High Resolution Audio: Does It Matter?

[RayDunzl]

However human hearing has a measurable s/n of close to 130dB

Maybe.

But I think I've read something more like 60dB in eveyday use. Someone here will have that data at their fingertips.

Blast the stereo for a while. Now, how long does it take to hear that faucet drip in the other room?

I don't think it happens right away.

Ok, I'm a little off:

Wiki:

"The dynamic range of human hearing is roughly 140 dB,[8][9][not in citation given] varying with frequency,[10] from the threshold of hearing (around −9 dB SPL[10][11][12] at 3 kHz) to the threshold of pain(from 120–140 dB SPL[13][14][15]). This wide dynamic range cannot be perceived all at once, however; the tensor tympani, stapedius muscle, and outer hair cells all act as mechanical dynamic range compressors to adjust the sensitivity of the ear to different ambient levels."

 

[March Audio]

Not to mention the issue of auditory masking where louder sounds mask quieter sounds that are close in frequency. The whole premise of mp3.

 

[RayDunzl]

In a few days, I'll see if I can hear my footsteps in the grass (30dB) while I mow it (90dB)...

 

[ceedee]

It may be a flawed comparison, but I do not recall the playback of a hybrid SACD ever sounding better from the CD layer than from the DSD stereo layer on the same player with reasonably careful level matching. Yet, they are from the same master.

In my experience, it most certainly could be a flawed comparison. There are many cases of the two layers having different mastering.

Even if the mastering engineer claims they are the same, there still could be a lot of variables. Some mastering engineers (Steve Hoffman is one) admit to using some sort of analog "split feed" in making the DSD and PCM masters at the same time. He would still say they are the same, but they're going through different pieces of analog equipment. Overall, the much better comparison is to do the conversion ourselves from the high-res – either with an A/D/A loop or by downsampling and upsampling back to the original format.

Then there's the question of fast switching: how many SACD players let you switch instantaneously from the DSD to PCM layers? Some may, I'm not sure. But if it requires stopping the disc and re-starting the other layer, then that would likely exhaust any auditory memory we may have had.

 

[Fitzcaraldo215]

In my experience, it most certainly could be a flawed comparison. There are many cases of the two layers having different mastering.

Even if the mastering engineer claims they are the same, there still could be a lot of variables. Some mastering engineers (Steve Hoffman is one) admit to using some sort of analog "split feed" in making the DSD and PCM masters at the same time. He would still say they are the same, but they're going through different pieces of analog equipment. Overall, the much better comparison is to do the conversion ourselves from the high-res – either with an A/D/A loop or by downsampling and upsampling back to the original format.

Then there's the question of fast switching: how many SACD players let you switch instantaneously from the DSD to PCM layers? Some may, I'm not sure. But if it requires stopping the disc and re-starting the other layer, then that would likely exhaust any auditory memory we may have had.

I am doubting that the RBCD and stereo DSD layers of a hybrid SACD are mastered differently "in many cases", as you suggest. That would cost the producers considerable extra money to do, and for what purpose?

In your example, I also doubt there is any difference in the analog "split feed" to simultaneous DSD and PCM recordings. It should just be a set of Y cable connections. Where there is a difference is in the a-d and the subsequent digital post processing.

Agreed about fast switching on a typical single disc player. PC rips or downloads can potentially be a bit better in that regard, though still a bit tricky depending on player software.

 

[TBone]

I recall a few of my early SACDs (2001) which had obvious extra compression added to the redbook layer, even if only a song or 2, sometimes a hit
single (perfect for those tilted sacd vs cd demos). PF Money comes to mind ...

 

[amirm]

Do high resolution files sound better than say a quality red book one

The answer is YES and NO really? How can that be after all the physical (PCM data) amount of information contained in a 24/192Khz file is substantially more than a red book standard 16/44.1Khz right?

There are two questions at hand here:

1. Does one sound better than the other as you ask.

2. Does it matter?

My opinion is focused on #2. We can argue 'till cows come home about #1. But as a practical matter, that question is moot. When Meyer and Moran test was conducted, a new physical format in the case of SACD/DVD-A was being introduced, both with limiting copy protection measures. Gearing up to produce all new physical format(s) is a huge undertaking, ultimately costing hundreds of millions of dollars which consumers would have had to pay in one form or the other (most likely through more expensive discs). So whether there was technical merit was a huge deal. The reason for existence had to be established first.

