Why 24 bit audio and anything over 48k is not only worthless, but bad for music.

[danadam]

It will become "less accurate" after each conversion, I suspect.

Looks like it loses a bit less than 2 bits after 10x dithers and a bit more than 3 bits after 100x dithers:

It looks the same for shaped dither, only the graphs are, well, shaped

 

[Steve999]

i.e. youre biased prejudiced and lazy? Surely not admitting such. I have an idea, you could just IGNORE, or try to focus intellect on what was said and, if your registration date makes you a higher caste over my lowly race of newbies, then no need for civil respect or any kind of higher culture. OR forget registration date, embrace humanity respectfully and answer on the content of the content, not the colour of the registration date.

Let a new nobility arise from registration date and let the rest eat cake! A new nobility has arisen!

I have the best registration date. I am in the highest caste. I am the nobility. I do not embrace or respect humanity or any higher culture. The rest of you eat cake. I win.

The rest of you are two-bit, and I do mean exactly two-bit, punks.

 

[eq1849]

Looks like it loses a bit less than 2 bits after 10x dithers and a bit more than 3 bits after 100x dithers:


It looks the same for shaped dither, only the graphs are, well, shaped

nice work!

Started with 24 bit?

It appears to me, every conversion raises the noise floor to some extent.

 

[71 dB]

Random idea...

Cycle 1
Take a 24/44 native file.
Convert to 16/44 with dither.
Then upconvert it to 24/44

Cycle 2
Convert back to 16/44 with dither.
Then upconvert it back to 24/44

Cycle 3
Convert back to 16/44 with dither.
Then upconvert it back to 24/44

And so on...
I'll null test the test file against the native file in Aadacity and check for differences.

It will become "less accurate" after each conversion, I suspect.

You don't keep dithering. Dithering is used when you go to lower bitrate you have been "before". So, cycle one needs dither, because it is the first time going from 24 to 16 bit, but after that we have 16 bit dither in the signal and we can go back and forth without dithering again and again. It is just adding ( ---> 24 bit) and removing ( ---> 16 bit) eight zero bits.

If you keep dithering, you add noise every cycle, but otherwise nothing bad happens. Danadam's fine post illustrates this well.

 

[eq1849]

You don't keep dithering. Dithering is used when you go to lower bitrate you have been "before". So, cycle one needs dither, because it is the first time going from 24 to 16 bit, but after that we have 16 bit dither in the signal and we can go back and forth without dithering again and again. It is just adding ( ---> 24 bit) and removing ( ---> 16 bit) eight zero bits.

If you keep dithering, you add noise every cycle, but otherwise nothing bad happens. Danadam's fine post illustrates this well.

 

[eq1849]

Results so far on DAC 1:
Correct when 24/44 versus unknown X (A):...................13/33
Correct when 24/44 versus unknown X (B):.. 22/28
35/61

I had a run, by chance because I'm using Itunes shuffle. where I did terrible simply comparing 24/44 reference to the unknown X which turned out to be A a surprising amount of times...felt like almost every time, lol. So comparing the file to its unknown self is an act of confusion even.

 

[gregorio]

Cycle 1
Take a 24/44 native file.
Convert to 16/44 with dither.
Then upconvert it to 24/44
Cycle 2
Convert back to 16/44 with dither.

Why would you add dither a 2nd time? 1st time sure but then you convert back up to 24bit, which just gives you your 16bits of audio data and adds 8 zeros to the last 8 LSBs. If you want to convert back to 16bit again, just truncate (without dither) those added 8 zeroes again. It doesn’t matter how many cycles you do after the first conversion, you’re just adding 8 zeroes and then removing them again. The result is identical no matter how many times you add and delete those padded zeroes and if there is any difference at all, then you’re doing something wrong, a volume change or something else on each conversion.

So comparing the file to its unknown self is an act of confusion even.

Not sure what you’re testing here.

G

 

[Maxx134]

But notice I'm saying these are "my perceptions", not factual, so as not to call others wrong if they don't have the same experience

Your perceptions of sound and the whole world are factual to you. It is true others will have different perceptions guaranteed for countless reasons.

