Beyond stereo?

[AJ Soundfield]

Continuing what I think may be an interesting topic without further polluting the "What uncontroversial audible differences..." thread

Starting here

Quote

 

 

Quote

 

  6 hours ago, Ralph Glasgal said:

Crosstalk is removed by using a barrier between two speakers directly in front of you

 

 

Yes, I was aware of that one and headphones.

 

Quote

or now much more practically using the same two speakers with a processor that does the same thing as the barrier.  

 

Well, I have some doubts there, unless you mean reduced, or even perceptually removed (via reduction) to the equivalent of the barrier. perhaps with some head tracking and HRTF info.

 

Quote

So now almost all movie dialog is not stereo but mono fed to a center speaker which you can localize to no matter where you are seated.

 

Understood, but again, for music, how much central "solidity" is needed to perceptually recreate an original soundfield?

I should admit to being a classical music attendee, but I enjoy jazz and other smaller/closer up ensembles as well.

 

 

 

 

Quote

 

Take a look at http://analoghighend.blogspot.gr/2017/04/stereo-in-3d.html for a rave review of one XTC system.  If you go to www.ambiophonics.org you can click on other reviews.  There was another rave review for Amtra but it is in German.  The secret of modern XTC is that you must not use head tracking or HRTFs.  The RACE software outperforms the barrier since it works down to lower frequencies and it is really hard to make a barrier that truly isolates the speaker from the wrong ear.  But those that use a barrier are indeed devoted to it.

 

 

Ralph, that is a purely subjective review. I would need actual measured data or controlled listening data to support your contention the crosstalk is "removed"

 

Quote

All imaging, central or not, benefits from psychoacoustic verisimilitude.  This means a coherent set of localization cues including those hall reflections from the sides, the rear, the front, and to a much lesser extent overhead (because these are mono).  You can compromise as non audiophiles do all the time, but you can do a lot better than traditional stereo quite easily while sticking to audiophile frequency response or resolution concerns. 

I didn't ask about all imaging. I asked specifically about the perceptual need for a "solid" central image as created by the crosstalk reduction you advocate, to replicate a real soundfield.

Such as those created by orchestras, or small, say jazz ensembles in smaller acoustic spaces.

TIA

 

 

EDIT: Adding some basics of what I'm really referring to:

http://www.aes.org/e-lib/browse.cfm?elib=9136 The description for the paper begins to explain the limitations of the stereo format and one possible MCH solution approach

http://www.linkwitzlab.com/Recording/acoustics-hearing.htm This explains the stereo is limited concept much further.

There is really no question MCH is a must for recreation of a real acoustic soundfield in spaces. But the  other reality is 98% of recorded music is stereo.

 

[AJ Soundfield]

10 hours ago, STC said:

 

 

I think it will be useful if you go through the technical papers in Ralph's website as he already answered them. 

 

No he has not. He provided a link to a subjective rave. If you can provide the links to the papers addressing the specific claim that the crosstalk has been removed, please do so, thanks.

This paper linked on the site mentions "attenuation", so again, perhaps "removed perceptually" might be the claim, but that is up to Ralph to clarify.

This graph has no units

[AJ Soundfield]

3 hours ago, semente said:

What Ralph was saying is that if you add listening room reflections these will be "similar" to the reflections for both direct sound and ambience cues and that ruins the illusion: "human hearing expects that the direct sound and reflected sound be logical and physically possible".

Johnstons paper and slides I linked in first post addresses this issue in much greater detail

[AJ Soundfield]

8 minutes ago, STC said:

As far as I know crosstalk cancellation only applies to certain frequency range and around 20 or 30dB depending on the algorithm. So you could be right. 

That might be enough, so I don't want to belabor a minor point.

Just noticed your link, interesting system you have there. Too many speakers for me, and I sell them! 

[AJ Soundfield]

6 hours ago, esldude said:

Some other articles which are relevant to the topic.

 

http://www.desena.org/multichannel/Ambisonics_2_Int_Symp_2010.pdf

 

Describes and compares two versions of Johnston and Lam's perceptual sound field methods and Ambisonics.

 

http://www2.ensc.sfu.ca/~ljilja/cnl/guests/cvetkovic.pdf

 

A powerpoint presentation of related material.  The video of that presentation is embedded below.

 

Finally a paper from earlier this year.