Note that Record labels then were fighting piracy as their #1 enemy. So for them, the new copy protection mechanisms was very important. And higher retail prices for discs.

Today, the situation is far, far different. High-resolution is distributed through online means. There is no new physical format to be invented. And cost of storage is a fraction of the cost of the music itself in the form of your NAS or other favorite form of music storage. Online bandwidth is also plenty and in the case of US, unlimited so there are no barriers here.

Most importantly, record labels have given up on protecting the high-resolution content so we get it as nicely as the CD in that we can copy it, make it smaller, compress it, etc.

So sitting here as an audiophile, do I want the files that were mastered in stereo at 96/24 or do I want to demand that they first truncate it to 16/44.1? My answer as I mentioned in the article is simple: give me the master. I can truncate if I want, thank you very much.

And the CD is there should someone want the bargain although for how long, that remains a big question.

Technically and a question you asked, 16 bits is insufficient to represent the ear's dynamic range (which is closer to 116 db than 130). So we want 24 bit to have full transparency. That we can prove although how much it matters in practice is harder to explain. The sampling rate choice makes this harder yet again as vast majority of listeners fail the test of 320 kbps AAC versus CD! But again, if it comes as a full meal, let's take it.

Ultimately if we get the original master, we are assured to not have left anything on the table. Any other lossy conversion in bit depth and sample rate is just that: lossy. I like to have the choice to not have that lossy transform applied.

 

[Fitzcaraldo215]

I recall a few of my early SACDs (2001) which had obvious extra compression added to the redbook layer, even if only a song or 2, sometimes a hit
single (perfect for those tilted sacd vs cd demos). PF Money comes to mind ...

Possibly so. I do not have many early SACDs in my collection. There was not much going on that attracted me to hi rez in 2001. Mch hi rez, my main interest, really did not get rolling much until years after that. And, I do not collect much rock, pops, jazz. I still question how common your finding might be, especially today. At least, you are not saying hi rez sounds worse.

Of course, we are dealing with imperfect and subjective, anecdotal listening impressions from whatever biased sample subset of recordings we have chosen. Our music choices and typical engineering practices, which likely differ by genre, play a big role, I believe, in our views of whether hi rez has any useful significance to us individually. As I said before, there does not seem to be nearly as much popular music recorded, mastered and distributed in true hi rez. So, if I were a big rock fan, I might not be too interested in hi rez myself. Just sheer availability of hi rez recordings would be a huge deterrent.

But, I think anyone might find a more consistent upside with classical music natively recorded in hi rez, as opposed to remasters from analog. Again, I am not suggesting a gee whiz, black/white difference, just a small but noticeable one in careful comparison. I cannot prove it scientifically, nor am I trying to, but that is my subjective impression. That difference may also be increasingly obscured as more and more new recording is done in hi rez, even for release downrezzed to CD.

As I said elsewhere, the difference is perhaps most noticeable, to me at least, in spatial and imaging differences between RBCD and hi rez, among other subtleties. And, I think that many pop recordings would not necessarily reveal much of that. They tend often to be done in relatively dead studios, spatially panned, processed and mixed down from one or more mike per performer setups. Note that I am not putting pop music itself down or trying to be elitist about my music tastes or classical music engineering. I am just generalizing about the typical engineering that is done to create and master different genres. And, if this is and remains true, hi rez might never develop significant interest among pop listeners.

 

[ceedee]

I am doubting that the RBCD and stereo DSD layers of a hybrid SACD are mastered differently "in many cases", as you suggest. That would cost the producers considerable extra money to do, and for what purpose?

Maybe not "many" but there are cases of this. I guess the difficult thing is, how do you know for sure? Perhaps they reused an existing CD master for the RBCD layer, or gave it some extra limiting. Maybe with an old PS3 and a lot of effort, one could find out for sure.

In your example, I also doubt there is any difference in the analog "split feed" to simultaneous DSD and PCM recordings. It should just be a set of Y cable connections. Where there is a difference is in the a-d and the subsequent digital post processing.

