High Resolution Audio: Does It Matter?

[ceedee]

The shell game is thriving in all sectors, as long as people assume that 'this sounds better to me...because it's hi rez!' As the industry and its cheerleaders have always encouraged them to do. Including the enablers who say 'listen and decide for yourself', as if that was a valid way to decide whether hi rez 'matters'.

Well said. Even as Meyer and Moran pointed out, the hi-res releases often sounded very good in their tests, but that was because of more care in the recording and mastering process. *That* is where I think the focus should be: allowing more dynamics, transferring analog recordings from the correct tapes, etc.

The industry would prefer to sell increasingly larger files at larger and larger price$. Those files may contain good masters, or they may be the same compressed masters used on CD. If they focused on better mastering, it would likely yield an improvement that probably all (who care about such things) could hear.

 

[Phelonious Ponk]

Well said. Even as Meyer and Moran pointed out, the hi-res releases often sounded very good in their tests, but that was because of more care in the recording and mastering process. *That* is where I think the focus should be: allowing more dynamics, transferring analog recordings from the correct tapes, etc.

The industry would prefer to sell increasingly larger files at larger and larger price$. Those files may contain good masters, or they may be the same compressed masters used on CD. If they focused on better mastering, it would likely yield an improvement that probably all (who care about such things) could hear.

This.

 

[Fitzcaraldo215]

Oh sure you can tell if they're from the same master. Why wouldn't you be able to? But that's not the point; comparing SACD layers is invalid … way too many possible variables. As we covered previously.

Oh and Krabapple was being sarcastic in the FLAC vs. WAV thread … obviously. He doesn't claim to hear differences.

We keep tripping over this "same master" issue. It is important, so let me explain.

You make a raw, multitrack music recording in, say, 88k 24 bit resolution. The raw recording is mixed down for stereo and processed - EQ, dynamic compression, added reverb, etc. - into a final stereo master at 88k/24. From that stereo master, a BD-A, say, production master is made at 88k/24 for our example. Also from that same final hi rez stereo master, a CD production master is made via downconversion at 44k 16 bit, possibly using dither of the least significant bits.

I hand those two discs to you and I say compare them with no additional information about the mastering process. There is no way to compare the two digitally to ensure they used the same final master prior to production because of the different sampling rates and bit depths. The bits on the two discs are entirely different, even if it is the same songs from the same recording session. The only way to know that the two used the same final stereo master is to watch the mastering process yourself or take the word of the engineers who mastered the recording and prepared it for final production.

If you listen to both versions and you hear a difference, is that just because one is hi rez and the other one is not? Or, is it possibly because there was not a single final stereo master because the multitrack to stereo mix, the EQ, the dynamic compression, etc. were different between two separate stereo masters - one for the hi rez, another for the CD? If you do not know the provenance of the recording, you have no way of knowing. And, you yourself brought up the possibility that there might be separate stereo masters for the hi rez and the CD some time ago.

In many of the studies of hi rez vs. CD, this problem was avoided by using hi rez music tracks downconverted to RBCD resolution on-the-fly to assure that there was no possibility of multiple stereo masters in the comparison. In other cases, specially prepared versions at different resolutions from the same hi rez master were made prior to the tests by the test administrators to avoid a possible problem.

As I said before, it is unlikely that hi rez and CD mastered and produced at the same time from a hi rez recording session used different stereo masters. Such is also often the case with hybrid SACDs, which contain a CD layer, but it might not always be true, especially with remastered for hi rez re-releases of older CDs. They might have used the original CD master for the CD layer and a remastered version for the hi rez layer. Again, though, trying to hear a difference between hi rez and RBCD is pointless with material originally recorded in analog or RBCD resolution.

Incidentally, Cookie Marenco at Blue Coast normally releases downloads simultaneously at RBCD, FLAC96, WAV96 and DSD formats/resolutions. If you wish to do your own comparisons, this might be a resource, because it is highly likely all came from the same hi rez stereo master. But, if in doubt, just ask hem.

 

[amirm]

Incidentally, Cookie Marenco at Blue Coast normally releases downloads simultaneously at RBCD, FLAC96, WAV96 and DSD formats/resolutions. If you wish to do your own comparisons, this might be a resource, because it is highly likely all came from the same hi rez stereo master. But, if in doubt, just ask hem.

Adding on to this, I bought a PCM version of their tracks and later Cookie was kind enough to give me high rate DSD version. I played the latter with on the fly DSD to PCM conversion and compared it to PCM version I downloaded. At first, the DSD to PCM converted version sounded a lot better. I then realized the volume was higher and Roon was applying a default amount of amplification to it (6 db?). I dialed that down a couple of db to match the levels and all the difference went away.

