Search:
Submit Search


Beyond Stereo to multiple Speakers

Speakers, amplifiers, headphones, multi-channel audio, reverb units, mixers, wiring, ...

Re: Beyond Stereo to multiple Speakers

Postby mdyde » Mon Sep 27, 2010 3:56 pm

Hello David,

For such people in their living rooms, multi channel is not necessary and top rate results are available in other ways.


Perhaps you could explain how your modified speakers are able to mitigate/address the 'signal mixing effect' and beating of note intervals within the mixed-down signal fed to them?
Best regards,
Martin.

[Please use email or the Contact page if you need to contact us privately, rather than private forum messages.]

Image
User avatar
mdyde
Moderator
 
Posts: 10644
Joined: Fri Mar 14, 2003 2:19 pm
Location: UK

Re: Beyond Stereo to multiple Speakers

Postby David Pinnegar » Mon Sep 27, 2010 5:40 pm

mdyde wrote:Hello David,

For such people in their living rooms, multi channel is not necessary and top rate results are available in other ways.


Perhaps you could explain how your modified speakers are able to mitigate/address the 'signal mixing effect' and beating of note intervals within the mixed-down signal fed to them?


Dear Martin

I'm very willing to assist people privately but do not wish to provide secrets in public to be viewed and exploited by the commercial competition to Hauptwerk and the enemies of pipe organs as to how to make electronic organ simulations sound brilliant.

Suffice to say that there are ways of ensuring that amplifiers work only in super-linear regions and that speakers work only within linear constraints and without phase shifts and delays between frequencies.

I had an amusing experience a couple of weeks ago when using some interesting speakers for public address at the village fete: my wife was walking the dogs whilst I was testing the microphone and we had a true "his master's voice" event before our eyes, the dogs becoming very confused. Good sound coverage was required over an area of about 5000m2 in the open air. A child walking past asked me if there were real musicians playing an accordian and flute in the porticos and after the fete someone remarked about the speakers as he had observed that those nearby weren't deafened but that the high quality sound had carried into the distance.

Best wishes

David P
http://www.organmatters.co.uk
David Pinnegar, B.Sc., A.R.C.S.
User avatar
David Pinnegar
Member
 
Posts: 437
Joined: Thu Jan 03, 2008 8:13 pm
Location: Sussex UK

Re: Beyond Stereo to multiple Speakers

Postby Eric Sagmuller » Mon Sep 27, 2010 6:58 pm

David Pinnegar wrote:The real extraneous frequencies that _are_ audible on thirds in particular are not IMD introduced frequencies but the inharmonic beat frequency between the two notes. Equal temperament is the bane of musical life: a major third should be the 4th and the 5th harmonic. This means that one note will be 4 times a fundamental frequency and another note will be 5 times the frequency, the difference being the fundamental frequency itself.

But in equal temperament, the thirds are wider, sharp, further apart, so the difference between the frequencies, the beat frequency, is higher. It's audible when you're tuning a piano with your head close to the strings, and it's heard when you're near to organ pipes and speakers alike. It happens to be two octaves below the lower note of the third but because the interval isn't pure and the beat frequency is higher, the note is a quarter tone sharp. Foul. It's why certain organ voicers of whom I'm aware consider that anyone who wants a tuning system which does not include 8 perfect thirds in the scale should be burnt at the stake.

David P


Yes, this is what I hear that bothers me quite badly on certain days more than others. Oddly enough with my real pipe organ here, I don't hear these beats nearly as much.

What temperament do you suggest David? I've tried many of them in HW and although some may sound better in a certain key, in other keys they sound quite bad. So far out of them all I find myself coming back to the original organ tuning which is equal in the case of the Hinsz, Bovenkerk organ.

Someone on the list had some other temperament they came up with that they claim mitigates many of these beats, but I haven't had the time to download it and try it.

Eric
Eric Sagmuller
Member
 
Posts: 879
Joined: Fri Apr 11, 2008 5:18 pm
Location: Bellefonte, PA USA

Re: Beyond Stereo to multiple Speakers

Postby toplayer2 » Mon Sep 27, 2010 7:47 pm

To demonstrate that IMD is far from subtle, I have created a small demonstration. Sound Forge was used to generate two notes, C5 and E5, using pure sine waves. In the first case, the notes are mixed to mono and thus will share the same audio channels during playback:

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdMixed.wav

The second case is to separate the two notes so that C5 is only in the left channel and E5 is only in the right channel.

