Is high-resolution audio audible or not audible and a waste of data?

[SIY]

Another thing is that the transients whose rise rate is lower than 22usec are diluted in time, before and after the event, as reported in the graph of the first post. And since our ear has a higher sensitivity than this value (6-10usec), 44KHz is not physically sufficient to preserve transient events for our ears.

That is completely untrue. And particularly galling since the video linked earlier had an actual real-time demonstration that timing is absolutely unrelated to sample rate. You either didn't bother to look at it or you did and completely missed the point.

 

[Waxx]

That link with timing and frequency was already debunked by Dan Lavry (the infamous ad/da convertor builder) more than a decade before in an infamous series of papers on his company website. He claims that 60kHz is the ideal sample rate for audio, so he prefer 88.2 and 96kHz (as they are the closed) for recordings, and even 44.1 and 48kHz should be more than enough. Most of the papers are here and more in depth studies of him are found on the (paying) site of AES. And he was probally not the first one who debunked that (but the one i know, and the one that explained it clear enough that non-electronic engineers like me can understand it)

 

[antcollinet]

I think there is some confusion.
One thing is that for a signal limited in audio bandwidth, the sampling rate of 44KHz is able to reproduce with high precision the instant in which a transient occurs (110psec, although I should elaborate on how this value is calculated), which exceeds the level of temporal accuracy (ITD/ILD) of our ear.

Another thing is that the transients whose rise rate is lower than 22usec are diluted in time, before and after the event, as reported in the graph of the first post. And since our ear has a higher sensitivity than this value (6-10usec), 44KHz is not physically sufficient to preserve transient events for our ears.

The other issue is then on whether or not these speeds are present in real musical signals, the recordings are able to record them, the reproduction systems are able or not to return such precision, the environment, etc. all factors that can affect listening.

Are we talking about sensitivity to time difference between ears, or sensitivity of a single ear to a transient rise time.

If the former, then sample rate is irrelevant, becuase the time difference of a band limited signal can still be maintained to much higher sensitvity than our ear/brain.

If the latter then sample rate (above 44.1kHz) is irrelevant because our ears are also band limited to around 20Khz (or much less for most) and even if the signal entering the ear is not band limited, the signal reaching the brain is band limited by the physical bandwidth limitations of the ear, and the transient event arriving at the brain will be slew limited similar to as shown in your second waveform.

 

[fpitas]

Like all extraordinary claims, OPs premise demands exceptional proof. I'm not seeing anything of the sort, sorry.

 

[TonyJZX]

i love this kind of thread because its a bit schizophrenic given the current buying habits of dacs

obviously 24/192 is old and we all need 32/768 dacs... and further this spec starts at USD$100

and that's before we get into debates about dsd and... mqa

and 'hi-res certification'

when it comes down to it, mastering seems to be the biggest determinent of quality (apparently)

i'm going to throw something else out there.... youtube even does good quality music these days

 

[voodooless]

Like all extraordinary claims, OPs premise demands exceptional proof. I'm not seeing anything of the sort, sorry.

... but he read it somewhere. Isn't that good enough?

 

[fpitas]

... but he read it somewhere. Isn't that good enough?

Yeah. It's that kind of thread

 

[Blumlein 88]

I was tested at the University of Ghent (Belgium) because i could hear "exceptional" high they said, but my first test at the age of 10 was just above 19kHz. They started to test me because i complaint about high pitched sounds of old tv's and other electronic devices that nobody hears. Never heared about people who hear above 20kHz. Do you have links to those studies?

Btw, now at the age of 43 i still can hear until 16854Hz (test done a few months ago). And that is also exceptional they say at the university, especially for a former dj and sound engineer who worked often in very loud envirroments. But i don't see myself as a golden ear, it's more a problem because i hear all kind of stuff i don't want to hear. I can also hear very silent sounds, that normal humans can't, and that is a big problem. I don't want to hear the insects that live in my house (in every house) when i'm in my bed, i want silence.

I couldn't find the older paper on it, though it is out there. One thing I hate is unless you keep yourself a list such things are sometimes hard to find.

