Hearing 099

[j_j]

This coming Wednesday, I will be giving a tutorial on the very basics of human hearing at Digipen as part of the AES local Chapter's activities.

You can get more information at www.aes.org/sections/pnw

This is an updated talk, much like one I did about 10 years ago, but with some new material and better illustrations.

It does not require a lot of mathematics, and is at the conceptual level, not at a mathematically stringent level.

 

[Juhazi]

Thank you, obviously we will see it listed here soon!
http://www.aes.org/sections/pnw/jj.htm

 

[Blumlein 88]

Will there be a video of the talk? Or at least an audio recording? I hope so, and hope it is available.

 

[j_j]

Will there be a video of the talk? Or at least an audio recording? I hope so, and hope it is available.

I think so.

 

[amirm]

Gosh, love to attend but have come down with a cold and would have to drive 2 hours each way. I have seen your old PPT and referenced it at times. Would be great to hear you explain it in person.

I will see if I can drag myself up there somehow.

 

[amirm]

I sent my son to attend but he could not get in. Door was locked and no security person there. Did this get cancelled or was there a special way to get in?

 

[j_j]

Huh. That's not how it's supposed to work. There is supposed to be a security guy at all times, and that should not have happened.

There is no special way. Thump on the door has always worked.

 

[amirm]

Thanks JJ. My memory is hazy of how I got in last year. I am thinking I found a side door and walked in.

Do you know if a video was captured?

 

[j_j]

There was a recording. It may be a while until it's edited and ready.

 

[j_j]

http://www.aes-media.org/sections/pnw/pnwrecaps/2019/apr2019/

For those of you who could not be there. No video, but at least audio.

 

[amirm]

Excellent. Thanks JJ.

 

[MRC01]

There's video there now. I wish I heard about this sooner, I would have like to attend in person.

 

[andreasmaaan]

Deepest thanks @j_j for the excellent lecture!

 

[amirm]

It is definitely a great talk. I created a stand-alone link for it in the home page: https://www.audiosciencereview.com/forum/index.php?threads/audio-how-we-hear.7545/

 

[andreasmaaan]

@j_j, I've also gone over your Hearing 101 slideshow and have been mulling over your comments on phase and pre-echo (Slides 29-32 in particular).

The basic concept of a "leading edge detector" seems fairly clear to me - i.e. (correct me if I'm wrong) that the leading edge of a waveform is critical in terms of instantaneous loudness, in contrast to loudness which is summed for up to 200ms, and that pre-echo is a problem because it can start the compression on the basilar membrane before the signal arrives, thereby reducing the loudness (or should that be instantaneous loudness?) of a transient.

I would love to hear your further thoughts on this. In this case, my interest is also quite practical, as I'd like to be able to estimate some ballpark audibility thresholds for pre-echo caused by linear phase FIR filters.

In particular, it would seem to me that a number of active linear phase loudspeakers and DSP implementations on the market would create relatively gross pre-echo as a result of their very long tap lengths.

Previously, I had looked to Zwicker and Fastl's work on pre-masking for guidance as to what might be an acceptable (i.e. likely inaudible) degree of pre-echo. But now I'm wondering if your work doesn't actually problematise their findings (which TBH were fairly ambiguous in the first place)?

Anyway, basically anything further you have to say on this would be very interesting to me

 

[MRC01]

There's one section I didn't quite follow. JJ mentions Elephant hearing and atmospheric pressure around thunderstorms and estimates the change at 1 part in 30 which would be a 30 dB difference: 20 * log (1/30) = -29.5 dB.

However, it seems like he reversed the ratio. Air pressure doesn't drop by 30:1 when a thunderstorm comes. It might drop by 1 part in 30, about 3%, which would be 20 * log(29/30) = -0.3 dB. An extreme TStorm might drop air pressure from 29.92" to 28" which would be a -0.6 dB difference.

Did I misinterpret what he was saying?

 

[MRC01]

... pre-echo is a problem because it can start the compression on the basilar membrane before the signal arrives, thereby reducing the loudness (or should that be instantaneous loudness?) of a transient.
... my interest is also quite practical, as I'd like to be able to estimate some ballpark audibility thresholds for pre-echo caused by linear phase FIR filters

I too am interested in this topic, for the same reasons & applications.

 

[j_j]

There's one section I didn't quite follow. JJ mentions Elephant hearing and atmospheric pressure around thunderstorms and estimates the change at 1 part in 30 which would be a 30 dB difference: 20 * log (1/30) = -29.5 dB.

However, it seems like he reversed the ratio. Air pressure doesn't drop by 30:1 when a thunderstorm comes. It might drop by 1 part in 30, about 3%, which would be 20 * log(29/30) = -0.3 dB. An extreme TStorm might drop air pressure from 29.92" to 28" which would be a -0.6 dB difference.

Did I misinterpret what he was saying?

You misinterpreted what it meant. The air pressure drops by 1 part in 30. Sounds pressure is defined from the MEAN air pressure. In terms of SOUND PRESSURE it's (milliHz or Microhz) incredibly high energy. In terms of absolute pressure, it/s still 1/30.

Since "air pressure" is 194dB SPL (well except you can never reach that level, air isn't linear, which is one of the reasons we have thunderstorms) the change in pressure when the storm passes is 164dB SPL.

And it's really 1 part in 31.6, but we're talking approximately.

 

[j_j]

@j_j, I've also gone over your Hearing 101 slideshow and have been mulling over your comments on phase and pre-echo (Slides 29-32 in particular).


Anyway, basically anything further you have to say on this would be very interesting to me

The answer is annoyingly difficult, varies with frequency (due to cochlear filter impulse), level (due to nonlinearities) and the frequency content of the "filter pre-ripple".

There isn't one answer.

 

[andreasmaaan]

The answer is annoyingly difficult, varies with frequency (due to cochlear filter impulse), level (due to nonlinearities) and the frequency content of the "filter pre-ripple".

There isn't one answer.

Fair enough J_J, thanks Are they papers/books etc. you could recommend please?