Hello,
Are you refering a speaker having a perceived size, and this 'size' would invalid perception of phantom image?
A normal speaker have bass at the bottom, midrange above that and tweeter at the top and it is not difficult at all to hear different freqs comming from different hights. Is this the thing you are refering to?
I think the main problem is not in the speaker (as it can be a point source to the certain limit) but the problem is in the stereo principle itself. It simply does not work consistently over the whole audio band and the ear+brain combination is not satisfied but concludes something is not real.
- Elias
Tom Danley said:
Are you refering a speaker having a perceived size, and this 'size' would invalid perception of phantom image?
A normal speaker have bass at the bottom, midrange above that and tweeter at the top and it is not difficult at all to hear different freqs comming from different hights. Is this the thing you are refering to?
I think the main problem is not in the speaker (as it can be a point source to the certain limit) but the problem is in the stereo principle itself. It simply does not work consistently over the whole audio band and the ear+brain combination is not satisfied but concludes something is not real.
- Elias
Hello,
Graaf, from memory (and my memory until now was still good) I retain the transition frequency of 4000Hz between the frequency domain for which audition can manage phase and the frequency domain for which it is inaudible.
Earl, phase contribution to the enveloppe of a given signal is important. If, as you said correctly, our audition is sensible to the enveloppe of the signal then phase coherence must be kept.
IMHO we have to keep the enveloppe of the sound the most correct possible and this means that we have to kept the group delay variations very small throughout the largest ferquency interval.
In aparte, I'll add that when high frequency components are removed from a signal (even musical signal) its enveloppe is often less smooth and the sound is more crisp. To give a simplistic image of that, someone has to think about the shape of a square wave which came through a low pass (example: a square wave at the output of a CD player), the top and bottom of the signal presents ondulations. That's why few apparatus (harmonizer?) try to recover the missing high frequencies components and to add them to the signal outcoming from the CD. ( Japanese people seems very sensible to that).
I don't want to maintain that dimensions of the instruments can be audible for everyone. But I guess it is like phase distortion audibilty, a small percentag of the people is sensible to it.
Best regards from Paris, France
Jean-Michel Le Cléac'h
Graaf, from memory (and my memory until now was still good) I retain the transition frequency of 4000Hz between the frequency domain for which audition can manage phase and the frequency domain for which it is inaudible.
Earl, phase contribution to the enveloppe of a given signal is important. If, as you said correctly, our audition is sensible to the enveloppe of the signal then phase coherence must be kept.
IMHO we have to keep the enveloppe of the sound the most correct possible and this means that we have to kept the group delay variations very small throughout the largest ferquency interval.
In aparte, I'll add that when high frequency components are removed from a signal (even musical signal) its enveloppe is often less smooth and the sound is more crisp. To give a simplistic image of that, someone has to think about the shape of a square wave which came through a low pass (example: a square wave at the output of a CD player), the top and bottom of the signal presents ondulations. That's why few apparatus (harmonizer?) try to recover the missing high frequencies components and to add them to the signal outcoming from the CD. ( Japanese people seems very sensible to that).
I don't want to maintain that dimensions of the instruments can be audible for everyone. But I guess it is like phase distortion audibilty, a small percentag of the people is sensible to it.
Best regards from Paris, France
Jean-Michel Le Cléac'h
graaf said:
Hello,
If you want to have low waveform distortion, then you must avoid room reflections!
Consider stereo speakers operating as Blumlein intended them to be working, i.e. amplitude panned waves from speakers summing at the listening position and forming an interference field where the brain detects the phase differences between the ears, this works below, say, below 700Hz. If you include room reflections below 700Hz all this collapses and brain cannot form a image in the intended phantom location.
Thus this should have some advice to loudspeaker design as the worst possible speaker has a monopole midrange producing a lot of room reflections. Best is to use highly directivity source below 700Hz. Just the opposite that is commonly employed!
- Elias
Russell Dawkins said:
If you want to have low waveform distortion, then you must avoid room reflections!
Consider stereo speakers operating as Blumlein intended them to be working, i.e. amplitude panned waves from speakers summing at the listening position and forming an interference field where the brain detects the phase differences between the ears, this works below, say, below 700Hz. If you include room reflections below 700Hz all this collapses and brain cannot form a image in the intended phantom location.
Thus this should have some advice to loudspeaker design as the worst possible speaker has a monopole midrange producing a lot of room reflections. Best is to use highly directivity source below 700Hz. Just the opposite that is commonly employed!
