Concert Hall sound

[gmgraves]

5 hours ago, Fitzcaraldo215 said:

Eliminating that reflected component by deadening all reflections produces sound that is unpleasant and not preferred, even with music.

True, but in the case of my "thought experiment" you would be replacing the deadened listening room with the acoustics of the venue. IOW, it would be no longer dead but chameleon-like in that the space would take on the characteristics of whatever venue the original performance was performed in. If this could be done, I would think that it could be a giant step forward in the SOTA of music reproduction. Of course, record companies would have to restrain themselves from the technical temptation of putting instruments all around the room!

I heard something similar to this years ago at Bert Whyte's* house. He had constructed a dedicated listening room in what he called "live end, dead end" His custom speakers were arranged  at the live end of the room which was highly reflective with hard, irregular surfaces. His listening chairs were at the dead end of the room in which every building surface was covered in acoustic insulation five inches thick with a tasteful light tan grill cloth material covering the insulation. I have never heard such pinpoint imaging and such a bloom of acoustical space. Of course, Bert was playing his own master tapes, so the program material was about as good as it could be in the late 1970's.  

 

* For those of you who don't remember Bert, he was a recording engineer and an early practitioner of recording in stereo. In the late 1950's Bert was the sole recording engineer for a classical label called "Everest" Records (many of these recordings are still available as CDs and high-resolution downloads). Starting in the 1960's and up until his death in the late 80's, Bert had a regular monthly column in Audio magazine.

[gmgraves]

2 hours ago, semente said:

 

I only listen to music with speakers, and I hate jazz and other recordings ( i.e. The Beatles) where the vocals are pulled fully to one of the speakers. It sounds very weird.

Image inside head is one of the reasons why I don't like headphones...

I enjoy jazz so much, that I've learned to put up with the traditional three-channel-mono recording technique pioneered by jazz recording engineer Rudy Van Gelder back in the early days of two-channel "stereo" recording. Broadly and generally speaking, the soloist, either vocal or instrumental, was (is) placed in the center (by equally mixing them into both the left and right channels with the pan-pot) and the rest of the ensemble was was grouped into either the left channel and the right channel which resulted in the drums and the stand-up bass, say, on the left, the vocalist or say, the lead sax (on a Coleman Hawkins record, for instance) in the center and the piano and trumpet on the right. Like you, as a recording technique, I don't much care for this and I've tried to avoid it with less than optimal results. Unlike with classical music, jazz does not respond well to an overall real stereo miking technique. Believe, me I've tried it, with variable results; none of them ideal. There is something about the nature of jazz that requires a close-up approach. This is especially true with small jazz ensembles: trios, quartets, quintets and sextets although even big band jazz is not all that happy with a stereo microphone technique that would make a classical ensemble sound absolutely real! Such ensembles just sound better with close-up multi-miking. I don't like it, but there it is.... Of course, I've never heard a "stereo"  Beatles album, so I can't say for sure, but it sounds like they used a similar final mix technique. 

[fas42]

1 hour ago, gmgraves said:

 

I heard something similar to this years ago at Bert Whyte's* house. He had constructed a dedicated listening room in what he called "live end, dead end" His custom speakers were arranged  at the live end of the room which was highly reflective with hard, irregular surfaces. His listening chairs were at the dead end of the room in which every building surface was covered in acoustic insulation five inches thick with a tasteful light tan grill cloth material covering the insulation. I have never heard such pinpoint imaging and such a bloom of acoustical space. Of course, Bert was playing his own master tapes, so the program material was about as good as it could be in the late 1970's.  

 

* For those of you who don't remember Bert, he was a recording engineer and an early practitioner of recording in stereo. In the late 1950's Bert was the sole recording engineer for a classical label called "Everest" Records (many of these recordings are still available as CDs and high-resolution downloads). Starting in the 1960's and up until his death in the late 80's, Bert had a regular monthly column in Audio magazine.

