20 bit vs 24 bit remastering

[Gary]

Purely unscientific thought here.... acknowledging that Steve can't master or remaster everything that we all want, we have to settle for second best. Third best? Whatever....

Just picked up three Mastersound Gold CDs Joplin / Pearl, Sly / Stand & Brubeck. They are 20 bit mastered, by Bob Irwin. Very nice sounding! No screaching highs.... and the Sly is the original mix!

Then I thought to the 24 bit remasters of the Bowie catalogue. Ughhh. And other albums that are screachers! Not the Newfie Screech, unfortunately ;)

Why is this?

Does 24 bit remastering lend itself to overly bright CDs? Or is 20 bit technology so old that the "super high treble" era had not started yet? Or is it that Bob Irwin actually cares....?

[Steve Hoffman]

20, 24, 56 bit, doesn't matter. The tonality stays the same. It is the mastering engineer who chooses the final sound of a master tape. It is this person who has his or her hand on the big bad "tone" switch!

[Gary]

So is 24 bit more accurate or something? Or just a different console?

And Steve - did you actually find some "tone" in some new remasters? Oh goodie - which ones? ;) :D ...just kiddin...

[Steve Hoffman]

24 bit. A "bit" more resolution. Heh heh.

[Doug Hess Jr.]

Steve,
Please continue on that thought. Even if you mastered at 96 bit-- dont' you lose the higher resolution you gained when it is converted back to 16 bit for regular CD players to play? Of course it would be great if you could buy 24 and 96 bit CD's (OK SACD makes that impractical) actually encoded that way our players would then just play whatever sampling was used, but we can't so what's the point?
Dough

[Sckott]

It's actually a way of using a high-bit transfer master for the purposes of capturing a higher resolution "sound" to 16bit. Ready for the excitement? If the master sounds one way, it's gonna sound basically the same at 16 bit. There's nothing wrong with mastering at 16 bit, ya know..but....

Yes, it's true that higher bit rate CAN capture a more accurate sound, but to coin a Peter Gabriel/Genesis phrase, Eggs is eggs. And as sure as eggs is eggs, what most masterings try to achieve is some EQUing and "fixing" after the transfer into the 20 to 24 bit realm. When you stuff that into a downsample to 16, not only do waveforms lose some of that data, but smoothing during the transfer is also done, as well as some data trunkation. After all, it can only BE 16 bit, acting like it came from a bigger place.

But to be honest, the best way of capturing the egg, is not to scramble it into a powder. It's best preserved as it was in the wild, using very simple and gentle moves to digital. Less fuss, no mess (or very little).

There's a lot of high-bit stuff out there, like SACD and DVD-a, et al, but the mastering is where the magic and backbone lies.

But if you want a simple answer to why 20-24 bit xfers are used, concider it data compression, trying to fit 3 suitcase fulls into a single bag. Somewhere along the line, although you gain storage in the bitrate, it's mooshed into tight places that 16bit doesn't do normally. There are things lost and gained this way. Some people think this is a good thing, some people would rather try and preserve the egg in the most natural state possible, sunny side up, rather than burn it, vaccum pack it, and freeze dry it.

Yes, it's an egg...or....was....Yuk!

Some 20-24 bit masterings DO sound great, but this, as Steve said, is usually happy accidents counting on what the sound puree will do at playback level. It's really part snake oil, yes. It doesn't mean you're hearing more of the master. It's about how it's mastered in the more effective stages.

Whew. Did I make any sense there??

[Steve Hoffman]

Yes you did, Sckott. But no where else than this forum will you be making any friends!

So, to recap, you do gain a little from a higher bit rate, (very little in 16 bit IMO).

I've mastered in 94/26, I've mastered in this and that and onto a good old Ampex analog tape machine. It's pretty much the same.

What is more important, is the correct master tape, correct playback deck, correct mastering techniques, etc. That is the key to getting a great sound. I know; same old story, but it's a true story nonetheless!

