• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Can we hear the bottom bits of 24?

MakeMineVinyl

Major Contributor
Joined
Jun 5, 2020
Messages
3,558
Likes
5,871
Location
Santa Fe, NM

MakeMineVinyl

Major Contributor
Joined
Jun 5, 2020
Messages
3,558
Likes
5,871
Location
Santa Fe, NM
My test—you mean the 24-bit file. Not really a test, just a reference.

Some people (from discussing on other boards) are 100% certain they can hear the bottom bit, some aren't sure, but they know that skilled recording engineers have said they can—it's just a reference that people can play with for their own interest. Same reason I included 16. Nothing to be proved, just a known reference to play with privately.
How is it even possible for 'hearing 24 bits' not to be a moot point when even the very best DAC which has ever been tested on this forum can only achieve something like 22 bits of resolution?
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
My test—you mean the 24-bit file. Not really a test, just a reference.

Some people (from discussing on other boards) are 100% certain they can hear the bottom bit, some aren't sure, but they know that skilled recording engineers have said they can—it's just a reference that people can play with for their own interest. Same reason I included 16. Nothing to be proved, just a known reference to play with privately.

BTW, references are good for something: One discussion maybe two years ago included a guy who claimed that we really only needed 12-bit audio. Fundamentally, the guy knew what he was talking about, and we agreed on a number of things, but I thought he took it too far with 12-bit. (16-bit, you've got an argument, even 14-bit for consumer is not unreasonable to argue, but 12-bit, nope—for tones maybe, not music.) I don't think I gave him a tone, I think I told him to maybe truncate to 12-bit and null, or whatever. He actually came back and said, "oops, I didn't realize the 12-bit noise floor was that loud", and modified his thinking a little.

So, I do think it's good to have something for listening perspective. For instance, for me, if I listen to the 16-bit tone, or the residual from nulling 16-bit dithered or truncated audio, at moderate to low listening levels it's amazing how good 16-bit is. Though if I crank to loud but not unreasonable listening levels, there are the flaws. Nothing proven, just for perspective.

One problem I come across often, though, is some people will concede that the 24-bit tone and 24-bit truncated/dithered nulls probably can't be heard, yet truly believe that if that tiny error level is added to perfect music, somehow the soundstage will collapse, etc.
My point is you cannot play that file without going some sort of a conversion. There are no DACs that can convert a 24-bit digital file to an analogue signal. What people are hearing is the difference of each DAC‘s conversion process.
 

Robin L

Master Contributor
Joined
Sep 2, 2019
Messages
5,211
Likes
7,590
Location
1 mile east of Sleater Kinney Rd
What about it? We're talking about noise floor and dead musicians.
Not to mention deaf musicians. And I guess the number of dead musicians depends on where the cannons [a canon is a "round" FWIW] are aimed during the performance.
 
OP
E

earlevel

Addicted to Fun and Learning
Joined
Nov 18, 2020
Messages
545
Likes
776
How is it even possible for 'hearing 24 bits' not to be a moot point when even the very best DAC which has ever been tested on this forum can only achieve something like 22 bits of resolution?
Well, yes. They simply don't believe it's true that a 24-bit DAC can't produce 24 bits of precision audio.
 
Last edited:

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
What about it? We're talking about noise floor and dead musicians.
Think again. What is the SPL difference between background noise of a concert hall and a canon ball exploding? :)
 
Last edited:

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
Well, yes. They simply don't believe it's true that a 24-bit DAC can't produce 24-bit of precision audio.
They also think placing holly water next to a turntable makes a difference? We don’t tend to care about those people.
 
OP
E

earlevel

Addicted to Fun and Learning
Joined
Nov 18, 2020
Messages
545
Likes
776
My point is you cannot play that file without going some sort of a conversion. There are no DACs that can convert a 24-bit digital file to an analogue signal. What people are hearing is the difference of each DAC‘s conversion process.
Ah—you were missing a "no" ("there are no 24-bit DACs") in your original message.

Anyway, of course it's pointless if you understand that. But it's not a pointless reference for someone who thinks they might be able to hear differences in the bottom couple of bits of 24-bit audio.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
Ah—you were missing a "no" ("there are no 24-bit DACs") in your original message.
It was a typo that I have now corrected. However, on my later posts the declaration was correct.
Anyway, of course it's pointless if you understand that. But it's not a pointless reference for someone who thinks they might be able to hear differences in the bottom couple of bits of 24-bit audio.
I rather do things that tend to teach people instead of confusing and tricking them.
 
OP
E

earlevel

Addicted to Fun and Learning
Joined
Nov 18, 2020
Messages
545
Likes
776
I rather do things that tend to teach people instead of confusing and tricking them.
You think that showing people the signal level being discussed is a trick? or confusing? If so, OK, I disagree, but maybe I'm not reading you.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
You think that showing people the signal level being discussed is a trick? or confusing? If so, OK, I disagree, but maybe I'm not reading you.
You are giving them a file that cannot be reproduced as intended.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
No recordings actually capture the full volume peak of a cannon blast. I can attest to this from experience when I was working in motion picture sound on a munitions-heavy film (I'm dead sure you've seen it). What was quickly found is that even when a suitable microphone is set right next to a cannon (these were recorded at the 29 Palms Marine Corps Base and Camp Pendleton, CA) and the levels were set for no clipping, the cannon sounded like a cap gun when reproduced on speakers.