 

The "pinpoint imaging and such a bloom of acoustical space" is what switches on when the playback chain is sufficiently sorted; using acoustic manipulation and dressing is another method for getting our hearing systems to accept the illusion ... I just happened to trigger this subjective experience via the setup refinement route, because I was that way inclined - if I had talked a lot to other audiophiles, back then, who used room treatments to switch this experience on then I probably would never have chanced upon the alternative, and to my hearing, better method.

[STC]

1 hour ago, gmgraves said:

True, but in the case of my "thought experiment" you would be replacing the deadened listening room with the acoustics of the venue. IOW, it would be no longer dead but chameleon-like in that the space would take on the characteristics of whatever venue the original performance was performed in. If this could be done, I would think that it could be a giant step forward in the SOTA of music reproduction. Of course, record companies would have to restrain themselves from the technical temptation of putting instruments all around the room!

I heard something similar to this years ago at Bert Whyte's* house. He had constructed a dedicated listening room in what he called "live end, dead end" His custom speakers were arranged  at the live end of the room which was highly reflective with hard, irregular surfaces. His listening chairs were at the dead end of the room in which every building surface was covered in acoustic insulation five inches thick with a tasteful light tan grill cloth material covering the insulation. I have never heard such pinpoint imaging and such a bloom of acoustical space. Of course, Bert was playing his own master tapes, so the program material was about as good as it could be in the late 1970's.  

 

* For those of you who don't remember Bert, he was a recording engineer and an early practitioner of recording in stereo. In the late 1950's Bert was the sole recording engineer for a classical label called "Everest" Records (many of these recordings are still available as CDs and high-resolution downloads). Starting in the 1960's and up until his death in the late 80's, Bert had a regular monthly column in Audio magazine.

 

Are you sure Bert's LEDE (Live end -dead end ) arrangement is correct? In listening room ( for two speakers stereo) the speaker front should be dead and the other end should be live. It cannot sound good with the arrangement you are describing.

[STC]

23 hours ago, gmgraves said:

I think you're missing my point, my friend. It wouldn't be a "dead room" when the music is playing because recordings designed to be played in such an ideal listening room would bring with them an approximation of the venue acoustics in which the performance was played. That is to say, that the acoustics of a standard living room would be replaced by the acoustics of the place where the recording was made. Ostensibly, the room would then be a place that transported the listeners to the performance rather than bringing the performance into one's living room where the acoustics of the performance are overlaid by that room's acoustics (good or bad).

 

With today's recording paradigm, that is a useful way to do things. What I was referring to is an ideal listening environment where the room acoustics don't have to be dialed out of the equation with DSP technology because such a listening room would not require that due to the fact that it has no acoustics. My room is, like yours, "neutralized" (to the extent possible) by DSP. The main thing that I find that this does is make the crossover between my main speakers and my powered subwoofers absolutely seamless. As far as altering my room acoustics, switching the DSP in and out of my system while listening yields only a very subtle and often indistinguishable difference (except for the subwoofers).

 

Playing a recorded music in dead room would still sound dead despite having as much ambience cues captured in the recording. I can demonstrate that and there are many literatures out there explaining this point. Reverbs must come from around us. Room reflection helps to fill in the absent cues that were intentionally left out in the recordings.

 

 

 

[STC]

9 hours ago, Fitzcaraldo215 said:

Thank you.  

 

As far as channel count is concerned, yes, we know that real images from speakers are more precise than phantom images between speakers, and also that phantom images are better and more precise over a smaller distance = a narrower angle between them.  Hence, 5.1/7.1 offer an improved frontal soundstage over stereo or '70's Quad due to the center channel.  That is in addition to better handling of dialogue in relation to the screen on video material than can be done with phantom center imaging.

 

But, there are sharply diminishing returns from adding additional channels.  Toole's experiments show the channel count has to be increased substantially in order to make much perceptable difference to listeners.  So, your guess of 30 channels would be consistent with that.  The downside, of course, is the difficulty of recording that way discretely and of distributing files with such large channel counts.  

 

Artificial simulation of that increased surround channel count is possible, except can it be done in a natural way without artifacts or an imposed single sonic signature?  So far, I don't find artificial upmixing algorithms to be the equal of discrete Mch recordings, and I distrust upmixing.

