24bit vs 16bit, the myth exploded!

[nick_charles]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif . The two files vary slightly in maximum amplitude as well as with arbitrarily tested sample points, but the deviation is within 0.04 dB.

On average yes, however there is an artifact at the beginning of the 2496 files that is a give-away, I was looking at these files a couple of years ago and was able to ABX them to 13/15.

If you take the segment from 0.07s to 0.17s on both files (unaltered) and analyse them you get some massive variation viz

Beyond this point the files are pretty much identical but this blip ruins the files for ABXing. If I just gave you this data you would have to conclude that these were not the same recordings at all.

 

[JaZZ]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif On average yes, however there is an artifact at the beginning of the 2496 files that is a give-away, I was looking at these files a couple of years ago and was able to ABX them to 13/15. If you take the segment from 0.07s to 0.17s on both files (unaltered) and analyse them you get some massive variation viz... (...) Beyond this point the files are pretty much identical but this blip ruins the files for ABXing. If I just gave you this data you would have to conclude that these were not the same recordings at all.

That's odd. I haven't listened to the first few seconds, though.

Unfortunately I have already inserted some silence at the start (my standard procedure with sound editing), so I can't locate the corresponding position. However, for your purposes it would have been easy to just retouch the issue.
.

 

[nick_charles]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif However, for your purposes it would have been easy to just retouch the issue. .

Indeed, but the point is that the website says here are two samples that just show the difference between 24/96 and 16/44.1 and nothing else, same mix same recording i.e they should be identical apart from the minor differences between the formats.

In fact without the blip they are functionally near identical and I would think very hard to tell apart in that test that we can mention here. But the blip is a dead give-away and makes the comparison easy(ish).

 

[JaZZ]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif Indeed, but the point is that the website says here are two samples that just show the difference between 24/96 and 16/44.1 and nothing else, same mix same recording i.e they should be identical apart from the minor differences between the formats.

True -- but probably said flaw is not threre on purpose (I can't imagine what for). So for people without a general mistrust agaist audiophile websites, attitudes and approaches the two samples are an interesting playground for format comparisons. Personally I have waited for such an occasion -- since my multiformat player usually doesn't output hi-rez signals.
.

 

[nick_charles]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif True -- but probably said flaw is not threre on purpose (I can't imagine what for). So for people without a general mistrust agaist audiophile websites, attitudes and approaches the two samples are an interesting playground for format comparisons. Personally I have waited for such an occasion -- since my multiformat player usually doesn't output hi-rez signals. .

I am not suggesting it is deliberate, I see cockup not conspiracy, and also I work on the basis that audiophile recording websites are being sincere. I have no gripe against this site and I agree it is nice to have High res samples for folks to try out.

 

[scompton]

Linn records has test samples for downloading. I have no idea if they are the exact same sample.

Download our testfiles

 

[JaZZ]

Quote:

Originally Posted by scompton /img/forum/go_quote.gif Linn records has test samples for downloading. I have no idea if they are the exact same sample. Download our testfiles

Thanks for the Linnk!

I can't discern the original 24/192 file from a 16/44.1 downsampled version. Apparently this piano duet is not critical enough a recording for my ears.
.

 

[gregorio]

Quote:

Originally Posted by nick_charles /img/forum/go_quote.gif Beyond this point the files are pretty much identical but this blip ruins the files for ABXing.

Strange, if the blip is on the 24/96 but not the 16/44.1 this would imply that the 16bit version was not made directly from the 24/96 master, otherwise it too would have the blip. Of course, for the perposes of this thread a more accurate comparison would be 24/44.1 and 16/44.1 or 24/96 and 16/96.

If we have moved on to to mentioning 192kFs/s, there are some problems with this sample rate that if any difference can be detected it is likely to be a deterioration in SQ, rather than an improvement. Have a read of this:
http://www.lavryengineering.com/docu...ing_Theory.pdf

Unfortunately, my system is setup for a project I'm working on at the moment and I can't actually test the files which have been posted.

