Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: the recent 2-channel 3D sound formats and their viability for actual f (Read 8757 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

the recent 2-channel 3D sound formats and their viability for actual f

(I posted this on the [SurSound] list but i'd like to get more takes on it.)
I'm interested in the idea of 360 horizontal surround sound but more interested in a sphere recreation.
-------------
Folks,

I've been reading up on the various proposals for 3D sound from a set of
stereo speakers. The 3D Audio Alliance is working on such a system.
Astound Surround is getting ready to market, Edward Choueiri is working
on the same idea (see:
http://www.studio360.org/2011/apr/29/adventures-3d-sound/ ) and there
are others. I used to have a Carver pre-amp with Carver's Holography
button but i could never get it to do much.

Has anyone heard a truly 3D/360 surround effect from 2 speakers using
this stuff? Ever heard a fly buzzing around your head, or an object in
the back far-left of you or some such?  Can any of this do as good of a
job as Ambisonics?  Is all of this just related to head transfer
function mathematics?

I've listened to some of the headphone applications of this like
binaural and whatever these folks are doing here
http://www.3d60.co.uk/index.php  That demo on the 3D60 page sounds
really cool, however nothing ever sounds like it's more than a foot from
my head and nothing is ever right in front of me. Why can't they create
an effect of something coming from a long distance away and getting
closer and closer behind me? If it's all related to head transfer
function you'd think you could create any sound your ears can hear.

I'm looking at my audio system building options and I'd love to throw my
money/space decor at 2 really good speakers and a good 2 channel pre-amp
instead of 12 speakers in an ambisonic system with all the associated
electronics.

Any thoughts on all this 3D through 2 channel stuff?

Thanks,

---------
Further thoughts i've had are that i don't understand how this works given crosstalk, the differences in everyone's rooms, the differences in everyone's hearing, outer ear/head shapes ect.  If this is relying on head position to create the same sounds as they would be after HRTF isn't that one hell of a sweet spot?

In review (and i've only read one) of the system put out by the 3D Audio Alliance the person listening to it described hearing the sound of a dog off to his back right.  From two speakers. That's really awesome..if it works.

Bearcat
Music lover and recovering high end audiophile

the recent 2-channel 3D sound formats and their viability for actual f

Reply #1
I've listened to some of the headphone applications of this like
binaural and whatever these folks are doing here
http://www.3d60.co.uk/index.php  That demo on the 3D60 page sounds
really cool, however nothing ever sounds like it's more than a foot from
my head and nothing is ever right in front of me. Why can't they create
an effect of something coming from a long distance away and getting
closer and closer behind me? If it's all related to head transfer
function you'd think you could create any sound your ears can hear.


Try using IEMs if you haven't.

the recent 2-channel 3D sound formats and their viability for actual f

Reply #2
It is possible. The problem is that the sweet spot (the place between the speakers where you experience the effect) is quite small and fussy.

the recent 2-channel 3D sound formats and their viability for actual f

Reply #3
I'm very skeptical that this can ever work as well as 5.1 or 7.1 surround.  When you want to  accurately pinpoint the source of a sound, you move your head around to trianglulate multiple readings.  Unless the system is tracking the listeners head movements, I just can't imagine how this is going to work. 

...Once they get it perfected with 2 speakers, maybe they'll figure-out a way to to get surround sound from one speaker! 


the recent 2-channel 3D sound formats and their viability for actual f

Reply #5
I've listened to some of the headphone applications of this like
binaural and whatever these folks are doing here
http://www.3d60.co.uk/index.php  That demo on the 3D60 page sounds
really cool, however nothing ever sounds like it's more than a foot from
my head and nothing is ever right in front of me. Why can't they create
an effect of something coming from a long distance away and getting
closer and closer behind me? If it's all related to head transfer
function you'd think you could create any sound your ears can hear.


I've heard pretty realistic stuff from microphones in the ears of a dummy head.  ZBS media  has some samples (not all their CD's, so look under binaural.) 

