Hello,
I'm interested in Ambiophonics and I tried it long time ago already
http://www.diyaudio.com/forums/showthread.php?s=&threadid=36254
But there are some practical barriers (pun intended) to tackle in cross talk cancelling.
I don't get it how the 'box' you describe could provide some cancelling? If you listen from distance, say, 2 meters from speaker the angle difference of the paths from the speaker to each of your ears is so small the directivity of the speaker do not change enough to provide any amplitude difference.
- Elias
graaf said:
I'm interested in Ambiophonics and I tried it long time ago already
http://www.diyaudio.com/forums/showthread.php?s=&threadid=36254
But there are some practical barriers (pun intended) to tackle in cross talk cancelling.
I don't get it how the 'box' you describe could provide some cancelling? If you listen from distance, say, 2 meters from speaker the angle difference of the paths from the speaker to each of your ears is so small the directivity of the speaker do not change enough to provide any amplitude difference.
- Elias
Meanwhile I have found a very simple solution for my problem: Listen to such pop-style recordings mono. Not only do I get rid of this three-source stereo, but also of the psychedelic spaceousness I experienced with conventional setups. Now I also won´t speak any longer of a strange signature these digital reverbs produce. The problem seems to be not in the reverberation itself, but in the dirty tricks like crosstalk-cancelling or stereo-basis broadening. We must not forget that a very small minority listenes to pop music in a correct stereo triangle. Usually the situation is either a ghetto-blaster with a stereo-basis of ca. 50 cm or two speakers at places where they are least disturbing optically. Fortunately these recordings are produced mono-compatible to make them suitable for simple kitchen radios.
Elias,
since you seem to have my problem even with a conventional speaker setup, maybe mono is your solution, either with two speakers or with one. Nobody should be afraid of making himself laughable by choosing this way.
Elias,
since you seem to have my problem even with a conventional speaker setup, maybe mono is your solution, either with two speakers or with one. Nobody should be afraid of making himself laughable by choosing this way.
MethMan said:
I would agree with adding decorrealetd reverb in the modal region because this helps to randomize the modes, its actually like adding modes in this modal weak frequency region. But above about 200 Hz. I don't see a technical reason to do this and I think that it might just be a personal preference of DG.
The damped walls do not add reverb, in fact they take it away, hence the desire to add back some reverb at the low end to compensate.
Elias said:
Our difficulty in localizing frequencies below about 500 Hz is well know and by 100 Hz it is virtually impossible. So clearly the localization capacity is falling starting somewhere below 1 kHz. There is no exact transition point, its a gradual thing. Thus, if the directivity gradually decreases (widens) below say 500 Hz. then this could not be a problem for imaging.
Hello,
Your claim is simple not true. There is nothing correct in your statements. Localisation accuracy does not decrease at lower frequencies despite common believe that some may have. See the pic from Blauert Spatial Hearing, it shows that localisation blur is getting smaller at lower freqs making localisation actually _improving_, not getting more difficult as you claim! Having localisation blur of only ONE degree at 500Hz hardly makes these freqs 'virtually impossible' to localise.
I think you need to revise your understanding of sound localisation and spatial hearing.
- Elias
gedlee said:
Your claim is simple not true. There is nothing correct in your statements. Localisation accuracy does not decrease at lower frequencies despite common believe that some may have. See the pic from Blauert Spatial Hearing, it shows that localisation blur is getting smaller at lower freqs making localisation actually _improving_, not getting more difficult as you claim! Having localisation blur of only ONE degree at 500Hz hardly makes these freqs 'virtually impossible' to localise.
I think you need to revise your understanding of sound localisation and spatial hearing.
- Elias
Attachments
Elias said:
well - frankly speaking I don't get it either
therefore I have nowhere never said that there is crosstalk cancelling
rather I have said that": "crosstalk is diminished" or perhaps even more precisely: "negative effect of crosstalk is diminished"
BUT this is of course just a hypothesis
in fact my aim was to create an ambiopole that would require minimal physical barier according to the formulas given by Ralph Glasgal
BUT to my surprise I discovered that in case of such "ambiopole" adding theoretically required physical barier leads to no perceptually significant change!
I was surprised but that was what I had found then
so frankly speaking I have no certainty as to the principle operating here
I supposed that it has to have something to do with directivity of box speaker, of a dynamic driver on a baffle (see picture attached)
best!
graaf
Attachments
Elias said:
I will have to look that up in Blauert as something doesn't seem right.
