I wouldn't specify it as the "non" ITD range. IMO the so-called range for time delay is significantly broader than the common "ILD to ITD" transition. I think they are both broad-band effects under normal conditions (that aren't/haven't been tested).
As for a monophonic emphasis as a part of the mix above 3 kHz.. Maybe, or perhaps greater emphasis.
It's not a hard and fast transition with Head-shading effects. Only around 5kHz (on average) does it result in complete L & R separation (..and again at 10 kHz). (..some freaky stuff centers around 7-7.5 kHz that makes that severe separation of L & R less severe.)
It's one octave below this at 2.5 kHz (again, on average) where the monophonic emphasis (or "center") *starts* to "slip". And it's gradual until it reaches that next higher octave (..where L & R become increasingly dominate and center less so).
Interesting hypothesis, is that due to too little cross talk the speakers become perceived as sources at high freqs. The cure ? Create cross talk by means of very wide directivity and room reflections
It works in practise, too (whatever the actual reasoning lol )- Elias
Monophonic at 4.7-5 kHz up, likely yes.
Something "less than" monophonic from 2.5 - "5" kHz.
Standard panning 2.5 kHz and below.
Interesting hypothesis, is that due to too little cross talk the speakers become perceived as sources at high freqs. The cure ? Create cross talk by means of very wide directivity and room reflections
- Elias
Yup.
Actually most "cures" I've seen suggested (in papers) are progressively narrower spread for a multi-way driver.
i.e. (stereo loudspeaker spread): tweeters 4 kHz about a foot apart. Upper mid.s about 4 feet apart. Lower mid.s about 8 feet apart. Bass about 12+ feet apart. (..or something like that.)
Most Loudspeaker designs are considered "wide dispersion". I think even Earl's products would fall under that category under this sort of research (..particularly given the fact that most use a stereo triangle with toe-in to on-center.)
Don't forget the effect of the pressure gradient however.. In the stereolith style the spread is both nominal at higher freq.s between speakers AND the gradient is "broken". The adaptation to improve freq. response at higher freq.s via the mono tweeter, does not have it's gradient "broken", but is of course both mono and centered.
I did this for a level panned tom:
https://dl.dropbox.com/u/21936387/audio/Tom_monophonic_highs.zip
I get better localization with the standard panning. Head shadowing (i.e. higher "natural" crosstalk cancellation) seems to help with localization, not the other way around.
No, horns are never "wide" in dispersion. Onmi is "wide". To set the references
At least it is possible to preserve:
"The phase coherence of harmonics in the vocal formant range, ~630Hz to 4000Hz"
with the stereolith style approach with mono signal
- Elias
At least it is possible to preserve:
"The phase coherence of harmonics in the vocal formant range, ~630Hz to 4000Hz"
with the stereolith style approach with mono signal
- Elias
Unfortunately the image projected by the Stereolith isn't stereo at all. Are we now back to "mono is better" because it has less problems with localization of the speaker according to Griesinger's hypothesis?
What does each represent?
Tom, Tom 1 kHz, Tom 2.5 kHz?
I'm assuming just "Tom" is the standard panning. I've no idea what the 1 kHz is - we never discussed 1 kHz. Nor do I really no what 2.5 kHz represents.
When you say you are getting better localization, what do you mean by that?
Head shadowing does indeed help with localizing near the source - so it should present a greater L and R. But that's the problem: to much L & R, so much so that its much more identifiable than center. Center on just "Tom" was more difficult to identify than on either the 1 or 2.5 kHz versions, while "Tom" was easier to identify as you got closer to L & R.
The interesting thing I got out of it was that it's not substantially different for an odd reason - onset and decay seem clipped on the 1 and 2.5 kHz versions, and that seemed to diminish overall accuracy no matter the panning/level difference.
Scott,
I've applied a 24dB high pass at 2.5kHz to a mono mixdown of the "Tom.wav" stereo track, then applied a 24dB low pass at 2.5kHz to each stereo channel and added the mono mixdown to each stereo channel.