My point is that we know there is a different signal path, including the A/D converter. Maybe there's a channel imbalance in the wiring, or slightly different gain-staging, etc. These variables shouldn't exist when doing a proper comparison. I'm not saying these things would necessarily be audible of course, but I'd expect them to be a heck of a lot more audible than the difference in formats.

 

[ceedee]

So sitting here as an audiophile, do I want the files that were mastered in stereo at 96/24 or do I want to demand that they first truncate it to 16/44.1? My answer as I mentioned in the article is simple: give me the master. I can truncate if I want, thank you very much.

I might agree with you if not for the cost difference. Based on the actual cost differences, perhaps they should charge a dollar or two more for hi-res. Yet we see the price tends to go up substantially with each "upgrade" in resolution.

I do find the higher-res files useful if I want to do my own processing before converting back to 16/44.1. But if not, I'd rather not pay a premium (in purchase price and storage space) for a difference that I can't hear. Currently, I only purchase high-res files when they have a superior tape source and/or better mastering than what's available in 16/44.1.

 

[amirm]

I might agree with you if not for the cost difference. Based on the actual cost differences, perhaps they should charge a dollar or two more for hi-res. Yet we see the price tends to go up substantially with each "upgrade" in resolution.

That's true but with some research and practice, I managed to get around substantial amount of that premium. Most high-res sites have routine weekly promotions. And then there is Pono which almost always undercuts the others (see my review of Pono music store here: http://www.audiosciencereview.com/f...pono-music-store-bargain-shoppers-rejoice.20/). The ability to download a track right then and there, and not have to rip it, has value to me over the price of the CD.

Alas, right now I have stopped almost all high-res buying given my subscription to Tidal. I now have so much music to listen to that I am just not interested in shopping one album at a time.

 

[RayDunzl]

Technically and a question you asked, 16 bits is insufficient to represent the ear's dynamic range (which is closer to 116 db than 130). So we want 24 bit to have full transparency.

16 bits is insufficient to represent the ear's dynamic range

Yes, if you arbitrarily assign a value of 6dB per bit.

Do we really do that in practice?

Don't we have a volume knob which we rotate to assign however many (even fractional) dB of in-room SPL to the so-called 96dB for the 16 bits on the disc?

If I turn the knob a little more to the right, for peaks of 116 dB in room, playing redbook, do I not get 7.25dB per bit from a 16bit source? (16 x 7.25 = 116)

If I play a 24 bit source at that same 116db peak in room, did I not select 4.83 dB per bit?

24 bit = 144dB at 6dB per bit, 16 bit = 144dB at 9dB per bit.



Straighten me out.

 

[amirm]

Well, anytime the technology makes you do something, it is not doing its job right. Since content is almost always created in 24 bits, then it makes sense to get that version and not worry about messing with the volume control all the time.

 

[Thomas savage]

That's true but with some research and practice, I managed to get around substantial amount of that premium. Most high-res sites have routine weekly promotions. And then there is Pono which almost always undercuts the others (see my review of Pono music store here: http://www.audiosciencereview.com/f...pono-music-store-bargain-shoppers-rejoice.20/). The ability to download a track right then and there, and not have to rip it, has value to me over the price of the CD.

Alas, right now I have stopped almost all high-res buying given my subscription to Tidal. I now have so much music to listen to that I am just not interested in shopping one album at a time.

How do you find the fidelity of tidal compared to say the same recording( same mastering etc ) but in 24/96 or higher from a download site?

 

[RayDunzl]

Well, anytime the technology makes you do something, it is not doing its job right. Since content is almost always created in 24 bits, then it makes sense to get that version and not worry about messing with the volume control all the time.

I am unsatisfied with your answer. 24 bit systems don't have a volume knob? Mine does.

1. There are no stairsteps.

2. 16bit is linear (low distortion) - we stretch and compress it all the time (linearly) with the volume control.

3. The littlest bit is said to be as linear at the rest of them when we have proper dither.

So. What's the deal?

 

[amirm]

How do you find the fidelity of tidal compared to say the same recording( same mastering etc ) but in 24/96 or higher from a download site?

I was going to do this test but have forgotten. Will put it on my todo list when I get a chance.

Someone also asked a related question of whether they actually stream identical copies to CD or have some other master.