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdLeftRight.wav

The level of IMD that will be heard in the first case is dependent on each listener's particular audio set up, but I am confident that with even the world's finest audio system, there will still be very audible IMD that will probably sound like a dissonant triad rather than just two notes. This effect completely disappears when the notes are segregated into independent channels. No IMD was introduced by recording the demonstration WAV files as they were generated completely in the digital domain. If anyone has lingering doubts or suspects trickery, just mix the left and right signals from the second file and the exact same result as the first case will be had.

There is nothing wrong with people doing the best they can given budgetary and space constraints, but to deny the indisputable superiority of multiple channels for reducing IMD (and signal mixing effects) strikes me as simply not being willing to accept some fairly obvious and basic facts.

Best regards,
Joe
User avatar
toplayer2
Member
 
Posts: 1055
Joined: Mon Jul 23, 2007 3:15 pm
Location: Michigan, USA

Re: Beyond Stereo to multiple Speakers

Postby Jim Reid » Mon Sep 27, 2010 8:20 pm

Joe,

I hear absolutely no difference between your two recordings, the
mixed or the left, right separate speakers.

Perhaps that is why I am so happy with my two channel stereo
HW audio system. Guess David P.s system would do me no good
or improvement for me at all. My ears are plumb IM deaf.

Though my dynamic hearing still seems superb, at least
based upon wide dynamic range CD music listening.
Jim Reid
User avatar
Jim Reid
Member
 
Posts: 1764
Joined: Mon May 17, 2004 7:50 pm
Location: Kauai, Hawaii

Re: Beyond Stereo to multiple Speakers

Postby David Pinnegar » Mon Sep 27, 2010 9:48 pm

Jim Reid wrote:I hear absolutely no difference between your two recordings, the
mixed or the left, right separate speakers.


Hi!

Yes - through my computer speakers I find the same, except for some reason the MajorThirdMixed file is less loud.

If the separate tones through separate speakers really is that, the IM distortion is that which happens in our ears on loud notes. It is this triad effect that makes equal temperament so foul and strident as if the bottom note is sounded, a discord occurs. This is the joy of playing Couperin and de Grigny in Meantone where Tierces and Grand Tierces on the pedals can be added at will created great sweetness. I have always said that Tierce en Taille refers as much to sweetness in the body as technically the Tierce in the Tenor.

Perhaps that is why I am so happy with my two channel stereo
HW audio system.


Yes - that's what old Percy Vickery used to say and, within limitations, I have come to agree with him.

A few weeks back, I was bothered by beat-notes from the electronic organ and thought that there was something wrong with my amps or speakers . . . as it's something that never _normally_ troubles my system and then realised that the beats were only coming through the speakers through which electronic reverb had been processed. So added reverb at a high level will sound artificial and if one wants reverb, wet samples will be much more successful.

Guess David P.s system would do me no good
or improvement for me at all.


To the extent of avoiding the beat notes, no, my speakers won't help because the beats happen in your ears, sometimes even in your brain - http://en.m.wikipedia.org/wiki/Binaural_beat

However, conventional speakers tend to have a series of related flaws of which you'll have taken for granted until you've heard otherwise. I'm not going to detail the root of this in public, but a pop-group did a recording session here earlier in the year and, becoming so annoyed by the sound of their studio monitors, I plugged in my speakers to be able to enjoy what they had created and after that, the record producers also became annoyed at the sound of their monitors. One of the things that was very obvious was that having recorded samples of drums and cymbals in the acoustic they were looking for, when reproducing those sounds, you can't get a cymbal out of a studio monitor 1 inch tweeter!

Temperament:
What temperament do you suggest David? I've tried many of them in HW and although some may sound better in a certain key, in other keys they sound quite bad.


Playing Couperin in Meantone is very instructive. It's very unforgiving of wrong notes and one starts to find chords and harmonies that express love and the breadth of love in a purity that Equal temperament cannot achieve, especially using pieces requiring the plein jeu, and other chords which in contrast express angst. He moves in this way between crisis points and you can hear this with the Premier Kyrie played on Plein Jeu of the deux Masses, l'une a l'usage des paroisses . .. , l'autre propre pour les Couvents, page 49 of the Guilmant edition which is the one that you'll find available free online.

