ITD is about ~22cm, ~660us or ~1500Hz wavelength by Interaural time difference - Wikipedia
With 3m listening distance that would be basically the distance on the baffle. Then distance from driver center to the baffle edge 22cm would be the most disturbing sound? We gotta remember most baffles are not round but rectangle so the diffraction pattern and time delay is "smeared".
With 3m listening distance that would be basically the distance on the baffle. Then distance from driver center to the baffle edge 22cm would be the most disturbing sound? We gotta remember most baffles are not round but rectangle so the diffraction pattern and time delay is "smeared".
If anyone is interested in diffraction - which is also very important when it comes to a good midrange sound:Hi TNT,
yeah there are many ways to think the up and down sides. Let me write more about the line of thought I hada diffraction "sound source" at the baffle/speaker edge won't sound the same as the driver like you say, but if the edge is as close as possible (the driver size) the diffracted sound interferes as little as possible with the direct sound causing off axis anomalies since the diffraction "source" is as close as the driver edges itself. It is a simplified thought experiment that considers a mid driver on a multiway speaker.
Chat with Dr. Earl Geddes of GedLee Audio - YouTube
I can’t quite follow this. ITD can be seen as the interval that determines above which frequency phase differences cannot be distinguished anymore. Above the frequency level differences due to shading dominate our sense of direction of the source.
What I wanted to say is that there might be a time-range of delayed radiation that disturbs imaging. But my assumption that delays around the average listener's ITD might be the worst could very likely be wrong when I give it a second thought. It is most probably the range between some two digit microseconds and about 660 microseconds (ITD).
Regards
Charles
Regards
Charles
digitalthor, Geddes has a lot to say Masking and temporal masking particularly was very interesting phenomenon I've never thought about. Seems very relevant for loudspeaker design, gotta study some. Quick googling gives hint that microsecond scales are used in the studies listening tests regarding temporal masking but the temporal masking is often visualized on a graph that has millisecond scale. Another thing caught my eye was that temporal masking is "more effective" when both signals have similar spectrum.
There seems to be pre- and post-masking period which maybe relates to the audibility of diffraction. Post-masking period changes with frequency and duration of the masking signal suggesting that transients have way shorter post-masking period? Temporal masking is a nonlinear effect.
Some stuff
Forward masking as a function of frequency, masker level, and signal delay - PubMed
http://eemedia.ee.unsw.edu.au/contents/elec9344/LectureNotes/Chapter 13.pdf
Looks like some speaker design tips could be drawn from audio codec papers? Audio codecs use psychoacoustics to lose large part of audio data because brain doesn't hear it, there must be lots of studies on this.
Masking and temporal masking particularly was very interesting phenomenon I've never thought about. Seems very relevant for loudspeaker design, gotta study some. Quick googling gives hint that microsecond scales are used in the studies listening tests regarding temporal masking but the temporal masking is often visualized on a graph that has millisecond scale. Another thing caught my eye was that temporal masking is "more effective" when both signals have similar spectrum. There seems to be pre- and post-masking period which maybe relates to the audibility of diffraction. Post-masking period changes with frequency and duration of the masking signal suggesting that transients have way shorter post-masking period? Temporal masking is a nonlinear effect.
Some stuff
Forward masking as a function of frequency, masker level, and signal delay - PubMed
http://eemedia.ee.unsw.edu.au/contents/elec9344/LectureNotes/Chapter 13.pdf
Looks like some speaker design tips could be drawn from audio codec papers?
Audio codecs use psychoacoustics to lose large part of audio data because brain doesn't hear it, there must be lots of studies on this.
Last edited:
YES!!! And every baffle designer must know this !!!
Every double impulse appearing between 0ms and approx. 660 microseconds (== ITD) potentially has a specific psychoacoustic information about localisation: It is the time laps for a single impulse heard from frontally (=0ms) to exact laterally at 90° (0.666ms or even a little bit more). So a double impulse not only affects the frequency response, but also misleads localizations to a lateralization information within this 0us ... 666us range. And therefore, a misshaped baffle will have the potential to stongly blur location discrimination within e.g. a stereo setup.
