 25, 2001

EARS-ON ANALYSIS

After reading this article, you may get a sense that Im a surround-sound enthusiast. Well, youre right. Im surrounded by surround sound in my home, office, and car. My living room unfortunately and typically has less-than-ideal acoustics. It does have a 5.1channel speaker setup, comprising Dahlquist front speakers, a KLH (www.klhaudio.com) center and direct radiating rear surround set, and an AudioSource (www.audiosource.net) SW Fifteen subwoofer. Driving the speakers is a Technics (www. panasonic.com) SA-DX1050 audio/video receiver, supporting Dolby Digital and DTS decoding, and connected to a Toshiba (www.toshiba.com) SD-2108 DVD player (with Spatializer Labs (www.spatializer.com) N-22 virtualization) and Mitsubishi (www.mitsubishi-tv.com) SRHD5 HDTV receiver via Belkin (www.belkin.com) coaxial and optical cabling. The four desktop PCs in my house contain Analog Devices (www.analog.com), now-defunct Aureal, Creative Labs (www. creative.com), and Zoltrix (www. zoltrix.com) audio subsystems, each providing virtual-surround capabilities. My Labtec (www. labtec.com) LCS-2414 speakers include a 3-D-stereo adjustment knob, and Ive installed a variety of audio-enhancement plug-ins for MusicMatch Jukebox, Real Player, Real Jukebox, WinAmp, and the Windows Media Player. Many of the PCs DVD software packages support Dolby Headphone, all can virtualize Dolby Digital and Dolby Stereo, and InterVideos (www. intervideo.com) WinDVD 3.0 even decodes DTS. QSounds (www.qsound.com) UltraQ and SRS Labs (www.srslabs.com) WOW Thing deliver audio enhancement in hardware. In my car, I have a 4.1-channel speaker system. I have no desire to cut into the dashboard; therefore, I have no center speaker. I also have a Sherwood (www.sherwoodusa.com) XCM7370 AM/FM radio and CD player, X-DTS80 audio decoder, and an AX-6275 multichannel amplifier. The X-DTS80 decodes DTS audio CDs, and it also supports SRS Labs CircleSurround algorithm for one- and two-channel audio sources. It almost makes me wish I had a long commute to work.