Soongsc, ITD is just the difference in path length from both ears to the sound source, divided by the speed of sound. Wave length, shape or spectral content do not enter the equation.
The matter that ITD and interaural loudness differences often do not match up for the same recorded instrument is a serious one imo, you hit the nail on the head. Hence my question to Pano if it it is possible to pan in delay as well as loudness.
Thanks guys for taking the time to listen to those tracks. I found it a very interesting test, and now understand better the dominance (or lack of) spacial clues >700Hz. There is certainly more to it than just that test.
My point being; the spectrum <700Hz in not unimportant. I find it plays an important roles in imaging and location. Yes, the spectrum above that is somewhat more prominent, but not all important. At least for stereo systems. It may be quite different in a concert hall.
It is conventional audio wisdom to say that the lower frequencies don't matter for location, thus we can use satellite + sub systems without fear of compromise. I agree that these systems do work surprisingly well and can fool the ear. But they don't work nearly as well as full bandwidth stereo. The images created by full bandwidth stereo give me a much more realistic and satisfying image and illusion of reality. Full bandwidth stereo is often not practical or affordable, however. So we compromise where we can.
Again, thanks for listening. FYI
Yes, I hear that too. I hear it when > 700Hz moves to the side, but hear it and feel it even more when the < 700Hz moves to one side.
It is a strange effect and strange sensation. Something sounds wrong, even if you can't put your finger on it.
Not all that low. These are 256Kbs MP3, which isn't going to hurt spoken voice all that much. Most people have a hard time identifying a music MP3 with this high a bit rate. I did find a slight reduction in stereo effect going to MP3, but it was really tiny. I think what you are hearing is not an MP3 effect. The file was dual mono before the panning effects. There is only amplitude difference between tracks.
Yes, again. Even if you can't exactly identify what's going on, you can hear that something is happening. For me, I can usually hear the split.
My point being; the spectrum <700Hz in not unimportant. I find it plays an important roles in imaging and location. Yes, the spectrum above that is somewhat more prominent, but not all important. At least for stereo systems. It may be quite different in a concert hall.
It is conventional audio wisdom to say that the lower frequencies don't matter for location, thus we can use satellite + sub systems without fear of compromise. I agree that these systems do work surprisingly well and can fool the ear. But they don't work nearly as well as full bandwidth stereo. The images created by full bandwidth stereo give me a much more realistic and satisfying image and illusion of reality. Full bandwidth stereo is often not practical or affordable, however. So we compromise where we can.
No, but I can probably figure it out. Will have to see how well I can implement it in software. I'd really like to try it.
I fixed it.
Again, thanks for listening. FYI
- A is mono
- B is >700Hz panning
- C is fullrange panning
- D is <700Hz panning
I actually looked at the response after listening a few times. One of the reasons is to verify the direction of the pan. Seems like all always shifted in the direction of the increased volume. But also the sound quality seemed pretty poor with lack of detail normally associated with certain throat, tongue and teeth. So I checked the spectrum of the recording the bit rate.
I can't claim it's the world's best recoding, just a rather dry recording I had easily at hand. 
No one recognized the voice?
Strangely, I have another edition of this same recording where they packed a ton of reverb onto the voice. Sounds awful, like he's in a sewer pipe.
The spectrum you are seeing in the pans I can't explain, but the file was definitely split at 700Hz before manipulation, then recombined.
I did some experiments in which reverb was added to the voice full bandwidth, >700 only and <700 only. All three sound quite different.
No one recognized the voice?Strangely, I have another edition of this same recording where they packed a ton of reverb onto the voice. Sounds awful, like he's in a sewer pipe.
The spectrum you are seeing in the pans I can't explain, but the file was definitely split at 700Hz before manipulation, then recombined.
I did some experiments in which reverb was added to the voice full bandwidth, >700 only and <700 only. All three sound quite different.
I can't claim it's the world's best recoding, just a rather dry recording I had easily at hand.
Sounds like they had to throw the pop filter into the dryer after recording
Nobody is saying that there isn't "more to it than that". But the fundamental and principle effect is > 700 Hz - lets not extrapolate the statements beyond what they say.
I'm pretty sure they were in done in a large auditorium, well away from any boundaries/surfaces but the floor.
He does however have gating employed. 50ms for the Revels and 25 ms for the CBT's.. while not "equal", I don't think a 50ms time would have been much different for the CBT's.
Yes, below 1 kHz they do look great *accepting the limited low freq. extension* - particularly how it deals with the floor in the mid-band. (..and the Revels are actually designed to deal with this better than most designs on the market.)
Above 1 kHz, not great, though still fairly remarkable given the design constraints. Note that a "perfectly reflective" floor surface is not normal in most domestic conditions - here the Revel's results are "skewed" in favor of the CBT. (..that's even somewhat true for freq.s below 1 kHz - the "ripple" is exaggerated for a common 1/12th octave smoothing.)
http://www.linkwitzlab.com/Keele - Introduction to CBT Loudspeaker Arrays.pdf
Many people claim to be able to hear "woofer bloom" (the frequency dependent transition from directional to omni-directional exhibited by some otherwise "controlled directivity" loudspeakers). And it is common that such claim is not "listening position specific", but that the perception (often described as "sounds like a loudspeaker") persists well outside the "stereo listening area", and for mono signals as well.
So it is a reasonable speculation, and possibly at least part of what Pano is demonstrating, that when the sound(s) from what we expect to be a single source come from different, frequency dependent, directions it sounds . . . wrong . . . and that changes in speaker directivity, when combined with reflections in a small room, would do that.