There were two methods to get around this. The most successful was to let the digital recorder (a Mackie ADAT in this case) go deeply into clipping. The practical effect of this was to act as a crude compressor which emphasized the lower level sounds of the blast. The recording sounded a lot more like the sound of actually standing next to the cannon. By the way, the Marines let me fire the cannon myself a couple times - I have a picture of it. :)

The other method of capturing a realistic sounding cannon, which is the method most commonly used, is to simply get far enough away from the cannon so that there is more indirect sound in the capture. This in effect limits the dynamic range, yet makes the cannon sound more 'real'. I used this method too (there were multiple microphone setups feeding the multi-track ADAT), but the close up with clipping sounded best.

Many other munitions were also recorded from rocket launchers to C4 - these too suffered from the same poor results with the microphones close to the blast and capturing the 'correct' dynamic range. The solution was the same as with the cannon recording.
You are preaching the preacher. Check my profile. I was joking as it is theoretically possible to generate a 24-bit SPL difference.
 
OP
E

earlevel

Addicted to Fun and Learning
Joined
Nov 18, 2020
Messages
545
Likes
776
Hi I think both 16- and 24-bit sample files are empty.
I clicked the link and listened to the 16-bit version, licked play, I can hear the sweep tone easily on my computer's sound system (USB to Topping DX7 Pro, balanced outs to ILoud-MTM speakers, seated about 28" from speakers) at modest music-listening levels. I'm confident the 24-bit tone is there to, you can download it and look at it in an audio editor (or cheat and give it digital gain before the conversion to hear it).

Of course, in real listening situations, it would be music plus low level errors that are not sweeping tones but noisy errors. The point of the sweep tones is to give an idea of what "a" signal at that level sounds like on its own. I make no representation that it's the best or most meaningful test tone ever. But note that you can't just dream up an arbitrary test tone sweep that you think would be good, because it would be subject to truncation distortion and you'd have to dither it...with noise at the same level as the signal, completely obliterating the entire idea. These tone are digital and don't have that issue, they are noise free because they simply are what they are, not a sampling of something else.
 
OP
E

earlevel

Addicted to Fun and Learning
Joined
Nov 18, 2020
Messages
545
Likes
776
You are giving them a file that cannot be reproduced as intended.
And the bottom bits of a 24-bit file can't be reproduced as intended—that's the point. There is nothing wrong with the file itself, just no way to reproduce it accurately. The 16-bit file can be reproduced fine, and of course the 5-bit. Somehow, you're misreading my intent.
 

Robin L

Master Contributor
Joined
Sep 2, 2019
Messages
5,211
Likes
7,590
Location
1 mile east of Sleater Kinney Rd
Think again. What is the SPL difference between background noise of a concert hall and a canon ball exploding? :)
Depends on the hall. There's recordings of major-league orchestras with the subway audibly coating the bottom octaves. Zellerbach Hall at UC Berkeley is never really "silent". This is the case more often than not. I know of a major league album recorded in a place where the ventilation system couldn't be shut off. Thank heaven for parametric EQ. Skywalker Ranch is probably the one recording venue suitable for an orchestra with the least "self-noise". About as far from road traffic as they can get and George Lucas has some arrangement with the powers that be to prevent flightpaths over the place. That place was quiet. But still, not "dead silent". Not 24 bits worth.
 
Last edited:

MakeMineVinyl

Major Contributor
Joined
Jun 5, 2020
Messages
3,558
Likes
5,871
Location
Santa Fe, NM
You are preaching the preacher. Check my profile. I was joking as it is theoretically possible to generate a 24-bit SPL difference.
Sometimes theory doesn't sound all that great in the real world.
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
And the bottom bits of a 24-bit file can't be reproduced as intended—that's the point. There is nothing wrong with the file itself, just no way to reproduce it accurately. The 16-bit file can be reproduced fine, and of course the 5-bit. Somehow, you're misreading my intent.
Even you admit the file is not reproducible. What is your objective then?
 
Last edited:

spacevector

Addicted to Fun and Learning
Forum Donor
Joined
Dec 3, 2019
Messages
553
Likes
1,003
Location
Bayrea
I clicked the link and listened to the 16-bit version, licked play, I can hear the sweep tone easily on my computer's sound system (USB to Topping DX7 Pro, balanced outs to ILoud-MTM speakers, seated about 28" from speakers) at modest music-listening levels. I'm confident the 24-bit tone is there to, you can download it and look at it in an audio editor (or cheat and give it digital gain before the conversion to hear it).

Of course, in real listening situations, it would be music plus low level errors that are not sweeping tones but noisy errors. The point of the sweep tones is to give an idea of what "a" signal at that level sounds like on its own. I make no representation that it's the best or most meaningful test tone ever. But note that you can't just dream up an arbitrary test tone sweep that you think would be good, because it would be subject to truncation distortion and you'd have to dither it...with noise at the same level as the signal, completely obliterating the entire idea. These tone are digital and don't have that issue, they are noise free because they simply are what they are, not a sampling of something else.
Interesting - I opened the 5 bit and 16 bit files in Audacity and see the 16 bit file is blank. The 5-bit file plays OK - high to low sweep. May be a download error on my part - will re-download.

1635891270268.png


1635891295707.png
 

sarumbear

Master Contributor
Forum Donor
Joined
Aug 15, 2020
Messages
7,604
Likes
7,314
Location
UK
Depends on the hall. There's recordings of major-league orchestras with the subway audibly coating the bottom octaves. Zellerbach Hall at UC Berkeley is never really "silent". This is the case more often than not. I know of a major league album recorded in a place where the ventilation system couldn't be shut off. Thank heaven for parametric EQ. Skywalker Ranch is probably the one recording venue suitable for an orchestra with the least "self-noise". About as far from road traffic as they can get and George Lucas has some arrangement with the powers that be to prevent flightpaths over the place. That place was quiet. But still, not "dead silent". Not 24 bits worth.
Re-read what I wrote. I said the SPL difference between a quite hall and the canon ball explosion.
 
Last edited:
Top Bottom