Anyway, I found one I had not seen before. Some of its results are slightly different than the other, but largely in agreement with it. 2AFC method using 3 down 1 up on the choices.

Hearing threshold for pure tones above 20 kHz

Access full-text academic articles: J-STAGE is an online platform for Japanese academic journals.

www.jstage.jst.go.jp

Note in the upper right hand of that page you can download the entire article as a PDF.

From some of the sources in that article:
Here is one I also had not seen, which shows some sensitivity to even 28 khz.

Cookie Absent

 

[Blumlein 88]

That link with timing and frequency was already debunked by Dan Lavry (the infamous ad/da convertor builder) more than a decade before in an infamous series of papers on his company website. He claims that 60kHz is the ideal sample rate for audio, so he prefer 88.2 and 96kHz (as they are the closed) for recordings, and even 44.1 and 48kHz should be more than enough. Most of the papers are here and more in depth studies of him are found on the (paying) site of AES. And he was probally not the first one who debunked that (but the one i know, and the one that explained it clear enough that non-electronic engineers like me can understand it)

Lavry got the idea about 60 khz from J.J Johnston who posts here. J.J. thought the sample rate should have been 60 khz, to cover the few who hear 25 khz, and allow a more generous 5 khz band for the filter to work. There are according to him edge cases where the filters could have audible effects from being too narrow. Though I think with better filters he would consider the filter part not a problem now (though I may be interpreting him wrongly). The idea was a certain completely blameless digital format. He also has expressed the idea 48 khz with its 4 khz dead band is probably right on the edge or may be effectively blameless.

And yes Lavry even in his 96 khz AD DA units rolls off the response at either 30 or 35 khz.

 

[ahofer]

I had some success telling hi-res MP3 from redbook, but only with classical music. With pop/rock I was no better than random.

I’ve read that some studio techs can tell higher sample rates from redbook, using very specific ‘tells’ - i.e. trained listeners (@amirm ? ).

I think the higher sample rates are useful for DSP and the multiple ADC-DAC conversion environment of studio mastering, but at our level useless. I have Roon downsample all the silly files (to 96, which leaves me a lot of room) before sending it around the house just to make the multispeaker feature work a little better.

The OP’s technical theories have been addressed here, but the important thing is seeing if you can hear the difference yourself. From my own tests, I may be able to hear up to redbook (with certain kinds of music), but no way I can reliably distinguish higher sample rates than 44k.

 

[Blumlein 88]

One of the sources for the article I link above was research on threshold levels vs age vs frequency. They used a large number of people to test. Dozens in each age group. I think for 50 years the thresholds above 15 khz were 40 db higher than young adults.

I was recording a group once and they used some maraca shakers on a couple songs. These are the things that look like small gourds with a wooden handle. Make a shoosh-shoosh sound when shaken. So when mixing the track I upped the level on them a fair bit. You could hear them, but without a boost they were almost lost. Several people liked that better and one fellow said it could use some more. I wasn't sure about that, but sent them a file with it boosted a bit more. Well the youngest member was a 14 year old girl. And she said boy we had the maracas really loud. Another gal around 25 said they were more than enough. The fellow who wanted it louder was the 68 year old guitarist and said they were just noticeable. I wondered about that so had the maracas recorded with my Earthworks microphone. The sound covered a half octave band centered on 13 khz. So no wonder different people had different reactions. How do you mix and master such a sound? Who is your target audience? I ended up going back and using them slightly lower than my initial boosted level with some compression.

 

[Waxx]

I couldn't find the older paper on it, though it is out there. One thing I hate is unless you keep yourself a list such things are sometimes hard to find.

Anyway, I found one I had not seen before. Some of its results are slightly different than the other, but largely in agreement with it. 2AFC method using 3 down 1 up on the choices.

Hearing threshold for pure tones above 20 kHz

Access full-text academic articles: J-STAGE is an online platform for Japanese academic journals.

www.jstage.jst.go.jp

Note in the upper right hand of that page you can download the entire article as a PDF.

From some of the sources in that article:
Here is one I also had not seen, which shows some sensitivity to even 28 khz.