Clearly localization at LF has to get worse because ALL theories of localization depend on interaural differences which have to go to zero at LF. There is no significant phase difference at the ears at LFs nor a level difference - hence there cannot be localization in that situation.
interesting PhD dissertation:
http://ses.library.usyd.edu.au/bits...524302whole.pdf
"Spatial Hearing with Simultaneous Sound Sources: A Psychophysical Investigation"
best!
graaf
http://ses.library.usyd.edu.au/bits...524302whole.pdf
"Spatial Hearing with Simultaneous Sound Sources: A Psychophysical Investigation"
best!
graaf
gedlee said:
This confusion comes up freqeuntly about sensitivity to time vs phase. It is best to think in units of time.
The general statement is that we are sensitive to interaural time rather than phase. One rule of thumb is that we are sensitive to about 25usec, or so, which would be a smaller interaural phase difference at 500 Hz than at 800 Hz (of course).
I am not sure where this notion of spatial acuity being so poor below 1000Hz comes from. That is simply not true.
I carefully reread those sections of Blauert - it raises more questions than it answers.
First, there is no data below 500 Hz so what happens at those frequencies is not supported in any way by Blauert's book. Also the chart shown is for real sources in free space and not the localzation of phantom sources from a stereo signal in a small room. Hence it is almost totally inapplicable to the question at hand.
It is clear that the localization of a sound in a small room must vanish at very low frequencies due to several effects - the room modes and the lack of interaural signal differences. Common experince also suggests that this is true.
Thus, while there may be a localization blur minimum at 500 Hz and hence some directivity at this frequency may be a benefit, this is not really realisitic in a small space. It is also clear that at "low enough" frequencies there would be no benefit to source directivity because the room would completely dominate the perception issue not the loudspeaker.
I must admit that I am now very curious about our ability to localize sounds below 500 Hz. Is it room dominated or source dominated? And interesting topic.
First, there is no data below 500 Hz so what happens at those frequencies is not supported in any way by Blauert's book. Also the chart shown is for real sources in free space and not the localzation of phantom sources from a stereo signal in a small room. Hence it is almost totally inapplicable to the question at hand.
It is clear that the localization of a sound in a small room must vanish at very low frequencies due to several effects - the room modes and the lack of interaural signal differences. Common experince also suggests that this is true.
Thus, while there may be a localization blur minimum at 500 Hz and hence some directivity at this frequency may be a benefit, this is not really realisitic in a small space. It is also clear that at "low enough" frequencies there would be no benefit to source directivity because the room would completely dominate the perception issue not the loudspeaker.
I must admit that I am now very curious about our ability to localize sounds below 500 Hz. Is it room dominated or source dominated? And interesting topic.
Hello,
In the case the link given by Graaf doesn't work, try this one:
frontpage:
http://ses.library.usyd.edu.au/bitstream/2123/576/2/adt-NU20041221.13524301front.pdf
thesis:
http://ses.library.usyd.edu.au/bitstream/2123/576/1/adt-NU20041221.13524302whole.pdf
Best regards from Paris, France
Jean-Michel Le Cléac'h
In the case the link given by Graaf doesn't work, try this one:
frontpage:
http://ses.library.usyd.edu.au/bitstream/2123/576/2/adt-NU20041221.13524301front.pdf
thesis:
http://ses.library.usyd.edu.au/bitstream/2123/576/1/adt-NU20041221.13524302whole.pdf
Best regards from Paris, France
Jean-Michel Le Cléac'h
graaf said:
To me the following quote from the above (most interesting!) work says what I have been saying:
"This work demonstrated that speech contains high frequency
energy that is essential for accurate three-dimensional localisation."
It is the HF that determine spatial imaging, not the LFs. This is completly consitent with my philosophy and completly contradicts the satements made that LF are important to imaging. Get the HF right and the problem is 95% solved reagrdless of the directivity of the LFs.
"This work demonstrated that speech contains high frequency
energy that is essential for accurate three-dimensional localisation."
It is the HF that determine spatial imaging, not the LFs. This is completly consitent with my philosophy and completly contradicts the satements made that LF are important to imaging. Get the HF right and the problem is 95% solved reagrdless of the directivity of the LFs.
