Beyond the stereo sound Part 2 - Never 2 speakers without the 3rd dimension...
Nov 10, 2003 - by Franck ERNOULDFollowing the first part dealing with the multi-channel sound playing, we will now talk about the 3D-sound universe applied to music : the art and the way to create the illusion of 3 dimensional music on 2 speakers or even in the headphones... Let's put a definition : the 3D sound is dealing with the process that enables, owing to the reproduction of 2 transductors, the simulation of a spatialized sonic field by playing with the phasing relations and the psycho acoustic properties of our hearing system. For a man, the sonic perception on 3-dimensions is based on 2 parameters, the time and inter-aural intensity difference (between the 2 ears if you prefer). The first depends on the location of the head according to the source, whereas the second relies on the mask effect from the head and shoulders. Reflexions on the pavilion of the ear are also of importance : we can imagine that a sound coming from above the head is not filtered by the pavilion the same way as a sound coming from the front or the back. Hence the timbral signature which, combined with the other information of time, enables to our brain to find the origin of the signal. All this to say that the multi-channel playing is not the only solution to go beyond the stereo... Some other solutions do exist : they use algorithms simulating time-lag/mask phenomena that we just told about, in order to produce some more or less realistic spatialization effects. Some work better with headphones, the others in the speakers, some let you choose... Let's have a look at the most famous.
|My stereo is wide.|
Within the sixties, making some stereo monophonic recordings using comb filters was a fashioned sport. We used to cut the audio spectrum into very thin bands alternatively sent on the left and on the right, thus creating some pleasant width effect, but usually wholly artificial. At the beginning of the seventies, with the progress of stereo hi-fi systems came some raw "stereo enhancers" (a process used widely on stereo-cassettes in order to compensate the tight width between the speakers !).
Some might remember with emotion the "Dimension D" of Roland at the beginning of the eighties : it was a magical box that gave an amazing width to a mono signal - coming from a synthesizer or an organ for instance. Playing with the phase, adding a slight chorus and a reverb, it is still used in many studios. Stereo enhancers must have a strong life, as long as they are still being developed, especially as software. For instance there is the Wave PS22 Stereo Maker plug-in.
At the end of the sixties the first search attempt for a "real" 3-dimension sound appeared... especially in the laboratory of the so called Zuccarelli. Guess who is trying the thing owing to a widespread commercial use ? Pink Floyd of course, who are always on top of the improvements - right in 1967, the group was playing on a quadraphonic sound system with "azimuth coordinator" ! Nevertheless, on the notes of the layout from "The final cut", released in march 1983, it shows "holophonics by Zuccarelli Labs". With the headphones, some effects (planes, shells), some arpeggiated guitars... give the amazing impression to "bath into the sound". The bassist/singer Roger Waters is fond of the process, so that he is using it again on his own album. Mickael Jackson, Stevie Wonder, Lionel Richie, Coca-Cola, Wes Craven or the MGM will follow.
The technique is still employed, as we can notice on the web site www.holophonics.com. According to the inventor, the process is to the ear what the hologram is to the eye. Thanks to that, "the sound speaks straight to the brain, thus it extends the perception to some others sensitive areas. The holophonic sound takes theoretical and practical applications within many fields of human life - from physiology to psychiatry, to religion or to philosophy".... What a large program ! As an anecdote, the recording is done using a so called system Hugo, which replaces the microphones and probably contains electronic circuitry for processing, Zuccarelli remaining very discreet on the topic.
Roughly at the same time, we can see the first work of what will become the QSound. Roger Waters (again...) is one the enthusiast adepts of the process... His last solo album released in 1992, "Amused to death" contains a lot of this, but also some albums from Madonna ("Ultimate Connection") or Sting ("Soul Cages"). There we are dealing with the processors which treat some pre recorded signal, whether it is monophonic (a single instrument) or stereophonic (a whole mixing for instance). The engineers from QSound declined their knowledge into many ways : multiple integrated circuitry for TV sets, arcade games, personal computers..., but also software versions (TDM plug-in, DirectX, or on the internet with the iQSound), not even talking about the development kit QMDX for Windows. The sound engineer Ludovic Lanen (Francis Cabrel, Didier Lockwood, National Orchester of Jazz...) is actually using the QSound plug-in QSYS TDM on his Pro Tools. " I'm using it sparing and punctual unless the treated sound pops out of the arrangement, which is useful only on small gimmicks, noise effects, or straight for huge effects. On a mixing, things such as the Waves PS22 are more easily used. With a too thick recorded instrument, the sound does not pot out of the speakers, but is quietly spreading, without putting the focus on it". Let notice that when listened in Surround mode, a monophonic guitar treated with the PS22 gains a surprising volume ! Furthermore, in the category "do yourself your psycho-acoustic effects", Ludovic gave us a tip which he recently used with the National Orchester of Jazz : "As the ONJ got 2 drummers, I had put 2 ambiences, one quite "thin" and the other very wide. Mixing both, I noticed that some quite odd phase relationships appeared, and with dosage refinement we could hear some sounds popping out of the speakers : the effect is less pronounced than with QSound, but it is worth being tested.
