[dalethorn]

 
Quote:

You're going to have to link me to something specific.... <snip>
 

 When someone invents the 48-hour day and I then have an abundance of time on my hands, I will happily get the link.  In the meantime, Google will get you there as fast as it gets me there.
 
This is not sarcasm, it simply means that my biggest contribution here with limited time is to give you tips - names, events, dates if possible - to help you go forward and get the answers yourself.
 
Oh, and one other thing - to help point out when you're pursuing the wrong approach to a problem.  Like going to major effort to justify 16-bit dithering when the simple solution is to quit dithering and save everyone a whole lot of time.

 

[Satellite_6]

 
Quote:

 
 When someone invents the 48-hour day and I then have an abundance of time on my hands, I will happily get the link.  In the meantime, Google will get you there as fast as it gets me there.
 
This is not sarcasm, it simply means that my biggest contribution here with limited time is to give you tips - names, events, dates if possible - to help you go forward and get the answers yourself.
 
Oh, and one other thing - to help point out when you're pursuing the wrong approach to a problem.  Like going to major effort to justify 16-bit dithering when the simple solution is to quit dithering and save everyone a whole lot of time.

And waste a lot of memory? It doesn't seem like the wrong approach. Your approach seems like the wrong one. To say that there is a great difference without any solid evidence is like saying someone is guilty until proven innocent. I think I understand were you are coming from - better to be safe than sorry - but your argument just seems unreasonable.
 
I'll leave it at that.

 

[dalethorn]

 

And waste a lot of memory? It doesn't seem like the wrong approach. Your approach seems like the wrong one. To say that there is a great difference without any solid evidence is like saying someone is guilty until proven innocent. I think I understand were you are coming from - better to be safe than sorry - but your argument just seems unreasonable.
 
I'll leave it at that.

What about all those comments about measurement? Surely you could measure the difference between 16 and 24-bit files?

 

[danroche]

Quote:

Quote:

 

And waste a lot of memory? It doesn't seem like the wrong approach. Your approach seems like the wrong one. To say that there is a great difference without any solid evidence is like saying someone is guilty until proven innocent. I think I understand were you are coming from - better to be safe than sorry - but your argument just seems unreasonable.
 
I'll leave it at that.

What about all those comments about measurement? Surely you could measure the difference between 16 and 24-bit files?

I'm going to be sorry for even jumping in to this thread, but I believe it's been said several times here that this isn't about mathematically testing the difference between 16 and 24 bit files, it's about detecting any difference whatsoever with our ears. You seem to be stuck on the idea that ANY difference in the files will translate to sonic differences, though an abundance of solid research IN ADDITION TO a common-sense understanding of the concept of bit depth would suggest otherwise.
 
If you take a 24 bit file, copy it, dither (or don't dither, heck) it down to 16 bits, upsample that file back up to 24 bits (basically adding 8 zeroes onto the end of each word), invert it, and sum it with the original 24 bit file, you will hear.... wait for it.... nothing.  Those last eight bits will have been hard at work capturing hiss and random interference from every stage of the recording chain. There will be no musical information there.
 

 

[Cataphract]

Quote:

On the other hand I'm not going to say 500PS of jitter is audible, because there's nothing out there to suggest it is (and in fact there's evidence contradictory to such a claim).

500 petaseconds of jitter is not audible?

 

[haloxt]

Quote:

 

I would agree to participate in a test, like a poll watcher, just to observe the conditions.  It's kind of amusing to recall many of those tests, and summaries of them afterward, noting all of the imperfections and compromises in the equipment that may have skewed the results.
 
Some people here have already noted that very noticeable differences have been proven with loudspeaker cables where the cables differed from very thin to heavy-duty.  Then they say that with small wires, like in headphones, such differences in audibility have not been proven.  Which means that we're talking about audibility thresholds now, not absolutes.  And when it comes to audibility thresholds, things change.  The equipment changes, and new generations of audiophiles learn to spot differences that previous generations either didn't hear, disregarded as irrelevant, or just hadn't developed the experience basis to identify new or newly-discovered distortions.
 
