Head-Fi.org › Forums › Misc.-Category Forums › Music › 24bit/96 khz vs 16 bit/41.1 khz
New Posts  All Forums:Forum Nav:

24bit/96 khz vs 16 bit/41.1 khz

post #1 of 20
Thread Starter 

In terms of subjective sound quality, how big is the difference to you personally?

This song is in 16 bit/41.1 khz now, will be released on the album in 24 bit/96 khz, how big would the difference be?http://www.npr.org/blogs/allsongs/2013/08/01/207932250/vikings-choice-kayo-dot-sums-up-a-decade-of-sonic-transgression-in-thief

post #2 of 20
Quote:
Originally Posted by clairyvic View Post

In terms of subjective sound quality, how big is the difference to you personally?

This song is in 16 bit/41.1 khz now, will be released on the album in 24 bit/96 khz, how big would the difference be?http://www.npr.org/blogs/allsongs/2013/08/01/207932250/vikings-choice-kayo-dot-sums-up-a-decade-of-sonic-transgression-in-thief

It's impossible to say what the difference will be without more information about the production path.  Simply taking a 16/44.1 file and converting it to 24/96 doen't change anything.  The potential might be if the master was originally 24/96 through the entire production chain, then down-sampled to 16/44.1, then a 24/96 version from the original (without the down-sampling) might...might be a bit better. 

 

What we don't know is what they did to get to the 24/96 version.  If there's "re-mastering" going on, then anything can change, for better or worse (most unrelated to just the resolution of the file), but without knowing the complete path you can't know what to expect.  The differences between original 16/44.1 and 24/96 files (no up or down sampling) is actually shockingly, actually vanishingly small.  But when people go for a 24/96 release they often do something deliberately different so there will be a clearly audible difference, making the 24/96 version "worth the money" or trouble.

post #3 of 20
If the master isn't better (and it probably won't be) then the 24-bit 96kHz version will probably be audibly worse because there are lots of reasons and tricks for downsampling to 16/44.1, like dither, noise shaping, removal of ultrasonic distortion, etc.
post #4 of 20
Quote:
Originally Posted by Tus-Chan View Post

If the master isn't better (and it probably won't be) then the 24-bit 96kHz version will probably be audibly worse because there are lots of reasons and tricks for downsampling to 16/44.1, like dither, noise shaping, removal of ultrasonic distortion, etc.
None of those are "reasons or tricks for down sampling" though. If the master is 24/96, and the file remains 24/96 all the way to the DAC, it will sound just like the master, no better or worse. Downsampling is dictated by the need to release on CD, or some file with a lower bit depth and rate than the master, and is only necessary if the master is an anything higher than the target rate and depth. If down sampling is done, then yes, all that you listed need to be handled correctly or degradation could result. But degradation due to down sampling is not a given, or a fixed result. It is possible to do it well, without degradation. It's done all the time, every day, and even on the fly in some devices and software.
post #5 of 20
Quote:
Originally Posted by jaddie View Post

None of those are "reasons or tricks for down sampling" though. If the master is 24/96, and the file remains 24/96 all the way to the DAC, it will sound just like the master, no better or worse. Downsampling is dictated by the need to release on CD, or some file with a lower bit depth and rate than the master, and is only necessary if the master is an anything higher than the target rate and depth. If down sampling is done, then yes, all that you listed need to be handled correctly or degradation could result. But degradation due to down sampling is not a given, or a fixed result. It is possible to do it well, without degradation. It's done all the time, every day, and even on the fly in some devices and software.

You responded quite well to a completely different argument than the one I made, lol.

Also, this is my 1,500th post, and that means I'm right. biggrin.gif
Edited by Tus-Chan - 8/8/13 at 4:28am
post #6 of 20
Quote:
Originally Posted by Tus-Chan View Post


You responded quite well to a completely different argument than the one I made, lol.

Also, this is my 1,500th post, and that means I'm right. biggrin.gif

Well, 1500!  I'm humbled!

Quote:
Originally Posted by Tus-Chan View Post

If the master isn't better (and it probably won't be) then the 24-bit 96kHz version will probably be audibly worse because there are lots of reasons and tricks for downsampling to 16/44.1, like dither, noise shaping, removal of ultrasonic distortion, etc.

Obviously, I didn't understand your point.

 

"If the master isn't better" (Better than what? It's the master, therefore, the reference. A down-sampled or up-sampled copy? I think it would be arguably better, though my point is, perhaps not audibly.  A bit-perfect copy? Nope, not better, but exactly equal.  It seems a master would always be slightly better or equal to a copy of any kind, certainly equal to a 24/96 copy if the master was 24/96, and these days, many are.  Some 24/96 files originate from analog masters...a whole other issue, but the digital intermediate would be audibly identical to the master in that case)

 

"then the 24-bit 96kHz version will probably be audible worse" (Perhaps I didn't understand...worse than what, the master?  If that's your point, that's what I was disagreeing with.  IF that's not your point, then it blew over my head...probably at 1500 feet.)