The worst baffle shape and size would indeed be be a circular baffle with a fixed radius between 0cm and r=c*ITD. So you may deliberatly shape your baffle to have the best evenly energy distribution between the values corresponding to r_driver and r=c*ITD. From my experience with dipole baffles, this is very effective and seems the very first priority in baffle design to me. No matter what the frequency response will look like (it will look very smooth anyway ...).
So do not shape you baffle for the best possible frequency range behavior. Shape it for best /most even possible distribution of it's diffraction sound energy, especially in the range up to approx. 0.7us. You may correct the frequency response to your needs by apply filtering later on.
Last edited:
I don't and I can't find any information about it. Please post a more precise reference.
Yes. I do. And I would try to correlate psychoacoustic ITD phenomens along with physical lagged signal duplication from the baffle diffraction.
wiki said:
We really don't know much yet. And I am talking of (I)TD only here, because this is what you can model when performing some trigonometry between the ears and along with speculations about the baffle diffraction effects on two-dimensional hearing. I am strictly modest and reducing complexity, only trying to understand one single effect (time lapse between arrival of the original signal and the of arrival of the baffle diffraction's signal), and a simplified model or a two-dimensional hearing along the axis between the two speakers locations. And the speculation would be, that even simultaneously feeding the two receiving systems at both side of the head psychoacoustically evokes a (or two antisymmetric) location interpretations. Which would have the potential to blur an intended stereo illusion by the program material itself.
If it bothers that neither ILD, nor IPD is included in this simple trogonometric reasoning of delays and distances, then just let's talk for the beginning only about the frequency range < 1kHz where ITD is predominant for localisation and ILD, and/or IPD are secondary phenomens.
So, if anyone has some hard data about baffle diffaction's psychoacoustic effects on localisation perception, then I would be very eager to read about it.
Attachments
The time delay caused by diffraction seems constant to me. Be it not constant over the frequency band due to the diffraction itself and the intensity of it, determined by the directivity of the driver on the baffle. But since we listen in stereo on two identical loudspeakers, the effect would occur in both channels and the phantom image would stay in the same perceived position.
There could be a slight shift if the spectral pattern would be different in both channels. Something that is not very likely with coincident or semi-coincident miking techniques, which are the only ones that are able to record a realistic stereo image. So that leaves me with the question, what could be the link between ITD and baffle dimensions?
There could be a slight shift if the spectral pattern would be different in both channels. Something that is not very likely with coincident or semi-coincident miking techniques, which are the only ones that are able to record a realistic stereo image. So that leaves me with the question, what could be the link between ITD and baffle dimensions?
This was interesting https://physicstoday.scitation.org/doi/10.1063/1.3366228
Says that hearing system detects direction from sound delay (diffraction around head) between about 500-4000Hz above which diffraction doesn't occur anymore but then head has switched to be a barrier and attenuates sound and hearing system can use that o detect location (ILD). Transition from ITD to ILD is between 1500Hz - 4000Hz. All due to head dimensions.
I think this could mean that hearing system is sensitive to phase between 500-4000Hz and frequency response above 1500Hz, because this is the info that uses lots of processing power in the brain. No wonder that is the bandwidth where ear is most sensitive (fletcher munson)?
From this we could draw that diffraction below the ITD is bad. Interestingly diffraction seems to be worse on low frequencies than high up, somehown I thought it was the other way around before
Says that hearing system detects direction from sound delay (diffraction around head) between about 500-4000Hz above which diffraction doesn't occur anymore but then head has switched to be a barrier and attenuates sound and hearing system can use that o detect location (ILD). Transition from ITD to ILD is between 1500Hz - 4000Hz. All due to head dimensions.
I think this could mean that hearing system is sensitive to phase between 500-4000Hz and frequency response above 1500Hz, because this is the info that uses lots of processing power in the brain. No wonder that is the bandwidth where ear is most sensitive (fletcher munson)?
From this we could draw that diffraction below the ITD is bad. Interestingly diffraction seems to be worse on low frequencies than high up, somehown I thought it was the other way around before
True that.
I have experienced that myself.
Had a cold, blew my nose too hard and clogged up one ear.
While that ear was clogged ie the perceived volume on it greatly reduced compared to the non-clogged ear listening to my stereo was basically a mono affair.
However while going for a walk in the park I had no difficulties at all pinpointing from which direction noises like bird song were coming from.
I have experienced that myself.