As to false positives, I recently downloaded as many test tones as I could find on the Web, and ran through them with my headphones listed below.  What was amazing was the quality of those tones, all supposedly sinewave (ignoring pink and white noise and other such things).  They sounded quite different, and I could clearly hear a 12 khz tone with one sample and not another, regardless of volume.  So things can get very complex in testing, and compromises have to be made, which almost always invalidates the tests.
 

You bring up a bunch of points I don't disagree with, but my question was to try to get you to realize that to "prove something beyond a doubt", or to really reduce that doubt, is the name of the game for the skeptics. You could provide a million examples such as the electrical differences between thin and heavy gauge speaker wire, but until a DBT has been done and published in Nature scientific magazine, then your claims of proof are not, or at least not yet, scientifically accepted.
 
Quote:

/img/forum/go_quote.gif


And waste a lot of memory? It doesn't seem like the wrong approach. Your approach seems like the wrong one. To say that there is a great difference without any solid evidence is like saying someone is guilty until proven innocent. I think I understand were you are coming from - better to be safe than sorry - but your argument just seems unreasonable.
 
I'll leave it at that.

I think a much more accurate analogy would be that someone owns a sharp knife, therefore he killed those people who are presumed dead.

 

[dalethorn]

 

And waste a lot of memory? It doesn't seem like the wrong approach. Your approach seems like the wrong one. To say that there is a great difference without any solid evidence is like saying someone is guilty until proven innocent. I think I understand were you are coming from - better to be safe than sorry - but your argument just seems unreasonable.
 
I'll leave it at that.

"Innocent until proven guilty" is a great logic test, and relevant to this discussion. IUPG applies to individuals acting purely on their own, but when we grow up and become more sophisticated, we understand that people also try to apply that to their official position rather than just to them as an individual. Which goes to the authority thing and tesing of psychoacoustics. Ideally the blind tests described here would work irrefutably, but they never do. Too complicated. And the people who have the resources and time to conduct those tests with the best of all possible equipment invariably have an agenda, be it their publishers or equipment manufacturers. So I would suggest that anyone interested collect all of the points made here and distill them down to a page or two of text, and save them for future reference.

 

[dalethorn]

I'm going to be sorry for even jumping in to this thread, but I believe it's been said several times here that this isn't about mathematically testing the difference between 16 and 24 bit files, it's about detecting any difference whatsoever with our ears. You seem to be stuck on the idea that ANY difference in the files will translate to sonic differences, though an abundance of solid research IN ADDITION TO a common-sense understanding of the concept of bit depth would suggest otherwise.
 
If you take a 24 bit file, copy it, dither (or don't dither, heck) it down to 16 bits, upsample that file back up to 24 bits (basically adding 8 zeroes onto the end of each word), invert it, and sum it with the original 24 bit file, you will hear.... wait for it.... nothing.  Those last eight bits will have been hard at work capturing hiss and random interference from every stage of the recording chain. There will be no musical information there.
 

As a purely academic argument, we could pick those nits until the cows come home. Fun for some.... But I got the sense of the discussion early on and moved to the relevant points. Using the lessons of history, the only *reasonable* solution is to stop dithering and just put out the 24-bit file. Those who don't know the history are repeating the mistakes of history. You cannot determine audibility in controlled tests, unless I (to paraphrase the former President) control the tests. My equipment, my selection of people, etc.

That may sound silly to you, but believe me, that's what happens in real-life tests, which are expensive and time-consuming (with the best equipment), and require sponsorship.

 

[dalethorn]

 
Quote:

You bring up a bunch of points I don't disagree with, but my question was to try to get you to realize that to "prove something beyond a doubt", or to really reduce that doubt, is the name of the game for the skeptics. You could provide a million examples such as the electrical differences between thin and heavy gauge speaker wire, but until a DBT has been done and published in Nature scientific magazine, then your claims of proof are not, or at least not yet, scientifically accepted.

 
The problem with magazines is they can lie, and very badly.  JAMA has done it, all major newspapers have, and don't even ask about Popular Mechanics.  The real point is, all of these tests have been conducted for decades, to little avail.  I would say that audio manufacturers can benefit, and use the measurement tech for developing better products in the future.  But the tests prove little or nothing.  Who on Earth is free of agendas, has the best audio and test equipment, and time and unbiased people to conduct those tests?