I might have forgotten to reduce levels before summing. Will post new versions tomorrow.
The low/high pass for the other version is 1kHz.
"Tom.wav" is the original.
I've applied a 24dB high pass at 2.5kHz to a mono mixdown of the "Tom.wav" stereo track, then applied a 24dB low pass at 2.5kHz to each stereo channel and added the mono mixdown to each stereo channel.
I might have forgotten to reduce levels before summing. Will post new versions tomorrow.
The low/high pass for the other version is 1kHz.
"Tom.wav" is the original.
Just checked the file, levels are fine, so no new version needed.
With the 1kHz version the drum is panned from left to right with the "tick" coming from the center.
In the 2.5kHz version the "tick" isn't that obviously coming from the center but overall localization is a bit blurry.
The unaltered sound is very focused and it's direction is unambiguous.
That's not what I get. (moved speakers out into room, stereo triangle, me closer to speakers than speakers to any wall.)
I get a better sense of localization for the 2.5 kHz track at and near center, though only slightly so because the onset "tick" sounds "clipped" ("thrum" sounds clipped as well, but it doesn't seem to alter localization).
"Depth" is slightly better at all positions, though particularly at L for the 2.5 kHz track. (whereas the unaltered version seems stuck at the drivers.)
The upper (freq.) end of the "thrum" seems to move with the "tick" better on the 2.5 kHz track. On the unaltered track that upper freq. portion of the "thrum" seems to move *before* the "tick" when moving from Left to Center. Near center the upper freq. portion of the "thrum" doesn't seem to move at all with the unaltered track, but it does with the 2.5 kHz track.
L seems to be "stuck" to the loudspeaker with the unaltered track, but curiously R isn't. In either R or L the localization is slightly better with the unaltered version, but in a negative fashion (..to close to loudspeaker, to "up-front").
For both tracks ONLY L & R seemed to have the lower freq. portion of the "thrum" sound localized to that area. At all other positions it sounded "amorphous" with a "sort of center" sound.
Still, a *very* similar sound overall.
It sounds like you guys have effectively isolated the lower midrange imaging timing cues, by turning the upper midrange to mono, and been able to hear and feel what most systems are missing, due to inter-aural crosstalk at the listening end. I think Griesinger argued in favor of reduced spread above about 1-2kHZ, but not full mono.
It was actually that set of arguments that convinced me to go back to doing bass (below 100HZ) in stereo after decades of doing it in mono, in my speaker systems. Sure, most recordings do the bass in mono anyway, but not the really good recordings.
My understanding about lower mid timing cues is that due to the distance between the eardrums (about 6 inches?), the ear-brain mechanism can't be sure what period it's comparing, once the half wavelength is shorter than this distance between the ears (could be full wavelength - not sure). So as a survival instinct, the ear-brain mechanism switches over to comparing only amplitudes between about 1.5kHZ and about 6kHZ, to determine image location. Above about 6kHZ things change again. Above 6kHZ, variations in the pinea come into play, due to the size of the wavelengths, and tell us up/down info based on individual experience, apparently especially in the 8kHZ region.
It was actually that set of arguments that convinced me to go back to doing bass (below 100HZ) in stereo after decades of doing it in mono, in my speaker systems. Sure, most recordings do the bass in mono anyway, but not the really good recordings.
My understanding about lower mid timing cues is that due to the distance between the eardrums (about 6 inches?), the ear-brain mechanism can't be sure what period it's comparing, once the half wavelength is shorter than this distance between the ears (could be full wavelength - not sure). So as a survival instinct, the ear-brain mechanism switches over to comparing only amplitudes between about 1.5kHZ and about 6kHZ, to determine image location. Above about 6kHZ things change again. Above 6kHZ, variations in the pinea come into play, due to the size of the wavelengths, and tell us up/down info based on individual experience, apparently especially in the 8kHZ region.
Last edited:
Yes, not full mono. Perhaps only full mono around 5 kHz and up.
But Markus was kind enough to do what he has done, so why not take a listen!