[feinstein]

What's funny to me is the fact that Hoffman's work from the late 80's and early '90's ("Buddy Holly -- From the Original Master Tapes" for example) which was done WITHOUT the benefit of 20 bit and 24 bit or DSD still sounds markedly superior to 95% of the stuff that's done now. I think that the sensitivity and talent of the remastering engineer is the MAJOR factor in what sounds good or bad, not the bitrate or the technology.

[Sckott]

I know this sounds dumb, but a lot of people don't believe how Steve did/does it. Heck, I think if he mentioned how he does 80% of his work, people woulnd't understand.

It's not modern. It can't be possible. It ain't that simple....etc...

The other thing is, another technical button push to market remasters is money spent back to the labels for the same CD/material. Remasters on LP weren't as common back in the day. Now, a lot of hooplah is being touted so people drop their well-done mastering for something in re-colored technicolor.

New paint, new canvas, same music.

Don't get me wrong, some of the Jazz issues on Verve (not all) sound great, but that's because they're doing a much better job with better tapes. You could just take the same CD and re-EQ it here and there and say "remastered" for sales. I'm not allergic to hi-lo xfered Cds, but the product will always stand out for what it is. If it sounds awesome and it was mastered on Mars, great!

Ears are your better judge. Just don't always believe the hype. Some of the greatest music I've heard are on LP even still, not yet conquored by hyper-bitted CDs.

"It's all in the mind" - George Harrisons charactor in 'Yellow Submarine'

[Grant]

You've got a friend here, Sckott, and are pretty much on the money.

[wes]

Amen brotha,........CAN I GET A WITNESS?!!


God damn...............ya know, I wish that more people knew what we know about good mastering and bad mastering, or mastering period.....................We need a bigger market somehow get an awareness thing goin on this stuff....


Most people don't know about finding the "right" tapes, or leaving the sound "PURE" and being faithful to the master tapes sound..........Most people don't even think about that.........It's just us and a few of our friends and loved ones that will get to hear it like it is..........


-Wes



=:rolleyes:

[tomcat]

Steve said: "What is more important, is the correct master tape, correct playback deck, correct mastering techniques, etc."

If the Japanese XRCD pros are right, the "etc." may be as important as the tape research and the rest of the CD production line. Are the differences of different glass mastering and pressing subtle or more than that? I have read somewhere that many Japanese CDs use exactly the same remasters than their US counterparts, but sound better because of the tighter controlling of the process...
Best to all
Thomas

[Beagle]

Nobody here can duplicate what Steve Hoffman does but you can get some idea of what he is talking about:

If you have a decent cassette deck, a good blank tape and an original clean copy of an LP (that you also have on regular run of the mill CD), and a decent analog setup, you can do an experiment. You make sure your turntable/cartridge is set up correctly, the LP is clean, and the tape deck is biased/calibrated properly for the tape formulation.

Then set your levels based on the loudest passages. Record the LP. Now compare that cassettes sound with the CD issue of same title. Chances are that cassette sounds closer to what you might wish for in a good CD remaster.

If you haven't got a good source to work with, then nothing on earth will make it sound great.

[YaQuin]

First the waveform can be divided with respect to time. This is usually refered to as sampling rate. 44.1kHz means the processor takes 44.1 thousand 'snapshots' per second. 96kHz means the processor takes 96 thousand 'snapshots' and so on ad infinitum.

Then each 'snapshot' is represented by a number of bits. This is usually refered to as bit-depth. Theoretically the more bits the better. 16-bits allows 65,536 possible representations of any given 'snapshot' (note 2^16 = 65536). 20-bit allows for 2^20= 1,048,567 possible values, 24-bit allows for 2^24 = 16,777,216 possible values and so on ad infinitum. So, kindof like video card, if it has a higher bit-depth then it can represent more colors.

So, in theory only, any technology that uses a higher sampling rate and an higher bit depth (above and beyond the minimun 44.1 kHz 16-bit for human ear pleasing) SHOULD technically be better. However, ultimately I do not believe any technology can outwit the artistic ability of a mastering artist.