Having become used to this, one starts to relish those chords which make you cringe, to create tension, only to be release in the next stroke by a chord of total purity. A friend was playing some Bach on meantone earlier today which did exactly this too. Bach exploited meantone tuning and it's good to try it out to see what he was doing and where he was applying it. It's for this reason that when Mark Shepherd performed the Dorian toccata and fugue here, in view of the modal nature of the piece, we tried it in Meantone
http://www.youtube.com/watch?v=VbwXpBcGm6Y
http://www.youtube.com/watch?v=3Uj9MORwoF0 (Don't like the registration of this so much as the fugue)

The discords of unequal temperament were intended and we are simply not used to the colour. However whilst Bach encouraged "well temperament" and personally I like Kellner, many instruments remained tuned to Meantone, so one should try it - and the results can be surprising - as well as trying Kellner too. One should remember that all composers up to and including Mendlessohn and Cesar Franck, and perhaps a few beyond, would be working with unequal temperaments whether by choice or by dint of circumstances, and we cannot properly understand their music if we don't try hearing their music on the temperaments they would have heard it played on. The French temperament may be different - D'Alembert is good and effective and among the Hauptwerk choices, St Maximin is superb.

What is really interesting about the St Maximin temperament is that it does have key colour and is capable of expressing Couperin's angst, but it's mild enough for Boellmann's Priere a Notre Dame in Ab to be utterly charming. One just wonders . . . whether perhaps Boellmann might even have been using such a temperament.


I have wanted to do a series of recordings of repertoire in different temperaments to see what repertoire is precluded by a choice of an unequal temperament were it to be specified for a new pipe organ, in the spirit of using electronics as the experimental platform, as an aural CAD as it were, for the encouragement of building pipe instruments. (If anyone would like to do an Unequal Temperament demonstration concert, then a platform is always available here . . . (OK - sorry for the downright complexity of the hardware here - I know there is technology not far away from this forum to make life a whole load simpler and possibly a good deal better))

Of course this diverges from the stereo/multi channel mythbusting thread in the domestic arena controversy, but it does go towards the sweetness and concord of the instrument that we play. Playing it at close quarters hearing beatnotes, or beatnotes generated by a reeverb process, certainly puts temperament under the spotlight.

Best wishes

David P
http://www.organmatters.co.uk
David Pinnegar, B.Sc., A.R.C.S.
User avatar
David Pinnegar
Member
 
Posts: 437
Joined: Thu Jan 03, 2008 8:13 pm
Location: Sussex UK

Re: Beyond Stereo to multiple Speakers

Postby kwbmusic » Mon Sep 27, 2010 11:06 pm

My goodness, did my simple query really start this extremely erudite discourse!
Keith
User avatar
kwbmusic
Member
 
Posts: 485
Joined: Thu Mar 29, 2007 6:06 am
Location: Melbourne, Australia

Re: Beyond Stereo to multiple Speakers

Postby jkinkennon » Mon Sep 27, 2010 11:37 pm

Some months ago when this subject was being discussed with reference to the Pykett demonstration I made the point that the demo falls apart if one listens with one ear plugged. Was anyone listenning? Did anyone actually try what I suggested? Jim Reid, it's not surprising that the two recordings sounded alike to you. It isn't necessarily any defect in your hearing as the "effect" is only apparent when considerable care is taken to not allow the left and right channels to mix in the air. Guess what, with headphones the two signals don't beat, but that's not normal and in the real world this retired piano tuner expects major thirds to beat. They should, but they don't if you listen with headphones.

Now try listening with a pair of small speakers that you can move around a bit. You will find that as the speakers are moved closer together they sound more and more like the "mixed" recording. The beat is present as it should be unless you separate the speakers a lot and listen in a "near field" mode where the signals don't mix as much in the AIR. Yes, the effect happens wherever the signals mix, in air or in an audio channel.

I submit that multiple channels help a little but think the major improvement is in moving from mono to stereo. Please try the modified test before rejecting these thoughts out of hand.

