Is earthworks > m23 unnecessary? (recording above 23kHz)

[boogerlad]

I understand that above 23kHz is still useful for scientific purposes, but their FAQ says this:

Why do you make microphones that exceed 20kHz?
Earthworks does not make microphones that go up to 50kHz because of a belief that you can hear tones at these frequencies. Due to research conducted over a long period of time, it has been determined that sound in an echoic environment has multiple pressure waves which converge on the listener nearly simultaneously. Its these wave fronts, which may only have a low kHz center frequency and bandwidth, but when they are spaced very closely in time, the human ear is sensitive to time relationships between the wavefront arrival times down to a resolution of around 20 microseconds. This time relationship corresponds to a frequency of 50kHz. Its not the pure tonal sounds you look for with a 50kHz microphone, its an accurate representation of the space the sounds you are capturing occurred in. This will result in a vastly improved, more true to life sound than traditional 20kHz band limited microphones. You really have to hear it to believe it.

Is this bullshit? I'm only interested in what is in the range of human hearing.

 

[Keith_W]

Need to tag people who know about psychoacoustics ... @j_j and @Thomas Lund

If you could link to their FAQ it might be useful.

Welcome to ASR, by the way!

 

[boogerlad]

Frequently Asked Questions — Earthworks Audio

01 / Service & Repair Who should I contact for repairs within or after the warranty period? You

earthworksaudio.com

Thanks for the warm welcome

 

[j_j]

I understand that above 23kHz is still useful for scientific purposes, but their FAQ says this:


Is this bullshit? I'm only interested in what is in the range of human hearing.

Not bullshit.

No, you don't hear above 20kHz give or take,but you can resolve time differences at lower frequencies down to maybe 5 microseconds. It has to do with 'signal to noise ratio vs. slew rate'.

Now, this does not mean that the microphone has to work well up to 50kHz, but it's (*&(*&*( hard to get good phase response at 20khz if the microphone does not respond well above 20kHz. Physics is a (*&(&*

 

[voodooless]

You really have to hear it to believe it.

That should really tell you al you need to know about this

 

[JanRSmit]

That should really tell you al you need to know about this

I am not sure what you really mean, perhaps you can elaborate what you ment with that remark?

 

[j_j]

I am not sure what you really mean, perhaps you can elaborate what you ment with that remark?

The "hear it to believe it" statement is very often a lead-in to imaginary results.

 

[Sjkrabbe]

for instance take recording of Bats flying around and screaming. You make a recording at 192kHz and afterwards fold it digitally down to the human hearing spectrum. Thats just one of many purposes

 

[Mart68]

Not bullshit.

No, you don't hear above 20kHz give or take,but you can resolve time differences at lower frequencies down to maybe 5 microseconds. It has to do with 'signal to noise ratio vs. slew rate'.

Now, this does not mean that the microphone has to work well up to 50kHz, but it's (*&(*&*( hard to get good phase response at 20khz if the microphone does not respond well above 20kHz. Physics is a (*&(&*

This bit ''This will result in a vastly improved, more true to life sound'' is bullshit though, surely?

 

[boogerlad]

Even if it is not bullshit, aren't we limited by the speakers?

 

[Miley_B]

Not bullshit.

No, you don't hear above 20kHz give or take,but you can resolve time differences at lower frequencies down to maybe 5 microseconds. It has to do with 'signal to noise ratio vs. slew rate'.

Now, this does not mean that the microphone has to work well up to 50kHz, but it's (*&(*&*( hard to get good phase response at 20khz if the microphone does not respond well above 20kHz. Physics is a (*&(&*

Any further evidence based reading you could point me towards?

 

[j_j]

This bit ''This will result in a vastly improved, more true to life sound'' is bullshit though, surely?

Well, if some other microphone has a huge phase shift (beyond delay) starting at 10k, maybe not compared to that, but ...

 

[j_j]

Any further evidence based reading you could point me towards?

How about you study physics, and find out what kind of phase shift you will get with a sharp IIR cutoff, like say a mike that's down 3dB at 20k, and 30dB at 30k. Then compare that to the published results for interaural and intraaural sensitivities.

I'm not going to interrupt my dinner to cite you well-known results, and I do NOT appreciate your tone using the phrase "evidence based". Read the refereed literature, ok?

Edited to add:
On this site there are several lengthy discussions about the sensitivity of the hearing apparatus to both interaural and intraaural time differences. Rather than go through all of that again, please read the discussions first.

As to the phase response of a microphone, I'm quite astonished that a simple understanding of resonances and diaphragm movement, and how they relate to relative phase, is in any way surprising.

 

[EERecordist]

Ultrasonic microphones are used for scientific research, recording insects, bats, and the like. Sometimes sound designers for movies, or experimental musicians take ultrasonic source recordings and slow them down to create a unique sound.

For example: https://www.sameliaaudio.com/blog/u...ding,hyper-real soundscapes enhance immersion.

 

[McFly]

j_j coming at you like Gandalf at Pippin

Also to accurately measure distortion at a given frequency, you need response out to the fundamental of the harmonic i.e. 3rd order at 20khz = 60khz.

 

[j_j]

j_j coming at you like Gandalf at Pippin

Also to accurately measure distortion at a given frequency, you need response out to the fundamental of the harmonic i.e. 3rd order at 20khz = 60khz.

However, what effect do you expect from a third harmonic of 20kHz, perceptually, unless the same nonlinearity results in low-frequency IMD?

That, of course, can be determined without extended bandwidth by using a "buzz tone" such as the ones presented here a while ago.

 

[Keith_W]

How about you study physics, and find out what kind of phase shift you will get with a sharp IIR cutoff, like say a mike that's down 3dB at 20k, and 30dB at 30k. Then compare that to the published results for interaural and intraaural sensitivities.

To illustrate @j_j 's point, here are two simulations. Both were created by making a minimum-phase low-pass filter. Upper graph = magnitude, lower graph = phase.

In red, the f0 is 20kHz and Q is 0.707. It is -3dB at 20kHz, and -8dB at 30kHz.

In green, the f0 is 48kHz and Q is 0.707. At 20kHz, it is -0.1dB. At 30kHz, it is -0.5dB.

 

[Thomas Lund]

Right. In general, a production signalpath with more precision in time and level than needed is a good starting point. Apply suitable dither and lo-pass filtering at the end, including for instance a gentle gaussian filter.

Re. audibility, we did systematic listening tests (Yamaha C7 and human voice, analog vs digital path), leading to bypass of converter chip down/up sampling filters at the time, including instead a selection of custom filters. The most favoured filters used considerable bandwidth (18-36 kHz) for a gaussian roll-off.

 

[j_j]

View attachment 463761

To illustrate @j_j 's point, here are two simulations. Both were created by making a minimum-phase low-pass filter. Upper graph = magnitude, lower graph = phase.

In red, the f0 is 20kHz and Q is 0.707. It is -3dB at 20kHz, and -8dB at 30kHz.

In green, the f0 is 48kHz and Q is 0.707. At 20kHz, it is -0.1dB. At 30kHz, it is -0.5dB.

The results in 's' domain would be more microphone-like, but the point holds.