EDIT: This thread was split off from Lynn Olson's "Beyond the Ariel thread here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=100392&perpage=25&pagenumber=1
END EDIT
Just read your post on the "Nathan" thread. You wrote:
"A small percentage of audiophiles can't seem to hear conventional phantom-image stereo at all - when you push them a little further, it seems that they never experience image fusion, and always hear two speakers, no matter how well-set-up the system or where they sit."
This is fascinating! ... I'm well described here!
Perhaps a suitable short description for the affliction could be: "A-Summation Syndrome"? (Hmm, come to think of it, A-SS ... think not!)
Cilla
http://www.diyaudio.com/forums/showthread.php?s=&threadid=100392&perpage=25&pagenumber=1
END EDIT
Just read your post on the "Nathan" thread. You wrote:
"A small percentage of audiophiles can't seem to hear conventional phantom-image stereo at all - when you push them a little further, it seems that they never experience image fusion, and always hear two speakers, no matter how well-set-up the system or where they sit."
This is fascinating! ... I'm well described here!
Perhaps a suitable short description for the affliction could be: "A-Summation Syndrome"? (Hmm, come to think of it, A-SS ... think not!)
Cilla
Hello,
This quote is from this thread but apparently the words are originally said elsewhere.
However, the point I'd like to make is this is very true! This is possible because of two things:
First and main reason is the pinna localisation of human hearing mechanism. Very hard to fool hearing above, say, 1kHz to make phantom images with two loudspeakers in stereo triangle. Pinna will know better. Anyway Blumlein stereo is working correctly only below, say, 700Hz i.e. where amplitude differences of two channels are interpreted as phase differences in ears formed by the interference field at the listening position. How high in freq you can trust interference? Due the pinna, high frequencies are located to the locations of individual sources, rather than between to form a phantom.
Second reason is that spatial hearing really is a learned behaviour/phenomena of humans. At birth human has the 'tools' of hearing but has to train himself to use them. Since baby visual aids are used in hearing training.
Interesting is one can actually train himself to hear pinna localisation. I did that and I've regreted that day since
Now it's very hard to be fooled to hear phantom image! 
Instead, I hear three different sources: one is the low freq phantom, two others are the loudspeakers at high freqs. Now how it can be convincing representation of real sound event when the image is split in three?!?
I haven't seen these issues disscussed a lot in audio forums.
- Elias
This quote is from this thread but apparently the words are originally said elsewhere.
However, the point I'd like to make is this is very true! This is possible because of two things:
First and main reason is the pinna localisation of human hearing mechanism. Very hard to fool hearing above, say, 1kHz to make phantom images with two loudspeakers in stereo triangle. Pinna will know better. Anyway Blumlein stereo is working correctly only below, say, 700Hz i.e. where amplitude differences of two channels are interpreted as phase differences in ears formed by the interference field at the listening position. How high in freq you can trust interference? Due the pinna, high frequencies are located to the locations of individual sources, rather than between to form a phantom.
Second reason is that spatial hearing really is a learned behaviour/phenomena of humans. At birth human has the 'tools' of hearing but has to train himself to use them. Since baby visual aids are used in hearing training.
Interesting is one can actually train himself to hear pinna localisation. I did that and I've regreted that day since
Now it's very hard to be fooled to hear phantom image! Instead, I hear three different sources: one is the low freq phantom, two others are the loudspeakers at high freqs. Now how it can be convincing representation of real sound event when the image is split in three?!?I haven't seen these issues disscussed a lot in audio forums.
- Elias
pdan said:
Hello,
I think one reason can be confusion in the brain, as if brain cannot find a believable image location of the sound event it can conclude the sound is above since that is the area of less location accuracy.
Another reason can be wierd spectral content of the source material, as pinna localisation is based on spectral notches. By manipulating spectral content one can move perceived image position.
