jneutron
10-04-2004, 08:57 AM
Geeze, away from a week and both threads I was interested in are locked down, and a moderator is gone.....a record..at least my post is still there..
OK..here goes..
I said...dbt's using instant switching has a problem..because of human hearing soundstage reconstruction..
But who said anything about instantaneous switching being used for such a purpose? (besides Risch, and maybe Curl) The majority of people are going to be concerned with FR changes which is done instantaneously, because science has already shown that is what we are the most sensitive to. I would agree that you use different methods appropriate to what you are seeking.
I am not speaking of fr issues..and not single channel phase change issues..and, from what I recall (perhaps too hazily), JR specifically puts dbt down because of the rapid switching...
Laterization, as far as I am concerned is a non-issue, except in controlling the large group delay variances in speaker systems during the manufacturing process so you get consistent results unit to unit so you can avoid the costly need to "match" pairs. Once connected, even if the image is slightly skewed(which could be attributed to many things), it is constant and therefore goes unnoticed
You would be wrong if that is your assertion...:-) (c'mon, put up your dukes!!)
I am not speaking of manufacturing variations of speakers..
If you sit in front of the pair, the human mind will adjust and reconstruct the audible image in time..
Lateralization is all important when sitting in the sweet spot and listening to a stereo signal..how else do you image a human voice in the center of the speakers, when there is no source at that location?? Same with off axis imaging..
I also don't see people sitting in a chair anchored to their listening room floor with a vice that holds their head perfectly still.
It is not necessary to do so to localize either a spot source or a virtual one constructed by two speakers.
Consider two bells ten feet away, five feet apart right to left..For small angular head deviations, say 20 degrees, it is not necessary to keep the head perfectly aligned to determine the relative location of each source...no head vice required...the brain takes positional deviation into account.
So, I suppose in the real workd, the only place one would be concerned with lateralization might be in headphones
NO. lateralization is what the brain does for all aural input..headphones are a special construct...no pics here today, hope the following explanation is adequate..I don't have a scanner operational at work at present..so Tony won't have a grade school level drawing to criticise...:-)
Consider two impulse sources, one center, one off axis 1 foot to the right, 10 feet away...for center source, the ears hear simultaneously, giving front image..for the side one, it pings, and each ear detects the result after a time of flight...the right ear first, 45 microseconds before the left..you can point in the general direction of the sounds based on reception timing.....
Next, let's reproduce that using stereo..two speakers, at 12 feet spacing.
First, the center source:
Given that speaker/listener geometry.. the time of flight difference from ear to ear for one speaker is 226 microseconds, meaning that if you produce an impulse in both at the same time, right ear will hear the right impulse 226 microsec before the left..but, there is an inherent problem with this...the left ear will hear the left speaker 226 usec before the right ear does..and this means what?? (glad you asked)
It means that a mono signal, produced by two speakers spaced that far apart will have four distinct auditory reception events..
1... the center image produced when the impulse initially arrives at the intended ear from it's speaker..
2... the image caused by right speaker impulse reaching the left ear 226 microseconds later, telling the brain that a sound is six feet to the right, concurrently with the left speaker sound reaching the right ear, with it's image six feet to the left.
3...lastly, the secondary center image caused by the 226 microsecond delayed concurrent arrival of opposite speaker/ears...
And yet, the human brain has the ability to present a center image from these two sources...it is a learned response...and one that headphones do not require of the brain because they eliminate temporal crosstalk...the channels remain separate..but for speakers, imagery reconstruction requires the brain ignore the side images, what I call sidebands, and the secondary image of the delayed, angularly inverted one..(it must be noted that the secondary center image deconstructs opposite the primary during head rotation..but, I'll leave that confounding issue alone for now, to keep the argument simple..)
Now, how to reproduce the source off to one side???what, delay one speaker 45 microseconds? what happens to the images from case 2 and 3??? don't forget, the other ear is listening....now what....inject an inverted, delayed image to the other speaker? create a dipole, with each ear in center plane? That can be done, but the math is non trivial...and geometrically determined by your stereo...
So, what delay to use? How to mix it into both channels to accomodate the crosstalk...did anyone really believe the pan pot is correct?? Amplitude based positioning is the farthest from reality..
This effect also occurs in three d movies..there, the requirement is to prevent each eye's intended picture from reaching the other..requiring the glasses. For most, it takes a little effort and time to realign the eyes to construct the 3 d image..but it is fooling the brain into believing the image is there...it is unfortunate there is a distance sweet spot for correct imaging, and I never seem to be able to sit there..
That is what high end audio is about...fooling the brain into constructing an image where no source is located, even though there is 100% crosstalk ear to ear, albeit with temporal delays based on speaker angular placement and distance..
To add to the model, consider what happens to the visual image of 3-d when you rotate your head, causing the polarizing filters of the glasses to de-align w/r to the vertical and horizontal image content...you get cross bleedthrough, and that provides the central 3 d image, as well as the obvious independent right and left flatscreen images, confusing the picture..
As for the head vice thing....humans do not need static head position to image the virtual imaging...it accomodates it...am I gonna have to derive the equations of sensitivity to show that???? I really didn't want to insult anyone with something that is so easily discerned by inspection... (slap, slap...feel like your in a crowd there Bruce???)
The work being done by Ingenia is a start..tis unfortunate their goal will be unrealizable beyond one sweet spot in 2 d space....but they do provide an excellent picture of their technology, from which all these derivations are easily culled.
My apologies to all for keeping the discussion so simple and elemental..I'm not sure this forum is ready for the heavy stuff..
Enjoy your vacation.
Thank you very much...I did enjoy it, quite a bit...
Did a hot air balloon thing at Napa Valley...wow..2200 feet up..
Cheers, John
PS...sorry I'm so punchy...just having fun...