- Elias
In human beings the nuerons, the fundamental elements of our nervous system have a response time of about 1 ms. It tkes this long for a nerve cell to chemically recover enough energy to fire again. In our hearing this means that individual hairs cells cannot possible fire sychonous to the wavefor above 1 kHz. and it is believed that this point is more like 500 Hz. This is clearly seen in the Loudness curves where the nature of our hearing changes dramatically below 500 Hz. and is nearly flat, with some resonances above 500 Hz. The transition point to phase inaudibilty is AT MOST 1 kHz and most people believe that it is more like 500 Hz.
To Jean-Michels comment about phase coherence required in an impulsive signal, this is completely untrue. I can detect an impulse in a mean squared signal even though all phase information is lost. How fast a transient that an envelope detector can track all depends on the integration time of the detector. There is much data on this, but it is very clear that phase coherence IS NOT REQUIRED to detect an impulse. Now if the phase change is sufficient to cause the group delay to exceed the integration time of the detector then there can be an effect on the output. But not until sufficient phase change has occured will this be detectable.
"IMHO we have to keep the enveloppe of the sound the most correct possible and this means that we have to kept the group delay variations very small "
This is completely true, and I couldn't agree more, but it has no bearing on whether phase in and of itself is detectable. Only that group delay needs to be kept below some threshold. And according to our resesrch this threshold varies with absolute playback level (and of course it is frequency dependent).
To Jean-Michels comment about phase coherence required in an impulsive signal, this is completely untrue. I can detect an impulse in a mean squared signal even though all phase information is lost. How fast a transient that an envelope detector can track all depends on the integration time of the detector. There is much data on this, but it is very clear that phase coherence IS NOT REQUIRED to detect an impulse. Now if the phase change is sufficient to cause the group delay to exceed the integration time of the detector then there can be an effect on the output. But not until sufficient phase change has occured will this be detectable.
"IMHO we have to keep the enveloppe of the sound the most correct possible and this means that we have to kept the group delay variations very small "
This is completely true, and I couldn't agree more, but it has no bearing on whether phase in and of itself is detectable. Only that group delay needs to be kept below some threshold. And according to our resesrch this threshold varies with absolute playback level (and of course it is frequency dependent).
Hello,
There are more into it obviously, but I'm also thinking about ITD at low freqs which can be translated to interaural phase difference as it is what stereo is based on in theory. However phase only works up to about 700Hz until head shadowing effects and wavelength beeing too short.
But I think the most important mechanism of spatial hearing is the pinna localisation. This dominates somewhere above 3kHz. Some other cues beeing present at the same time, pinna will determine the perceived location, or confusion in the brain.
- Elias
markus76 said:
There are more into it obviously, but I'm also thinking about ITD at low freqs which can be translated to interaural phase difference as it is what stereo is based on in theory. However phase only works up to about 700Hz until head shadowing effects and wavelength beeing too short.
But I think the most important mechanism of spatial hearing is the pinna localisation. This dominates somewhere above 3kHz. Some other cues beeing present at the same time, pinna will determine the perceived location, or confusion in the brain.
- Elias
Hello,
This is very important indeed.
The most easiest loudspeaker 'disappearing act' is performed so that you make full range omnidirectional speakers and place them in a room far away from the listeming position. In this way you place yourself in the diffuse field conditions where sound is coming from every direction at equal strength. Cannot localise the speaker in this way. BUT there is no imaging either! All you hear is the room reflections. How can you do image in a diffuse field?
- Elias
gedlee said:
This is very important indeed.
The most easiest loudspeaker 'disappearing act' is performed so that you make full range omnidirectional speakers and place them in a room far away from the listeming position. In this way you place yourself in the diffuse field conditions where sound is coming from every direction at equal strength. Cannot localise the speaker in this way. BUT there is no imaging either! All you hear is the room reflections. How can you do image in a diffuse field?
- Elias
gedlee said:
There has been some work from van Hemmen and others that show that ensembles of neurons can fire at least an order of magnitude more exact than single neurons can do.
http://www.ncbi.nlm.nih.gov/pubmed/...nel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum
Hello Earl,
I never spoke about pulse detection.
Pure sine signals or intermodulated signals often used for psychoacoustical analysis may present if they possess a HF component such parts with a high slope our internal ear will difficultly track. But you surely knows, like me, signals (e.g. triangular) that don't show high speed slope while they possess high frequency components. Such slow speed signals according to your comment should be easily tracked by the internal ear. If we remove one HF component or if we turn its phase drastically, the shape of the signal may presents high speed slope that' our ear will track less easily. So we cannot generalize and such a HF limit for phase audibility is IMHO questionnable.
