Hello,
Yes, of cource Ambiophonics does not create the original soundfield, instead it does something that will satisfy human psychoacoustics to the great lengths. In that way it is very satisfying experience.
Wavefield synthesis methods on the other hand try the opposite, they don't care about psychoacoustics but rely to the assumption that if the original waveform is accurately copied to the listening space, human hearing is also satisfied.
The problem with WFS is it has severe limitations when freq is going higher, due to simple fact that when you insert the head to the listening point the wavefield is altered by diffraction. Since most WFS methods try to copy the waveform in a singular point!
Ambisonic is a derivative of WFS but some psychoacoustics included and is very interesting method and for example Jerome Daniel has done some good work in higher order Ambisonics.
But the practicality of WFS in home environment... well, I would say WAF is negative
- Elias
syntheticwave said:
Yes, of cource Ambiophonics does not create the original soundfield, instead it does something that will satisfy human psychoacoustics to the great lengths. In that way it is very satisfying experience.
Wavefield synthesis methods on the other hand try the opposite, they don't care about psychoacoustics but rely to the assumption that if the original waveform is accurately copied to the listening space, human hearing is also satisfied.
The problem with WFS is it has severe limitations when freq is going higher, due to simple fact that when you insert the head to the listening point the wavefield is altered by diffraction. Since most WFS methods try to copy the waveform in a singular point!
Ambisonic is a derivative of WFS but some psychoacoustics included and is very interesting method and for example Jerome Daniel has done some good work in higher order Ambisonics.
But the practicality of WFS in home environment... well, I would say WAF is negative
- Elias
Elias said:Hello,
The problem with WFS is it has severe limitations when freq is going higher, due to simple fact that when you insert the head to the listening point the wavefield is altered by diffraction. Since most WFS methods try to copy the waveform in a singular point!
Ambisonic is a derivative of WFS but some psychoacoustics included and is very interesting method and for example Jerome Daniel has done some good work in higher order Ambisonics.
But the practicality of WFS in home environment... well, I would say WAF is negative
- Elias
Hi Elias,
I agree, ambiophonics is the most practicable way to fault the brain today. Also Ambisonics deliver an excellent reproduction, today the only available way for reproduce in all three room dimensions. But only in a small listening area, however WFS is a real spatial solution. The realized plants are reduced upon the horizontal plain 2D solution and the speaker rows all around the listener of course the ultimate reason for a WAF war.
But you know the 3D proposal from www.holophony.net ? It uses only speakers behind the screen wall, invisible for the mistress and protected from daily pouring process by the screen sheet.
H.
syntheticwave-
Is that your site? If so, I see no mention of VBAP. I've read a few papers on the comparison of Binaural/transaural/ambiophonics, ambisonic, and VBAP and it seems to be a variation on the ambisonic approach. What are your thoughts? Have you heard such a system?
Do you feel ambiophonic/ambisonic/VBAP/WFS is needed if just using a simple line array? The basic concept is the same-over-ride the room's spacial effects via a "wall of sound" type concept increasing listening sweet spot area and improving dynamics to make it sound more life like?
Is that your site? If so, I see no mention of VBAP. I've read a few papers on the comparison of Binaural/transaural/ambiophonics, ambisonic, and VBAP and it seems to be a variation on the ambisonic approach. What are your thoughts? Have you heard such a system?
Do you feel ambiophonic/ambisonic/VBAP/WFS is needed if just using a simple line array? The basic concept is the same-over-ride the room's spacial effects via a "wall of sound" type concept increasing listening sweet spot area and improving dynamics to make it sound more life like?
durwood said:
Hi Durwood,
yes, www.holophonics.net is my site. I am a real hard core fan of wave field synthesis. Its right, Vector based Amplitude Panning, Ambiophonics, Ambisonics and wave field synthesis related approaches for overcome the phantom source problems, which formidable described on Ralph Glasgals Ambiophonics website.
I was able to listen wave field synthesis loudspeakers many times, ever the spatial depiction inside the horizontal plain of the loudspeakers has enthused my. But until today all realized approaches are reduced upon the horizontal plain of the loudspeaker rows around the listener. As mentioned, the WAF is below zero for such approach; besides the playback room acoustics remain disturbing, because all, the direct wave, first reflections and reverberation tail become restored completely by the commonly used impulse response based approach.