 

There is another aspect to this also.  Contrary to popular belief, the surround channels in discrete Mch recordings do not just contain diffuse reflected sound.  If you think about it from the perspective of a distant surround mike, what is picked up is inseparably both the direct sound from that perspective plus the diffuse reflected sound from that perspective.  This is clear from listening to surround channels up close.  On playback, that combination of direct plus reflected sound creates phantom images heard by the listener as a result of interaction with the front channels and other surround channels.  

 

The effect is to not only bring the diffuse hall sound to the listener, but also to enhance the frontal soundstage in terms of width, depth and dimensionality.  I personally find that discrete Mch recordings pull the frontal soundstage out into the room somewhat in front of the plane of the front speakers.  A sense of added natural sounding depth and dimensionality over stereo playback is notably apparent to me, since the front speaker array can, like stereo, simultaneously produce phantom depth behind the front speaker plane.

 

My point is it may not be quite so simple to introduce added algorithmic channels that naturally convey the sound in the hall.  One must consider both the direct sound output of those channels and their interaction via phantom imaging with other speaker channels in the array.

 

Quote

As far as channel count is concerned, yes, we know that real images from speakers are more precise than phantom images between speakers, and also that phantom images are better and more precise over a smaller distance = a narrower angle between them.

 

I maybe wrong but I always thought where the phantom image is further towards the side the more accurate they are.

 

Quote

But, there are sharply diminishing returns from adding additional channels.  Toole's experiments show the channel count has to be increased substantially in order to make much perceptable difference to listeners.  So, your guess of 30 channels would be consistent with that.  The downside, of course, is the difficulty of recording that way discretely and of distributing files with such large channel counts.  

 

I am not sure which Toole's experiment you are referring to. For front sound reproduction, the more channels will allow you you to precisely recreate the actual soundfield . However, it is impossible to have a similar setup as that means an accurate representation will only be possible with similar arrangement of the speakers as the recording. It is not practical and hard to be consistent with the varying number of instruments.

 

When I mention 30 channels, I was referring to the ambience cues for recreating the hall's ambiance. That's to recreating the reflecting walls of the concert hall. This require as many channels as possible to replicate the actual reflecting sound coming from multitude angles in the concert hall.  Right now, I am only using reflecting for every 15 degrees at 23 degree elevation. Prof Farina created one for every 10 degrees. Technically, with these impulse response you can create infinite number of virtual concert halls. IMO, 30 channels is good enough. I use 144* channels. Ideally, when I get another Motu I will increase another 144 channels provided my computer and Reaper could handle this. The more the channels are the smother the frequency response is as it flatten the dips and peak caused by the room.

 

* Previously, I mentioned 72 channels. It should be 72 times two as each set of IRs caries both left and right channels. So in total it should be 144 channels.

 

Quote

There is another aspect to this also.  Contrary to popular belief, the surround channels in discrete Mch recordings do not just contain diffuse reflected sound.  If you think about it from the perspective of a distant surround mike, what is picked up is inseparably both the direct sound from that perspective plus the diffuse reflected sound from that perspective.  This is clear from listening to surround channels up close.  On playback, that combination of direct plus reflected sound creates phantom images heard by the listener as a result of interaction with the front channels and other surround channels.  

 

A good diffuse reflected sound will actually sound like the rear ambiance speakers of multichannels.  That's the reason why a concert hall sound is unique. Here is a quick video where at 1:00 I have mute all the speakers except for the ambiance speakers. In this video, I was using 36 channels of different IRs.

 

Quote

My point is it may not be quite so simple to introduce added algorithmic channels that naturally convey the sound in the hall

 

Actually, it is as simple as that. leave whatever recorded sound as it is and pay attention to the room acoustics. The IRs is dealing with the room/hall acoustics. The front channels remain as it is -untouched. Think this as more elaborate and precise method of doing room treatment. 

[fas42]

1 hour ago, STC said:

 

Playing a recorded music in dead room would still sound dead despite having as much ambience cues captured in the recording. I can demonstrate that and there are many literatures out there explaining this point. Reverbs must come from around us. Room reflection helps to fill in the absent cues that were intentionally left out in the recordings.