This information about image elevation in this link applies also for humans, by the way
http://jn.physiology.org/cgi/content/full/90/1/525
I cannot explain exactly how to train one's hearing to hear something particular. The basic thing is however you need first to be aware of the existence of the phenomena before you can understand it. After awareness comes hearing of it. As a new born baby cannot conclude the location of a specific sound event unless there is a visual aid for back up at first. I guess you can start by listening and looking at the speakers at the same time
- Elias
mige0 said:
I think one reason can be confusion in the brain, as if brain cannot find a believable image location of the sound event it can conclude the sound is above since that is the area of less location accuracy.
Another reason can be wierd spectral content of the source material, as pinna localisation is based on spectral notches. By manipulating spectral content one can move perceived image position.
This information about image elevation in this link applies also for humans, by the way
http://jn.physiology.org/cgi/content/full/90/1/525
markus76 said:
I cannot explain exactly how to train one's hearing to hear something particular. The basic thing is however you need first to be aware of the existence of the phenomena before you can understand it. After awareness comes hearing of it. As a new born baby cannot conclude the location of a specific sound event unless there is a visual aid for back up at first. I guess you can start by listening and looking at the speakers at the same time
- Elias
I've asked because I've never read about that phenomenon in the literature. I've also never heard of anybody not being able of summing localization as brought in by Lynn. My best guess is that you have strong reflections in your room that disrupt summing localization and/or plausibility and therefore lead to that kind of phenomenons.
I encourage you to take your speakers to the outside where only the floor reflections are present. Are you still able to hear each speaker individually?
Another explanation is that the sense of hearing is highly dominated by the sense of sight. Other expectations of "what to hear" can play a big role too.
Best, Markus
I encourage you to take your speakers to the outside where only the floor reflections are present. Are you still able to hear each speaker individually?
Another explanation is that the sense of hearing is highly dominated by the sense of sight. Other expectations of "what to hear" can play a big role too.
Best, Markus
Hello,
You never read about pinna localisation? Then how to understand.
I didn't say I'm not able to hear summing localisation, I said it is very hard to have phantom image but not impossible. Sometimes phantom is strong when the spectral content of the source material is right meaning not much high freq.
One should understand that summing localisation can only happen below, say, 700Hz where wave length is long enough. When going higher the head start to shadow and the interference field becomes so rugged that the phase at the ear canals becomes something absurd and cannot represent the phase of the wavefront coming from the intended phantom location. Consider also that Blumlein stereo assumes no head in the listening position!
Above, say, 3kHz pinna localisation will provide the dominant cues for spatial localisation. You cannot form a phantom image at these freqs with traditional stereo triangle where speakers are placed about 60 degrees angle in front of you. (However there are some more advanced methods to form phantom images at high freqs.)
If any doubt about pinna localisation do the following test: Can do this with electronic cross over like Behringer or such.
Connect one channel only from a good quality recording to both of your stereo speakers through a high pass filter about 1kHz so you will only hear the treble and the signal at the speakers is identical. Sit in the sweet spot. Observe if you are able to hear stable phantom image between the speakers. If you are about to hear phantom image, then summing localisation is working at high freqs too. If you cannot hear phantom, you can blame/praise your pinnas for not letting you be fooled!
If you want to test more you can try to tune the corner freq of the high pass filter and try to find the highest freq where you are still able to hear the phantom. Then report back.
- Elias
markus76 said:
You never read about pinna localisation? Then how to understand.
I didn't say I'm not able to hear summing localisation, I said it is very hard to have phantom image but not impossible. Sometimes phantom is strong when the spectral content of the source material is right meaning not much high freq.
One should understand that summing localisation can only happen below, say, 700Hz where wave length is long enough. When going higher the head start to shadow and the interference field becomes so rugged that the phase at the ear canals becomes something absurd and cannot represent the phase of the wavefront coming from the intended phantom location. Consider also that Blumlein stereo assumes no head in the listening position!
