I only know for sure with the counting voice, because I did that pan. It is amplitude only. The Firesign Theatre clip is most likely amplitude only, as that was very common at the time. As for the Time Machine clip, who knows? Common practice these days would be to give each voice a mic and a track, then place then place them wherever needed in DAW software. Hard to say if delays were used, tho those plugins do exist now and that's the way that I would do it.
How much would it matter - do you think - between amplitude only and delay based panning?
Yes, simple.
Two identical* signals arriving from a left and a right speaker, will cause cancellation and comb filtering at the human head. This results in notches in the frequency response at each ear. See the attached chart.
Those notches fall at approximately 2K and 6K, resulting in a sound that is duller than one arriving from a single source. The phantom center is formed by two identical signals arriving from left and right. Thus the phantom center sounds duller than the sides alone.
*identical is important
Attachments
Not really sure how much it would matter yet. But a think it really takes some experiment to find out. The important thing is what changes can be heard with the different mixes under different conditions.
I was listening to a Reference Recording test CD track where Keith Johnson moves to the sides of a cross cardioid mic set, the sound turned hollow midway of the center and extreme sides, whereas center position was not so well focused. I think this is also an interesting track to use for testing.
Sent from my iPhone using Tapatalk
Please do test it. I'll have to leave it up to you and a few other people here, as at the moment I do not have a system or room clean enough to reveal this effect.
I would suggest to the test files I posted early on, both straight and convolved, to find if you can hear the difference. If you can't, then your system/room is not subject to it.
I would suggest to the test files I posted early on, both straight and convolved, to find if you can hear the difference. If you can't, then your system/room is not subject to it.
To model 2 ears (HRTF) verse mono micophone into REW take say your left speaker measurement and open it twice, then for one of them delay IR -0,27mS and on "All SPL" tab pick them as (A+B)/2 and hit "Generate". You welcome share curves here
below is a result from perfect IR-wav file created in Rephase.Attachments
Last edited:
Thanks for summarizing. Appreciated.
Is it the El Greco fallacy once again? It often shows up here where knowledge of perception is scanty. To illustrate, the same argument could be made for pinna shadowing or every other sensory non-linearity that typically has no consequence for perception.
And as someone who has been following Daniel Kahneman's research*, this is a very well-fertilized opportunity for erroneous human judgment.
Ben
*Nobel prize
Last edited:
Thank you for explaining the physical theory... as it would apply to a bowling ball with microphones where humans have two ears. But see my previous post about El Greco fallacy.
Are we talking about the Benjamin Bauer analysis of 1955(?)? Important to understand how (when listening to conventionally recorded music) the Bauer circuit does apply to headphones but not to speakers, as in this thread.
Ben
Last edited:
I don't think so, no.
- It's measurable
- It's audible
- The cause of the effect was found after the effect was noticed, not vice versa
That said, the effect is not audible on all systems. As has been noted by Wesayso and others, you need a clean, phase coherent system and low reflection amplitude to hear it. I don't know, but suspect, that most system/room combos will not reveal the notches. Certainly my 1966 Clairtone console in our untreated living room does not.
I have posted a number of test files that are both clean and treated. You may wish to give those a listen, to hear for yourself. Some people hear it, others do not.
Nobody is questioning the physics of bowling balls and mics.
But is it always hearable* by careful listeners on good quality systems?
Ben
*in a low-reverberant room
It has been heard by some people here, yes. Others hear no difference at all. Also it would seem by others not here, as per the paper on the subject. Their concern is mostly dialog in the phantom center, not even tonal balance.
Test it yourself. You will hear it or you won't, it's simple. Start with the counting voice, that's an easy one.
Test it yourself. You will hear it or you won't, it's simple. Start with the counting voice, that's an easy one.
Measure and show the first 20 ms of your IR at the listening spot, either from a single speaker or both, to show your level of early reflections. That will tell us a lot if you can or may hear it.
I'll show you mine:
Impulse and STEP response as measured at the listening spot from a stereo pair.
(this one is a combined "A + B" from a left + right measurement in REW, but you can find many more in my thread actually measuring the summed result:
http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array.html)
Notice the absence of early reflections...
To show the frequency response and phase belonging to that IR:
Also as measured (same as above left + right combined in REW) at the listening spot, shown with a frequency dependent window of 6 cycles.
A different view on this data by APL_TDA:
This one is measured at the listening position, a stereo pair,
recorded "live" at the listening position, contrary to both examples above where the left and right channels were summed in REW.
Now I don't know if this qualifies as a good quality system in your book, but on this system I've heard it.
I also picked up on the difference the phase shuffler made, and noticed another tonal balance problem it caused in my system.
After checking what's going on, based on what I heard during the listening tests, I could find that new problem in measurements.
The explanation is in this thread.
Right now I don't use phase shuffling. I do still deal with this problem by using other means. (mild mid/side EQ and ambient "Haas Kicker" speakers)
I'll show you mine:
Impulse and STEP response as measured at the listening spot from a stereo pair.
(this one is a combined "A + B" from a left + right measurement in REW, but you can find many more in my thread actually measuring the summed result:
http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array.html)
Notice the absence of early reflections...
To show the frequency response and phase belonging to that IR:
Also as measured (same as above left + right combined in REW) at the listening spot, shown with a frequency dependent window of 6 cycles.
A different view on this data by APL_TDA:
This one is measured at the listening position, a stereo pair,
recorded "live" at the listening position, contrary to both examples above where the left and right channels were summed in REW.
