• Welcome to ASR. 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!

Is there any music that actually requires 24 bits for replay?

No hostility - but we will robustly correct falsehoods stated as facts.

Telling you that you are wrong != hostility.
Well, yes and no. That's a cultural thing.

Kilgore Trout ably expressed the situation as reported in Breakfast of Champions.

Ideas on Earth were badges of friendship or enmity. Their content did not matter. Friends agreed with friends, in order to express friendliness. Enemies disagreed with enemies, in order to express enmity.​


But I learned from experience working in certain other cultures that this is not correct. Trout is right about UK and US, and to an extent about some other places I worked in Europe but in a couple of other countries I was pleased to find that I didn't need to be so careful in explaining how someone is wrong. All that work I was accustomed to doing to show more senior colleagues that such-and-such was their belief all along, or as close as I could manage, was just a waste of time with these people. I loved it. But I also had to get used to being bluntly told I was wrong.
 
Soft sounds are recorded with less bits than loud sounds in the same recording with PCM digital, isn't that true? Isn't that how PCM works? You guys are going to tell me that this is incorrect? I described fewer bits as less resolution, OK so perhaps not the correct terminology. DSD does not work like this, and neither does analog recoding.

It seems at least reasonable to me because of these facts that a higher bit depth recording would allow more bits for the softer sounds. I think that is true. What I got from this exchange is that dithering combined with noise shaping is so very effective that the differences between 16-bit and higher resolution recordings cannot be heard under normal listening conditions. I thought not initially, because Amir, (whose opinion I've come to trust due to his focus on objectivity and testing, which I very strongly believe in) had stated that he could tell the difference between CD quality and higher res quality in blind testing. I didn't know that he had to "cheat" (i.e., listen to only the very soft parts at very high volume) to do this. That is an important revelation.

If that is what it takes to tell the difference, then I agree that it's not necessary, and that 16-bit PCM is sufficient. I'm not interested in using a stethoscope on my tweeters to listen to music. But as I said earlier, high-res recordings are here, and in my case, I don't pay a premium price for their availability. Enough said.
 
Soft sounds are recorded with less bits than loud sounds in the same recording with PCM digital, isn't that true? Isn't that how PCM works? You guys are going to tell me that this is incorrect? I described fewer bits as less resolution, OK so perhaps not the correct terminology. DSD does not work like this, and neither does analog recoding.

It seems at least reasonable to me because of these facts that a higher bit depth recording would allow more bits for the softer sounds. I think that is true. What I got from this exchange is that dithering combined with noise shaping is so very effective that the differences between 16-bit and higher resolution recordings cannot be heard under normal listening conditions. I thought not initially, because Amir, (whose opinion I've come to trust due to his focus on objectivity and testing, which I very strongly believe in) had stated that he could tell the difference between CD quality and higher res quality in blind testing. I didn't know that he had to "cheat" (i.e., listen to only the very soft parts at very high volume) to do this. That is an important revelation.

If that is what it takes to tell the difference, then I agree that it's not necessary, and that 16-bit PCM is sufficient. I'm not interested in using a stethoscope on my tweeters to listen to music. But as I said earlier, high-res recordings are here, and in my case, I don't pay a premium price for their availability. Enough said.
What would convince you of dithered 16 bit being enough? I posted a file with a very low (-118 dbFS) tone you could hear amplified and without. Is 118 db not enough range for you?

The other aspect is there are nearly no recordings with even 16 bits of range on them due to environmental noise. Purely electronic recreations would be the only exception. Even if you picture fewer bits having less resolution, if the noise level is not less than -70 dbFS does it really matter in what you will hear?
 
You didn't finish reading what I wrote, I agree that 16-bit PCM is sufficient.
 
Soft sounds are recorded with less bits than loud sounds in the same recording with PCM digital, isn't that true?
I'd rather say: with fewer bits set to 1. You still need all 16 bits, regardless if the sound is soft or loud. You can't just get rid/ignore those most significant bits that are set to 0.

The number of bits determines the noise floor. Which bits are set to 1 determines the signal level, or how far from the noise floor the signal is.

Btw, you can always take your higher-bits signal and the lower-bits signal and subtract them from each other. What will be left is just a broadband noise. That noise is the noise floor of the lower-bits signal. There's nothing else there, no "higher resolution".

(of course I assume dither is used)
DSD does not work like this, and neither does analog recoding.
Both DSD and analog recordings have noise floor, so imo that's very similar.
 
Soft sounds are recorded with less bits than loud sounds in the same recording with PCM digital, isn't that true? Isn't that how PCM works? You guys are going to tell me that this is incorrect? I described fewer bits as less resolution, OK so perhaps not the correct terminology. DSD does not work like this, and neither does analog recoding.
Vinyl, tape and PCM all have noise (floor). The noise in vinyl comes from the roughness of the groove, in tape it comes from the magnetic structure of the tape, the Weiss' domains, and in PCM the noise comes from the dither.
Signals below the noise floor will be swamped the lower their level is.
The maximum level for vinyl is a mechanical limit when the pickup can no longer follow the groove or there is only a limited space available for the groove deflection. For tape it is limited by the saturation of the tape.
In both cases, the distortion increases towards the maximum level.
If you do not need the maximum level (because of "soft sounds") then not the full space for deflection is used or the magnetisation is varying below the possible maximum.
With PCM, the maximum level is mathematically defined (number of bits) and there is no principle-related distortion.
If you do not need the maximum level (because of "soft sounds") then not all bits are used.
The room between noise floor and the maximum level is the available dynamic.
It is all quite similar, PCM is just better, much better. Math beats matter.
 