And of course David is right (al least this time...) in suggesting looking to other temperments as a solution. Isn't that why they existed all along?
jkinkennon
Member
 
Posts: 1044
Joined: Thu May 07, 2009 10:43 am
Location: Vancouver, WA

Re: Beyond Stereo to multiple Speakers

Postby mdyde » Tue Sep 28, 2010 4:59 am

Hello Joe,

To demonstrate that IMD is far from subtle, I have created a small demonstration. Sound Forge was used to generate two notes, C5 and E5, using pure sine waves. In the first case, the notes are mixed to mono and thus will share the same audio channels during playback:

https://sites.google.com/site/vtheatreo ... dMixed.wav

The second case is to separate the two notes so that C5 is only in the left channel and E5 is only in the right channel.

https://sites.google.com/site/vtheatreo ... tRight.wav

The level of IMD that will be heard in the first case is dependent on each listener's particular audio set up, but I am confident that with even the world's finest audio system, there will still be very audible IMD that will probably sound like a dissonant triad rather than just two notes. This effect completely disappears when the notes are segregated into independent channels. No IMD was introduced by recording the demonstration WAV files as they were generated completely in the digital domain. If anyone has lingering doubts or suspects trickery, just mix the left and right signals from the second file and the exact same result as the first case will be had.


I'm afraid I don't have time to do it myself, but why not record the output from each of your two tests using a good microphone/A-D, good speakers (e.g. your Mackies) and keeping all other factors constant.

Then you could use an FFT simply to measure how much IMD was present by looking at the levels of the inharmonic frequencies introduced, relative to the levels of the harmonic frequencies. That should easily give you a reasonable estimate for the IMD level in decibels, and thus whether it's of a sufficient level to be objectionable or noticeable to people with normal hearing.
Best regards,
Martin.

[Please use email or the Contact page if you need to contact us privately, rather than private forum messages.]

Image
User avatar
mdyde
Moderator
 
Posts: 10644
Joined: Fri Mar 14, 2003 2:19 pm
Location: UK

Re: Beyond Stereo to multiple Speakers

Postby toplayer2 » Tue Sep 28, 2010 5:27 am

In order to eliminate the distraction of the beats (tremolo-like undulations) that are heard with an equal tempered major third interval, two new files have been posted, this time using Just Intonation (the E is flattened slightly so that the ratio is exactly 5:4). Perhaps the absence of beats will make the discordant spurious tone caused by IMD become more obvious. The mixed version will sound like three notes playing instead of two, the third frequency (which could be louder or softer or pitched differently depending on the listener's audio equipment) is the product of IMD.

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdJustIntonationMixed.wav

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdJustIntonantionLeftRight.wav

Repeated, but using square waves:

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdJustIntonationSquareMixed.wav

https://sites.google.com/site/vtheatreorgan/Home/files/MajorThirdJustIntonantionSquareLeftRight.wav

All files are set to -20.0 dB on both channels.

Joe
User avatar
toplayer2
Member
 
Posts: 1055
Joined: Mon Jul 23, 2007 3:15 pm
Location: Michigan, USA

Re: Beyond Stereo to multiple Speakers

Postby toplayer2 » Tue Sep 28, 2010 5:48 am

mdyde wrote:I'm afraid I don't have time to do it myself, but why not record the output from each of your two tests using a good microphone/A-D, good speakers (e.g. your Mackies) and keeping all other factors constant.

Then you could use an FFT simply to measure how much IMD was present by looking at the levels of the inharmonic frequencies introduced, relative to the levels of the harmonic frequencies. That should easily give you a reasonable estimate for the IMD level in decibels, and thus whether it's of a sufficient level to be objectionable or noticeable to people with normal hearing.


Thanks for this suggestion, Martin. That would certainly demonstrate the IMD generated by my particular audio equipment. To demonstrate that there is nothing spurious in the digitally generated waveforms, I have posted the following spectrograph which shows the C5 and E5 (Just Intonation) frequencies with no IM by-products.

https://sites.google.com/site/vtheatreorgan/Home/files/IMD_Demo_Spectrum_Analysis.jpg

Joe
User avatar
toplayer2
Member
 
Posts: 1055
Joined: Mon Jul 23, 2007 3:15 pm
Location: Michigan, USA

Re: Beyond Stereo to multiple Speakers

Postby mdyde » Tue Sep 28, 2010 5:48 am

Hello David,

Dear Martin

I'm very willing to assist people privately but do not wish to provide secrets in public to be viewed and exploited by the commercial competition to Hauptwerk and the enemies of pipe organs as to how to make electronic organ simulations sound brilliant.