 

"because there are lots of reasons and tricks for downsampling to 16/44.1"  (I don't see lots of reasons to down-sample.  Only one: the release formant is required to be of lower bit depth or sampling rate than the master.  Did I miss one..or two?  If by tricks you mean "how to do it right or wrong", we agree on that point.)

 

"like dither, noise shaping, removal of ultrasonic distortion, etc." (Those would be the "tricks" but not "reasons to down-sample", and you've missed a few besides, but no matter)

 

It seemed like you were making the point that a 24/96kHz version would always be worse than the original, which is not true as it depends on what the original was, and if and how down-sampling was done.  OR that at 24/96 up-sampled version would always be worse than the original.  Again, not true.  Or that down-sampling always causes audible degradation, unless the proper "tricks" are applied (which they usually are).  We agree on that, except for the implication that audibly degraded downsampling is the norm, rather than the exception.

 

If I got all of this wrong, I apologize.  I clearly have about 500 more posts to go before I'm right!  biggrin.gif

post #7 of 20

You are waaaaay overthinking this.

 

First, nowhere in my post did I use the word "always". You're (unintentionally) misrepresenting my argument by changing the context of its meaning; basically, a strawman. There are variables that can be changed and scenarios that can happen to nullify the differences, but these differences are not common or practical, which is why I used "probably" (once I excluded the most common, but technologically irrelevant, reason for sounding better).

 

The "high-res" or vinyl versions of music are quite often mastered differently than the regular .mp3 or CD version of the music. The CD gets the loudness-war crippled version, and the vinyl and high-res version will get the edition with full dynamic range. This is a benefit of "audiophile" formats that comes quite independently from the technical strength of the format; ironically, vinyl has many audible and practical weaknesses compared to digital that make it vastly inferior for full audiophile treatment, and high-res tracks (specifically ones with higher sampling rates and therefore the capability to store higher frequencies) will inherently be somewhat audibly worse because they can't take advantage of the benefits of downsampling, and if they take advantage of the "benefits" of higher sampling rates.

 

My point is, in the absence of a better master, the "low-res" 16/44.1 version will usually be audibly better than the "high-res" version of the music because downsampling from the studio 24/96 version gives you the opportunity to do quite a lot of things that improve sound quality, like dither and noise shaping; not to mention the fact that many playback systems are not inherently fully capable of reproducing 20kHz+ frequencies, which can cause theoretically audible distortion.

 

For some reason, you seem to think I'm arguing against downsampling, when I am actually arguing in favor of it. There's no good reason to have a 24/96 version of music over a 16/44.1 unless the master is better, because 24/96 is inferior in terms of how it is finished, as well as in the context of how it is reproduced on your average/above-average sound system.


Edited by Tus-Chan - 8/8/13 at 8:13am
post #8 of 20

Thanks for (finally) writing something that explains your position so that I can understand it.  I apologize for being so stupid.

 

You see, if you include some detail in the post, stupid people like me start to actually get the point through our thick skulls and then we (I) wouldn't disagree with you.

post #9 of 20
Quote:

Originally Posted by Tus-Chan View Post

 

My point is, in the absence of a better master, the "low-res" 16/44.1 version will usually be audibly better than the "high-res" version of the music because downsampling from the studio 24/96 version gives you the opportunity to do quite a lot of things that improve sound quality, like dither and noise shaping; not to mention the fact that many playback systems are not inherently fully capable of reproducing 20kHz+ frequencies, which can cause theoretically audible distortion.

 

Dither and noise shaping do not improve the sound compared to the 24-bit original. They just add noise, which should however not be audible at 16-bit resolution under normal circumstances. In theory, downsampling to 44.1 kHz could be an improvement because of avoiding the ultrasonic distortion you mentioned, but in practice it will most likely sound the same on reasonably performing equipment.

post #10 of 20
The more is better crowd will pick 24/96. If the master is crap, both rez are wasted. If the master is good, 16/44 is all that's needed. So I believe you only need 16/44.
post #11 of 20

I've been exploring this, in response to a thread on the Head-Fi group on Facebook. I remember when I started buying 24/96 albums from HDTracks.com, I was amazed at how much better they were. After downsampling a few of them to 16/44, and being unable to distinguish the two in any way (with ABXer on my iMac), I'm beginning to think that the remastering from the original tapes is responsible for the improved sounds, NOT the 24/96 digital sampling. But I haven't listened to enough different music yet to decide for sure. I'll happily save the disk space if it isn't really doing anything for me.