Above, say, 3kHz pinna localisation will provide the dominant cues for spatial localisation. You cannot form a phantom image at these freqs with traditional stereo triangle where speakers are placed about 60 degrees angle in front of you. (However there are some more advanced methods to form phantom images at high freqs.)
If any doubt about pinna localisation do the following test: Can do this with electronic cross over like Behringer or such.
Connect one channel only from a good quality recording to both of your stereo speakers through a high pass filter about 1kHz so you will only hear the treble and the signal at the speakers is identical. Sit in the sweet spot. Observe if you are able to hear stable phantom image between the speakers. If you are about to hear phantom image, then summing localisation is working at high freqs too. If you cannot hear phantom, you can blame/praise your pinnas for not letting you be fooled!
If you want to test more you can try to tune the corner freq of the high pass filter and try to find the highest freq where you are still able to hear the phantom. Then report back.
- Elias
Elias, either me or you hasn't understood what summing localization (or fusion) means. It's the phenomenon to hear only one sound source even when there are 2 or more coherent sound sources arriving within 1 ms at the listening position. The sound source "widens" and becomes less localizable (don't know if it's the right word in english?) when each sound source becomes less coherent or delayed strong reflections arrive at the listening position (i.e. sound becomes more spacious).
Summing localization occurs with all coherent sound sources higher 150 Hz arriving within a timeframe of 1 ms.
The effects you described in your earlier posting are covered by the precedence effect. Litovsky describes it in more detail in JASA 106.
Best, Markus[
Summing localization occurs with all coherent sound sources higher 150 Hz arriving within a timeframe of 1 ms.
The effects you described in your earlier posting are covered by the precedence effect. Litovsky describes it in more detail in JASA 106.
Best, Markus[
Hi Lynn, Markus
Lynn wrote;
“I am much more interested in energy storage and spurious emission - from the diaphragm, from the surround, from the horn, and the large surface areas of the cabinet. Although open-baffle certainly has major drawbacks in terms of requiring very large diaphragm areas, it has the significant advantage of minimizing the emissive area of the cabinet. I consider it important that conventional speaker cabinets have the greatest spurious emission in the same frequency band (200 ~ 800 Hz) where direct-radiation drivers are actually quite well-behaved and are operating in their piston band.
These spurious artifacts, like spurious artifacts in amplifiers, do not appear in conventional measurements, and may require unconventional design (and measurement techniques) to really address the problem. Removal of these artifacts falls outside the scope of standard modelling techniques (which assume perfection in the real world), and is a good reason why the physical device has to be built and measured.”
…………………………………………………………………………
This is an area has intrigued me for some time. Understand these are my views based what I see / have seen, not some “universal truth”.
Measurements don’t tell you what you hear, at best they are in effect like an exact foreign language that has a few words that are commonly known, where most of the value comes from associating what you hear with funny marks on paper or computer screen, developing your own mental dictionary about what does “this” mean.
Unfortunately, The big problem (or part of it) I think is that a microphone “listens” from one point in space, you hear from two points and then have processing attached so this isn’t the same IF what one hears from one ear is different than what one hears from the other (with the exception of the differences your head and angle cause)
I believe that It is partly the “things” which make the differences between ears, that give a single loudspeaker an identity in space if this makes sense.
The source radiates clues as to “where it is” and these clues are counter productive / harmful so far as preserving the recorded stereo image. These are in the same category as close reflections IE; bad juju
An actual point source doesn’t result in a difference between one ear and the other (except for head acoustics)
I observed the phenomena developing the Unity and then Synergy horn speakers, as time went on, these got better and better in time and spatially, they became more like a single driver and then more like a Manger.
If you have or have a Manger driver to play with or borrow, you can hear what I mean.
This will also work with a small Fostex full range driver on a large baffle.
Place a speaker at say 5 to 15 feet away from you, in the middle of the room (no close reflections) or outside and play something with natural voices. Now face the speaker with your eyes closed.