You've probably heard it before "Never Trust Your Life To A Computer!";)

[Sckott]

Let's say for years, a 4th generation tape was the only master for something. You'll have many people try their hand on it, and then they'll be the 20-24 bit CDs done by the label. Lots of EQuing because the dub wasn't that great, maybe even gate compression like no-noise.

And you'll say something like "It sounds great. Not bad."

Oh, and then the Master gets found. If it's done with strict attention to detail, maybe even on correct playback and mastering techniques, you'll flip. "

"OMG! It sounded like this?!?"

[SVL]

[QUOTE]Originally posted by Sckott
[B]some of the Jazz issues on Verve (not all) sound great, but that's because they're doing a much better job with better tapes.

_________________

There should be a big emphasis on SOME, I'm afraid. I like the job they did on Kenny Burrell's Guitar Forms; Ben Webster Meets Oscar Peterson is not bad; some of Ella Fitzgerald's songbooks are OK, and some are rather poorly remastered; Ben Webster and Associates is horrible.

Verve's transition from 20 to 24 bit did not uniformly result in better quality remasters IMO.

Wonder which VMEs people like, and which they do not like so much.

[Matt]

Have you heard the VME of Ellington songbook? I'm thinking of buying it, but I'm a little hesitant due to their track record with that particular album (very bright mastering; the out-of-phase "swishing" artifact on past issues, etc.)

Anyway, I was tempted to buy the VME of the Gershwin songbook, but decided against it. The album is mixed so poorly in stereo, I'm holding out for a good mono edition, or a good stereo remix. Some of the Gershwin tracks just drench Ella's voice in echo in the stereo mix.

Outside of these two, the only ones I'm considering buying is "Sarah Vaughan" (with Clifford Brown) and Oscar Peterson's "Night Train."

[SVL]

Originally posted by Matt
Have you heard the VME of Ellington songbook? I'm thinking of buying it, but I'm a little hesitant due to their track record with that particular album (very bright mastering; the out-of-phase "swishing" artifact on past issues, etc.)

Anyway, I was tempted to buy the VME of the Gershwin songbook, but decided against it. The album is mixed so poorly in stereo, I'm holding out for a good mono edition, or a good stereo remix. Some of the Gershwin tracks just drench Ella's voice in echo in the stereo mix.


I had a one-CD Best of Ellington songbook compilation that sounded rather bright. I think the VME uses the same remaster, but I am not sure. The one I did not like for certain was the Irving Berlin songbook; my general impression from several songbook VMEs (Ella's Irving Berlin and Rogers & Hart, and Sarah Vaughan Sings Gershwin) was that the brass section of the orchestra is nearly always too hard on the ear. On Ella's Cole Porter set on DCC it is prominent, but almost never too bright.

The fourth CD of the Gershwin VME has mono versions of 11 tracks.

[Paul C.]

Tomcat, you raise an interesting point, one that i have been trying to come to grips with. The XRCD guys seem to have a big emphasis on keeping jitter to a minimum. As I understand it, the impact of jitter is to some extent dependent on the playback system, and it's ability to accurately read the stream of bits. If the disc is well manufactured, most players will be able to read the bits accurately.

I was wondering, are gold CDs less likely to cause unwanted jitter compared to aluminium, because of the lesser number of pinholes and dodgy pits?

Re the VMEs - 24-bits is not necessarily better in the hands of the Verve folks. The recent "Oscar Peterson Plays the Harold Arlen Songbook" could have sounded better, I'm sure. And quite apart from bitrate issues, it has problems with the first half of the disc, the 1952 tracks mastered from disc sources. I've no problem with them using disc sources if the tapes are ruined or lost, but Verve actually released some of the tracks on a 2-disc compilation in 1996, and they sounded a zillion times better than the disc transfers on the 2001 CD. I asked them about this on their board, and got a rather unconvincing answer - that was all they could do with tose tracks, because the tapes were lost. The 1996 version of a couple of the same tracks sounds like it never came from a disc at all. I am thinking that perhaps the Verve production and research people are not doing their work properly these days.