Now I don't know if this qualifies as a good quality system in your book, but on this system I've heard it.
I also picked up on the difference the phase shuffler made, and noticed another tonal balance problem it caused in my system.
After checking what's going on, based on what I heard during the listening tests, I could find that new problem in measurements.
The explanation is in this thread.
Right now I don't use phase shuffling. I do still deal with this problem by using other means. (mild mid/side EQ and ambient "Haas Kicker" speakers)
Last edited:
Ben, human beings' brains can partially equalize for the phantom center. Play mono pink noise, sit down in center position and hear, how the dips and peaks sweep thru the frequency band, while you move your head. But it is only a partial fix and ties up mental resources. The error is not natural but only 70 years old, so we have not much evolutionary gear for it, and I do not want to breed that gear any further.
I've had visitors at various points in my experiments, some of them were kind enough to supply feedback:
Review by Jan Fekkes, without ambient Haas Kicker but with mild mid/side EQ
Review by DIYaudio member "Boden' with virtual Haas kicker, mild mid/side EQ and cross talk cancelation in place
Review by DIYaudio member 'xrk971' with altered Haas Kicker and mild mid/side EQ
Review by DIYaudio member 'BYRTT' with altered Haas Kicker and mild mid/side EQ
To judge some of this stuff you need time, a lot of listening time. For instance, the experimental state as reviewed by Eelco (member name: 'Boden') was very convincing, but more listening fatigue/strain in the long run as experienced by myself. Basically it gave me a headache (neck and shoulder lock up) after longer sessions. In other words, it wasn't a good permanent solution. This was a negative IR cross talk cancelling band passed signal added to the front left and right speakers, 0.27 ms after the main impulse. Loosely based on the work of Dr. Edgar Choueiri: https://www.princeton.edu/3D3A/
Obviously my implementation wasn't up to par.
The last incarnation (altered Haas Kicker) has been in place for a long time and no listening fatigue has been noticed in any of these sessions.
It features natural sound and uses the Haas Effect to fill in some of the dips created by the cross talk. This idea was based on the measurements of member jim1961 as shown on this thread. For me this works.
I still need to work on the EQ for multichannel sources, I do notice a tonal difference in the centre channel of 5.1 recordings as I only use a phantom centre. I'm convinced I can solve that because prior versions of my experiments worked very well for movies/5.1 recordings. So I need to alter/copy an older mid/side EQ scheme for movies and use that. You'd be surprised what we can hear, I know I am.
Review by Jan Fekkes, without ambient Haas Kicker but with mild mid/side EQ
Review by DIYaudio member "Boden' with virtual Haas kicker, mild mid/side EQ and cross talk cancelation in place
Review by DIYaudio member 'xrk971' with altered Haas Kicker and mild mid/side EQ
Review by DIYaudio member 'BYRTT' with altered Haas Kicker and mild mid/side EQ
To judge some of this stuff you need time, a lot of listening time. For instance, the experimental state as reviewed by Eelco (member name: 'Boden') was very convincing, but more listening fatigue/strain in the long run as experienced by myself. Basically it gave me a headache (neck and shoulder lock up) after longer sessions. In other words, it wasn't a good permanent solution. This was a negative IR cross talk cancelling band passed signal added to the front left and right speakers, 0.27 ms after the main impulse. Loosely based on the work of Dr. Edgar Choueiri: https://www.princeton.edu/3D3A/
Obviously my implementation wasn't up to par.
The last incarnation (altered Haas Kicker) has been in place for a long time and no listening fatigue has been noticed in any of these sessions.
It features natural sound and uses the Haas Effect to fill in some of the dips created by the cross talk. This idea was based on the measurements of member jim1961 as shown on this thread. For me this works.
I still need to work on the EQ for multichannel sources, I do notice a tonal difference in the centre channel of 5.1 recordings as I only use a phantom centre. I'm convinced I can solve that because prior versions of my experiments worked very well for movies/5.1 recordings. So I need to alter/copy an older mid/side EQ scheme for movies and use that. You'd be surprised what we can hear, I know I am.
Last edited:
This is a good test. I did it before. The sweep should be audible with a good hi-fi stereo in a good room. To compare, use a barrier like post #8, which instantly removes the comb filter effect very well. A chair cushion is OK.
Or just use mono recorded material. Solo violin is good.
I've had a visitor on my thread (osterhammer-mar) who used a "Mohawk like" anti cross talk device with his concave arrays :
Link: http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array-255.html#post4845579
:Link: http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array-255.html#post4845579
Attachments
To quote from a part of his reply:
So I'm guessing: yes!
You are right, my setup is not usable for more than one listener. The sound becomes dull if you move up or down more than 15 cm. tolerance in lateral direction is +-10 cm, and axial +- 40 cm.
But I expected this effect and do not have any problem with it.
Inside this sweet spot of course the sound is better than everything I ve heard.
Exactly like a one point source, but with the advantage of a big driver-surface-You know what I mean.
Btw:Crossover cancellation was a theme for me ,too.
I made som experiments with recussive HRTF base filters, but the most convincing way to improve the sound was quite simple:
Thank you once more for your inspiration you gave me!
Best wishes
Martin
So I'm guessing: yes!
You would get exactly the same astonishing demonstration without even moving your head with just one ear. Now isn't that amazing proof of something or other?
Ben
Last edited: