Thread: 24/96 or 24/192

audiophile65
Given that they are both the same price, and assuming that your player is able to downsample the 24/192 version to 24/96, why not get the 24/192 version ? Even if your present DAC can only go to 24/96., a future DAC that you own may be able to take advantage of those higher resolution files.
Alex

Some of Barry Diament's 24/192 recdordings have musical content as high as 57kHz due to his use of wideband stereo microphones.We often do not know what microphones are used in various recent high resolution recordings.

Oh Bill, still playing the audiophool, I think it was Barry Diament from Soudkeeper Recordings who found that when making digital copies from the mastertape that he could not distinguish between the sound of the master tape and the digital copy until the digital copy got to 192/24.

Barry please forgive me if I misquoted, if it was not you, it was someone else in the recording industry.

Cheers from Australia.

Originally Posted by audiophile65

I'm getting ready to download my first hi-res download (album) from Linn Records -- it's a solo piano recoding with jazz accompaniment.
I would like to download the Studio Master (choice of 24/96 or 24/192, both the same price at $24.99), but I just wanted to ask if I would be OK with ordering/downloading just the 24/96 Studio Master, given the fact that the difference between 96 and 192 are very subtle.

Thoughts please.

Same price, why go for less information? Are you running out of hard-drive space?

Agree that the differences are going to be subtle at best, but isn't subtle differences a lot of what constitutes this hobby? Especially when we're talking about a true studio master.

It's a different story when you're asked a premium obviously.

Originally Posted by wgscott

Briefly, if there is no data above 48 kHz, there is no point in going above a sampling frequency of 2X48=96kHz. If there is data above 48 kHz, there may be a case for doing as Sandy suggests. The main reason for wanting a sampling frequency higher than 2X human hearing (generously 44.1 kHz) is to avoid folding artifacts. 96kHz sampling should be more than adequate for that.

In the abstract, that may or may not be true. Re frequency response, of course Nyquist works. Re avoiding "folding artifacts" and other unavoidable consequences of filtering, whether 96kHz is enough to allow sufficiently good/gentle filtering is not settled in the abstract. But in practicality, in terms of what the vast majority of DACs on the market today actually do (even many advertised as non-oversampling), the digital signal is going to be oversampled to 352.8 or 384kHz before being filtered, not 96kHz. (In other words, the industry itself long ago moved past a debate between 44.1, 96, and 192kHz, and said heck with trying to filter these resolutions adequately, we're going to 352.8/384 then converting to analog, so we can be sure we're filtering out audible artifacts reasonably well.) The oversampling is almost always done in rounds of 2x, so 44.1/48kHz signals will have to go through 3 rounds, 88.2/96kHz 2 rounds, 176.4/192kHz 1 round. The filters used for oversampling are not perfect - they can't be made perfect because the math used, called Fourier transforms, does a worse job on frequency response the better it does on time accuracy, and vice versa. So the filters are a compromise, and lower resolutions are subject to more rounds of that compromise before they get to the 352.8/384kHz resolution the DAC actually uses for conversion.

There is more to this and it can easily be proven by means of any random volume control.
It also needs a DAC (like the one I use) which doesn't alter a single bit itself. So, say a bit contrary to Jud's examples;

We can well say that a properly organized digital volume (working with sufficient decimals) is sufficient to counter rounding errors.
OK, but rounding errors will still be there. Now notice that at sample rates like 384K this *will* create patterns. Dither that properly ? it still does because no 100% random dither exists.
Compare this to a lossless digital volume control (which I use) and see how the sound improves.

Also (try to) compare this to what's all involved with up and downsampling and see that this is more devistating than this simple volume control. So what I try to say is that the volume control (not lossless) is your proof of how nothing should happen to the sound. Our brain can just detect the inconsistencies and no matter we can reason out it's 140dB down where it all happens, we perceive it. If you think this is not so, do the experiment (which most can't because of the prerequisites mentioned).

Jud, you will recall that I had some problem with Peak Extension needing to be on, because otherwise rounding errors beyond the audible band would be audible. Everybody agreed that PE needed to be on or otherwise we'd hear "wrongness". You will also recall that I worked around that so-called compiler bug by means of multiplying the whole level stuff with 1.001. Consistently. I thought I could get away with this, and at least the beyond audio band "errors" were out of the way now.
That version lasted for two days I think, because everybody (owning a NOS1) heard what I did. It's all on the forum ...

So for me one thing is (and always was) key : we are not allowed to change even the least siginificant bit, because it will be audible. And hey, this is what XP did, by standard dithering, even when no sample rate conversions where there. And another hey : this is where the so called necessity for "bit perfect" came from.
Make it inconsistent and you will hear it.

That this leads to it not being allowed to downsample any higher resolution file is ... okay, the subject.
But there's also something like older albums which were natively sampled at 16/44.1 (there was't any higher sampling ADC) while it is a sort of fact that albums from that era sound the best of all eras. To me this leads to something like : sampling at that low rate is low all right, but at least that means of creating the file/data is a consistent one. And otherwise it's always one step less of conversion.

2c
Peter

Originally Posted by PeterSt

That this leads to it not being allowed to downsample any higher resolution file is ... okay, the subject.
But there's also something like older albums which were natively sampled at 16/44.1 (there was't any higher sampling ADC) while it is a sort of fact that albums from that era sound the best of all eras. To me this leads to something like : sampling at that low rate is low all right, but at least that means of creating the file/data is a consistent one. And otherwise it's always one step less of conversion.