3d360 is a little vague about their process, but since they seem to be reworking existing material, I'd guess it's done with DSP from a multitrack recording rather than a dummy head.

How well it works depends on how well the model (either the dummy or the DSP) matches your head.  Your sense of sound placement, especially height and sounds behind vs in front, is tuned to the shape of your ears, which likely won't be an exact match to what the producer used.

I have doubts about how well you could do surround with  2 loudspeakers.  Even assuming it works, there would be a pretty limited amount of material available, whereas there is tons for conventional 5.1 and up surround sound.

I've never tried it, but there's at least one system that converts Dolby 5.1 to binaural on the fly.

the recent 2-channel 3D sound formats and their viability for actual f

Reply #6
The first problem in 2 speaker surround and 3D audio is that both ears hear both speakers.  One common approach (and the one used by Carver, BTW), is to produce a crosstalk cancellation signal that's mixed into the opposite channel.  In the late 1970s there was a brief rash of devices that did this, the Carver Sonic Holography Generator was one.  Carver used a sample from one channel, inverted it, and delayed it with an all-pass filter network, and some response shaping filters, then mixed it into the opposite channel in hopes it would cancel the sound from, for example, the left speaker arriving a the right ear.  It worked quite well, but you had to eliminate all possible early reflections with a combination of spacing the speakers away from walls, and applying absorption on any reflecting surface between you and the speaker.  Unless all early reflections were reduced, the effect was inaudible.  But when it worked, it worked startlingly well.  However, there was no way to place a sound behind a listener in any predictable way, though you did get a sense of 3D audio.

Another approach was to develop the crosstalk cancellation signal by deriving it from an L minus R signal which was then eq'd and delayed.  The idea was that the "enhancement" would not be present for mono signals.  While that was true, the system had the undesirable effect of placing vertical groove distortion from vinyl records out in the room with you.  That negative effect notwithstanding, it also did present a somewhat 3D soundfield, though again, positioning a sound predictably wasn't possible. 

Both the Carver concept and the L-R concept had by necessity a very tiny sweet spot, and the 3D effect collapsed if the listener's head was moved even an inch out of center.  The L-R idea did recover a lot of hidden reverberation and room ambiance though, even if heard outside the sweet spot.

The creators of Astound Surround claim to use technology that was developed through brain research, that it has a large sweet spot, and works incredibly well for every listener.  You can download the post-processor version in plug-in form and install it as a free demo.  It works with iTunes.  However, in my tests it falls very short of a desirable effect.  The plug-in has a control to vary the amount of "effect".  In my personal tests, I kept backing the control down until things sounded good, only to find I'd turned it off completely.  Also, the demo has a highly irritating "nag" function that will get you annoyed enough to uninstall it, almost without regard to what it does.  I felt it had too much impact on the balance of the mix, making radical changes like burying the vocal for example, rather than just providing a surround effect.  Their demo clips, which differ in that they were mastered using Astound Surround technology rather than applying it as a post process, are interesting, but not as "astounding" as one might wish.

The big issue with binaural recordings is that they must be made with mics in simulated ears (pinna at very least) in a head with a chest cavity, both of which must be similar in acoustic density to that of a real human.  This makes recording cumbersome.  Using a real head and ears is possible, but has its own issues.  The real problem is that no two human ears are identical, particularly the pinna shape with its folds and curves.  This means that everyone has a slightly different head related transfer function, and is why binaural recordings don't always work well from one listener to the next.  In fact, creating a universal HRTF (head related transfer function) is probably impossible because of listener to listener variations.  Binaural recordings also cannot be played well on conventional speakers, and only work well on headphones, typically open on-ear types.

So far, attempts and two speaker surround have all fallen a bit flat.  It's a compromise at best, but if you realize that, some two channel surround can be though of as an improvement over simple stereo. At least there's some attempt at surround placement, and though far from the performance of real surround speakers, can provide listeners with at least an improved experience.  But, in practical terms of real surround, most phantom images are fragile at best.  "If you want a sound there, you have to put a speaker there."