Had a cold, blew my nose too hard and clogged up one ear.
While that ear was clogged ie the perceived volume on it greatly reduced compared to the non-clogged ear listening to my stereo was basically a mono affair.
However while going for a walk in the park I had no difficulties at all pinpointing from which direction noises like bird song were coming from.
My 2ct:
Phase recognition becomes problematic where the wavelength of sound is smaller than the distance between the ear drums (I believe 17cm is often used). Per definition. And since the head is in the way of signals that do not originate right in front of us, the effective frequency at which we could use phase differences to localize sound off axis is even lower. That is because even more wavelengths fit in the detour path. Think of our stereo speakers in the proverbial listening triangle and you know my idea about these things. We can't defeat physics.
Phase recognition becomes problematic where the wavelength of sound is smaller than the distance between the ear drums (I believe 17cm is often used). Per definition. And since the head is in the way of signals that do not originate right in front of us, the effective frequency at which we could use phase differences to localize sound off axis is even lower. That is because even more wavelengths fit in the detour path. Think of our stereo speakers in the proverbial listening triangle and you know my idea about these things. We can't defeat physics.
It might. Strong diffraction leads to ditto interference patterns, there is a time domain issue and it leads invariably to poor directivity control. The last (leading to irregular lateral reflections) could be the most problematic.
Interestingly diffraction seems to be worse on low frequencies than high up, somehow I thought it was the other way around before
I think you might draw a wrong conclusion regarding the relation between how human detects localization and speaker quality perception.
^I'm not sure what you mean. The post was my thought based on the referenced text that sound diffracting around head makes the ITD region of hearing sensitive to delays. Line you've quoted is about my personal conceptualization of diffraction as phenomenon in general Might be wrong still, but now it seems logical for me that diffraction has worse effect on low than high frequencies I mean it is more mid than treble issue if thinking loudspeakers. This is visible even in the by various frequency response plots demonstrating baffle diffraction, diffraction has less amplitude modulation with rising frequency. Maybe my writing causes confusion?
I've conceptualized (to myself) diffraction of head works like so: example frequenzy 1000kHz head is not big enough to attenuate signal so both ears hear it different time as it arrives. In addition the ear detects delayed sound which is, the sound diffracts around the head literally goes over your face
Might be wrong still, but now it seems logical for me that diffraction has worse effect on low than high frequencies I mean it is more mid than treble issue if thinking loudspeakers. This is visible even in the by various frequency response plots demonstrating baffle diffraction, diffraction has less amplitude modulation with rising frequency. Maybe my writing causes confusion?I've conceptualized (to myself) diffraction of head works like so: example frequenzy 1000kHz head is not big enough to attenuate signal so both ears hear it different time as it arrives. In addition the ear detects delayed sound which is, the sound diffracts around the head literally goes over your face
Last edited:
Please correct me if my conclusions are wrong since I'm literally thinking this through as I write and am eager to learn since it is interesting topic Maybe this is common knowledge for many already?
Before yesterday my concept about speaker diffraction was built on notions like "roundovers reduce diffraction", "roundovers should be big enough to reduce diffraction lower in frequency", "diffraction on low frequencies is not as detrimental than high" and some other similar but all these are were without context. To me it looks logical that hearing system has evolved to take advantage of head geometry. Literally the head provides predictable anomaly for incoming sound and brain has learned how to compute information about it. Head geometry relates to diffraction related delay about at the below 4000Hz so I think this would be the range delay related issues in loudspeakers "distract" hearing system the most.
Sometimes intuition is just wrong, especially when there is not enough acquired knowledge so take my posts as entertainment than truth
Maybe this is common knowledge for many already?Before yesterday my concept about speaker diffraction was built on notions like "roundovers reduce diffraction", "roundovers should be big enough to reduce diffraction lower in frequency", "diffraction on low frequencies is not as detrimental than high" and some other similar but all these are were without context. To me it looks logical that hearing system has evolved to take advantage of head geometry. Literally the head provides predictable anomaly for incoming sound and brain has learned how to compute information about it. Head geometry relates to diffraction related delay about at the below 4000Hz so I think this would be the range delay related issues in loudspeakers "distract" hearing system the most.
Sometimes intuition is just wrong, especially when there is not enough acquired knowledge so take my posts as entertainment than truth
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.