 

 

 

 

Which is what I disagree with - in an extremely dead room the sound coming from the speakers, provided you're standing in a position where you can pick up the direct sound to some degree will provide an experience which has very satisfying ambience - it will mimic being in front of a huge open window onto the "concert stage", where you are sitting in an open field.

 

The cues are in the recording - the engineers may have thought they were clever and left them out, "intentionally" - but it is quite easy to pick the space where the sounds, if natural, were recorded. People might be upset with this, because it doesn't conjure up a uniform acoustic - I don't mind, it's interesting to note the balance between the spaces one "sees"; doesn't get in the way of "enjoying the spectacle".

[fas42]

11 hours ago, Fitzcaraldo215 said:

Thank you.  

 

As far as channel count is concerned, yes, we know that real images from speakers are more precise than phantom images between speakers, and also that phantom images are better and more precise over a smaller distance = a narrower angle between them.

 

Ummm, no ... with convincing sound, there are no "real images from speakers" - it's all phantom images, nothing else - it's absolutely impossible to see an image "in the speaker", no matter how hard you try to place it there, in your mind. When I have a rig that's slipping between the convincing and unconvincing modes, that's what happens - if the sound degrades, the "speaker images" come to life, and the illusion is lost.

[STC]

3 minutes ago, fas42 said:

 

Which is what I disagree with - in an extremely dead room the sound coming from the speakers, provided you're standing in a position where you can pick up the direct sound to some degree will provide an experience which has very satisfying ambience - it will mimic being in front of a huge open window onto the "concert stage", where you are sitting in an open field.

 

The cues are in the recording - the engineers may have thought they were clever and left them out, "intentionally" - but it is quite easy to pick the space where the sounds, if natural, were recorded. People might be upset with this, because it doesn't conjure up a uniform acoustic - I don't mind, it's interesting to note the balance between the spaces one "sees"; doesn't get in the way of "enjoying the spectacle".

 

I have a room which is extensively covered with 4 inch Roxul Rockwool, including the ceiling. The floor was cover with 1 inch thick mat (not carpet) with excellent absorbing coefficient. And I also added another five 2x4 portable Roxul Rockwool which can be used to block any reflection around me. If I remember correctly, the RT60 was 0.28s. It was so dead that even a person on the other end of the phone could tell the difference when I speak to them  in that room. In the end, I have to use twisted aluminium foils to cover the rockwool to liven the room to bring the RT higher.  I even arranged all favorite whisky bottles in the room to create more reverbs (they are still there as it add nice touch of reverberation to the convolution) as I couldn't remove the rockwool without demolishing the walls.

 

That's enough first hand experience to confidently say that you are very wrong. Forget about the published research papers on this subject. Recordings only contains the frontal and a little of surrounding cues for the front channel and a bit of the rear cues for multichannel recordings. 

 

[fas42]

Yes, you've addressed the room ... but not the replay chain. And I have done absolutely nothing to any room where a system is set up that I've worked on.

 

There is a step change in the presentation of recordings - in a positive direction - if the SQ from the speakers is good enough. If someone hasn't experienced this transition, occurring to some degree at least, then they will probably never understand what this is like - it goes from ordinary, to Woowww!!! And there is nothing in the research that talks about this - the quality of playback of the systems used for investigating has never been good enough ... QED.

[STC]

13 minutes ago, fas42 said:

Yes, you've addressed the room ... but not the replay chain. And I have done absolutely nothing to any room where a system is set up that I've worked on.

 

You mentioned elsewhere that you are still chasing the elusive sound some 30 year ago that you accidently discovered. By now you should have known you are going the wrong direction. More than half of the sound that determines fidelity/realism is not even in the recording. It is the ambiance. Even with mp3 SQ you can make great sound if you address the room acoustics.