G

 

[JaZZ]

Quote:

Originally Posted by gregorio /img/forum/go_quote.gif If we have moved on to to mentioning 192kFs/s, there are some problems with this sample rate that if any difference can be detected it is likely to be a deterioration in SQ, rather than an improvement. Have a read of this: http://www.lavryengineering.com/docu...ing_Theory.pdf

Note that this paper is 5 years old. Moreover it's just one opinion -- stated as a fact.
.

 

[gregorio]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif Note that this paper is 5 years old. Moreover it's just one opinion -- stated as a fact. .

Actually there are a lot of people who agree with him. Also, this opinion is worth more than most as he's regarded professionally as a leading expert in the field. Some of the principles I laid out in my original post are 30 years old, doesn't make them any less true now though.

G

 

[JaZZ]

Quote:

Originally Posted by gregorio /img/forum/go_quote.gif Actually there are a lot of people who agree with him. Also, this opinion is worth more than most as he's regarded professionally as a leading expert in the field. are 30 years old, doesn't make them any less true now though.

My objection isn't addressing «some of the principles you laid out in your original post», but rather the generalization deduced from Lavry's pretension that a sampling rate of 192 kHz will forever be too high a challenge for any future ADCs or DACs -- this from a perspective from 2004. Even if he still may have some following in this matter (such as you, among others), that doesn't make the hypothesis any more reliable or plausible.
.

 

[gregorio]

Quote:

Originally Posted by JaZZ /img/forum/go_quote.gif My objection isn't addressing «some of the principles you laid out in your original post», but rather the generalization deduced from Lavry's pretension that a sampling rate of 192 kHz will forever be too high a challenge for any future ADCs or DACs -- this from a perspective from 2004. Even if he still may have some following in this matter (such as you, among others), that doesn't make the hypothesis any more reliable or plausible. .

It's not just me, many, many others have respect for one of the leading experts on the planet. I realise this counts for nothing here on head-fi though, where everyone appears to know more than those who do it for a living or indeed know more than the leading experts.

Of course, it's not just the ability to process the datastream but also the fact that there is nothing in those frequencies to capture. There has been arguments that audiophiles here on head-fi can apparently hear beyond 20kHz and are therefore different from normal human beings. Are we now going to have a discussion that head-fiers can now hear beyond 48kHz and need the frequency of digital audio to go up to 96kHz? If so you are wasting your time, no microphone in any recording studio goes anywhere near 96kHz, in fact very few of them go much beyond 20kHz, what about your speakers or cans, do they have a freq response of 96kHz? Anyone thinking there is anything that can either be captured or heard up there is completely fooling themselves.

Just to make it clear, there could (in theory) be some benefit to 96kFs/s under certain conditions. 192kFs/s is a complete waste of time, it's even a waste of time for recording, let alone for listening.

G

 

[ILikeMusic]

Quote:

Are we now going to have a discussion that head-fiers can now hear beyond 48kHz and need the frequency of digital audio to go up to 96kHz?

Only if it can be done with vacuum tubes.

 

[mark_h]

Quote:

Originally Posted by manaox2 /img/forum/go_quote.gif Your never safe from the recording engineer, especially in today's music.

I have had a really hard weekend at work, this thread has put me to sleep. Just what I needed, thanks G!

Writes in sleep...trust your ears, ignore sound science...zzzzzzz

 

[JaZZ]

Quote:

Originally Posted by gregorio /img/forum/go_quote.gif It's not just me, many, many others have respect for one of the leading experts on the planet. I realise this counts for nothing here on head-fi though, where everyone appears to know more than those who do it for a living or indeed know more than the leading experts.[/i]

That's not what I intended. Please read my post again. I do have respect for knowledgeable people. You're simply -- and arbitrarily -- misinterpreting my statement.


Quote:

Just to make it clear, there could (in theory) be some benefit to 96kFs/s under certain conditions. 192kFs/s is a complete waste of time, it's even a waste of time for recording, let alone for listening.

Maybe -- but you can't be sure about that. Not more sure than other -- knowledgeable -- people. Or people with positive experience with 192 kHz.
.