My idea was to build up a real 3D solution, including the playback room acoustic into the synthesis by a two dimensionally speaker field behind the screen. By reduction upon the first reflections instead upon the horizontal plain the computing power becomes bearable for 3D. It works in principle as like Yamahas sound projectors, but isn’t searching for restore virtually loudspeakers, but the source itself, including its first reflections in the recording room. Unfortunately you would need up to 1296 single speakers, each driven by an own small amp for widely aliasing free reproduction. But then breathtaking possibilities arise, you can align the sound waves of the virtually sources nearly at will. Per example, if your living room ceiling high 2,7 m, but you will fake a 12 m concert hall all what you need is retarding the radiation in the direction of your living room ceiling until the wave arrive by the listener in the same temporally condition after the direct wave as in the concert hall. If you do so with all wall reflections the genuine sound field should become restored completely. The reverberation tail by convolution as in the ambiophonics approach, its direction hardly does matter psycho acoustically. The playback room acoustics would remain hardly important by such solution because of the near field effect of the huge speaker wall. Besides the dynamic increasing strongly, the huge resulting membrane is adapted onto the air resistance as good as a horn of the same dimension!
The idea for including the playback room properties into the wave field synthesis in a common system I have patented in Germany, but nobody wants to realize it until because of the effort. As you know possibly, Mark Twain says all people with a new idea regarding spinner, until the procedure is commonly used.
H.
syntheticwave said:
Hi Durwood,
yes, www.holophonics.net is my site. I am a real hard core fan of wave field synthesis. Its right, Vector based Amplitude Panning, Ambiophonics, Ambisonics and wave field synthesis related approaches for overcome the phantom source problems, which formidable described on Ralph Glasgals Ambiophonics website.
I was able to listen wave field synthesis loudspeakers many times, ever the spatial depiction inside the horizontal plain of the loudspeakers has enthused my. But until today all realized approaches are reduced upon the horizontal plain of the loudspeaker rows around the listener. As mentioned, the WAF is below zero for such approach; besides the playback room acoustics remain disturbing, because all, the direct wave, first reflections and reverberation tail become restored completely by the commonly used impulse response based approach.
My idea was to build up a real 3D solution, including the playback room acoustic into the synthesis by a two dimensionally speaker field behind the screen. By reduction upon the first reflections instead upon the horizontal plain the computing power becomes bearable for 3D. It works in principle as like Yamahas sound projectors, but isn’t searching for restore virtually loudspeakers, but the source itself, including its first reflections in the recording room. Unfortunately you would need up to 1296 single speakers, each driven by an own small amp for widely aliasing free reproduction. But then breathtaking possibilities arise, you can align the sound waves of the virtually sources nearly at will. Per example, if your living room ceiling high 2,7 m, but you will fake a 12 m concert hall all what you need is retarding the radiation in the direction of your living room ceiling until the wave arrive by the listener in the same temporally condition after the direct wave as in the concert hall. If you do so with all wall reflections the genuine sound field should become restored completely. The reverberation tail by convolution as in the ambiophonics approach, its direction hardly does matter psycho acoustically. The playback room acoustics would remain hardly important by such solution because of the near field effect of the huge speaker wall. Besides the dynamic increasing strongly, the huge resulting membrane is adapted onto the air resistance as good as a horn of the same dimension!
The idea for including the playback room properties into the wave field synthesis in a common system I have patented in Germany, but nobody wants to realize it until because of the effort. As you know possibly, Mark Twain says all people with a new idea regarding spinner, until the procedure is commonly used.
H.
Maybe it is not the same thing but i know the IRCAM in France is trying with speakers all over the room + phase modification to create holographice sound
Maybe ther is a link somewhere
TomatoBangBang said:
Hi, TomatoBangBang
its in principle the same think as by IRCAM http://recherche.ircam.fr/equipes/salles/WFS_WEBSITE/Index_wfs_site.htm and others, the largest approach until is build on the TU Berlin with more as thousand speakers, but all are using horizontal speaker rows around the listener. Two dimensional fields were building up by Prof. Ono on the Tokyo University in 1998 or so. It was too much effort on this time.
Elias said:
Hi Elias,
Mikels calculation was right in principle for a Holophony approach by all around speakers. If you use the reflecting surfaces as sound source you need 1/6 of that = 66 000. That would work aliasing free up to 20 kHz. But you can enlarge the space between the single speakers up to 2 inch, if you limits by 3,6 kHz. Above the ITD`s becomes ambivalent, for ILDs you can use phantom source detection according Witteks OPSI (Optimized Phantom Source Imaging ) approach by some single tweeters.