Obviously you can localize the speaker BUT… the stronger the acoustic source “clues” are, the easier it is to identify how far away the speaker is with your ears.
These “clues” or the various things that can produce them are what makes what arrives at one ear different than the other and so provides a “distance” and makes it easy to hear how far away it is as you walk closer.
On the other hand, what I found (being stuck with a multi-way speaker) was the closer I could get all the sources to combine into one acoustic point source, the smaller the clues became.
As these source interferences and re-radiations went down, it became harder and harder to “hear” how far away the speaker was.
In stereo, the effect was very positive and the phantom image (when in the listening sweet spot) appears to be seamless from one side to the other.
With less “radiated speaker identity”, like with less reflected sound compared to direct sound , more stereo image gets through it seems.
Anyway, in your quest for the audio grail I am afraid I can’t help much, I never played with open back baffles and in installed sound in large rooms, your always fighting to minimize the reverberant field so projecting energy to the rear would be undesirable.
I would suggest that if you haven’t a way, figure out how to take polar plots as your developing the crossovers. The greater the acoustic spacing between sources, the larger the number of lobes are produced at xover.
Also, in hifi there is a great concern about edge termination / refraction for horns, the presence or absence of this energy can be seen in polar plots as well.
Lynn, unless you need to go down low with that big white horn (and if you can buy one more driver haha) try the 1.4 inch version of the BMS 4550 (I forget the number) but it is a smaller format driver that works very well and wouldn’t be likely to have significant HOM’s on that horn (given its origin is about 3/8 inch dia). Generally larger drivers are more problematic up high.
Best,
Tom Danley
Lynn wrote;
“I am much more interested in energy storage and spurious emission - from the diaphragm, from the surround, from the horn, and the large surface areas of the cabinet. Although open-baffle certainly has major drawbacks in terms of requiring very large diaphragm areas, it has the significant advantage of minimizing the emissive area of the cabinet. I consider it important that conventional speaker cabinets have the greatest spurious emission in the same frequency band (200 ~ 800 Hz) where direct-radiation drivers are actually quite well-behaved and are operating in their piston band.
These spurious artifacts, like spurious artifacts in amplifiers, do not appear in conventional measurements, and may require unconventional design (and measurement techniques) to really address the problem. Removal of these artifacts falls outside the scope of standard modelling techniques (which assume perfection in the real world), and is a good reason why the physical device has to be built and measured.”
…………………………………………………………………………
This is an area has intrigued me for some time. Understand these are my views based what I see / have seen, not some “universal truth”.
Measurements don’t tell you what you hear, at best they are in effect like an exact foreign language that has a few words that are commonly known, where most of the value comes from associating what you hear with funny marks on paper or computer screen, developing your own mental dictionary about what does “this” mean.
Unfortunately, The big problem (or part of it) I think is that a microphone “listens” from one point in space, you hear from two points and then have processing attached so this isn’t the same IF what one hears from one ear is different than what one hears from the other (with the exception of the differences your head and angle cause)
I believe that It is partly the “things” which make the differences between ears, that give a single loudspeaker an identity in space if this makes sense.
The source radiates clues as to “where it is” and these clues are counter productive / harmful so far as preserving the recorded stereo image. These are in the same category as close reflections IE; bad juju
An actual point source doesn’t result in a difference between one ear and the other (except for head acoustics)
I observed the phenomena developing the Unity and then Synergy horn speakers, as time went on, these got better and better in time and spatially, they became more like a single driver and then more like a Manger.
If you have or have a Manger driver to play with or borrow, you can hear what I mean.
This will also work with a small Fostex full range driver on a large baffle.
Place a speaker at say 5 to 15 feet away from you, in the middle of the room (no close reflections) or outside and play something with natural voices. Now face the speaker with your eyes closed.
Obviously you can localize the speaker BUT… the stronger the acoustic source “clues” are, the easier it is to identify how far away the speaker is with your ears.