 

[rroseperry]

The first few pages of the 24-bit myth thread from 2009 should really be referenced here.  The limit to hearing a difference is not the equipment, it's the nature of digital sampling. The short take-home from gregorio's first post:
 

Hopefully you're still with me, because we can now go on to precisely what happens with bit depth. Going back to the above, when we add a 'bit' of data we double the number of values available and therefore halve the number of quantisation errors. If we halve the number of quantisation errors, the result (after dithering) is a perfect waveform with halve the amount of noise. To phrase this using audio terminology, each extra bit of data moves the noise floor down by 6dB (half). We can turn this around and say that each bit of data provides 6dB of dynamic range (*4). Therefore 16bit x 6db = 96dB. This 96dB figure defines the dynamic range of CD. (24bit x 6dB = 144dB).

So, 24bit does add more 'resolution' compared to 16bit but this added resolution doesn't mean higher quality, it just means we can encode a larger dynamic range. This is the misunderstanding made by many. There are no extra magical properties, nothing which the science does not understand or cannot measure. The only difference between 16bit and 24bit is 48dB of dynamic range (8bits x 6dB = 48dB) and nothing else. This is not a question for interpretation or opinion, it is the provable, undisputed logical mathematics which underpins the very existence of digital audio.

So, can you actually hear any benefits of the larger (48dB) dynamic range offered by 24bit? Unfortunately, no you can't. The entire dynamic range of some types of music is sometimes less than 12dB. The recordings with the largest dynamic range tend to be symphony orchestra recordings but even these virtually never have a dynamic range greater than about 60dB. All of these are well inside the 96dB range of the humble CD. What is more, modern dithering techniques (see 3 below), perceptually enhance the dynamic range of CD by moving the quantisation noise out of the frequency band where our hearing is most sensitive. This gives a percievable dynamic range for CD up to 120dB (150dB in certain frequency bands).

 

[dalethorn]

The first few pages of the 24-bit myth thread from 2009 should really be referenced here.  The limit to hearing a difference is not the equipment, it's the nature of digital sampling. The short take-home from gregorio's first post:
 

Hopefully you're still with me, because we can now go on to precisely what happens with bit depth. Going back to the above, when we add a 'bit' of data we double the number of values available and therefore halve the number of quantisation errors. If we halve the number of quantisation errors, the result (after dithering) is a perfect waveform with halve the amount of noise. To phrase this using audio terminology, each extra bit of data moves the noise floor down by 6dB (half). We can turn this around and say that each bit of data provides 6dB of dynamic range (*4). Therefore 16bit x 6db = 96dB. This 96dB figure defines the dynamic range of CD. (24bit x 6dB = 144dB).

So, 24bit does add more 'resolution' compared to 16bit but this added resolution doesn't mean higher quality, it just means we can encode a larger dynamic range. This is the misunderstanding made by many. There are no extra magical properties, nothing which the science does not understand or cannot measure. The only difference between 16bit and 24bit is 48dB of dynamic range (8bits x 6dB = 48dB) and nothing else. This is not a question for interpretation or opinion, it is the provable, undisputed logical mathematics which underpins the very existence of digital audio.

So, can you actually hear any benefits of the larger (48dB) dynamic range offered by 24bit? Unfortunately, no you can't. The entire dynamic range of some types of music is sometimes less than 12dB. The recordings with the largest dynamic range tend to be symphony orchestra recordings but even these virtually never have a dynamic range greater than about 60dB. All of these are well inside the 96dB range of the humble CD. What is more, modern dithering techniques (see 3 below), perceptually enhance the dynamic range of CD by moving the quantisation noise out of the frequency band where our hearing is most sensitive. This gives a percievable dynamic range for CD up to 120dB (150dB in certain frequency bands).

The logic is still backwards here. We do in fact encode 24-bit now, and so the proper question is do we need to decode that to 16 bits, based on the theory of audibility? Stereophile, who has a 50 year solid reputation, says no - use the better sampling we already have. Since I am an audiophile and trust Stereophile, along with their extensive history and reputation, I go with them.