PPS..forgot a graph...gotta have a graph...:-)
OK..here goes..
I said...dbt's using instant switching has a problem..because of human hearing soundstage reconstruction..
But who said anything about instantaneous switching being used for such a purpose? (besides Risch, and maybe Curl) The majority of people are going to be concerned with FR changes which is done instantaneously, because science has already shown that is what we are the most sensitive to. I would agree that you use different methods appropriate to what you are seeking.
I am not speaking of fr issues..and not single channel phase change issues..and, from what I recall (perhaps too hazily), JR specifically puts dbt down because of the rapid switching...
Laterization, as far as I am concerned is a non-issue, except in controlling the large group delay variances in speaker systems during the manufacturing process so you get consistent results unit to unit so you can avoid the costly need to "match" pairs. Once connected, even if the image is slightly skewed(which could be attributed to many things), it is constant and therefore goes unnoticed
You would be wrong if that is your assertion...:-) (c'mon, put up your dukes!!)
I am not speaking of manufacturing variations of speakers..
If you sit in front of the pair, the human mind will adjust and reconstruct the audible image in time..
Lateralization is all important when sitting in the sweet spot and listening to a stereo signal..how else do you image a human voice in the center of the speakers, when there is no source at that location?? Same with off axis imaging..
I also don't see people sitting in a chair anchored to their listening room floor with a vice that holds their head perfectly still.
It is not necessary to do so to localize either a spot source or a virtual one constructed by two speakers.
Consider two bells ten feet away, five feet apart right to left..For small angular head deviations, say 20 degrees, it is not necessary to keep the head perfectly aligned to determine the relative location of each source...no head vice required...the brain takes positional deviation into account.
So, I suppose in the real workd, the only place one would be concerned with lateralization might be in headphones
NO. lateralization is what the brain does for all aural input..headphones are a special construct...no pics here today, hope the following explanation is adequate..I don't have a scanner operational at work at present..so Tony won't have a grade school level drawing to criticise...:-)
Consider two impulse sources, one center, one off axis 1 foot to the right, 10 feet away...for center source, the ears hear simultaneously, giving front image..for the side one, it pings, and each ear detects the result after a time of flight...the right ear first, 45 microseconds before the left..you can point in the general direction of the sounds based on reception timing.....
Next, let's reproduce that using stereo..two speakers, at 12 feet spacing.
First, the center source:
Given that speaker/listener geometry.. the time of flight difference from ear to ear for one speaker is 226 microseconds, meaning that if you produce an impulse in both at the same time, right ear will hear the right impulse 226 microsec before the left..but, there is an inherent problem with this...the left ear will hear the left speaker 226 usec before the right ear does..and this means what?? (glad you asked)
It means that a mono signal, produced by two speakers spaced that far apart will have four distinct auditory reception events..
1... the center image produced when the impulse initially arrives at the intended ear from it's speaker..
2... the image caused by right speaker impulse reaching the left ear 226 microseconds later, telling the brain that a sound is six feet to the right, concurrently with the left speaker sound reaching the right ear, with it's image six feet to the left.
3...lastly, the secondary center image caused by the 226 microsecond delayed concurrent arrival of opposite speaker/ears...
And yet, the human brain has the ability to present a center image from these two sources...it is a learned response...and one that headphones do not require of the brain because they eliminate temporal crosstalk...the channels remain separate..but for speakers, imagery reconstruction requires the brain ignore the side images, what I call sidebands, and the secondary image of the delayed, angularly inverted one..(it must be noted that the secondary center image deconstructs opposite the primary during head rotation..but, I'll leave that confounding issue alone for now, to keep the argument simple..)
Now, how to reproduce the source off to one side???what, delay one speaker 45 microseconds? what happens to the images from case 2 and 3??? don't forget, the other ear is listening....now what....inject an inverted, delayed image to the other speaker? create a dipole, with each ear in center plane? That can be done, but the math is non trivial...and geometrically determined by your stereo...
So, what delay to use? How to mix it into both channels to accomodate the crosstalk...did anyone really believe the pan pot is correct?? Amplitude based positioning is the farthest from reality..
This effect also occurs in three d movies..there, the requirement is to prevent each eye's intended picture from reaching the other..requiring the glasses. For most, it takes a little effort and time to realign the eyes to construct the 3 d image..but it is fooling the brain into believing the image is there...it is unfortunate there is a distance sweet spot for correct imaging, and I never seem to be able to sit there..
That is what high end audio is about...fooling the brain into constructing an image where no source is located, even though there is 100% crosstalk ear to ear, albeit with temporal delays based on speaker angular placement and distance..
To add to the model, consider what happens to the visual image of 3-d when you rotate your head, causing the polarizing filters of the glasses to de-align w/r to the vertical and horizontal image content...you get cross bleedthrough, and that provides the central 3 d image, as well as the obvious independent right and left flatscreen images, confusing the picture..
As for the head vice thing....humans do not need static head position to image the virtual imaging...it accomodates it...am I gonna have to derive the equations of sensitivity to show that???? I really didn't want to insult anyone with something that is so easily discerned by inspection... (slap, slap...feel like your in a crowd there Bruce???)
The work being done by Ingenia is a start..tis unfortunate their goal will be unrealizable beyond one sweet spot in 2 d space....but they do provide an excellent picture of their technology, from which all these derivations are easily culled.
My apologies to all for keeping the discussion so simple and elemental..I'm not sure this forum is ready for the heavy stuff..
Enjoy your vacation.
Thank you very much...I did enjoy it, quite a bit...
Did a hot air balloon thing at Napa Valley...wow..2200 feet up..
Cheers, John
PS...sorry I'm so punchy...just having fun...
PPS..forgot a graph...gotta have a graph...:-)