Consider, the practically plants use up to 10 inch spacing between the speakers without the aliasing effekts become very disturbing.
I have not the possibility to realize it in practice. But why I should? It sounds like the genuine because it is a virtually copy!
H.
Hi graaf,
i wasn’t familiar with the Stereolith- setup, but now I have seen the
http://www.stereolith.ch/db/stereo/template=intro.html&var=ln=en
site. It isn’t a 3D approach, NO information regarding the sound source elevation and, what’s much more disturbing, no information regarding the elevation of the early reflection sources including in the signal.
Stereolith makes random early reflections from elevated directions; because of the directed radiation possibly more as the radio in my kitchen, but in principle I would consider neither as a 3D approach.
H.
syntheticwave said:
actually I refer rather to the so called "Stereolith-like set up" that is without mono tweeter - only speakers pointing sideways
I know of opinion of one very credible and very competent sound engineer, chief engineer at one of two Polish biggest nation-wide radio stations who had opportunity to examine the original Stereolith
he was not impressed and it took me a lot of effort and patience to persuade him to try Stereolith like set up
fortunately finally He did it and his opinion was very favourable
on the other hand read this: http://www.magazine-audio.com/essais-hi-fi/essais-enceintes-hi-fi/22
one of the very rare Stereolith reviews
They marvel at 3D and the vertical dimensions of the soundstage:
"Profondeur et hauteur d’image musicale hors du commun"
I don't know how it works I have to admit
but it works - try it
best!
graaf
Originally posted by graaf actually I refer rather to the so called "Stereolith-like set up" that is without mono tweeter - only speakers pointing sideways I know of opinion of one very credible and very competent sound engineer, chief engineer at one of two Polish biggest nation-wide radio stations who had opportunity to examine the original Stereolith he was not impressed and it took me a lot of effort and patience to persuade him to try Stereolith like set up
…pointing sideways improves the IACC, such attractions very important for spatial impression. But you need sufficient time delays, caused by sufficient spaced playback room dimensions to achieve a psychoacoustic benefit.
On the other hand, all spatial impression for the Stereolith-like approach is caused by the playback room. No difference in the arrival time for the engendered ceiling or wall reflections whether the recording room ceiling was in 3 meters or 30 meters high. A real 3 D approach would make a difference. But you have no appropriated source elevation information in a stereo or surround signal, you cannot restore the 3D impression of the recording room, only those of the playback room.
H.
syntheticwave said:
why? what are sufficient?
syntheticwave said:
in what sense it is "caused by the playback room"?
because I don't understand all this I don't know how to explain it BUT I can hear very clear differences between various recordings
acoustics I can hear with various recordings is certainly not acoustics of my listening room, absolutely!
jazz club, rock arena, classical auditorium, church - it all makes significant clearly audible difference
how can You explain this fact?
syntheticwave said:
are You sure? what about recorded floor reflections? aren't floor reflections main (apart from HRTF) spatial cue for sound source elevation?
best regards!
graaf
graaf said:
All walls, except floor, at least one, better two meters apart from the Boxes. Two meters are causing time differences between the direct wave and first wall or ceiling reflections of around ten milliseconds. Below that all addition of direct wave and delayed wave fronts causing comb filter effects ( 100 Hz period time is assigned 10 msec. ) , decreases the tonal accuracy and speech intelligibility and destroy the initial time delay gap.
But in the 10 to 60 ms window the fist reflections the best what can happen, establishing the room size impression, render possible the distance perception of the source and boost the perceived volume impression. Later reflections are need for reverberation caused room impression, but cannot support the detection of the source position and become disturbing for the speech intelligibility. You can see that in the calculation formula for speech intelligibility or I that little animation:
http://www.holophony.net/pictures/Stereo50.swf
You see clearly the utterly different impulse response, as far as the playback room very differ in size regarding the recording room. Possibly your boxes would work well in a balloon, because as you can adapt the early reflection times by air pressure inside.