I have learned when viewing a thread, that the "general consensus" of their observations is usually on point.

But did you now check the levels? If not equal within 0.1 dB (which you can not check by ear) it is no good.

So 0.1db will still equal a difference to the ear, but you cannot just check by ear. (!).
This is an admission that the ear can perceive, and then be biased by the difference. I can agree
I believe given most optimal conditions we have potential to hear differences to a certain level. The problem is that because of the mind, those optimal conditions are like chasing a ghost, lol.

So essentially you'll believe anything, right?

Blind test or you don't know what the heck you're hearing.

Just The introduction of even the word "blind" will put the mind in a state of doubt and confusion from the outset.
We need to change that term, especially if the goal is to perceive a difference. This is part of the problem.

Then we dismiss your claim and assume it is all placebo, which it must be in the first place.

By the way: you can do a double blind ABX with casual listening at busy times. It won't improve your chances to pass, but it is possible and "allowed". You can causually listen (while being busy) to A, B and unknown X for as long as you like and choose X=A or X=B. Repeat many times everytime with a new unknown X.

Results so far on DAC 1:
Correct when 24/44 versus unknown X (A):...................13/33
Correct when 24/44 versus unknown X (B):.. 22/28
35/61

I had a run, by chance because I'm using Itunes shuffle. where I did terrible simply comparing 24/44 reference to the unknown X which turned out to be A a surprising amount of times...felt like almost every time, lol. So comparing the file to its unknown self is an act of confusion even.

That's why I feel the ABX method is flawed. The simple introduction of any parameter which is "unknown" will take away the mind perceptions off of any based foundations.

"Uncertainty", will kill off any result. Introduction of uncertainty is the key to destroy any result. The user must be in full control, in order to have any foundation and basis of any assumptions.

This is what another person wrote about "ABX" testing :
"When you AB test every time you switch you improve your knowledge of the suspected differences between A and B. With ABX, however, your brain is required to “suspend” an opinion about the classification of X while your brain is trying to classify A and B. I think it is quite understandable that this is a more mentally complex task. I also could not deny that the complexity of suspending X classification affects one’s ability to classify A and B."
IMO, this is a downright sneaky evil ploy to get the results of not being able to get anything correct, especially for the mentally untrained.


Also, on another point;
I have been listening to some downloads, and found many 16bit recordings seem to be recorded & playing louder than my 24 it recordings. About half of them actually sound better(?)! Which makes me realize that I cannot dissmiss a file due to its Format. Too much variability going on.

Also, a last good point mentioned earlier by member @bigshot, opens the rabbit hole of how well or truthful to the bits, is my DAC preforming. So in some cases we have to rule that variable of gear, out by staying on the same setup.
My recent experience history has shifted from the desktop, to Daps, so I feel that they are hugely responsible for sound difference, and is Totally separate from the topic here of being about quality.
I have already been convinced here about the actual thread topic, hearing various quality of same actual song on different bitrates and found half of the lower bitrate files as good or better than higher, which also leads to the issue of how it was recorded.

 

[sander99]

So 0.1db will still equal a difference to the ear, but you cannot just check by ear. (!).
This is an admission that the ear can perceive, and then be biased by the difference. I can agree

Indeed. You can not consciously determine by ear that the volume is matched within 0.1 dB or just 0.2 dB, but a 0.2 dB difference can make it feel as if the louder one is sounding better. Also in the case that there were real audible differences besides level, then level matching by ear could potentially be even more difficult.

 

[sander99]

"When you AB test every time you switch you improve your knowledge of the suspected differences between A and B. With ABX, however, your brain is required to “suspend” an opinion about the classification of X while your brain is trying to classify A and B. I think it is quite understandable that this is a more mentally complex task. I also could not deny that the complexity of suspending X classification affects one’s ability to classify A and B."

You can first listen to A and B as often as you like to "develop your knowledge".

 

[gregorio]

Your perceptions of sound and the whole world are factual to you.