 

Has anyone else actually done this? Taken 24/96 recordings, down-sampled to 16/44, and been able to distinguish the two? If so, which music, DAC, amp, cans/speakers.

 

I'm using a NuForce µDAC-2, with Sennheiser HD 598 cans and Swan M200MkIII speakers. Not super high-end equipment, but pretty good.

post #12 of 20
Quote:
Originally Posted by Bill St. Clair View Post
 

I'm beginning to think that the remastering from the original tapes is responsible for the improved sounds, NOT the 24/96 digital sampling. But I haven't listened to enough different music yet to decide for sure. I'll happily save the disk space if it isn't really doing anything for me.

 

That is correct.

post #13 of 20
Quote:
Originally Posted by Bill St. Clair View Post
 

 I'm beginning to think that the remastering from the original tapes is responsible for the improved sounds, NOT the 24/96 digital sampling. But I haven't listened to enough different music yet to decide for sure. I'll happily save the disk space if it isn't really doing anything for me.

I asked the folks at HD Tracks point-blank about how their 24/96 versions came to be, if they were digitized at 24/96 from the original master analog tapes, or up-sampled from a digital intermediate, or some other way.  I was trying to get to how "authentic" the 24/96 versions are, and what care was taken in mastering them, as I don't think people are really interested in paying for up-sampled versions.  They answered me with sort of a cop-out (here it is, copy/paste from the email) "We receive the most up to date masters from our record label partners. Each record label has their own recording/mastering process. We are not a record company. We don't do anything to music provided to us by record companies. If something needs to be change or revised it is sent back to label for the process to be taken care of. We are just a retailer. We don't downsample or upsample.   Whenever possible we try to provide as much information as possible about the mastering process that is giving to us by the labels."

 

So, sometimes ya knows, sometimes ya doesn't, sometimes they is, sometimes they ain't.  You pays yer money and takes yer cherse. 

 

Last I checked, downloads weren't returnable, but perhaps there's a refund policy.  I know I've gotten refunds from iTunes when the download went poorly.  Perhaps HD Tracks would do a refund if the quality isn't any better.  Somehow I doubt it.

post #14 of 20
Quote:
Last I checked, downloads weren't returnable, but perhaps there's a refund policy.  I know I've gotten refunds from iTunes when the download went poorly.  Perhaps HD Tracks would do a refund if the quality isn't any better.  Somehow I doubt it.

 

I find HDTracks quality to be excellent, though some of the rock titles I've gotten from them don't seem markedly improved. But I haven't done ABX testing between their version and my old CD version, so that's all subjective at this point. My question, though, is whether it's even possible for 24/96 to sound better than 16/44, no matter how good a job is done. And I'm looking for someone who can give a definitive "Yes!" from personal experience with example digital source and signal path. So far I've heard of lot of statements to the effect of, "Well, of course it sounds better. It's higher bandwidth and more dynamic range. And I spent lots of money to get equipment to handle that." All well and good, but not a personal 10 out of 10 ABX test where you have proved you can distinguish the two.

post #15 of 20
Quote:
Originally Posted by Bill St. Clair View Post
 

 My question, though, is whether it's even possible for 24/96 to sound better than 16/44, no matter how good a job is done. And I'm looking for someone who can give a definitive "Yes!" from personal experience with example digital source and signal path. So far I've heard of lot of statements to the effect of, "Well, of course it sounds better. It's higher bandwidth and more dynamic range. And I spent lots of money to get equipment to handle that." All well and good, but not a personal 10 out of 10 ABX test where you have proved you can distinguish the two.

The problem is that there's really no way to do this unless you have control over the analog source, hopefully something live.  Then, you can't compare 24/96 and 16/44 without also including the converters and filters at both ends.  What you end up comparing is specific ADC and DAC technologies as a larger factor than the digital format itself. 

 

Even so, if you had theoretically equally excellent hardware for both formats, I feel quite confident that the raw differences between 24/96 and 16/44 would be completely inaudible.  The differences "heard" between released 24/96 and 16/44 files are all due to differences in the source, and production/mastering chain.  But since we can't control those differences, there may be files in either format that have positive merit.  I would never say that 24/96 guarantees better SQ every time.  But idealistically, you'd hope that people generating those files for premium sale to discerning customers would at least try for a better end result.  The reality is, it varies. 

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Music
Head-Fi.org › Forums › Misc.-Category Forums › Music › 24bit/96 khz vs 16 bit/41.1 khz