Suffice to say that there are ways of ensuring that amplifiers work only in super-linear regions and that speakers work only within linear constraints and without phase shifts and delays between frequencies.

I had an amusing experience a couple of weeks ago when using some interesting speakers for public address at the village fete: my wife was walking the dogs whilst I was testing the microphone and we had a true "his master's voice" event before our eyes, the dogs becoming very confused. Good sound coverage was required over an area of about 5000m2 in the open air. A child walking past asked me if there were real musicians playing an accordian and flute in the porticos and after the fete someone remarked about the speakers as he had observed that those nearby weren't deafened but that the high quality sound had carried into the distance.


It does make it somewhat difficult to give much credibility to the far-fetched anecdotal claims you repeatedly make for your self-developed speakers and techniques if you won't actually tell us what those techniques are, or give any actual data or evidence to back them up.

If you have genuinely managed to develop speaker technologies that the audio industry (which is a large industry with a lot of experts, research and funding) and/or organ industry haven't yet been able to develop, then there are standard and recognised routes for making them available (bringing them to market) without others being able to copy them, e.g. patent them first, publish them in the relevant journals, and license the technologies (patent) to a company that you want to produce and market them for you.

Perhaps you could either publish in detail what your techniques/technologies actually are (with valid supporting evidence), so that we can all asses their merits, or refrain from using this forum to proclaim their merits and promote them to Hauptwerk users until you're prepared to do that. Thank you.
Best regards,
Martin.

[Please use email or the Contact page if you need to contact us privately, rather than private forum messages.]

Image
User avatar
mdyde
Moderator
 
Posts: 10644
Joined: Fri Mar 14, 2003 2:19 pm
Location: UK

Re: Beyond Stereo to multiple Speakers

Postby David Pinnegar » Tue Sep 28, 2010 7:08 am

mdyde wrote:It does make it somewhat difficult to give much credibility to the far-fetched anecdotal claims you repeatedly make for your self-developed speakers and techniques if you won't actually tell us what those techniques are, or give any actual data or evidence to back them up.

If you have genuinely managed to develop speaker technologies that the audio industry (which is a large industry with a lot of experts, research and funding) and/or organ industry haven't yet been able to develop, then there are standard and recognised routes for making them available (bringing them to market) without others being able to copy them, e.g. patent them first, publish them in the relevant journals, and license the technologies (patent) to a company that you want to produce and market them for you.

Perhaps you could either publish in detail what your techniques/technologies actually are, so that we can all asses their merits, or refrain from using this forum to proclaim their merits and promote them to Hauptwerk users until you're prepared to do that. Thank you.


Dear Martin

Part of what I use is technology developed by a British manufacturer and I often use their output. It's mainly made for export and rarely marketed in the UK: anyone who has dealt with Diane will be amongst the cognoscenti of louspeakers and, having ordered through the company, know exactly who I mean. Indeed, the Tuba Magna on my organ would not be the same without that unit. Whilst as stereo pairs such speakers are brilliant domestically, additional considerations apply in choosing units to serve performance instruments, and possibly cheaper but effective units exist.

The best thing about electronic organs is that they generally don't sound as good as pipe organs. We need to keep it that way.

For that reason I'm happy to send details and ideas to forum members privately, as I have done to Keith, but clearly do not want to give an open ticket to Johannus, Rodgers, Allen, Content, Ahlborn, Makin, Viscount, Eminent, to name a few, to make their instruments shine as possibly they could and to pull the wool over the eyes of sheeploads of unsuspecting congregations all too willing to do away with the space requirement of a pipe organ as soon as a 100 year rebuild comes around.

However, I don't keep the information secret nor do I hide it away and use it for personal gain: I run a historic house which is open to the public and anyone interested is welcome to visit. Indeed, we operate an open house for musicians and anyone with enthusiasm for the arts, music and especially organs and visitors are often introduced to the finer points of temperaments, organs, and any speaker carcasses I happen to be working on in the course of tours and tea and after tea organists are welcome to explore organs and pianos, and evaluate speakers of course.

I have advised (gratis) a major Hauptwerk installation in America which, from the reports I read, has greatly benefitted in realism from my advice and will clearly give a lot of pleasure to a lot of people. He has not used British units nor enclosures I recommended but he has used American units made in the same spirit. I have advised Keith to look at Chinese units made by the same factory as the American units but possibly better, and I have the confidence to know that he'll find them brilliant.