2c
Peter

And also there is the thing that the older albums did not have to contend with the loudness wars. But a careful well done 24/192 remaster from analog tape of an older recording might show something. As I'm writing this, I'm thinking of the fact that such a remastering is available from HDTracks, as well as MoFi SACD, as well as (in a different but also well-thought-of remastering by DCC) on CD. That is the Beach Boys' epochal Pet Sounds. I have the CD. I've been debating between the download and the SACD. But perhaps (purely in the service of helping to research the questions of hi-res vs. CD and PCM vs. DSD, of course) I'll go ahead and get both, and listen to them all.

Jud,

If you like Pet Sounds, I suggest you read this interview, with the remastering engineer. The different releases are all "from the analog master tapes"; yet are not exactly based on the same source. Read the article and you'll see what that means.

Originally Posted by firedog

Jud,

If you like Pet Sounds, I suggest you read this interview, with the remastering engineer. The different releases are all "from the analog master tapes"; yet are not exactly based on the same source. Read the article and you'll see what that means.

I had, yes. A very good article. I think I remember the SACD and HDTracks download being from the same source, though I could be mistaken.

Originally Posted by Robert Hutton

Rubbish. I have tweeters that go flat to 60kHz and have still useable output up to 100kHz. Where my ears go to is irrelevant. There is significant content up there harmonically that reinforces the harmonics of the entire frequency spectrum. Any musician who has played an acoustic instrument knows that is the case. I hear the effects of that content above 20kHz on the entire signal below 15kHz.

For the purposes of this discussion, why not get 192k if it is available? It is far more future proof than 96k.

There was a time when many believed that 16 bit at 44.1k sampling was perfect sound forever, using these same arguments.

We would not be having this discussion today if that were the casewould we?

Robert, I concur.
I agree with every single word in your post.

FWIW, there are some recording samples recorded at different sample rates you can use to determine if a higher sample rate will have a SQ benefit for you.

SR001 Dragon Boats 1644 - Download - 4shared

Originally Posted by Robert Hutton

Rubbish. I have tweeters that go flat to 60kHz and have still useable output up to 100kHz. Where my ears go to is irrelevant. There is significant content up there harmonically that reinforces the harmonics of the entire frequency spectrum. Any musician who has played an acoustic instrument knows that is the case. I hear the effects of that content above 20kHz on the entire signal below 15kHz.

For the purposes of this discussion, why not get 192k if it is available? It is far more future proof than 96k.

There was a time when many believed that 16 bit at 44.1k sampling was perfect sound forever, using these same arguments.

We would not be having this discussion today if that were the casewould we?

Robert, I totally agree with you. Then listen to differences in the SQ from different sample rate, and/or bit rate, could be also listening equipment dependent. Not only including speakers, but DAC also.

I remember the time when CD was introduced. I remained on LP until some decent and listenable CD players where available to us. BTW, I still remember the first one who impressed me (and bought): Micromega from France. It was about US$500.- Then, exotic companies like Mark Levinson, entered the market with their also exotic prices. Better chipsets, cd drive mechanism, etc., brought the change, the 'definitive' change to digital music listening.

When I have musician friends in my music room they are fascinated with hi res recordings, and now with DSD. Of course real hi res. They can recognize between different instruments built, like guitars, piano, violin, et al. Why? I asked them. Because of the harmonic content, including timbre, etc.

This thread (and others) cite 24/192 Music Downloads are Very Silly Indeed where they make he own statement as Neil Young's one, and this is big lie, like "Unfortunately, there is no point to distributing music in 24-bit/192kHz format. Its playback fidelity is slightly inferior to 16/44.1 or 16/48, and it takes up 6 times the space." Mr. Young is a defender of hi res and of course an enemy of low res. From the same article when the title says "Gentlemen, meet your ears" I would suggest a new one, "Gentlemen, meet your brain".

wgscott said "Briefly, if there is no data above 48 kHz, there is no point in going above a sampling frequency of 2X48=96kHz". I could answer you, is not all about frequency limit and/or content, but harmonic content from hi res.

Nobody oblige us to buy hi res, if someone is happy with low res, is your choice. But, I do want to have the choice of hi res, hard disk space is cheaper now than ever. Plus a bonus, more people downloading hi res, could lower the price.

My 4 cents (double than Peter St., ha, ha, ha)

Roch

+1
-Paul

Originally Posted by Robert Hutton

Rubbish. I have tweeters that go flat to 60kHz and have still useable output up to 100kHz. Where my ears go to is irrelevant. There is significant content up there harmonically that reinforces the harmonics of the entire frequency spectrum. Any musician who has played an acoustic instrument knows that is the case. I hear the effects of that content above 20kHz on the entire signal below 15kHz.

For the purposes of this discussion, why not get 192k if it is available? It is far more future proof than 96k.

There was a time when many believed that 16 bit at 44.1k sampling was perfect sound forever, using these same arguments.

We would not be having this discussion today if that were the casewould we?

Originally Posted by Stiv

Be careful that the DCC is the original mono mix. The killer MFSL SACD is a newer stereo mix that created a new master tape. The SACD/DSD is better than the stereo download by more than a bit.

Yep, the DCC is the mono mix. Believe me, there is no mistaking the mono from the very first note. It's just as simple as all the music coming from midway between the speakers, because there's no difference between right and left channels.

Originally Posted by Musicophile

It's a different story when you're asked a premium obviously.

So is 192 worth 40% more ($7) than 96? Assuming there is any noticeable difference, how good does the rig have to be to have any hope of noticing? Let's assume the master warrants 192 consideration.

I suspect that I'll never put a rig together of that quality. And my hearing isn't getting any better. But I am launching a round of "once-and-for-all" investments in my music collection (lossless rips and HD purchases). I don't want to upgrade again.

My NAS is on its way, so I need to decide. I'm thinking, spend the 40% premium for 192 on more 96.