Michel Geiss (mastering engineer... and also member of the staff, who has worked with Jean Michel Jarre for many years), is more perplex : "I know what are those 3D processors and what they can do, but I almost never used them into final works : neither for Jean-Michel, in spite of many trying, nor for personal projects such as that of Gaumont Grand Ecran (see KBHSR ní2, note from the writing team)... Lately, I attended the live surround mixing of Michel Sardou, carried out by Bruno Mylonas. They tried with a RSS Roland, but they did not reach good results : there were too much phasing troubles because of the Dolby encoding... Therefore I noticed that those effects were mostly difficult to use "daily". They are unknown at the mastering stage... This did not prevent me from "enlarging" exceptionally the stereo of the mixing to please a group. But then the monophonic compatibility often deals with problems. Especially FM transmitters which alter the phase and usually put down the spatialization when those are subtle. To tell the truth, I think that those 3-dimension techniques are a bit old fashioned as the audio DVD is coming out as well as the real and not replaceable multi-channel. "
Some of you may have read an article from David Korn in the 2nd issue as it was still included in Keyboards magazine, dealing with the first process designed by Roland : Sound Space (well known as RSS). Thanks to proprietary algorithms, it created phase effects which the ear uses to define the location of the sound. Right in 1992, making abuse of the process in order to show its abilities our friends Jean-Bernard Celier, Alain Etchart (Univers Sons) and Christian Sal?s (chief for Roland products) did a quite astonishing CD called "The sound of the third dimension". We've listened to it again while writing this article, and we must admit that some tracks impressed ourselves : especially jets moving from left to right, helicopter or ambulances added with Doppler effects which are staggering when listened to alone !
Once these sounds are "drowned" into a mix, we are confronted to some mask effects that alter the efficiency of the RSS. From this beginning, Alain Etchart and Jean-Bernard Celier have published a few banks with 3D treated sounds... Following this, Roland diminished the volume taken by electronics and put it into a processors the size of one rack unity, the RSS-10, then simplified and democratised the applications of the process - they put it into multi-effects (especially SRV-330 and SDE-330) or other devices such as VS-1680 for instance. Anyway, that interesting concept is quite complex to operate, it needs the use of some mac or pc helper software in order to tell to the machine the desired moves. So is the RSS a gadget ? Not really... At the end of 1995, Europe 2 used it widely for its broadcast packaging and many disks have used it even not mentioning they were using it. The fact that the process still is on sale is a proof that it meets its public. On the contrary, as it was not declined in plug-in format, Roland is kind of exception.
The Spatializer is aimed at professionals and also at a wide public. The "hardware" studio version has been released for many years : interviewed few page further, the engineer/producer Elliot Scheiner used the Pro SpatializerTM on the album "The Glenn Frey Solo Collection" released in march 95, but also on "Tambu" from Toto. At that time he said "we were able to create a natural and noticeable width on the voice of Glenn. They come literally towards the listener, thus putting the right intimacy for the song spirit". Some other products have been treated by the SpatializerTM : "Le roi lion", "Q's Jook Point" from Quincy Jones, "Woodstock 94" and an album live from Barbra Streisand.
The Spatializer knowledge is also declined in Pro Tools TDM format stereo plug-in (the PT3D v1.2, a demo version is available for download on the web site www.spatializer.com), also Audio Suite and DirectX formats... even for the M5000 t.c. electronics... Talking about TDM, let notice that Aureal (www.aureal.com) also propose its 3-dimension treatment plug-in. Called A3D Pro, it is the right follower for the ProTron. With its Spectral Design division, Steinberg is not apart : the German producer is selling FreeD which simulates 3D location in real time (www.steinberg.net/products/freed.html). Still with Cubase VST, let's finally talk about the Ambisone 3D plug-in from Prosoniq, a demo version of which is once again available for download on the web site of the editor (www.prosoniq.com/html/mvstambisone.html). These are some of the most famous software's, the list may not be exhaustive !