This thread started on the question of stereo vs multi-channel and I'm sorry for offending anyone in busting the multi-channel myth in domestic situations. I have merely said that mere stereo can work and work well, and in being pressed for justification have detailed why it can work and how it can be achieved even on a low budget, all important for Hauptwerk users wanting to explore sample sets rather than buy have to buy further hardware for which any but a psychological result might be lacking.

I heard of a recent survey of wine appreciation that perception of the quality of wine was usually proportionate to knowledge of the cost per bottle.

I'm not convinced that the world of technology is much different and, having auditioned the infamous million dollar speakers with "diamond" tweeters made by a firm with Buddhist sounding connexions, and having heard bragging of teenagers whose uncles boast £300k speakers with computer controlled feedback, it's apparent that there's a lot of techno-babble around designed to appeal to the psychology of commerce which makes organists using Hauptwerk taking up the instrument in their retirement feel unable to compete. They don't need to.

It will be interesting to hear in a couple of weeks time how the pair of modified Celestion Ditton 200 speakers perform when installed on Hauptwerk in a home installation. Total cost will have been less than £120 and in a home environment in their modified form they will outperform most other alternatives. They could be modified further to better result for a larger environment but possibly failing the WAF test.

No doubt I would have been burnt at the stake many times in the past. :-)

Best wishes

David P

PS
You will find that as the speakers are moved closer together they sound more and more like the "mixed" recording. The beat is present as it should be unless you separate the speakers a lot and listen in a "near field" mode where the signals don't mix as much in the AIR. Yes, the effect happens wherever the signals mix, in air or in an audio channel.

I submit that multiple channels help a little but think the major improvement is in moving from mono to stereo. Please try the modified test before rejecting these thoughts out of hand.


Very interesting observations here. . .
http://www.organmatters.co.uk
David Pinnegar, B.Sc., A.R.C.S.
User avatar
David Pinnegar
Member
 
Posts: 437
Joined: Thu Jan 03, 2008 8:13 pm
Location: Sussex UK

Re: Beyond Stereo to multiple Speakers

Postby mdyde » Tue Sep 28, 2010 8:20 am

Hello David,

This thread started on the question of stereo vs multi-channel and I'm sorry for offending anyone in busting the multi-channel myth in domestic situations. I have merely said that mere stereo can work and work well, and in being pressed for justification have detailed why it can work and how it can be achieved even on a low budget, all important for Hauptwerk users wanting to explore sample sets rather than buy have to buy further hardware for which any but a psychological result might be lacking.

I heard of a recent survey of wine appreciation that perception of the quality of wine was usually proportionate to knowledge of the cost per bottle.


I very strongly disagree with your assertion that the existence of noticeable benefits of 'multi-channel in domestic situations' is a myth that needs busting (with reference specifically to dry sample sets).

Take a good pair of speakers, and play a dry rank of samples through them in a room (domestic or otherwise). Now repeat with the rank distributed evenly amongst eight of those same speakers, with the speakers scattered around the room. I think you'll find the improvement in realism very far from subtle or imaginary. (I've tried it myself, and also compared the results from stereo microphone recordings made of otherwise-matching stereo vs. multi-channel systems.)

If you want to eliminate any subjective bias, instead record the results of each test through a pair of fixed (e.g. binaural) microphones and analyse and compare the recordings (FFTs etc.). They will be very different because of the differences in room reflections.

You can alternatively perform the whole test virtually, using room modelling software to synthesize a separate impulse response for each speaker's position in the room.

Those differences in room reflections very significantly change spatial perception (the brain will be able to identify the sound as coming from more than two points in the room) as well as the magnitude and phase of each of the frequency components.

Wet sample sets already contain room reflections so don't need multi-channel audio for accurate near-field/anechoic reproduction (although some people still prefer multi-channel anyway).

However, dry sample sets don't have those room reflections (spatial cues), and so will be perceived as emanating from a single point in space (the speaker) unless you use multiple speakers, or apply multiple separate impulse responses to position the pipes separately in virtual space before you play the mixed-down (post-convolution) result back through a single speaker (effectively making them behave like wet samples).