Soft sounds are recorded with less bits than loud sounds in the same recording with PCM digital, isn't that true? Isn't that how PCM works? You guys are going to tell me that this is incorrect? I described fewer bits as less resolution, OK so perhaps not the correct terminology. DSD does not work like this, and neither does analog recoding.

It seems at least reasonable to me because of these facts that a higher bit depth recording would allow more bits for the softer sounds. I think that is true. What I got from this exchange is that dithering combined with noise shaping is so very effective that the differences between 16-bit and higher resolution recordings cannot be heard under normal listening conditions. I thought not initially, because Amir, (whose opinion I've come to trust due to his focus on objectivity and testing, which I very strongly believe in) had stated that he could tell the difference between CD quality and higher res quality in blind testing. I didn't know that he had to "cheat" (i.e., listen to only the very soft parts at very high volume) to do this. That is an important revelation.

If that is what it takes to tell the difference, then I agree that it's not necessary, and that 16-bit PCM is sufficient. I'm not interested in using a stethoscope on my tweeters to listen to music. But as I said earlier, high-res recordings are here, and in my case, I don't pay a premium price for their availability. Enough said.
I think you need to watch this. Hopefully you'll get a better understanding of how quantisation - and quantisation noise impacts the signal.

 
You didn't finish reading what I wrote, I agree that 16-bit PCM is sufficient.
I noticed that. The tone seemed to me as if you were taking this somewhat provisionally as if to say you guys tell me I had the wrong idea, so okay, but it seems wrong. So my question was about finding some way you could get it and believe in it for yourself. The idea of bit levels and resolution are less than intuitive once dither is included. Shaped dither can lower actual noise levels dramatically where we can hear most sensitively by moving it to where we don't hear very well. If you do go with higher sample rates it can do even more as it can park the noise above 20 khz where we won't hear it at all.
 
Nope. In a home audio setting, the 16 bitt 44.1kHz sample rate is as good as it needs to be. Make it technically better, and the audio improvement won't be audible.
 
A "song" consisting of alternating pin / toothpick drops and jackhammers / close up jet engines might need 24 bits played in a very quiet room for accurate reproduction.

I can't imagine any music I would want to hear needing more than 16 bits.
 
I think you need to watch this. Hopefully you'll get a better understanding of how quantisation - and quantisation noise impacts the signal.

I already posted this and more in my comment #254 But as is often the case, the user seems unwilling to engage with the facts laid out plainly in front of them, likely because it challenges their preconceived ideas. "Better just keep repeating and rewording my old opinions," seems to be the approach.

When this happens -when they either don’t respond at all or reply in a way that clearly shows they’re not open to learning or having their views challenged -you know with absolute certainty that continuing the discussion will be a waste of time. And yet, if you’re anything like me, you still keep going. Not for them, but for the silent readers who might actually take something useful away from it.
Am I right? ;)

1744517330032.png
 
Please read comment #282. Again. I read everything that everyone gave me to read, listened to everything and took the tests too. I wound up agreeing with you that 16-bit PCM is sufficient. You have a convert. Enough said.
I was answering the question you posed in your first paragraph of that post.

Your post massively failed to make it clear you were a convert and that you were no longer questioning what you were being told, seemingly for ma as for @Blumlein 88
 
Your post massively failed to make it clear you were a convert and that you were no longer questioning what you were being told, seemingly for ma as for @Blumlein 88

Time to leave him alone, I think. It took humility for him to admit he was wrong, and to his credit - he changed his mind when presented with evidence. We should applaud that.
 
Time to leave him alone, I think. It took humility for him to admit he was wrong, and to his credit - he changed his mind when presented with evidence. We should applaud that.
Agreed.
 
It seems at least reasonable to me because of these facts that a higher bit depth recording would allow more bits for the softer sounds. I think that is true.
That is your problem.
What seems reasonable to you (and plenty of others) is not the case.
 
That is your problem.
What seems reasonable to you (and plenty of others) is not the case.
But it is the case. "a higher bit depth recording would allow more bits for the softer sounds" is true pretty much by tautology, viz. more bits == more bits.

It may or may not be the case that quantization noise is well below the system/environment noise floor and/or threshold of hearing but that is an entirely different assertion. And it's a more interesting thing to teach and learn about since it raises a whole bunch of caveats, e.g. assumptions about the signals involved, mastering, and what not.
 
Thank you, Multicore.
Objectivity is supposed to be the keynote of this forum, is it not? Amir passed double blind testing in comparing 16-bit digital to higher resolution digital.
So, that is concrete, objective evidence that 16-bit PCM is not perfect, and that higher resolution digital has shortcomings for the very reasons that I cited.

What I didn't know is that he sort of "cheated", which to me is a very important, even revelatory insight. That makes a difference. That's why I now agree.

Some of you apparently get offended when a simple, straight-forward mathematical analysis shows quite clearly where the bones are buried in 16-bit PCM digital. Two of you couldn't even finish reading a short post before quipping back. Perhaps your literary skills are not yet as vast as your audio knowledge. Ditto for manners.
 
It may or may not be the case that quantization noise is well below the system/environment noise floor and/or threshold of hearing but that is an entirely different assertion.
Well if the noise from dither is well below the system noise floor, which IS the case in many recordings as your typical concert hall has (significantly) more noise, then it does not matter whether you can lower the dither noise even further with more bits. The number of "usable" bits will not change.
So it depends. But in theory you are right. ;)
 
Back
Top Bottom