[semente]

9 hours ago, gmgraves said:

I enjoy jazz so much, that I've learned to put up with the traditional three-channel-mono recording technique pioneered by jazz recording engineer Rudy Van Gelder back in the early days of two-channel "stereo" recording. Broadly and generally speaking, the soloist, either vocal or instrumental, was (is) placed in the center (by equally mixing them into both the left and right channels with the pan-pot) and the rest of the ensemble was was grouped into either the left channel and the right channel which resulted in the drums and the stand-up bass, say, on the left, the vocalist or say, the lead sax (on a Coleman Hawkins record, for instance) in the center and the piano and trumpet on the right. Like you, as a recording technique, I don't much care for this and I've tried to avoid it with less than optimal results. Unlike with classical music, jazz does not respond well to an overall real stereo miking technique. Believe, me I've tried it, with variable results; none of them ideal. There is something about the nature of jazz that requires a close-up approach. This is especially true with small jazz ensembles: trios, quartets, quintets and sextets although even big band jazz is not all that happy with a stereo microphone technique that would make a classical ensemble sound absolutely real! Such ensembles just sound better with close-up multi-miking. I don't like it, but there it is.... Of course, I've never heard a "stereo"  Beatles album, so I can't say for sure, but it sounds like they used a similar final mix technique. 

I much prefer Keepnews' recordings to those of Van Gelder (i.e. Bill Evans live).

I agree that you do need close-mic'ing for jazz to balance the levels.

 

Margot's voice at the Trinity Sessions had to be amplified for this reason.

[Fitzcaraldo215]

15 hours ago, gmgraves said:

I have no dispute with that, but ideally, the best result would be multi-channel recordings (pure stereo for the two front channels, ambience only in the side, rear and overhead channels [separate subwoofer channels optional])played in a room with no acoustical signal of it's own. Such a listening experience would be totally immersive, and from the standpoint of the concert experience startlingly real. Of course, binaural sound does that to some degree. I've been listening to the BBC Proms from the Royal Albert Hall in London streamed in binaural AAC sound. At times the result is almost spooky, but you can't move your head, or the illusion disappears like being awaken from a dream. 

George - a "room with no signal of its own" is an anechoic chamber that kills all reflections.  That is not what we want.  What we want is a speaker/room that has no sonic signature that imposes itself on the the signal via reflections.  There is an important difference.

 

Most of the issue is purely in the frequency domain.  We do not want the speaker/room to distort the frequency response, even if there are reflections in the room, and we don't want to or need to kill those reflections as long as they are not distorting frequency response.  The reflections are then actually desirable.  They help amplify and spread the sound.

 

Sometimes, there are also time domain issues, ringing, etc. These are sometimes harder to deal with, though often, for example, pulling down modal bass peaks in the frequency domain will tame them in the time domain, as well.  

 

So, killing all the reflections is draconian and leaves us with an unpleasant sounding, totally dead room.  Taming the frequency and time domain issues caused by reflections but allowing the reflections eliminates their imposition on the sound, which is what we want.  That allows us to better hear through the room to the acoustical cues on the recording.

 

[Fitzcaraldo215]

12 hours ago, STC said:

 

 

I maybe wrong but I always thought where the phantom image is further towards the side the more accurate they are.

 

 

I am not sure which Toole's experiment you are referring to. For front sound reproduction, the more channels will allow you you to precisely recreate the actual soundfield . However, it is impossible to have a similar setup as that means an accurate representation will only be possible with similar arrangement of the speakers as the recording. It is not practical and hard to be consistent with the varying number of instruments.

 

When I mention 30 channels, I was referring to the ambience cues for recreating the hall's ambiance. That's to recreating the reflecting walls of the concert hall. This require as many channels as possible to replicate the actual reflecting sound coming from multitude angles in the concert hall.  Right now, I am only using reflecting for every 15 degrees at 23 degree elevation. Prof Farina created one for every 10 degrees. Technically, with these impulse response you can create infinite number of virtual concert halls. IMO, 30 channels is good enough. I use 144* channels. Ideally, when I get another Motu I will increase another 144 channels provided my computer and Reaper could handle this. The more the channels are the smother the frequency response is as it flatten the dips and peak caused by the room.