 

http://epubs.surrey.ac.uk/813317/13/hacihabiboglu_etal2017.pdf

 

Good catch, forgot about the first one and the last one had not seen, checking out...

[AJ Soundfield]

This is my favorite local (USF), which can vary significantly with the adjustable cloud over the stage

 

Only someone who has never been into a concert hall, or a madman, could claim 2 frontal plane wave generators can recreate this (physically, not just perceptually impossible of course).

[AJ Soundfield]

7 minutes ago, STC said:

 

I don't think anyone saying that.

Not in this thread, no, but it was spawned from the morass of the other where such claims were being made.

Plus it's not to hard to find audiophiles discussing some magic widget and it's "accuracy" to some fantasy of 2ch rendering in their minds.

[AJ Soundfield]

3 hours ago, cfisher said:

 

Tried to get John Zorn to play here but administration couldn't come up with the money. Nice space but not utilized very well as far as I can see.

 

Shame about Zorn, but I'm perfectly ok with the performers they do have there, as I quite enjoy the sound as well.

[AJ Soundfield]

4 hours ago, STC said:

 

Too bad your speakers are too big for me

That is a form follows function thing based on the directivity characteristics desired for position independency (aka wider "sweet spot"), sans any head tracking etc

[AJ Soundfield]

9 hours ago, STC said:

It is actually less than that.

 

 

Thanks, that's in line with physics a I understand it.

I'm still pouring through the links esldude provided but there is some interesting stuff in the last one, as I'm recognizing many of the references.

Of course, as I noted earlier, the vast majority of music is 2ch, so my focus is mainly on how to maximize that.

I personally use a system (loosely) based on this for that end.

But I'm PSR et al ready, were such recordings and decoders to materialize.

[AJ Soundfield]

13 minutes ago, STC said:

 

You could have mentioned earlier and saved my time combing through his website ?

 

 

Hey, Ralph sent us on this goose chase! 

 

I was curious enough to dig a bunch too LOL. I'm always interested in finding things I don't, or may not quite understand. I was also puzzled by early his head tracking/HRTF response to my comment, but again, I think there is clarification now. I would still like to hear a good 22ch demo 

Not too many of those around these parts, or any audio show I've attended

[AJ Soundfield]

Quote

Ralph Glasgal said:

 

In a plain stereo system in a live room, everything about a speaker will make a difference but there is no real fix available due to human hearing properties and that the pattern of peaks and dips that a stereo system produces will be different for every listener, every speaker and every angle (60,61,59,etc.) and so there can be no universal fix that would be stable.

Well, there are methods for producing a wider listening area with stereo loudspeakers 

[AJ Soundfield]

1 minute ago, Ralph Glasgal said:

If you demo your speakers using RACE you will sell a lot more of them.

Thanks Ralph, appreciate that.

The problem is that most folks don't want or most like can't have 22 speakers in their listening room. Bless those who can!

Btw, some cross posting here as you might notice, since we were large "polluting" the other thread.

Thanks for the info so far, I'm still reading and digesting some of your work

[AJ Soundfield]

12 minutes ago, Ralph Glasgal said:

I don't know how many of you are AES members or are coming to The AES convention this October.

I am a member, but unfortunately can't make NYC due to my "real" work schedule, Soundfield being purely a hobby-business. Perhaps at the next one down towards FL.

[AJ Soundfield]

8 minutes ago, Ralph Glasgal said:

Yes.  I call them all Loudspeaker Binaural methods.  Ambisonics, Wavefield Synthesis, VMAX, BACCH, SRS, Sonic Holography, Lexicon Panorama Mode, etc. are some. 

PSR also (see first post in thread). This link was an interesting comparison vs a 2nd order Ambisonics system, though hamstrung by use of only 5 speakers I assume

[AJ Soundfield]

17 minutes ago, esldude said:

In the other MCH thread someone asked if they had an extra pair of speakers and amps what was a good path to get a MCH setup going.  Here was my reply to that:

 

Well somewhere you need a Dolby Digital surround decoder or DTS or better. 

 

If just for kicks and sticking your toe in the waters at this point, I would suggest hopping on the local Craigslist and finding a cheap AVR with digital input and pre-outputs. .