That’s not true. Take for example the McGurk Effect; like most other people, I perceive/hear both “baa” and “faa”, however, the “faa” is not “factual” to me, I know it’s just an illusion/perceptual error. And the same is true of many other aural illusions (stereo soundstage for example).

Just The introduction of even the word "blind" will put the mind in a state of doubt and confusion from the outset.

No it won’t, if anything, the opposite is true. Whose mind are you talking about? And if it’s different for different people’s minds, then your assertion isn’t applicable to everyone.

This is what another person wrote about "ABX" testing :
"When you AB test every time you switch you improve your knowledge of the suspected differences between A and B. With ABX, however, your brain is required to “suspend” an opinion about the classification of X while your brain is trying to classify A and B. I think it is quite understandable that this is a more mentally complex task.

And that is why you have to be careful when reading “what another person wrote”! This seems to be a common trait in the audiophile world, taking what other people wrote or said seriously, without knowing or finding out the actual facts. This is particularly problematic in the audiophile world because so much of what audiophiles write/claim is BS to start with!

I also could not deny that the complexity of suspending X classification affects one’s ability to classify A and B."

All the above that you quoted is nonsense! ABX is actually very simple: You have X, which is either A or B, so you listen to say X and A. If they are the same then X = A, if they are not the same then X = B. So you never have to listen to B or “suspend X classification”!

IMO, this is a downright sneaky evil ploy to get the results of not being able to get anything correct, especially for the mentally untrained.

No, it’s the opposite, a downright sneaky evil ploy to help you get it right! As explained, you only ever have to listen to X and A but if you’re not sure, you have the option of listening to B as well (if you want).

I have already been convinced here about the actual thread topic, hearing various quality of same actual song on different bitrates and found half of the lower bitrate files as good or better than higher, which also leads to the issue of how it was recorded.

It doesn’t lead to the issue of how it was recorded because it’s the same recording whether it’s a 16bit, 24bit or lossy release. They’re all derived from the same master, although additional compression maybe be applied to some of the formats. Of course you might be comparing a mix with a remix but even then, the actual recording is usually the same even though what was recorded is mixed differently, although some remixes have added parts/instruments that were not recorded at the time of the original recording.

It just goes to show how easily many audiophiles can be “convinced”, all it needs is a bit of suggestion in some marketing and ignorance of the actual facts!

G

 

[eq1849]

Your perceptions of sound and the whole world are factual to you. It is true others will have different perceptions guaranteed for countless reasons.

I have learned when viewing a thread, that the "general consensus" of their observations is usually on point.



So 0.1db will still equal a difference to the ear, but you cannot just check by ear. (!).
This is an admission that the ear can perceive, and then be biased by the difference. I can agree
I believe given most optimal conditions we have potential to hear differences to a certain level. The problem is that because of the mind, those optimal conditions are like chasing a ghost, lol.



Just The introduction of even the word "blind" will put the mind in a state of doubt and confusion from the outset.
We need to change that term, especially if the goal is to perceive a difference. This is part of the problem.





That's why I feel the ABX method is flawed. The simple introduction of any parameter which is "unknown" will take away the mind perceptions off of any based foundations.

"Uncertainty", will kill off any result. Introduction of uncertainty is the key to destroy any result. The user must be in full control, in order to have any foundation and basis of any assumptions.

This is what another person wrote about "ABX" testing :
"When you AB test every time you switch you improve your knowledge of the suspected differences between A and B. With ABX, however, your brain is required to “suspend” an opinion about the classification of X while your brain is trying to classify A and B. I think it is quite understandable that this is a more mentally complex task. I also could not deny that the complexity of suspending X classification affects one’s ability to classify A and B."
IMO, this is a downright sneaky evil ploy to get the results of not being able to get anything correct, especially for the mentally untrained.


Also, on another point;
I have been listening to some downloads, and found many 16bit recordings seem to be recorded & playing louder than my 24 it recordings. About half of them actually sound better(?)! Which makes me realize that I cannot dissmiss a file due to its Format. Too much variability going on.