Another way is shown I that animation:http://www.holophony.net/pictures/principe.swf
, for all directions:
http://www.holophony.net/pictures/WFS_transformation_principle.swf
graaf said:
Not alone the acoustics of your playback room, I admit, also the recorded early reflections and the reverberation tail in the recording remain audible. Reverberation can deliver room impression, but is hardly important for all directive effects. We haven`t the possibility for connect the late reflection source positions with any optically impression, no chance for a learning process.
The influence of the playback room overrates the direct radiated wave fronts usually in a distance below one meter in the playback room, as far as the playback room isn¡¦t specially dampened or the boxes radiate very directed.
graaf said:
The recorded reflections in all traditionally procedures like stereo or surround become includes in the signal, radiated from the boxes. Per example the first ceiling reflection in the recording room hits the listener in the concert hall from 60 degree elevation angle. Recorded from the main microphones its source direction during the playback process from your boxes is, like for the floor reelection, the Loudspeaker direction, 0 degree elevation level! All boxes aligned in a horizontal plain, but a horizontal plain is two dimensionally, I am sure.
H.
syntheticwave said:
on the other hand Toole claims that early reflections improve speech intelligibility?
syntheticwave said:
why did You mention "optical impression"? after all no audio (including WFS) is about "optical impression"
it is rather about "close Your eyes"
it is about an illusion
illusion including "optical illusion" would amount to total virtual reality
syntheticwave said:
yes that is correct, in most listening rooms the listener is in the reverberant field but what is wrong with it if as You admitted "early reflections and the reverberation tail in the recording remain audible"?
syntheticwave said:
it is true BUT only in case of conventional front-firing pseudo-point source speakers in conventional stereo triangle
this is not the case with Stereolith-like set up where the speakers completely acoustically dissapear
they are not perceived as real sources of sound
therefore neither impression of depth nor width nor height of the soundstage is limited by placement of the speakers
I understand that in conventional stereo the apparent soundstage width and height and depth of it are determined by placement of the real sound source - the speakers - plus comb filtering of sound reflected off the wall behind speakers (for depth)
this is not the case of Stereolith-like set up
I cannot explain it in an expert way but it is something completely different
best regards!
graaf
graaf said:
Yes, strongly, but only in certain time window. Is like by your woman a little, too early or too late is wrong.
Our spatial hearing isn’t an compleately inherent effect, most must be learned. Our HRTFs are so much individually because we have seen many times; a sound source in that direction in that environment sounds so. The source mostly was directly visible, for the first reflection our brain was able to calculate a mirror source position. But for the reflection of the reflection the task was too complicated. By that reason we are really able to connect a perception pattern with a source position only for the direct wave and its first reflections. Now also by closed eyes, its better at home but not in the concert hall in which the optically impression agree with the spatial audio impression.graaf said:
Isn’t wrong, it is very helpfully for the remaining accuracy. I wouldn’t go so far to say the perception of the tenor’s voice in the opera house would be utterly different regarding the phantom source between the loudspeakers. But we are very tolerant for all the lot of mistakes during the phantom source based audio reproduction. The only thing what remain really genuine like if we examine the conditions exactly is the Text of the song.graaf said:
graaf said:
I can imagine that and I was often able to listen reproductions in totally damped rooms. Boring, hardly spatial impression, the limitation upon the horizontal plain becomes very audible. Possibly you have read in the first captures in Ralph Glasgals www.ambiophonics.org site the description of the acoustic research experiments in the Carnegie Hall. As far as the reflection patterns agree with the recording room in certain degree a very realistic perception can occur. But as far as differ both rooms strongly the things going wrong. And you cannot change the detour of the first reflections in the playback room, neither wit the stereolith boxes and not for my 3D - kitchen radio.
H.
www.holophony.net
Could it be that crosstalk cancellation is impossible in real life situations? Maybe i got a knot in my thoughts, but it seems to be the case.
With software crosstalk cancellation, we add an inverse, slightly lower amplitude, delayed and mirrored in stereo position signal to the original signal (either recursive or not, depending on the algorithm). Now, imagine a mono signal fed into this processor. You get an inverse signal, which is identical on both speakers. Depending on the delay time, you will get a certain frequency (for a typical ambio setup that will be around 1-3khz) below which the inverse signal is less than 1/4 WL out of phase with the original signal, and both will cancel each other more or less completely. Above this frequeny, the cancellation will occur periodically like a comb filter (only for constant signals like sine waves, less so for time-variant signals). This process will only work for signals, which are panned completely to the left or right speaker and are above the frequency where this cancellation will happen.