SRS, standing for Sound Retrieval System, is also using psychoacoustics techniques in order to improve the spatial rendering. First included into the Vivid 3D Theatre, then to the Vivid 3D & 3D plus (see Home Studio 36 former formula, in Keyboards magazine 91), Sony was interested in it as well as Kenwood, RCA, Nakamichi, Sharp, Packard Bell, NEC, Kurzweil, Pioneer, Thomson, Seiko, Yamaha & Motorola. After that, the system reached the top of the interest as Apple, the firma from Cupertino bought the licence few years ago : the SRS Sound appeared on the Mac Performa 5500, then it was used on all the actual iMac (see the screenshot). While browsing the www.srslabs.com site, we've learnt that the process was cutting the stereo signal into elementary parts which could be isolated and treated in order to alter its location and replace it to a correct perspective according to the direct sound". We can't stop improvements...
|Some 3D for my PC.|
After this panorama of some of the most famous 3D processes, let's have a look at the computer universe. Let remember : at the beginning of the PC, the audio was the poor side of the games. To the rudimentary graphics corresponded raw sounds : that was the time when some poor piezzo-electric transductors of few square millimetres size were emitting with difficulty some anaemic "bips" in the best cases. The more graphics, animation and movement rendering were improving, the more the standstill of the audio side became inconvenient, and was therefore too much visible ! Thus we dug the brain in order to do better, thanks to real speakers, real sounds, real music and even stereophony... and more if possible. Indeed, as Windows 95 was already well there 3 years ago, a few companies such as Intel created some spatialization algorithm - so called codecs (coding/decoding).
That was a godsend for computer games makers, which ran wildly onto them : the movements of the sounds could actually follow those of the image. When immersion takes ourselves... With his 2 amplified speakers plugged in the stereo mini-jack plug of the sound card the user was "surrounded" by the sound. Hence the big sensations and the seasickness... Therefore, the editors doubled their efforts in order to create some sound more ready for the spatialization : the sound design followed the technology. In a different way, some internet sites also used the codecs - sometimes you can choose to listen into the headphones or the speakers - it is even included in the Shockwave plug-in. The next step started with the French Guillemot releasing the Maxi Sound cards (Home Studio 2, Home Studio Pro 64...)... with 2 stereo mini-jacks outputs aimed at plugging 4 speakers - 2 in the front, 2 in the back.
This is a kind of poor computer version of the surround sound, the same signal is sent to the two pairs of speakers, but slightly delayed at the rear in order to give a spatial sensation. It would have been pity to stay at there... The big Microsoft must have thought so, and after observing the 3D sound cards then multi-channel ones, a set of API (Application Programming Interface, routines used by software) were developed for Windows, dedicated to the location of the sound sources through the four speakers. Into the games for instance, some contextual audio events are added to the music (which is in stereo), that is to say events triggered by the actions of the player, and that the developers can thus place anywhere thanks to the routines, without worrying about the hardware of the player... or almost. The API itself is aimed at managing and communicating with the audio card drivers in order to let it do the job (unless there is DirectSound 3D compatibility and four audio inputs). Named DirectSound 3D Hardware Acceleration, the API are wholly included into the DirectX standard. Thanks to them, up to forty-eight sounds can be located in the 3 dimensions at the same time without using CPU resource...
Despite all those means, many people remained sceptical : people already have it hard to put two worthy speakers on their PC, why would they plug four of them ? Merely because they are small and cheap (around 30,5 euros the 2 speakers). That's why the Home Studio Pro family gained a real success. SoundBlaster soon released the Live soundcard !... also equipped with two stereo jack outputs. The producers even propose the 5.1 sound, thanks to a bundle including the card, fives speakers and a woofer, all this for around 300 euros. How to think about having a DVD at home ! Talking back about the games, all the best-sellers now available on the screens - Half Life, Tomb Raider III, Douarche Killer IV... - have got a great 3D sound. More and more outstanding, the sound cards start to include DSP functions, thus enabling some chorus, reverbs effects, etc. (especially thanks to the EAX standard), some are even creating a false 3D on old programs, precisely they were proposing a list in which we could choose the game in order to apply the preset created by the game editor - also with a personal equalization ! Thus the Isis Maxi Studio from Guillemot (se the trial in this issue) includes 400 optimisation patches, corresponding to games of 4 or 5 years old. This does not let us be younger...