Hi Dennis, yes, I mentioned something similar and linked a "stereo surround" setup in this post https://www.computeraudiophile.com/forums/topic/32135-beyond-stereo/?page=2#comment-677367

For movies I use a modern AVR, but for 2ch music I use a similar setup like this

 

I found Dr Greisingers Logic7 to have the best performance for the rear channels. The fronts remain "pure" stereo

[AJ Soundfield]

15 hours ago, esldude said:

Well somewhere you need a Dolby Digital surround decoder or DTS or better. 

 

If just for kicks and sticking your toe in the waters at this point, I would suggest hopping on the local Craigslist and finding a cheap AVR with digital input and pre-outputs.  One should be available now or soon for less than $100.  The newer the better, and better if it has HDMI.  Denon, Marantz, Onkyo, and Harman Kardon are usually good suspects among common brands. 

One of the reasons I went with the above setup, is that I found I preferred "pure" stereo LR channels vs any of the upmix algorithms of DD, DTS or Logic7. I intend to conduct some experiments with the local audio club to see if I'm deluding myself. The issue of whether to use a center or not is also an interesting one, especially with LR speakers that utilize this research http://www.linkwitzlab.com/Links/Optimized-listening-area-Davies.pdf and this http://www.google.ch/patents/US20090060236

No question a real center reduce timbre distortion from stereo combing, but the question is how much is actually preferred with music

[AJ Soundfield]

On 6/20/2017 at 3:48 AM, semente said:

Despite my relative ignorance in audio matters I suspect that for accurate reproduction of an instrument or singer one would need to record its sound in anechoic conditions with a large amount of mics distributed equally around it in a sphere-like grid, then feed each channel into the equivalently positioned driver of a sphere-like speaker.

That would not represent what 2 ears + head would hear, especially in a typical performance space.

I have yet to attend a music event in an anechoic chamber 

[AJ Soundfield]

36 minutes ago, semente said:

You room would be the venue so you wouldn't be listening in an anechoic chamber.

That would work great if you listen to live violins in your room and wanted to listen to a recording of it without it being there.

What would that have to do with how most folks hear violins?

[AJ Soundfield]

4 hours ago, Ralf11 said:

this thread will get -really- interesting when VR comes out

Well, there are issues to be dealt with first.

This is an interesting video about spatial audio and a bit of history of 2ch through to now possibly several hundred channel...and the problems that start to arise, like spatial aliasing and "realness"

https://www.youtube.com/watch?v=y82nth2Pnwk

 

[AJ Soundfield]

3 hours ago, semente said:

But how is it that most folks listen to violins?

I guess they sit in a room and hear the sound coming out from the violin.

And conventional recordings and speakers cannot reproduce the sound of a violin accurately, the can only recreate a two dimensional illusion of what the violin sounded in the room it was playing in mixed with the reflections produced by your own room.

It's helpful if you are aware of the field of psycho-acoustics and what has been found about perception of reflections, playing same instrument in different rooms to same people, etc, etc.

Try watching the Rumsey video I just linked, much of this is covered, or anything Dr Toole has written. All helpful here.

[AJ Soundfield]

5 minutes ago, semente said:

Reproducing the recording of a violin playing in a music hall with 2 or even 5 channels will not do it.

So how do you explain this? http://www.onhifi.com/features/20010615.htm

It seems clear you're either not reading any of the research linked, or not understanding

[AJ Soundfield]

3 minutes ago, semente said:

 

Listening is not helpful here.

Well there you have it

[AJ Soundfield]

2 minutes ago, semente said:

Yes. 

One of Ralph's posts a couple of days provides good information on what I mean.

So you didn't mean this

Quote

will not be able to provide a realistic holographic soundfield

Because that does involve listening.

Perhaps you could make up your mind?

[AJ Soundfield]

1 minute ago, semente said:

I meant realistic as in mimicking the radiation of a violin, which as shown previously in pics.

As long as the encoding captures the soundfield as 2 ears/head would hear said radiated polar field, that is irrelevant.

Once again, it appears you are not reading any of the literature.

Quote

Yes a pair of mics does capture both direct and reflected sound

 

A pair of mics is woefully inadequate. If you read, you would have seen Perceptual Soundfield Reconstruction techniques use 7 mics. The specific directional characteristics and reasons are explained.

Indeed one does want to capture both onset (aka direct) and indirect fields.

 

Quote

Perception studies are about listening and listening doesn't help here.

 I'm afraid you are quite lost here. That's what this thread is all about. If you want to know whether you have realistically encoded the violins radiated sound, listening tests are the final arbiter.