 

* Previously, I mentioned 72 channels. It should be 72 times two as each set of IRs caries both left and right channels. So in total it should be 144 channels.

 

 

A good diffuse reflected sound will actually sound like the rear ambiance speakers of multichannels.  That's the reason why a concert hall sound is unique. Here is a quick video where at 1:00 I have mute all the speakers except for the ambiance speakers. In this video, I was using 36 channels of different IRs.

 

 

Actually, it is as simple as that. leave whatever recorded sound as it is and pay attention to the room acoustics. The IRs is dealing with the room/hall acoustics. The front channels remain as it is -untouched. Think this as more elaborate and precise method of doing room treatment. 

 

[gmgraves]

13 hours ago, STC said:

 

Are you sure Bert's LEDE (Live end -dead end ) arrangement is correct? In listening room ( for two speakers stereo) the speaker front should be dead and the other end should be live. It cannot sound good with the arrangement you are describing.

That's the way I recall it. Of course, it's been close on to 40 years and human memory is quite fallible about details like that. I'm not insisting I'm right, but if I am wrong about Bert's LEDE arrangement, you can't prove it by me. I can ask Gene Pitts when next I speak to him (he was the Editor-in-Chief of Audio magazine for the last 20 years or so of it's life), he might remember more clearly, after all he interfaced with Bert a lot, while I just visited his house once.

[gmgraves]

40 minutes ago, Fitzcaraldo215 said:

George - a "room with no signal of its own" is an anechoic chamber that kills all reflections.  That is not what we want.  What we want is a speaker/room that has no sonic signature that imposes itself on the the signal via reflections.  There is an important difference.

 

Most of the issue is purely in the frequency domain.  We do not want the speaker/room to distort the frequency response, even if there are reflections in the room, and we don't want to or need to kill those reflections as long as they are not distorting frequency response.  The reflections are then actually desirable.  They help amplify and spread the sound.

 

Sometimes, there are also time domain issues, ringing, etc. These are sometimes harder to deal with, though often, for example, pulling down modal bass peaks in the frequency domain will tame them in the time domain, as well.  

 

So, killing all the reflections is draconian and leaves us with an unpleasant sounding, totally dead room.  Taming the frequency and time domain issues caused by reflections but allowing the reflections eliminates their imposition on the sound, which is what we want.  That allows us to better hear through the room to the acoustical cues on the recording.

 

Again, all you say would be quite correct were we not replacing the acoustics of the listening room with the acoustics of the space where the recording was made. To do this we must have a listening room that has no acoustical properties of it's own. All reflective surfaces must be first removed from the listening room before the acoustics can be transferred from another venue via multi-channel recordings and multiple speakers. How else can one have a listening room with a null acoustic signature of its own except in a room that's as much of an anechoic chamber as is practical. 

[gmgraves]

14 hours ago, fas42 said:

 

The "pinpoint imaging and such a bloom of acoustical space" is what switches on when the playback chain is sufficiently sorted; using acoustic manipulation and dressing is another method for getting our hearing systems to accept the illusion ... I just happened to trigger this subjective experience via the setup refinement route, because I was that way inclined - if I had talked a lot to other audiophiles, back then, who used room treatments to switch this experience on then I probably would never have chanced upon the alternative, and to my hearing, better method.

I see Frank's put "his" broken record on the gramophone again (mine is a different broken record)!

[gmgraves]

14 hours ago, STC said:

 

Playing a recorded music in dead room would still sound dead despite having as much ambience cues captured in the recording. I can demonstrate that and there are many literatures out there explaining this point. Reverbs must come from around us. Room reflection helps to fill in the absent cues that were intentionally left out in the recordings.

 

 

 

I never said that such an arrangement would be practical. It might take 8 or 10 discrete channels to make such a scheme work, and then maybe still, a lot of DSP computing power to boot. I mean this is just a "thought experiment" and, as such, is more of a question than an answer.

[gmgraves]

6 hours ago, semente said:

I much prefer Keepnews' recordings to those of Van Gelder (i.e. Bill Evans live).

I agree that you do need close-mic'ing for jazz to balance the levels.