Unlike some foreseeing, it seems that with the low price and the quite small size of the PC multimedia speakers, the wide public has easily joined the spatialization and multi channel rendering concepts. It will be more difficult with hi-fi... Let's hope that the computer does not replace the stereo system. Listening loud to music on 30 • multimedia speakers is not recommended. Even within the software field, miracles don't exist ! Talking back about the three-dimension sound, we must have faith in it first : we have to be a "good public". Let's notice again that this is not an illusion aimed at making the brain fail (such betrayals sometimes cause the sea sickness while listening to some brave tracks using RSS or other things : usually altered in matter of phases, the sound is going anywhere, and they can cause sometimes a physical inconvenient feeling). Anyway the 3-dimension audio illusion is somewhat like the books where you can see relief after looking at it fixedly for some time : some see them easily, the others never see them, like me.
That is just the same with audio 3D. Let also remember that when played on speakers, a good 3-dimensional sound needs a precise location of the speakers as well as the listener location according to the speakers, and a good acoustic design of the room too (see the precaution use of the RSS for instance, or the article "how to put the speakers within twelve lessons", Home Studio Ní28). Anyway, however developed our imagination may be, we will never feel the presence of the sound better than with multi-channel systems. This is not by fortune that the engineer Elliot Scheiner suddenly stopped to use 3D processing after tasting the 5.1. With these good words, we give you a third and last rendezvous : we will talk about some more serious aspect of those researches, the virtual acoustic which is also using lots of resources to go beyond the stereophony...
Soundfield : 3D sound taking!
From what we know, only Soundfield is building microphones such as the SPS422. Its 4 tetrahedral located capsules take as many modulations which enable to carry out the x, y and z components of the real 3-dimensional sound thanks to processing which are too complex to be written here (B-Format, something well appreciated in virtual acoustic). Mixing the signals taken by the capsules in different ways, we can also work with the traditional stereophony with a real advantage : it is possible to change from a distance the directivity or the angle of the sound taking thanks to the external processor included with the microphone. An example of use of it : putting it forever above the tennis court in Paris-Bercy in order to operate from a distance the width of the stereo/the directivity of the couple capsules and thus optimising the rendering of a Dolby Surround broadcast. Lately, after Abbey Road and CMS, the Real World studio bought a SPS422, which musical application are obvious.
Headphones with Surround.
The question is simple : how to listen to Dolby Surround sounds owing to the 2 ears of the headphone, sound which is theoretically played by 5 speakers at least (left, centre, right, left rear, right rear) ? But this way avoids to wake up its relatives and neighbours while watching "the return of the Jedi " or any James Bond around midnight ! What solution ? Recycle one of the algorithms described in this article in order to simulate the fact that the mono or stereo surround channels come from the back of the head of the listener and also to enable him to alter the width of the front stereo. This is a kind of "virtual Dolby" ! A musical mode can give the impression that the sound comes come the front of the head instead of the centre of the head skull - a usual drawback of headphones. One of those first available headphones, the Lucas Sennheiser, consists in a box with an inline input and a headphone output. We can even choose our location in the "virtual theatre room" : this is something aimed at appreciating quietly a Dolby Surround mix ! A brand new Sony wireless model, the MDR-DS5000 (about 534 • required) is even dealing with an analogue Dolby Pro-logic modulation or a digital AC-3 format. That is the most straight way from a DVD player to the ears !
Furthermore, the Dolby Labs just cooperated with Lake DSP, a leader in the field of virtual acoustic, in order to create the "Dolby headphone". Some even alter the physical concept of the headphones... Thus Sennheiser again has released the Surrounder. Designed specifically for the games, this "headphone" with 4 real transductors is put onto the shoulders (100dB SPL, bandwidth from 45 Hz to 20kHz). No longer need of 4 speakers to boost the sound ! There is a top version aimed at movie lovers too, with a Pro-Logic decoder included (for Dolby Surround). Again for a Dolby Pro-Logic broadcast, AKG is selling the Hearo-777 : the Austrian has recycled its IVA process (Individual Virtual Acoustics), which is used in engines such as CAP 340 M Creative Audio Processor, the BAP 1000 Binaural Audio Processor or the IVM-1 In Ear Monitoring System. No more trouble of optimal location... even if there will always lack the physical impact from the bass sound, that only a subwoofer is able to create while emitting enough energy in the room and received by the ribcage or the stomach of the listener.
On the 15th March, the software producer Human Machine Interface Inc., based in Oregon, presented InMotion 3D Audio Producer. As it is called, that software is aimed at studios, musicians or sound designers and web site sound makers dealing with 3D audio. It works in real-time, running on windows 95, 98, NT. It proposes location effects of course (you can put the virtual speakers anywhere in the virtual room), but also Doppler effects, air absorption, obstacle obstruction, making reverbs... The demo version can be downloaded at www.humanmachine.com !
(c) Franck ERNOULD