This topics is somehow related to phase distorsion audibility. Some of us can feel differences between LR (low-pass + high-pass 4th order) and Le Cléac'h crossover if the relay frequency is over 500Hz, even if the overall response is the same. IMHO this can only be explained by phase audibility over 500Hz.
Have a good weekend ( I am offline until monday).
Jean-Michel Le Cléac'h
I never spoke about pulse detection.
Pure sine signals or intermodulated signals often used for psychoacoustical analysis may present if they possess a HF component such parts with a high slope our internal ear will difficultly track. But you surely knows, like me, signals (e.g. triangular) that don't show high speed slope while they possess high frequency components. Such slow speed signals according to your comment should be easily tracked by the internal ear. If we remove one HF component or if we turn its phase drastically, the shape of the signal may presents high speed slope that' our ear will track less easily. So we cannot generalize and such a HF limit for phase audibility is IMHO questionnable.
This topics is somehow related to phase distorsion audibility. Some of us can feel differences between LR (low-pass + high-pass 4th order) and Le Cléac'h crossover if the relay frequency is over 500Hz, even if the overall response is the same. IMHO this can only be explained by phase audibility over 500Hz.
Have a good weekend ( I am offline until monday).
Jean-Michel Le Cléac'h
mige0 said:
Guys, you should come back to facts and avoid nonproven presumptions.
Here are the promised examples:
http://www.mehlau.net/audio/examples/spaciousness/
Best, Markus
Hello,
This is in great interes for me also.
I keep constanly coming accross some reference stating that Ando (if I remember correctly) defined that early reflections coming from +/-55 degrees front provides maximum spaciousness.
Of cource one can think intuitively that early reflections should come from somewhere frontal half space. But Ando actually defined the angle for maximising the effect. One can take advantage of that information and use it in a 'surround' system.
Secondly late reverberance tail should come from back half space or from sideways more than from front. This is another reason why stereo does not sound good - Late reverberance does not come from front in real life!
- Elias
markus76 said:
This is in great interes for me also.
I keep constanly coming accross some reference stating that Ando (if I remember correctly) defined that early reflections coming from +/-55 degrees front provides maximum spaciousness.
Of cource one can think intuitively that early reflections should come from somewhere frontal half space. But Ando actually defined the angle for maximising the effect. One can take advantage of that information and use it in a 'surround' system.
Secondly late reverberance tail should come from back half space or from sideways more than from front. This is another reason why stereo does not sound good - Late reverberance does not come from front in real life!
- Elias
Elias, 90° is "best":
Best, Markus
An externally hosted image should be here but it was not working when we last tested it.
Best, Markus
Markus
Reverberation and spaciuosness are seperate issues, I agree, but they are also closely linked as one cannot have spaciosness (in the room I mean) without reverberation.
Jean-Michel
Your examples are simply circumstantial evidence of something that contradicts the physics. For me to accept what you are claiming I would have to see some real data that supports it.
Reverberation and spaciuosness are seperate issues, I agree, but they are also closely linked as one cannot have spaciosness (in the room I mean) without reverberation.
Jean-Michel
Your examples are simply circumstantial evidence of something that contradicts the physics. For me to accept what you are claiming I would have to see some real data that supports it.
Elias said:
I said early on that there is a tradeof that one has to make in a small room to achieve spaciousness AND good imaging. Its a balancing act. And I have also stated that highly directional speakers are the main component of how this balance can be optimized. There appears to be a lot of background information that you are not aware of. Perhaps you should read my white paper at www.gedlee.com as I address many of these issues.
gedlee said:
"the ear is synchronous" to the input stimulus - there is the so called phase locking - up to approximately 5 kHz:
see: http://openlearn.open.ac.uk/mod/resource/view.php?id=287534
phase audibility? isn't it a completely different matter?
isn't it a matter of binaural hearing, phase differences between two ears?
in fact the so called phase Interaural Time Delay ("ongoing ITD") is useless in real life because it is ambiguous in echoic environment
see: http://ses.library.usyd.edu.au/bitstream/2123/576/1/adt-NU20041221.13524302whole.pdf
"Spatial Hearing with Simultaneous Sound Sources: A Psychophysical Investigation", s. 8-9
geez, this thread is strange, isn't it?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.