For hardware crosstalk cancellation, aka sound absorbing wall between the speakers, other drawbacks occur, or rather one should say, it doesnt do what it should do. The barrier theory assumes, that waves travel only in a straight line. Since waves bend, as soon as the end of the barrier is reached, even with a barrier, signals from the left speaker will reach the right ear and the other way round. The only thing such a barrier does is positioning the speakers at a virtual position, which is defined by a straight line from the ear to the edge of the barrier, where the sound wave diffracts around the barrier end.
So basically, software cancellation introduces alot of artefacts into the sound and hardware cancellation doesnt actually cancel anything but positions the stereo speakers at a virtually wider angle (which is actually quite a nice effect but not what it should do).
Now, please tell me where my errors are, since it seems to work for alot of people except me. Thanks
With software crosstalk cancellation, we add an inverse, slightly lower amplitude, delayed and mirrored in stereo position signal to the original signal (either recursive or not, depending on the algorithm). Now, imagine a mono signal fed into this processor. You get an inverse signal, which is identical on both speakers. Depending on the delay time, you will get a certain frequency (for a typical ambio setup that will be around 1-3khz) below which the inverse signal is less than 1/4 WL out of phase with the original signal, and both will cancel each other more or less completely. Above this frequeny, the cancellation will occur periodically like a comb filter (only for constant signals like sine waves, less so for time-variant signals). This process will only work for signals, which are panned completely to the left or right speaker and are above the frequency where this cancellation will happen.
For hardware crosstalk cancellation, aka sound absorbing wall between the speakers, other drawbacks occur, or rather one should say, it doesnt do what it should do. The barrier theory assumes, that waves travel only in a straight line. Since waves bend, as soon as the end of the barrier is reached, even with a barrier, signals from the left speaker will reach the right ear and the other way round. The only thing such a barrier does is positioning the speakers at a virtual position, which is defined by a straight line from the ear to the edge of the barrier, where the sound wave diffracts around the barrier end.
So basically, software cancellation introduces alot of artefacts into the sound and hardware cancellation doesnt actually cancel anything but positions the stereo speakers at a virtually wider angle (which is actually quite a nice effect but not what it should do).
Now, please tell me where my errors are, since it seems to work for alot of people except me. Thanks
MaVo said:
Hi MaVo,
as far as you are using virtually sources instead of phantom sources you doesn’t need crosstalk cancellation. The Interaural cross cancellation value of a virtual sound source doesn’t differ regarding the real source.
H.
www.holophony.net
MaVo said:Could it be that crosstalk cancellation is impossible in real life situations? Maybe i got a knot in my thoughts, but it seems to be the case.
With software crosstalk cancellation, we add an inverse, slightly lower amplitude, delayed and mirrored in stereo position signal to the original signal (either recursive or not, depending on the algorithm). Now, imagine a mono signal fed into this processor. You get an inverse signal, which is identical on both speakers. Depending on the delay time, you will get a certain frequency (for a typical ambio setup that will be around 1-3khz) below which the inverse signal is less than 1/4 WL out of phase with the original signal, and both will cancel each other more or less completely. Above this frequeny, the cancellation will occur periodically like a comb filter (only for constant signals like sine waves, less so for time-variant signals). This process will only work for signals, which are panned completely to the left or right speaker and are above the frequency where this cancellation will happen.
[cut]
Now, please tell me where my errors are, since it seems to work for alot of people except me. Thanks
You aren't actually wrong, although I think your post overstates the problem a bit. You are right though that the cancellation problem becomes ill-conditioned at the top end and at the bottom end it eats up headroom as more and more of your output signal is comprised of the cancellation signal rather than 'real' signal.
This is exactly the problem that the "Optimal Source Distribution" paper/approach addresses. The basic idea is that you take a multi-way approach and position the driver pairs at an angular separation that maximizes the xtalk-cancellation ability over the frequency range they cover. for a 3-way this gives you tweeters separated by a couple inches right in front of you, mids spaced at about +-15 degrees or so, and then woofers out to the sides.
I haven't yet tried this 'opsodis' approach, but it's up on my list of things to do.
As to why ambiophonics works despite the problems - my speculation is that the +-10 degree separation coupled with the bandpass filtering typically applied to the cancellation filters means that you are getting effective cancellation over enough of a range that it the ear accepts it due to the lack ofcompeting/conflicting spatial cues.