These “clues” or the various things that can produce them are what makes what arrives at one ear different than the other and so provides a “distance” and makes it easy to hear how far away it is as you walk closer.
On the other hand, what I found (being stuck with a multi-way speaker) was the closer I could get all the sources to combine into one acoustic point source, the smaller the clues became.
As these source interferences and re-radiations went down, it became harder and harder to “hear” how far away the speaker was.
In stereo, the effect was very positive and the phantom image (when in the listening sweet spot) appears to be seamless from one side to the other.
With less “radiated speaker identity”, like with less reflected sound compared to direct sound , more stereo image gets through it seems.
Anyway, in your quest for the audio grail I am afraid I can’t help much, I never played with open back baffles and in installed sound in large rooms, your always fighting to minimize the reverberant field so projecting energy to the rear would be undesirable.
I would suggest that if you haven’t a way, figure out how to take polar plots as your developing the crossovers. The greater the acoustic spacing between sources, the larger the number of lobes are produced at xover.
Also, in hifi there is a great concern about edge termination / refraction for horns, the presence or absence of this energy can be seen in polar plots as well.
Lynn, unless you need to go down low with that big white horn (and if you can buy one more driver haha) try the 1.4 inch version of the BMS 4550 (I forget the number) but it is a smaller format driver that works very well and wouldn’t be likely to have significant HOM’s on that horn (given its origin is about 3/8 inch dia). Generally larger drivers are more problematic up high.
Best,
Tom Danley
Why should I? You would play back two incoherent signals, i.e. the hearing can't fuse the two signals to one. But summing localization and hearing a phantom sound source IS ONE THING. There can't be a phantom sound source without coherent signals arriving within the timeframe of 1 ms. The phantom sound source can be disrupted with coherent signals within the timeframe of the precedence effect but this a totally different phenomenon.
Best, Markus
P.S. Summing localization is the one phenomenon that makes stereophony (and multichannel playback) possible. We still have no good understanding why our brain is fusing coherent signals and little research has been done in that field.
Markus, If you read again my post I'm suggesting to use only ONE channel from a recording to be played on both of the speakers, thus signals are identical. With identical signals you should hear phantom in the middle as assumed. The test will show you are NOT able to hear a phantom above, say, 1kHz. That is my point.
For the P.S. part I agree, but one should not forget that stereo as Blumlein intended works only below, say, 700Hz.
- Elias
For the P.S. part I agree, but one should not forget that stereo as Blumlein intended works only below, say, 700Hz.
- Elias
markus76 said:
Elias, of course is it possible to hear a phantom image coming from exactly the middle with a high pass filtered mono signal. As we speak I'm experiencing that with mono music coming from my MacBook Pro. Maybe I still have not understood your test or maybe you inverted the cables running to your speakers?
Best, Markus
Best, Markus
A high pass is per definition a filter that passes high frequencies and attenuates low frequencies. Do you agree with me?
I can sweep through the whole audible spectrum (using Protools). There's no change in imaging. There would be problems when setting the filter lower than 150 Hz but the MacBook speakers don't transmit anything usable below 200 Hz.
Best, Markus
I can sweep through the whole audible spectrum (using Protools). There's no change in imaging. There would be problems when setting the filter lower than 150 Hz but the MacBook speakers don't transmit anything usable below 200 Hz.
Best, Markus
Why shouldn't your test work with me sitting in an extreme nearfield situation? The stereo triangle is about 30 cm so no early reflections disturb imaging despite the fact that those drivers have serious limitations in maximum level and a lack of frequencies below 200 Hz. Following your thoughts this should help in creating the phenomenon you're trying to describe!?
Sorry, made a mistake in post Post #4477: As we're talking about a high pass filter there will be NO effect on summing localization when setting the cutoff frequency to about 150 Hz. By applying a low pass instead there would be of course localization problems because only frequencies below 150 Hz would be present.
Best, Markus
Best, Markus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.