I totally I agree with Earl but nonetheless there is useful spatial information in signals lower 500 Hz. Find out for yourself with the following two examples of a low-pass filtered male voice:
Low-Pass (24 db) at 150 Hz (AAC, 2 MB)
Low-Pass (24 db) at 300 Hz (AAC, 2 MB)
It also becomes pretty obvious how important higher frequencies are to understand one word at all
Best, Markus
Low-Pass (24 db) at 150 Hz (AAC, 2 MB)
Low-Pass (24 db) at 300 Hz (AAC, 2 MB)
It also becomes pretty obvious how important higher frequencies are to understand one word at all
Best, Markus
Upon further reading, I will rest my case with the following conclusions drawn in this paper:
"It was interesting to find that the separation of concurrent sounds at the frontal location was not impaired by high-pass filtering the signals. It seems that in this situation, the ongoing low-frequency ITDs are redundant, and the perception is maintained by high-frequency acoustical cues in this case. It has been shown previously that high-pass filtering does not significantly affect the accuracy of localisation of single-source stimuli (Butler and Humanski, 1992; Carlile and Delaney, 1999), and the present findings suggest that high-frequency cues can also be sufficient for identifying that more than one location is present in a multiple-source stimulus."
The LFs are "redundant" and add nothing to localization.
"It was interesting to find that the separation of concurrent sounds at the frontal location was not impaired by high-pass filtering the signals. It seems that in this situation, the ongoing low-frequency ITDs are redundant, and the perception is maintained by high-frequency acoustical cues in this case. It has been shown previously that high-pass filtering does not significantly affect the accuracy of localisation of single-source stimuli (Butler and Humanski, 1992; Carlile and Delaney, 1999), and the present findings suggest that high-frequency cues can also be sufficient for identifying that more than one location is present in a multiple-source stimulus."
The LFs are "redundant" and add nothing to localization.
markus76 said:
Markus
Lets be clear that I am not saying that there is not spatial information below 500 Hz only that our perception of "image" will be dominated by a sound system / rooms performance above 500 Hz. I simply said at the start of this discussion that one does not need to worry as much about directivity below about 500 Hz because it is not a major factor in localization - especially in a small room. I stand by that statement in light of the paper that Graaf highlighted (thanks).
gedlee said:
You are changing the arguments & context around and this confuses the conversation.
1. Spatial acuity is usually defined (unless otherwise noted) for the horizontal plane. this is where out acuity is best. The definitive work was done by Stevens in the the 1930s and more importantly by Mills in the late 1950s (in published in JASA).
2. Yes (of course) acuity in the vertical plane requires high frequency energy. However, even if there is high frequency energy, acuity will still be about 5-10 times poorer in the vertical plane compared to the horizontal plane (many references in JASA).
3. Spatial acuity in the depth plane (distance) is quite poor overall. There are multiple cues, but researchers generally agree the most critical cue is some sort of ratio (metric is not always agreed upon) between the direct-to-reverberant energy ratio. This reverberant energy could be in the recording or the what is introduced by the playback (room interactions). The issue of depth perception is rather complicated and I have made a very cursory overview of it. The best current research is probably done by Simpson and published in JASA.
BTW, a number of confusions were made by several folks regarding "imaging". I always recommend that folks should read the research on "stereophony" (as it is formally defined) by Ben Bauer and published in the early 1960s (in JASA). It is interesting stuff and unfortunately it seems to have been forgotten.
markus76 said:
Markus is right!
we really should distinguish between FOUR fundamentally different situations:
1) localization of multiple simultaneous real sound sources under free field/anechoic conditions
2) localization of multiple simultaneous real sound sources under reverberant conditions
3) localization of multiple simultaneous stereophonic phantom sources under free field/anechoic conditions taking into account that stereo loudspeakers themselves represent real sound sources for our hearing
4) localization of multiple simultaneous stereophonic phantom sources under reverberant conditions taking taking into account that stereo loudspeakers themselves represent real sound sources for our hearing
the fourth situation is precisely what we have in our listening rooms, it is our audiophile reality which is worth to be discussed
best regards!
graaf
Hello,
Wrong again. Not all the localisation is based on two ear signals. Humans are able to localise sounds with ONE ear only! Ever heard about pinna localisation? Seems to be you haven't. It allows you to localise sounds at high freqs, say above 3kHz, and with a little bit of training of one ear listening the accuracy improves. Remember spatial hearing is a learned process. Those people who lose their hearing ability of onother ear due to accidents or other reasons are able to gain some of the spatial hearing accuracy they had with two ears earlier.
Pinna+brain = the power!
- Elias
gedlee said:
Wrong again. Not all the localisation is based on two ear signals. Humans are able to localise sounds with ONE ear only! Ever heard about pinna localisation? Seems to be you haven't. It allows you to localise sounds at high freqs, say above 3kHz, and with a little bit of training of one ear listening the accuracy improves. Remember spatial hearing is a learned process. Those people who lose their hearing ability of onother ear due to accidents or other reasons are able to gain some of the spatial hearing accuracy they had with two ears earlier.
Pinna+brain = the power!
- Elias