 

[ceedee]

I hand those two discs to you and I say compare them with no additional information about the mastering process. There is no way to compare the two digitally to ensure they used the same final master prior to production because of the different sampling rates and bit depths. The bits on the two discs are entirely different, even if it is the same songs from the same recording session. The only way to know that the two used the same final stereo master is to watch the mastering process yourself or take the word of the engineers who mastered the recording and prepared it for final production.

I'm not sure why you think this, but I've done these comparisons many times. You can definitely take two files (or discs in your example) and find out if they are from the same digital master. The few high-res releases I own are different from any available CDs, so I had to make my own 16/44 version of a track from the 24/96 2012 remaster of Ram by Paul McCartney. (Yes, it's from analog tape, but the process is the same either way.) Here is the frequency content of the 24/96 file from HDTracks:

I have quite a bit of audio editing and production software, but for the purposes of this discussion, I did everything with freeware. I used the SoX plugin in Foobar2000 to go from 24/96 to 16/44 -- without dither for simplicity's sake. I opened both the 24/96 original and the "CD" master in Audacity, inverting the polarity of one. I exported the mix of these two tracks. The resulting file contains information outside of the audible range, which is what we would expect if the two files are from the same digital master. Everything from 0-20kHz (down to about -120 dBFS) is missing because it is comprised of exactly the same bits as our "CD" master.

In this case, even though the meters move when playing the "difference" file, I can't hear anything. This stuff is fun to do in a geeky way sometimes, but the best way to make a valid comparison between formats is to do the conversion yourself so that you are sure the only difference is the bit depth/sample rate.

If I do this comparison with the 2012 remaster released on CD, I don't get any kind of nulling. Extra processing was done to the CD version (slight limiting) which prevents a null. If reverb is added to one master and not the other, then you'll get a difference file that contains only reverb.

 

[krabapple]

Well said. Even as Meyer and Moran pointed out, the hi-res releases often sounded very good in their tests, but that was because of more care in the recording and mastering process. *That* is where I think the focus should be: allowing more dynamics, transferring analog recordings from the correct tapes, etc.

The industry would prefer to sell increasingly larger files at larger and larger price$. Those files may contain good masters, or they may be the same compressed masters used on CD. If they focused on better mastering, it would likely yield an improvement that probably all (who care about such things) could hear.

It's a quite contemptible diversion of resources. There are 4 obvious failure points in audio production and playback that militate against 'high fidelity' : recording/production, mastering, loudspeakers/room interactions, number of delivery channels. Different choices in these have huge qualitative and quantitative effects. Yet relentless Bob Stuart, the industry he represents , and possibly now the AES, choose to fuss and fret over 'high resolution audio' instead.

Who stands to gain most materially?

 

[krabapple]

As I said before, it is unlikely that hi rez and CD mastered and produced at the same time from a hi rez recording session used different stereo masters. Such is also often the case with hybrid SACDs, which contain a CD layer, but it might not always be true, especially with remastered for hi rez re-releases of older CDs. They might have used the original CD master for the CD layer and a remastered version for the hi rez layer.

It may be 'unlikely', but it has happened, as even Stereophile magazine noted in its 'expose' of Dark Side of the Moon. Different mastering on DSD vs CD layers (NB: they weren't the first to notice it, and the CD layer mastering was, AFAIR, specific to the SACD, not a retread of a previous release).

Again, though, trying to hear a difference between hi rez and RBCD is pointless with material originally recorded in analog or RBCD resolution.

That wasn't always the party line. Plenty of listeners tried, succeeded wildly, and proclaimed it in reviews as proof of the superiority of hi rez. Still do, I'm sure.

Incidentally, Cookie Marenco at Blue Coast normally releases downloads simultaneously at RBCD, FLAC96, WAV96 and DSD formats/resolutions. If you wish to do your own comparisons, this might be a resource, because it is highly likely all came from the same hi rez stereo master. But, if in doubt, just ask hem.

Cookie Marenco doesn't know what she's talking about whenever she talks about digital. So I would hardly trust her to 'get it right'.

 

[ceedee]

(NB: they weren't the first to notice it, and the CD mastering was, AFAIR, new, not a retread of a previous release).

No it wasn't a re-used CD mastering. It was specifically done for the SACD, for some unknown reason.

That wasn't always the party line. Plenty of listeners tried, succeeded wildly, and proclaimed it in reviews as proof of the superiority of hi rez. Still do, I'm sure.

Of course they do! Probably 95% or more of the "night and day" differences I've seen people talk about involved remasters from analog tape. Nobody tell Neil Young!

 

[Phelonious Ponk]

We keep tripping over this "same master" issue. It is important, so let me explain.