 

Margot's voice at the Trinity Sessions had to be amplified for this reason.

I find it's not just to balance the levels. Jazz, especially small ensemble jazz is a very intimate musical genre. The close up miking is to get that cozy, tiny Greenwich Village coffee house feeling into the recording. An over-all stereo mike just sounds too distant for jazz.

[Fitzcaraldo215]

14 hours ago, STC said:

 

 

I maybe wrong but I always thought where the phantom image is further towards the side the more accurate they are.

 

 

I am not sure which Toole's experiment you are referring to. For front sound reproduction, the more channels will allow you you to precisely recreate the actual soundfield . However, it is impossible to have a similar setup as that means an accurate representation will only be possible with similar arrangement of the speakers as the recording. It is not practical and hard to be consistent with the varying number of instruments.

 

When I mention 30 channels, I was referring to the ambience cues for recreating the hall's ambiance. That's to recreating the reflecting walls of the concert hall. This require as many channels as possible to replicate the actual reflecting sound coming from multitude angles in the concert hall.  Right now, I am only using reflecting for every 15 degrees at 23 degree elevation. Prof Farina created one for every 10 degrees. Technically, with these impulse response you can create infinite number of virtual concert halls. IMO, 30 channels is good enough. I use 144* channels. Ideally, when I get another Motu I will increase another 144 channels provided my computer and Reaper could handle this. The more the channels are the smother the frequency response is as it flatten the dips and peak caused by the room.

 

* Previously, I mentioned 72 channels. It should be 72 times two as each set of IRs caries both left and right channels. So in total it should be 144 channels.

 

 

A good diffuse reflected sound will actually sound like the rear ambiance speakers of multichannels.  That's the reason why a concert hall sound is unique. Here is a quick video where at 1:00 I have mute all the speakers except for the ambiance speakers. In this video, I was using 36 channels of different IRs.

 

 

Actually, it is as simple as that. leave whatever recorded sound as it is and pay attention to the room acoustics. The IRs is dealing with the room/hall acoustics. The front channels remain as it is -untouched. Think this as more elaborate and precise method of doing room treatment. 

Yes, phantom imaging between two speakers becomes more accurate as the apparent, panned sound source approaches the speaker to one side or the other.  Also, the centered phantom image between two speakers with a 60 degree angle between them will be less accurately conveyed than if the angle was 30 degrees.  

 

There is also a special case of sounds from front dead center and slightly either side of that.  Due to our head related transfer function and ears to the sides of our head, there is a large drop at upper middle frequencies when heard from phantom imaging by two front speakers.  The same signal when played by a single center channel speaker eliminates that.  That is important for imaging at the front of the soundstage for music, and also for dialog articulation on video.  It is a frequency response problem with center imaging in stereo we have just gotten used to.

 

I think all  this is in Toole's latest book: Volume 3 of Sound Reproduction.  But, phantom imaging is obviously vitally important to stereo and all forms of Mch.  They would not work adequately without it.  I also believe that phantom imaging, imprecise though it may be,  can be perfectly adequate to convey the diffuse reflected and reverberant sounds from Mch surround channels.  That is, I think, why Toole believes adding surround channels requires a huge increase in channel count in order to make much perceivable difference.  Adding height speakers is a different matter because that is adding additional location and dimensional information in the Z-direction.

 

Incidentally, Toole is a big fan of Mch sound.

 

I also disagree with you about what signal is present in Mch surround speakers on recordings.  I will go through it again.  In a simple Mch mic setup, there are stage mics for the front channels plus more distant hall mics, usually omnis, for surround channels.  Focus on one of those distant mics for a moment.  Imagine sitting where one of those mics is.  What will they pick up?  Obviously, they will pick up much reflected hall sound, given their location.  But, there is no way to block them from picking up or to suppress the direct sound from the stage at that more distant perspective.  Ergo, they contain signal that is an inseparable combination of reflected hall sound plus that more distant direct sound from the stage, just as you would hear if sitting there.  That direct sound component cannot be erased, nullified or removed from the signal via editing, mastering, etc. 