(..I spent more time on it than I should have, but that was mostly down to cross-checking again between the two tracks and listing comments for inclusion on the follow-up post, other than that it really doesn't take much time.)What kind of speakers? Size of stereo triangle? Asymmetric setup? Is your room treated?
2-way, 9" and cone tweeter (1st order), bass reflex (vent tuning around 35 Hz) no baffle step (though the woofer has about 3-4db naturally to compensate). Reasonably linear on and (marginally) off-axis - listened to on axis. 7 feet for the triangle (7-7-7). Room 18 by 22. Pretty much centered in room, with speakers spanning the long wall (..with me closer to one of the 18 foot walls).
(..I pulled the speakers out for this, often I'll just listen to some cheap headphones..
)Scott,
I can only speculate what's causing the difference in perception. My front wall is completely covered with absorption and speakers have high directivity (though there's no absorption at side walls and speakers are not toed-in). So it's probably the difference in the indirect sound field of our rooms.
Would love to hear other's results. Here's the link again:
https://dl.dropbox.com/u/21936387/audio/Tom_monophonic_highs.zip
I can only speculate what's causing the difference in perception. My front wall is completely covered with absorption and speakers have high directivity (though there's no absorption at side walls and speakers are not toed-in). So it's probably the difference in the indirect sound field of our rooms.
Would love to hear other's results. Here's the link again:
https://dl.dropbox.com/u/21936387/audio/Tom_monophonic_highs.zip
Hi Markus
I have listened to your recording a couple times, cool.
I was standing in front of the listening couch between two SH-50’s about 10 feet away, CD direct to amps etc.
Here is what it sounded like to me.
Track 1 a stepped pan from one side to the other.
That one was very clear like a dry recording and moved in very clear pretty even steps, an excellent stereo image, mono phantom perfect no audibility of separate left and right speakers.
I thought maybe I could hear the image waver slightly, like maybe it was an analogue pan pot moved in steps instead of generated electronically at level X.
Track 2 Same stepped locations, very similar as before but with a slight awareness of sound from the speaker opposite the location of the stereo image or some kind of upper content or “something”.
Track 3 Same stepped locations, very similar like before but much more noticeable hf something, now maybe sort of a fuzziness’ in it’s location in the stereo image.
That one was the least distinct stereo image of the three but still moved one side to the other clearly, while there was somewhat more of an awareness of the speaker opposite the side where the image was, it was not enough to stand out as a right and left source.
Per the title of the thread, here is a measurement of the same speakers or one anyway, this is at one meter in a livingroom with a bare hardwood floor and no room treatments. The electrical crossovers are around 300 and 1200 Hz. As one can see, there is no crossover phase shift, it looks like one very broad band driver in a large CD horn.
Best,
Tom
I have listened to your recording a couple times, cool.
I was standing in front of the listening couch between two SH-50’s about 10 feet away, CD direct to amps etc.
Here is what it sounded like to me.
Track 1 a stepped pan from one side to the other.
That one was very clear like a dry recording and moved in very clear pretty even steps, an excellent stereo image, mono phantom perfect no audibility of separate left and right speakers.
I thought maybe I could hear the image waver slightly, like maybe it was an analogue pan pot moved in steps instead of generated electronically at level X.
Track 2 Same stepped locations, very similar as before but with a slight awareness of sound from the speaker opposite the location of the stereo image or some kind of upper content or “something”.
Track 3 Same stepped locations, very similar like before but much more noticeable hf something, now maybe sort of a fuzziness’ in it’s location in the stereo image.
That one was the least distinct stereo image of the three but still moved one side to the other clearly, while there was somewhat more of an awareness of the speaker opposite the side where the image was, it was not enough to stand out as a right and left source.
Per the title of the thread, here is a measurement of the same speakers or one anyway, this is at one meter in a livingroom with a bare hardwood floor and no room treatments. The electrical crossovers are around 300 and 1200 Hz. As one can see, there is no crossover phase shift, it looks like one very broad band driver in a large CD horn.
Best,
Tom
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.