I would further offer the speculation that this explains a sometimes perceived flaw in the ORION midrange (which proved ultimately uncorrectable with that loudspeaker without external "room correction") . . . it's not that the radiated sound changed perceptibly at the MT crossover but rather that the sense of surrounding "space" and source location was changed by the changed reflection pattern (described as "tweeter bloom" in the ORION case).
It does not surprise me at all that this effect would persist an octave (or two, or three) lower in frequency . . .
Indeed. There is not much argument that the perception of the direction of a sound source is more acute for, and to some extent even dominated by, short transients above 700 Hz. Let's not "extrapolate" that into a justification for woofer bloom in loudspeakers . . .
Earl,
I don't think anyone is arguing with your that above 700hz that is the major contributor to localization. What I think we are trying to point to is that it is the composite waveforms that are working together that we hear. Pano's experiment, while not done under strict scientific conditions does point to the need to have a coherent waveform across the frequency spectrum. If the lower frequencies are not following a similar path of axis then even if it is not a localization issue we can hear something is not correct, something has skewed the image.
This would seem to be an argument against random placement of subwoofers or even bass sections in relationship to the main sound system. Pano's observation that a full range system is preferred to a distributed system points back to the single source concept involving as much of the full bandwidth as possible.
I don't think anyone is arguing with your that above 700hz that is the major contributor to localization. What I think we are trying to point to is that it is the composite waveforms that are working together that we hear. Pano's experiment, while not done under strict scientific conditions does point to the need to have a coherent waveform across the frequency spectrum. If the lower frequencies are not following a similar path of axis then even if it is not a localization issue we can hear something is not correct, something has skewed the image.
This would seem to be an argument against random placement of subwoofers or even bass sections in relationship to the main sound system. Pano's observation that a full range system is preferred to a distributed system points back to the single source concept involving as much of the full bandwidth as possible.
The measurements show a prototype. One slide says 100ms window.
Last edited:
That you feel yourself to have more expertise that Griesinger in this area is incredibly arrogant. That you did not continue reading to where he explains and answers his questions is even worse.
Well Griesinger would argue against this position stating that the fundamental is far less important than > 700 Hz and doesn't really have to "work together" with the upper harmonics.
And yes there have been statements here that claimed the < 700 Hz was more important than above, not everyone, but some.
Pano's conclusions are different than Griesingers, fine, I find Griesengers more well thought out - he explains his results - and they have been done more thoroughly. I am simply not going to write off an expert like Griesinger because it doesn't fit a few simple tests posted on the internet. If you want to that's fine, there is nothing that I can do about that, but I have not seen anything here that makes me question what I believe. And I should point out that reading Griesinger did make me question what I believed, and I changed my beliefs as a result. That's because my beliefs led to some details that did not fit the bigger picture, Griesingers did, so I changed - new data, new beliefs.
Earl,
It just seems that you have taken a position that only a narrow band of frequencies from above 700hz up to about 5Khz are all that are important in music reproduction. I don't feel that is necessarily true and I think that is where we differ in our thinking.
You never answered my question about the 500hz dip in your polar plot you showed earlier except to say that is was a general problem with all 15" speakers and blamed that on the spider, though I suspect you meant to say the surround.
It just seems that you have taken a position that only a narrow band of frequencies from above 700hz up to about 5Khz are all that are important in music reproduction. I don't feel that is necessarily true and I think that is where we differ in our thinking.
You never answered my question about the 500hz dip in your polar plot you showed earlier except to say that is was a general problem with all 15" speakers and blamed that on the spider, though I suspect you meant to say the surround.
There is always a trade off with these things. My understanding of use of multiple subwoofers is to overcome the sparsity of LF modes in a small room. It seems to me that if the mains roll off low enough and if the subs are not pushing out HF components (many do, especially BRs), and if the subs aren't playing too loud or mains too quietly, then there needn't be ambiguity or shiftiness in image location.
When year after year the largest loudspeaker company in the US sends you a check every quarter for your IP they use, you run the danger of becoming arrogant, so my profound excuse if I came across as such.
But seriously, if you know the literature, what to do with just another random quote from Giesingers site: "
The research presented in the preprint above shows that the ability to separate simultaneous sound sources into separate neural streams is vitally dependent on the pitch of harmonically complex tones. The ear/brain system can separate complex tones one from another because the harmonics which make up these tones interfere with each other on the basilar membrane in such a way that the membrane motion is amplitude modulated at the frequency of the fundamental of the tone (and several of its low harmonics). When there are multiple sources each producing harmonics of different fundamentals, the amplitude modulations combine linearly, and can be separately detected. Reflections and reverberation randomize the phases of the upper harmonics that the ear/brain depends upon to achieve stream separation, and the ampltude modulations become noise. When reflections are too strong and come to early separation - and the ability to detect the direction, distance and timbre of individual sources - becomes impossible. But if there is sufficient time in the brief interval before reflections and reverberation overwhelm the onset of sounds the brain can separate one souce from another, and detect direction, distance, and meaning. "
This is plain hogwash with little or no foundation in science. Some parts of it are plain wrong (the bit about the function of the basilar membrane), the idea that there should be an interval between direct and indirect sound for the brain to filter out the echo's (Haas research points out that the shorter the time between direct sound and reflection, the more the reflection gets inhibited), or conjecture without foundation (the whole bit about harmonics).
In short, if you know something about the science in this field, it is difficult to become a fan of Griesinger.
The opposite apparently is true as well.