Cookie Absent

ok, on very high volume that is (+100dB for 24Hz). That is not a tone where you listen to music if you want not to damage your ears i think.

But in my case they started measuring my ears because i could not watch a 1970's CRT TV because of the high pitch resonance of the screen. I became crazy from it. I don't remember what frequency that was exactly but that surely was higher than my parents and my brother and sister could hear. And that was not on a very high level. Modern flatscreens don't have that (luckely) and some more expensive CRT's neighter. The solution was to buy a very expensive at that time Sony TV. And later with computers we used screens like used in the graphic industry from Viewsonic that were also very epxensive, but didn't have that resonance in my hearing passband like ordinary CRT monitors had.

 

[fpitas]

ok, on very high volume that is (+100dB for 24Hz). That is not a tone where you listen to music if you want not to damage your ears i think.

But in my case they started measuring my ears because i could not watch a 1970's CRT TV because of the high pitch resonance of the screen. I became crazy from it. I don't remember what frequency that was exactly but that surely was higher than my parents and my brother and sister could hear. And that was not on a very high level. Modern flatscreens don't have that (luckely) and some more expensive CRT's neighter. The solution was to buy a very expensive at that time Sony TV. And later with computers we used screens like used in the graphic industry from Viewsonic that were also very epxensive, but didn't have that resonance in my hearing passband like ordinary CRT monitors had.

The old CRTs were around 15kHz (15.75kHz? I think) horizontal sweep. Some modern ones are at about 31kHz, but I very much doubt you hear that

 

[Blumlein 88]

I think there is some confusion.
One thing is that for a signal limited in audio bandwidth, the sampling rate of 44KHz is able to reproduce with high precision the instant in which a transient occurs (110psec, although I should elaborate on how this value is calculated), which exceeds the level of temporal accuracy (ITD/ILD) of our ear.

Another thing is that the transients whose rise rate is lower than 22usec are diluted in time, before and after the event, as reported in the graph of the first post. And since our ear has a higher sensitivity than this value (6-10usec), 44KHz is not physically sufficient to preserve transient events for our ears.

The other issue is then on whether or not these speeds are present in real musical signals, the recordings are able to record them, the reproduction systems are able or not to return such precision, the environment, etc. all factors that can affect listening.

Well you are mixing ITD, ILD with transients. I forget the details, but a single impulse has variable thresholds depending upon how loud the transient is. I think the time from memory is in the dozens of milliseconds range. Some very short transients may be audible if they are loud enough. I don't know if that is so short as 5 microseconds. I get the idea the spreading in time done by our ear which is essentially a filter to such things is why the perception of such things is related heavily to how loud it is. Such transients effectively have energy across all frequencies or very wide band energy. So you are only perceiving a portion of the wide band energy in the below 20 khz range. I don't know of info that says transient impulses or signal edges are perceptible at microsecond range levels. So such information is what everyone here would like for you to post.

 

[Tangband]

My own experience about this, making my own recordings of acoustic instruments with two good capacitor microphones and have tried different sampling frequencies between 44,1 , 48 kHz and 96, and have compared 16 bit with 24 bit many times, is this :

1. If the digital filter is very good in your dac - then the 44,1 kHz sampling rate is good enough, otherwise its gonna be a small improvement with 48 or 96 kHz , if the recording is 48 kHz or higher.

2. 24 bit is better than 16 bit because of this : in reality when recording real instruments you have a safe margin about 12 dB , - thats 2 bits of resolution. Recording at 16 bit is in reality 14 bit. Always record in 24 bit and why shouldnt you ? The option is there in every interface and its the year 2023….not 1990.

3. If using lots of plugin programes like peq, limiter, compression, exciter, eq , during the mixing process its better with 20-24 bit source material on all separate tracks. You can hear the difference If there is a lot of conversions during this process.

4. If using a digital volume control at home, a source material of 24 bits ( or converted to 24 bit ) are better than 16 bit.