 

On playback, that combined direct plus reflected signal from the distant mic goes to a surround channel.  It reproduces the desired hall ambience picked up from hall reflections.  But, that direct sound is also there, delayed, somewhat down in level, and altered somewhat in frequency response relative to the front main and center channels by air/distance attenuation.  But, it interacts with the sound from the main channels, producing a phantom image mainly in the front to rear direction into the room.  The diffuse reflected energy component in the surround channel does that, too, but to a lesser degree.  

 

That phantom image of the direct sound from front and surround speakers is closer to the front speakers, because the front channels are somewhat louder, having been recorded closer to the performers.  This is why I hear the sound being "pulled into the room" in front of the front speaker plane in a way stereo does not do.  It is also why I find that discrete Mch provides a greater apparent depth of the frontal soundstage.

 

Oversimplified, that is what I believe is going on in Mch.  It also exactly reflects what I hear when I listen to the signal from surround channels in discretely recorded Mch - diffuse reflected sound plus direct sound from the stage.  And, properly done, it works.   I think it is the best approach yet to recording and reproducing a successful illusion of live music in the concert hall.

 

[Fitzcaraldo215]

1 hour ago, gmgraves said:

I never said that such an arrangement would be practical. It might take 8 or 10 discrete channels to make such a scheme work, and then maybe still, a lot of DSP computing power to boot. I mean this is just a "thought experiment" and, as such, is more of a question than an answer.

George - you can test your theory with a bit of work.  Set up your system outdoors on a sunny day on a soft grassy surface.  There will be essentially no reflections, just like your ideal.  Let us know how you like the sound.

[phosphorein]

1 hour ago, gmgraves said:

That's the way I recall it. Of course, it's been close on to 40 years and human memory is quite fallible about details like that. I'm not insisting I'm right, but if I am wrong about Bert's LEDE arrangement, you can't prove it by me. I can ask Gene Pitts when next I speak to him (he was the Editor-in-Chief of Audio magazine for the last 20 years or so of it's life), he might remember more clearly, after all he interfaced with Bert a lot, while I just visited his house once.

 

I remember Whyte writing about this in Audio magazine and remember the Dead End being where the speakers were situated. I arranged my listening space based on Whyte's suggestion. My impression is that most mastering facilities are set up based on this concept.

[phosphorein]

1 hour ago, phosphorein said:

 

I remember Whyte writing about this in Audio magazine and remember the Dead End being where the speakers were situated. I arranged my listening space based on Whyte's suggestion. My impression is that most mastering facilities are set up based on this concept.

 

Whyte writes about the LEDE approach in the June, 1982 (p.91) issue of Audio.

[shtf]

On 7/23/2018 at 2:03 AM, STC said:

We have seen many videos of audiophiles listening room. But the real sound recorded at you listening position in concert hall will sound like the video below. Ideally, our listening room too should sound like that when recorded at our sweet spot. Can this be done?

 

 

 

I’m curious why, when listening to a beautiful recording exhibiting volumes of ambient info such as this performance, you would possibly think this ambient info is somehow associated to a given listening room and its acoustics and not the recording? 

 

Especially since ambient info is a very audible part of the live performance and embedded in the recording as we can clearly hear from this very recording regardless of where you watch/listen to this video.

 

 

[gmgraves]

3 hours ago, Fitzcaraldo215 said:

George - you can test your theory with a bit of work.  Set up your system outdoors on a sunny day on a soft grassy surface.  There will be essentially no reflections, just like your ideal.  Let us know how you like the sound.

Since I don't have any multi-channel recordings where all the channels but RF and LF are ambience signals nor the spare speakers and amps to set-up such an experiment, the results will have to remain conjecture. But you are right that would settle the debate. I just don't see why there is a debate. I keep insisting that the result would not be that of an anechoic chamber because enough of the ambience cues from the recording would be there to replace the room acoustics. so the room would be no longer dead and several of you keep talking about a dead room that would sound unpleasant. I'm sure that would be right if we were only playing back 2-channel material, but since we're bring the hall sound of the recording venue along with us, I don't see how there could be a dead room.