Is the dynamic range of CD sufficient?
- Thread starter Al M.
- Start date
Sigh,
"move on" as move on from the argument/to let go of the "discussion" without further swipes or digs.....
NOT move on and leave as you infer in your post 561
Please note you have taken the wrong context of my post by splitting it the way you have, even earlier on I mentioned about dropping this and why I stopped posting further technical information (that still disagree with you from my point-perspective).
Orb
"move on" as move on from the argument/to let go of the "discussion" without further swipes or digs.....
NOT move on and leave as you infer in your post 561
Please note you have taken the wrong context of my post by splitting it the way you have, even earlier on I mentioned about dropping this and why I stopped posting further technical information (that still disagree with you from my point-perspective).
Orb
Last edited:
Ain't no such thing as RIAA EQ'd tapes, master or otherwise.
RIAA EQ is applied to the cutter head coils.
What many appear to be confusing is the use of so-called "EQ'd limited copies" as the source tapes for many (but certainly not all) early CDs.
Some background:
When a lacquer is created for vinyl production, there are only so many pressings that can be derived from said lacquer. When an album was expected to sell in large numbers, more than one lacquer was needed. Sometimes the additional lacquer(s) might be needed for additional pressing plants. Sometimes the additional lacquer(s) might be needed at a later time.
An easy way to preserve all the mastering "moves" (i.e., processing, EQ, level changes, etc.) was to run a second tape recording in the mastering room. The first machine played back the master, with the signal being fed to the cutting console, any hardware inserted into the chain and ultimately to the cutting amplifiers for the cutter head. The output of the console (containing all the mastering changes but *not* any RIAA EQ, which as I said, is applied to the coils of the cutter head) is fed to the second tape machine, which captures the "mastered" sound onto a new tape. This new tape becomes the "EQ'd limited master" since it contains all the "EQ and limiting" applied during mastering.
When more lacquers are needed, the EQ'd limited tape can be used, even by another engineer at another mastering facility or perhaps by a night-shift trainee to (theoretically) create new lacquers that sound the same as the original.
Now, in the early days of CD, some folks in the record industry wanted the CD to "sound like" the vinyl -- in other words, to contain all the changes wrought in the mastering room when the lacquer was cut for vinyl. These changes could include things like changes to the speed of a given track (a slight speeding up or slowing down if the producer decides after the mix, that the song would be better served by a slight tempo change) in addition to all the other possible changes to the original that might have been decided upon in the mastering room.
It should be remembered that in most cases, the "master" that comes out of the mix room is not necessarily (and in fact more often than not, *isn't*) intended as the final master. Civilians often tend to get confused about exactly what a master is. Well, there is the original recording - the first time the sounds are captured, which is referred to as the "master". Most times, this is a multitrack intended to be "mixed down" to stereo (or sometimes surround), where instrumental balances, placement and tonality are manipulated to create the intended stereo (or surround) "master". Once this has been created, there will often be a session or two in the editing room to create (you guessed it) the "master". Then, when everyone is (mostly) happy, this is taken to the mastering room where the engineer will create the "master". (Please read all this again as it *will* be on the test. ;-})
Among the many problems that arise when taking the convenience of using an EQ'd tape made during vinyl mastering as the source for a CD master is that in addition to all the production decisions regarding music and sound, there are manufacturing decisions, many of which, in my opinion, work against getting the best possible results. For starters, many vinyl cutters will routinely apply certain processes because these make for easier lacquer creation (i.e. fewer "blown" lacquers"). Often (too often in my view), bass is rolled off, high top end is rolled off and the bass that remains is "matrixed" into mono. (I've known "mastering engineers" who would set up many of these things *before* they've even listened to what they are about to master!)
Ideally, the CD mastering engineer will be given access to *all* the tapes so the mastering engineer can determine which, if any, changes on the EQ'd tape are relevant to the CD, while still using the same source the vinyl mastering engineer used (i.e., the edited, mixed "master"). It if often useful for the mastering engineer to have access to the multitracks too.
***
As to the question asked in the topic at hand, in my view the answer is very distinctly no.
First, I find that in discussions of this sort, the signal-to-noise ratio (which is what we're really talking about here) is confused with dynamic range, as if the two were interchangeable. While they might be in theory, my experience has been that in practice, they are very, very different.
All one need do is record the same signal into a 16-bit system at many different levels. As the level goes down, quantizing issues go up. Instrumental harmonics become thin, bleached and not at all like the real thing -- or like they are captured at higher levels, where more of the 16-bits are put to use. That low level information, where the harmonics, spatial cues, etc. reside, gets increasingly coarsened as the recorded level goes down. So, while there is no hiss or other steady-state noise to get in the way, sounds recorded toward the bottom of the loudness range change for the worse and (to my ears) in not at all subtle ways.
As always, just my perspective.
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
RIAA EQ is applied to the cutter head coils.
What many appear to be confusing is the use of so-called "EQ'd limited copies" as the source tapes for many (but certainly not all) early CDs.
Some background:
When a lacquer is created for vinyl production, there are only so many pressings that can be derived from said lacquer. When an album was expected to sell in large numbers, more than one lacquer was needed. Sometimes the additional lacquer(s) might be needed for additional pressing plants. Sometimes the additional lacquer(s) might be needed at a later time.