5. Some SRC programes is worsening the sound In the DAW, but some are transparent to the ear.

6. If you think 320 kbit mp3 is good enough for studio masters, - good for you. Im sure no serious soundtechnician would agree. With my Genelec 8340 and good source material, its a clear difference between 24 bit 96 kHz and 44,1 kHz 320 kbit - with my own recordings. I have done this conversion from 24/96 kHz to 320 kbit 44,1 Khz many times in Logic proX and there is a difference If using good speakers.

Conclusion - 16 bit 44,1 kHz recordings are in most cases good enough at playback if the digital filtering is good enough at Nyqvist freq. and only if you dont do a mixing mastering process with plugins in the studio. In reality, 99 % of all recording are processed after the recording session, and here is the need for a higher resolution. I would say that minimum 20 bit and 48 kHz would be transparent in a mixing process in the studio.

 

[fpitas]

My own experience about this, making my own recordings of acoustic instruments with two good capacitor microphones and have tried different sampling frequencies between 44,1 , 48 kHz and 96, and have compared 16 bit with 24 bit many times, is this :

1. If the digital filter is very good in your dac - then the 44,1 kHz sampling rate is good enough, otherwise its gonna be a small improvement with 48 or 96 kHz , If the recording is 48 kHz or higher.

2. 24 bit is better than 16 bit because of this : in reallity when recording real instruments you have a safe margin about 12 dB , - thats 2 bits of resolution. Recording at 16 bit is in reality 14 bit. Always record in 24 bit.

3. If using lots of plugin programes like peq, limiter, compression, exciter, eq , during the mixing process its better with 24 bit source material on all separate tracks. You can hear the difference If there is a lot of conversions during this process.

4. If using a digital volume control at home, a source material of 24 bits ( or converted to 24 bit ) are better than 16 bit.

5. Some SRC programes is worsening the sound In the DAW, but some are transparent to the ear.

6. If you think 320 kbit mp3 is good enough for studio masters, - good for you. Im sure no serious soundtechnician would agree. With my Genelec 8340 and good source material, its a clear difference between 24 bit 96 kHz and 44,1 kHz 320 kbit - with my own recordings. I have done this conversion from 24/96 kHz to 320 kbit 44,1 Khz many times in Logic proX and there is a difference If using good speakers.

Conclusion - 16 bit 44,1 kHz recordings are good enough at playback If, and only If you dont do a mixing mastering process in the studio. In reality, 99 % of all recording are processed after the recording session, and here is the need for a higher resolution. I would say that minimum 20 bit and 48 kHz would be transparent in a mixing process.

I'm under the impression that even mastering is done at higher resolution than the final product. But that may depend on the particular studio and producer.

 

[Tangband]

I'm under the impression that even mastering is done at higher resolution than the final product. But that may depend on the particular studio and producer.

Yes, its usual so, because 16 bit are not good enough doing this. When the mix is done and normalized, the 2 channel master is usually downconverted to 16 bit 44,1 kHz for streaming services in the last step in the DAW. Totally unnecessary at 2023.
All streaming services should be at least 24 bit 48 kHz. 320 kBit mp3 is so 1990….

 

[markanini]

The argument that "I can't even tell MP3 and PCM apart" is out of place, it doesn't quite relate to hires vs, 16/44.1. It gives a kind of false sense to the main question as if it were a question of hearing ability and not physics.
For me, being aware of the existence of killer samples, that basically break lossy codecs even at high bitrate, it's pretty obvious that lossy audio doesn't need to be inserted into the discussion.

 

[fpitas]

The argument that "I can't even tell MP3 and PCM apart" is out of place, it doesn't quite relate to hires vs 16/44.1. It gives a kind of false sense to the main question that it's a question of hearing ability and not physics.

But, it is ultimately a question of hearing ability. Were our ears as precise as Amir's test equipment, higher resolution would be unquestionably better.

 

[Tangband]

The argument that "I can't even tell MP3 and PCM apart" is out of place, it doesn't quite relate to hires vs 16/44.1. It gives a kind of false sense to the main question that it's a question of hearing ability and not physics.

One could argue that 720p is good enough if you use a 50 inch screen for movies, but the question is why one should put up with this low resolution when there are 4K material for the same cost today ?