You make a raw, multitrack music recording in, say, 88k 24 bit resolution. The raw recording is mixed down for stereo and processed - EQ, dynamic compression, added reverb, etc. - into a final stereo master at 88k/24. From that stereo master, a BD-A, say, production master is made at 88k/24 for our example. Also from that same final hi rez stereo master, a CD production master is made via downconversion at 44k 16 bit, possibly using dither of the least significant bits.

I hand those two discs to you and I say compare them with no additional information about the mastering process. There is no way to compare the two digitally to ensure they used the same final master prior to production because of the different sampling rates and bit depths. The bits on the two discs are entirely different, even if it is the same songs from the same recording session. The only way to know that the two used the same final stereo master is to watch the mastering process yourself or take the word of the engineers who mastered the recording and prepared it for final production.

If you listen to both versions and you hear a difference, is that just because one is hi rez and the other one is not? Or, is it possibly because there was not a single final stereo master because the multitrack to stereo mix, the EQ, the dynamic compression, etc. were different between two separate stereo masters - one for the hi rez, another for the CD? If you do not know the provenance of the recording, you have no way of knowing. And, you yourself brought up the possibility that there might be separate stereo masters for the hi rez and the CD some time ago.

In many of the studies of hi rez vs. CD, this problem was avoided by using hi rez music tracks downconverted to RBCD resolution on-the-fly to assure that there was no possibility of multiple stereo masters in the comparison. In other cases, specially prepared versions at different resolutions from the same hi rez master were made prior to the tests by the test administrators to avoid a possible problem.

As I said before, it is unlikely that hi rez and CD mastered and produced at the same time from a hi rez recording session used different stereo masters. Such is also often the case with hybrid SACDs, which contain a CD layer, but it might not always be true, especially with remastered for hi rez re-releases of older CDs. They might have used the original CD master for the CD layer and a remastered version for the hi rez layer. Again, though, trying to hear a difference between hi rez and RBCD is pointless with material originally recorded in analog or RBCD resolution.

Incidentally, Cookie Marenco at Blue Coast normally releases downloads simultaneously at RBCD, FLAC96, WAV96 and DSD formats/resolutions. If you wish to do your own comparisons, this might be a resource, because it is highly likely all came from the same hi rez stereo master. But, if in doubt, just ask hem.

It's not unlikely that the difference will be obvious, because the process you describe above, skips the mastering process altogether. After the straight master of the stereo mix is made, a mastering engineer takes that stereo master, adds EQ and, increasingly, compression, to create the master from which the CDs or final file will be generated. As the "loudness wars" have heated up, more compression has been added at the mastering stage and often the differences between remasters and original, more gently mastered versions can be very obvious.

Tim

 

[ceedee]

Indeed. At least with SACD, differences in mastering would likely be attributed to the supposed benefits of hi-res. The only way to know they are from the same master would be to do a PS3 rip or record the analog outputs. Just looking at the waveform can easily tell you if there was extra compression or limiting added, as Stereophile demonstrated with the DSOTM SACD.

Unfortunately I don't think many have the inclination to do this!

 

[TBone]

Indeed. At least with SACD, differences in mastering would likely be attributed to the supposed benefits of hi-res. The only way to know they are from the same master would be to do a PS3 rip or record the analog outputs. Just looking at the waveform can easily tell you if there was extra compression or limiting added, as Stereophile demonstrated with the DSOTM SACD.

DSOTM sacd made the top 10 ... http://audiophilereview.com/sacddvd...phile-sacds-ever---many-are-out-of-print.html

Kinda wish I'd kept that Keb' Mo' pressing. Still own the DSOTM SACD ... very much remember it being used at shows to demonstrate SACD over CD. While my SACD collection is small, many don't have CD layers. The only reason I've kept 'em this long is because some offer excellent dynamic values. However, I've no sense of urgency to rip 'em, b/c it's often much easier/cheaper finding dynamically similar CD versions.

 

[ceedee]

DSOTM sacd made the top 10 ... http://audiophilereview.com/sacddvd...phile-sacds-ever---many-are-out-of-print.html

Interesting list. Forgot about the Norah Jones album "Come Fly With Me"... it's a good one!

Still own the DSOTM SACD ... very much remember it being used at shows to demonstrate SACD over CD.

Yes it definitely was. It's probably the #1 title that would come to my mind as an SACD demonstration disc that audiophiles were using back in the day. I was surprised to see the same crowd attempting to debunk Meyer & Moran's study because they used it (and other analog recordings) too.

I noticed from that "10 Best Audiophile SACDs Ever" list, most are from analog tape. AFAIK only two are digital recordings (Peter Gabriel and Keb Mo'), but who knows what the file formats were used originally.