 

the recent 2-channel 3D sound formats and their viability for actual f

Reply #7
That was a very informative post. Thank you. I used to have a Carver pre-amp with the "Sonic Holography". I couldn't get it to do much either.

One thing that i don't understand: recording engineers must be taking the cross-talk into account when they record things, right? So if you reduce the cross-talk, aren't you messing things up further?

I donno wether "3d" recordings are done with out any crosstalk or how they differ in that regard.
Music lover and recovering high end audiophile

the recent 2-channel 3D sound formats and their viability for actual f

Reply #8
That was a very informative post. Thank you. I used to have a Carver pre-amp with the "Sonic Holography". I couldn't get it to do much either.

The "Sonic Hologram Generator" was very sensitive to acoustic issues.  If you had a really good room, it was spectacular, but didn't work at all in a bad room.
One thing that i don't understand: recording engineers must be taking the cross-talk into account when they record things, right? So if you reduce the cross-talk, aren't you messing things up further?

I donno wether "3d" recordings are done with out any crosstalk or how they differ in that regard.


The monitoring environment in a mixing room is fairly standard, though hopefully attention has been paid to acoustics, in particular early reflections which strongly degrade a normal stereo image.  Since acoustic cross-talk is a fact of life, especially in two-speaker stereo, it's not compensated for during the mix.  A good mixing engineer knows that proper acoustic cross-talk compensation is highly situational, and it does little good to try to deal with it at all in the mix.  It's assumed the target audience will have the normal acoustic cross-talk issues. 

Some 3d sound processing has been done during mixdown.  It's usually an "effect", a temporary one.  I have a CD with one track where 3D sound processing was used.  Frankly, I never even noticed it until I read the booklet notes!  It's a fine line between making a 3D effect "noticeable" to everyone, and making it overpowering, which is why it's hardly ever done in the mix, but is better applied as a playback function.  However, this makes precision 3D placement and panning impossible. 

But, to the rescue comes multichannel audio!  A 5.1 channel system adds a tremendous amount of dimensional capability.  These days there are millions of 5.1 channel systems in homes, most of which are never used for music only entertainment.  But 5.1 music is really very good. 

But lets not stop at 5.1.  It seems every time the channel count is doubled (or slightly more than doubled), every listener can hear the difference without question.  Enter 10.2, and it's outgrowth 11.2.  To gain true 3D dimensionality, yet retain a huge listening sweet spot, you have to add channels and speakers.  Systems like Audyssey DSX, with three front channels, plus two wide channels, plus two height channels, then the usual surround compliment really does succeed in presenting a fully dimensional acoustic environment.  Yes, it's impractical.  5.1 is impractical too, but we do manage it because it's clearly an improvement over 2 channel.  Currently, Audyssey DSX is a synthetic post process, as there is no true 11.2 channel material easily available.  However, BD has the capability of lots of high bit rate, uncompressed audio channels, and the players are now connecting to existing systems.  There's your delivery medium, and your reproduction system.  All we really need is the content, and a nice way to fold 11.2 channels down to 5.1, which as I understand it, DSX is all set to do. 

So, the problems with 3D audio systems from two speakers are: a narrow listening window, poorly controlled directional placement, somewhat random amounts of reproduced 3D effects due to system design and acoustic issues, and more.  Multichannel systems solve the listening window issue, the placement issue, are consistent, easily auto-calibrated, and much more immune to acoustic issues. 

I doubt 3D from two speakers will ever really be widely accepted.  But 20 years ago there were only a few home 5.1 systems.  Today there are millions, and thousands of 7.1 systems, with numbers growing all the time.  Though none of these concepts, 3D or multichannel, present true "sonic holography", the multlichannel approach is, for the foreseeable future, the only reliable and practical way of presenting 3D sound at home.