Also, a last good point mentioned earlier by member @bigshot, opens the rabbit hole of how well or truthful to the bits, is my DAC preforming. So in some cases we have to rule that variable of gear, out by staying on the same setup.
My recent experience history has shifted from the desktop, to Daps, so I feel that they are hugely responsible for sound difference, and is Totally separate from the topic here of being about quality.
I have already been convinced here about the actual thread topic, hearing various quality of same actual song on different bitrates and found half of the lower bitrate files as good or better than higher, which also leads to the issue of how it was recorded.

Two years ago I bought an optical cable, tried and compared with HDMI a few days. I put the optical in the desk drawer for 2 years as HDMI was more "detailed".

6 weeks ago, found the optical cable, gave it another shot, but left it in about 3 weeks. Went back to HDMI and compared. The HDMI was more brittle and bright sounding, not more detailed. Then I went on battery power, removed the USB charging cable.

With every reduction in noise, I perceive an increase in sound quality.

 

[gregorio]

I put the optical in the desk drawer for 2 years as HDMI was more "detailed".

Was the HDMI actually more “detailed” or was it just your imagination/perception? If it was the latter then obviously that’s a function of your brain (your imagination/perception), has nothing to do with optical vs HDMI and your assertion is therefore false. If it’s the former, how did you ensure that difference was real rather than just your imagination/perception?

6 weeks ago, found the optical cable, gave it another shot, but left it in about 3 weeks. Went back to HDMI and compared. The HDMI was more brittle and bright sounding, not more detailed.

How did you ensure that the HDMI was more brittle and bright sounding rather than just your imagination/perception?

With every reduction in noise, I perceive an increase in sound quality.

How did you determine there was an actual/real reduction in noise, rather than just an assumption or your imagination/perception?

Time and again, we see people posting assumptions or perceptions, claiming that they’re real audio/sound phenomena. This is a fallacy and it should be obvious that fallacies are not a valid basis for science or factual claims but for some bizarre reason we don’t just keep getting this but we keep getting people actually arguing on this basis. It really is weird, you’d expect it from a 5 year old who hasn’t yet been taught what science is, but not from adults who’ve completed compulsory education.

G

 

[bigshot]

Two years ago I bought an optical cable, tried and compared with HDMI a few days. I put the optical in the desk drawer for 2 years as HDMI was more "detailed". 6 weeks ago, found the optical cable, gave it another shot, but left it in about 3 weeks. Went back to HDMI and compared. The HDMI was more brittle and bright sounding, not more detailed.

Isn't it interesting how your perception changes even when the things you're comparing don't. Expectation bias is fickle. You can't depend on it.

Of course a simple blind, line level matched, direct A/B switched comparison with multiple trials would have shown you they both sound identical. Heck! Thirty seconds of intelligent googling would have told you that and you wouldn't have had to waste all the time you've wasted on sloppy comparisons that have already been proven and established.

 

[castleofargh]

I think there's a terrible logical fallacy going on with blind testing and casual listening. Intuitively, with casual listening we do nothing special, it's "natural"/"true" listening, while a blind test imposes things onto the listener, affecting the outcome.
But in reality it's the other way around of course, because the key word should be listening. Casual listening is a full audio and non audio experience, the brain does not do well keeping them separated so we're never free of biases, memory mistakes as we typically allow way more time to pass, and just setup mistakes(just because you do nothing doesn't mean there are no issues with the questions about sound you're trying to answer with your non test).
A blind test on the other hand is an effort to reduce external factors so we can try to just listen without the worries of all other things potentially affecting the sound, the test, and the human brain when it's trying to interpret the experience it will consider sound.

Have fun with what's fun, but please don't lie to yourself and everybody else pretending that slightly paying attention to sound while still being conscious of everything else going on is the one true accurate way to know about sound.

I get that it's difficult. We keep telling overconfident people not to be. If they were able to correctly estimate a situation, they wouldn't be overconfident in the first place. Real willpower is necessary to try and change that. Their own. At the risk of sounding like an AA meeting, If you wish to fool yourself less often, the first step is you deciding that you want it.