Of course dry sample sets can sound perfectly acceptable with just a single stereo pair of speakers (home electric/electronic organs did that for decades), and of course you can potentially make great speakers yourself much cheaper than buying great speakers, if you have the time, skill and inclination.

In no way am I advocating that people need to or should buy lots of expensive speakers in order to be able to use and enjoy dry sample sets in a domestic environment. However, I cannot agree that multi-channel audio is irrelevant for dry sample sets in a domestic environment, or that its benefits (in terms of differences in frequency amplitudes, phases and spatial perception) are imaginary. They are very real, very obvious and very measurable.

Maybe somebody has the time to perform the above experiment and post the resulting recordings so that the difference can be demonstrated once and for all, and so that we don't have to keep spending more time on this topic?

You can of course also get an ultra-realistic (in terms of acoustic) domestic virtual organ for a comparatively low audio-hardware cost by using wet sample sets with a good pair of headphones.
Best regards,
Martin.

[Please use email or the Contact page if you need to contact us privately, rather than private forum messages.]

Image
User avatar
mdyde
Moderator
 
Posts: 10644
Joined: Fri Mar 14, 2003 2:19 pm
Location: UK

Re: Beyond Stereo to multiple Speakers

Postby mdyde » Tue Sep 28, 2010 10:32 am

P.S. I found some recordings made in 2006 as demos for the dry MDA Virginia WurliTzer sample set (played by Jelani Eddington) according to the experiment I described, i.e. providing a direct comparison of stereo output vs. multi-channel output with all other things being equal.

In this case some (low-cost and not especially realistic) room-modelling software was used to synthesize a set of 16 stereo impulse responses to model a virtual theatre acoustic with 32 virtual speaker positions in it and a single virtual stereo pair of microphones. The dry sample set was then fed to those virtual speakers, and what you hear is the output from the virtual microphones.

In this version, the dry output was mixed down to stereo and then fed to just two of the virtual speakers. I.e. this is a modelled acoustic with just a single virtual stereo speaker pair:

Virtual stereo:
http://downloads.hauptwerk.com/forum_temp_images/from-this-moment-on-v2-wet-stereo-only.mp3

In this second version multi-channel output was used to distribute the pipework evenly between all 32 virtual speakers. This uses the same MIDI recording, same dry sample set, same virtual room model but this time modelling 32-speaker multi-channel output:

Virtual multi-channel:
http://www.milanaudio.com/audio/mp3/virginia-v2/from-this-moment-on-v2-wet-stereo.mp3

It isn't the best of comparisons - the impulse responses aren't real or especially realistic (these just happened to be some existing recordings that I could find), and smaller registrations and a less-reverberant virtual acoustic would perhaps have been more instructive, but if you perform side-by-side listening comparisons of any specific sections using headphones (e.g. in an audio editor) then I would say that the improvement, especially in terms of spatial impression (and resulting increase in perceived clarity, since the brain can identify the virtual sound sources and their audio streams as separate), is still very evident. In particular, my brain perceives the second recording (multi-channel) as significantly more life-like and three-dimensional, with greater clarity.

Note that since the speakers, mics and room are all virtual (impulse responses) then there are no non-linearities, so this comparison doesn't test IMD.

For reference (not part of the test), this is a 'live' (mic'd) recording of the same MIDI file and same sample set but with the pipework distributed across a real multi-channel audio system in a real acoustic:

Real multi-channel:
http://www.milanaudio.com/audio/mp3/virginia-v2/from-this-moment-on-v2-live.mp3

(Unfortunately I don't have a matching recording played through just a single stereo pair of speakers in that real acoustic for comparison.)

Also for reference, this was the mixed-down dry stereo output, which might be of some marginal comparison interest for playing back through a single pair of real speakers, although your listening acoustic will inevitably be significantly different to that used in the live recording of the real multi-channel system above:

Dry stereo:
http://www.milanaudio.com/audio/mp3/virginia-v2/from-this-moment-on-v2-dry-stereo.mp3
Best regards,
Martin.

[Please use email or the Contact page if you need to contact us privately, rather than private forum messages.]

Image
User avatar
mdyde
Moderator
 
Posts: 10644
Joined: Fri Mar 14, 2003 2:19 pm
Location: UK

PreviousNext

Return to Amplification

Who is online

Users browsing this forum: No registered users and 1 guest