 

[Thomas savage]

I was going to do this test but have forgotten. Will put it on my todo list when I get a chance.

Someone also asked a related question of whether they actually stream identical copies to CD or have some other master.

The fact it's not so dramatic that you down tools and measured it tells me it's not musically as relevant as many would have us believe.

Or your bloody lazy

To the second point I would think some times yes sometimes no, there are so many slightly different CD issues ( versions) who knows what ends up being streamed.

It would be nice if they had the master ( maybe hi res) then chopped it down using best practice to the streaming 16/44.1.

 

[Phelonious Ponk]

I drove from Charlotte to Carrboro today for the inspection of my new place there; 2.5 hours. I have iTunes automatically compress files to 256kbps when transferring to my iPhone. Being the old dog I am, there were a lot of 60s/early 70s recordings on there. They sounded fine. Then on came Edith and the Kingpin from Herbie Hancock's River: The Joni Letters. This is an amazing DDD recording. And even at 256kbps, even through my bog-standard Ford audio system, it was still amazing. Really stood out from those older recordings. It's about the recording, gentlemen. Always has been.

Tim

 

[Fitzcaraldo215]

Maybe not "many" but there are cases of this. I guess the difficult thing is, how do you know for sure? Perhaps they reused an existing CD master for the RBCD layer, or gave it some extra limiting. Maybe with an old PS3 and a lot of effort, one could find out for sure.

You will get nowhere with a PS3 rip of an SACD because you still have the PCM vs. DSD issue.

Extra limiting? Possible, I suppose. But, why would a recording or mastering engineer limit or compress the top end of the dynamic range in order to preserve the bottom end toward the noise level? It makes no rational sense. Just chopping off the least significant bits, -100 dB or more down with or without dither, makes rational sense. But, then, not everything or everybody is always rational.

And, as the opening post of this thread describes, we do not "know" anything for sure. It is perceptual. We know how RBCD and hi rez measure, but there are still many open questions as to audibility. Given the uncertainty of all that, you just have to reach your own conclusions based on listening, unless you want to conduct your own comprehensive double blind tests with a large number of subjects and samples. I do not need to do that. I have reached my own conclusions which are good enough for me, if no one else. But, if you find there is no difference between RBCD and hi rez with your own music in your own system, you will have your own opinion to which you are entitled.

My point is that we know there is a different signal path, including the A/D converter. Maybe there's a channel imbalance in the wiring, or slightly different gain-staging, etc. These variables shouldn't exist when doing a proper comparison. I'm not saying these things would necessarily be audible of course, but I'd expect them to be a heck of a lot more audible than the difference in formats.

Yes, well, a lot of things "could" have happened. Maybe this, maybe that. If you doubt one recording try another. Or, better still, try a bunch of them. The main point is there are very few opportunities available to compare hi rez DSD vs. hi Rez PCM. That debate still rages in engineering circles and among audiophiles with no "killer" result or definitive study to end all debate. Personally, I take no side in that, and I believe the consequences of one hi rez format vs. another are very small.

But, on to hi rez vs. RBCD resolution, if you cannot hear a worthwhile difference in careful comparative listening with your music, fine. Don't buy it and enjoy your music choices. I will stick with my own choices. But, the infinite number of possible of hypotheticals - maybe this, maybe that - changes nothing.

 

[fas42]

I drove from Charlotte to Carrboro today for the inspection of my new place there; 2.5 hours. I have iTunes automatically compress files to 256kbps when transferring to my iPhone. Being the old dog I am, there were a lot of 60s/early 70s recordings on there. They sounded fine. Then on came Edith and the Kingpin from Herbie Hancock's River: The Joni Letters. This is an amazing DDD recording. And even at 256kbps, even through my bog-standard Ford audio system, it was still amazing. Really stood out from those older recordings. It's about the recording, gentlemen. Always has been.

Tim

The variation on that - superb recording on very ordinary playback system - which I pursue, is very ordinary recordings on superb playback system. The end result is similar - the musical message overcomes the "hurdles" in the way, the "ordinariness" completely disappears and you get the enjoyment hit ...

Which is where the RB vs. hi res thing is silly, and pointless - when the playback "works" that's all that matters - decimal points of extra technical performance is meaningless, unless it translates into a worthwhile subjective experience.