 

[haloxt]

Quote:

 
 
The problem with magazines is they can lie, and very badly.  JAMA has done it, all major newspapers have, and don't even ask about Popular Mechanics.  The real point is, all of these tests have been conducted for decades, to little avail.  I would say that audio manufacturers can benefit, and use the measurement tech for developing better products in the future.  But the tests prove little or nothing.  Who on Earth is free of agendas, has the best audio and test equipment, and time and unbiased people to conduct those tests?

That's like saying because politics is corrupt then anarchy would be an alternative method of governance. I think there's a lot of bs in scientific journals, but because of that doesn't mean we throw out DBT. I don't think it can be denied that DBT has the benefit of greatly reducing false positives when implemented properly, something I think is invaluable when we're treading an area as delicate as the limits of audibility. You can come up with a million examples of why experiments done to date can be flawed, wouldn't you agree that some of these mistakes might also occur in experiments claiming to prove differences?

 

[dalethorn]

That's like saying because politics is corrupt then anarchy would be an alternative method of governance. I think there's a lot of bs in scientific journals, but because of that doesn't mean we throw out DBT. I don't think it can be denied that DBT has the benefit of greatly reducing false positives when implemented properly, something I think is invaluable when we're treading an area as delicate as the limits of audibility. You can come up with a million examples of why experiments done to date can be flawed, wouldn't you agree that some of these mistakes might also occur in experiments claiming to prove differences?

I'm not asking for Audio Anarchy. Exactly the opposite. Instead of having all these self appointed "experts" reducing my signal, I want the original signal.

And we are treading a delicate area. The ear and brain are still better than the test gear, but unfortunately, the only valid way to test is in the large arena - millions of gear users and their experiences, digested over time by the community at large.

 

[Cataphract]

 
Quote:

Since I am an audiophile and trust Stereophile, along with their extensive history and reputation, I go with them.

But that is not enough in the way of proof. I hate taking people on their word; I'd much rather take them on their evidence. And the valid evidence to demonstrate an audible difference is DBT.

 

[rroseperry]

Quote:

Quote:

The first few pages of the 24-bit myth thread from 2009 should really be referenced here.  The limit to hearing a difference is not the equipment, it's the nature of digital sampling. The short take-home from gregorio's first post:
 

Hopefully you're still with me, because we can now go on to precisely what happens with bit depth. Going back to the above, when we add a 'bit' of data we double the number of values available and therefore halve the number of quantisation errors. If we halve the number of quantisation errors, the result (after dithering) is a perfect waveform with halve the amount of noise. To phrase this using audio terminology, each extra bit of data moves the noise floor down by 6dB (half). We can turn this around and say that each bit of data provides 6dB of dynamic range (*4). Therefore 16bit x 6db = 96dB. This 96dB figure defines the dynamic range of CD. (24bit x 6dB = 144dB).

So, 24bit does add more 'resolution' compared to 16bit but this added resolution doesn't mean higher quality, it just means we can encode a larger dynamic range. This is the misunderstanding made by many. There are no extra magical properties, nothing which the science does not understand or cannot measure. The only difference between 16bit and 24bit is 48dB of dynamic range (8bits x 6dB = 48dB) and nothing else. This is not a question for interpretation or opinion, it is the provable, undisputed logical mathematics which underpins the very existence of digital audio.

So, can you actually hear any benefits of the larger (48dB) dynamic range offered by 24bit? Unfortunately, no you can't. The entire dynamic range of some types of music is sometimes less than 12dB. The recordings with the largest dynamic range tend to be symphony orchestra recordings but even these virtually never have a dynamic range greater than about 60dB. All of these are well inside the 96dB range of the humble CD. What is more, modern dithering techniques (see 3 below), perceptually enhance the dynamic range of CD by moving the quantisation noise out of the frequency band where our hearing is most sensitive. This gives a percievable dynamic range for CD up to 120dB (150dB in certain frequency bands).

The logic is still backwards here. We do in fact encode 24-bit now, and so the proper question is do we need to decode that to 16 bits, based on the theory of audibility? Stereophile, who has a 50 year solid reputation, says no - use the better sampling we already have. Since I am an audiophile and trust Stereophile, along with their extensive history and reputation, I go with them.

It's not a theory.  If you read this, you'd see that the additional bits do not add resolution, but dynamic range that you can't actually use, unless you're planning death by sound.