 

[TBone]

Interesting list. Forgot about the Norah Jones album "Come Fly With Me"... it's a good one!

Wasted my money on the LP, so warped, it had its own gravitational pull.

Yes it definitely was. It's probably the #1 title that would come to my mind as an SACD demonstration disc that audiophiles were using back in the day. I was surprised to see the same crowd attempting to debunk Meyer & Moran's study because they used it (and other analog recordings) too.

Can't claim I was an early adopter of the format, still very much remember the intro, Sony suddenly became the audiophile medias darlings, w/their behemoth Sony SACD1/777 players. Heaviest 45 pounds I've ever had to lug, the 777 didn't strike the same audiophile funny bone, not only in my system, but quite a few others. I also figured the DSOTM demo tilted, the CD layer easily inferior to many an early CD.

I noticed from that "10 Best Audiophile SACDs Ever" list, most are from analog tape. AFAIK only two are digital recordings (Peter Gabriel and Keb Mo'), but who knows what the file formats were used originally.

I vaguely remember reading Keb recording something in native dsd?

Also think I still have a R2R to DSD copy of LZ2 R2R RL version, which I'm quite keen to rip/hear.

 

[TBone]

It's probably the #1 title that would come to my mind as an SACD demonstration disc that audiophiles were using back in the day. I was surprised to see the same crowd ...

Perhaps my fav demo, back in the day, ~mid 80's shows: some vendors were using "Audiophile" LPs which were orig rec@16bit. While they did sound demonstrably fine, few in the crowd (vinyl dominated era) would consider any advantage of its digital provenance; the LP sold like hotcakes.

I purchased the orig CD & LP; for good measure.

 

[TBone]

Song: Money (heard often in those demo's).

SACD layer ~dr10, and its CD layer (dr8) in comparison ...

SACD CD layer Can.CD Rem UK CD Harvest

DR l/r: 7.80/8.11 10.26/9.71 12.51/12.19

The rather ordinary sounding Can.remaster offers > dynamic contrast to the CD layer; and even superior SACD layer had < dynamic value compared to the orig UK.

No wonder those titled demo's convinced ...

 

[ceedee]

While they did sound demonstrably fine, few in the crowd (vinyl dominated era) would consider any advantage of its digital provenance; the LP sold like hotcakes.

There are several audiophile LPs from the early digital era that proudly touted their digital mastering. Many do sound good; I have a rip of a Nautilus pressing of Rumours which was digitally-mastered and sounds nice.

Digital was good before it was bad. As Monty said in the xiph.org Digital Show and Tell video, "That's why seeing 'DDD' on a compact disc used to be such a big high-end deal."

 

[TBone]

There are several audiophile LPs from the early digital era that proudly touted their digital mastering. Many do sound good

Yes, one could make the case; early digital recordings (well prior to compression wars) benefited their LP counterpart, consider Famous Blue Raincoat (87), Tracy Chapman (88), Trinity Sessions (88), to name a few back in the day, demo'd on LP.

I have a rip of a Nautilus pressing of Rumours which was digitally-mastered and sounds nice.

Curious, does it include the reverb mix?

Digital was good before it was bad.

Yes, it was, but ...

 

[ceedee]

Curious, does it include the reverb mix?

No, it doesn't. I just looked at the files, and it appears to have all of the original versions except for Gold Dust Woman, which has the "glass break" ending with the much quieter howling. That version had to be easier to cut to disk than the original.

If they did cut from a digital copy of the 1st generation masters, that probably explains why it sounds so good. The tapes were only three years old at the time!

 

[TBone]

I've seen that Nautilus pressing for sale, local, used, never cheap.

 

[krabapple]

Indeed. At least with SACD, differences in mastering would likely be attributed to the supposed benefits of hi-res. The only way to know they are from the same master would be to do a PS3 rip or record the analog outputs. Just looking at the waveform can easily tell you if there was extra compression or limiting added, as Stereophile demonstrated with the DSOTM SACD.

Unfortunately I don't think many have the inclination to do this!

Song: Money (heard often in those demo's).

SACD layer ~dr10, and its CD layer (dr8) in comparison ...

SACD CD layer Can.CD Rem UK CD Harvest

DR l/r: 7.80/8.11 10.26/9.71 12.51/12.19

The rather ordinary sounding Can.remaster offers > dynamic contrast to the CD layer; and even superior SACD layer had < dynamic value compared to the orig UK.

No wonder those titled demo's convinced ...

Measured 'DR' is affected by EQ choices.

For example if there's more bass, on A vs B, A will read as 'lower DR'.

As always one number rarely tells the whole (audio) story.