An easy way to preserve all the mastering "moves" (i.e., processing, EQ, level changes, etc.) was to run a second tape recording in the mastering room. The first machine played back the master, with the signal being fed to the cutting console, any hardware inserted into the chain and ultimately to the cutting amplifiers for the cutter head. The output of the console (containing all the mastering changes but *not* any RIAA EQ, which as I said, is applied to the coils of the cutter head) is fed to the second tape machine, which captures the "mastered" sound onto a new tape. This new tape becomes the "EQ'd limited master" since it contains all the "EQ and limiting" applied during mastering.
When more lacquers are needed, the EQ'd limited tape can be used, even by another engineer at another mastering facility or perhaps by a night-shift trainee to (theoretically) create new lacquers that sound the same as the original.
Now, in the early days of CD, some folks in the record industry wanted the CD to "sound like" the vinyl -- in other words, to contain all the changes wrought in the mastering room when the lacquer was cut for vinyl. These changes could include things like changes to the speed of a given track (a slight speeding up or slowing down if the producer decides after the mix, that the song would be better served by a slight tempo change) in addition to all the other possible changes to the original that might have been decided upon in the mastering room.
It should be remembered that in most cases, the "master" that comes out of the mix room is not necessarily (and in fact more often than not, *isn't*) intended as the final master. Civilians often tend to get confused about exactly what a master is. Well, there is the original recording - the first time the sounds are captured, which is referred to as the "master". Most times, this is a multitrack intended to be "mixed down" to stereo (or sometimes surround), where instrumental balances, placement and tonality are manipulated to create the intended stereo (or surround) "master". Once this has been created, there will often be a session or two in the editing room to create (you guessed it) the "master". Then, when everyone is (mostly) happy, this is taken to the mastering room where the engineer will create the "master". (Please read all this again as it *will* be on the test. ;-})
Among the many problems that arise when taking the convenience of using an EQ'd tape made during vinyl mastering as the source for a CD master is that in addition to all the production decisions regarding music and sound, there are manufacturing decisions, many of which, in my opinion, work against getting the best possible results. For starters, many vinyl cutters will routinely apply certain processes because these make for easier lacquer creation (i.e. fewer "blown" lacquers"). Often (too often in my view), bass is rolled off, high top end is rolled off and the bass that remains is "matrixed" into mono. (I've known "mastering engineers" who would set up many of these things *before* they've even listened to what they are about to master!)
Ideally, the CD mastering engineer will be given access to *all* the tapes so the mastering engineer can determine which, if any, changes on the EQ'd tape are relevant to the CD, while still using the same source the vinyl mastering engineer used (i.e., the edited, mixed "master"). It if often useful for the mastering engineer to have access to the multitracks too.
***
As to the question asked in the topic at hand, in my view the answer is very distinctly no.
First, I find that in discussions of this sort, the signal-to-noise ratio (which is what we're really talking about here) is confused with dynamic range, as if the two were interchangeable. While they might be in theory, my experience has been that in practice, they are very, very different.
All one need do is record the same signal into a 16-bit system at many different levels. As the level goes down, quantizing issues go up. Instrumental harmonics become thin, bleached and not at all like the real thing -- or like they are captured at higher levels, where more of the 16-bits are put to use. That low level information, where the harmonics, spatial cues, etc. reside, gets increasingly coarsened as the recorded level goes down. So, while there is no hiss or other steady-state noise to get in the way, sounds recorded toward the bottom of the loudness range change for the worse and (to my ears) in not at all subtle ways.
As always, just my perspective.
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
Just to be clear, are you talking about recording at 16 bits or producing discs/files for distribution at 16 bits?
And - so what do you think the problem was with, for example, the early Zep and Beatles CDs?
Tim
And - so what do you think the problem was with, for example, the early Zep and Beatles CDs?
Tim
Yes you can go back to the DIY forum and continue bashing and bad mouthing WBF over there. And reverse engineer the Trinity DAC.
Hi Tim,
Not sure if your questions are addressed to me.
As to your second question, I have my opinion. I'll keep one to myself. Since I was the mastering engineer for the other, I'll talk about that one. I'm referring to the first Zep CDs. Outside of the fourth album ("Zoso") which was mastered by Joe Sidore out at Warner Brothers in L.A., I mastered the original releases of the other Zep CDs.
While I've seen lots of posts on various fora touting these as the best digital Zep so far, I haven't seen anyone before your refer to them as "bright". If anything, most who haven't liked them thought them "dull". (Certainly, to my ears, appreciably *less* bright --and more dynamic-- than any subsequent remasterings of same.)
My own take is that there is a lot of midrange energy in the recordings themselves, at least the sources I was given to work with. At the time, I was told the original tapes were lost. (Based on what I've heard from colleagues, I have no reason to suspect this situation has changed.) What I was given to work with were, to the best of my knowledge, flat copies of the original mixes. That said, my experience has been that if you have three engineers create a flat copy from the same source, all three copies may well sound different from each other. (This has to do with how well the playback and record machines are set up and with the nature of the signal path between them.)
When I created the CD masters, I bypassed most of the room (as has always been my custom), keeping the signal path as clean and short as physically possible. When I deemed the application of EQ was warranted to make the best of the source (in my experience, 99% of the time with typical studio recordings), *only* the equalizer--analog in those days--was inserted into the chain. No elaborate mastering console, switching network, patch bay, etc. The signal was monitored *after* A-D conversion.
The conversions were as faithful to what I could get from the source tapes as the technology Atlantic had at the time allowed. This included Sony's 1630 A-D converters and DAE-1100A editor.
Were I to do the same albums again today, would I do them differently? Well the philosophy is essentially the same but the gear I have now is orders of magnitude better than what I had access to then. And, if I do say so myself, I'm a better engineer now. (Here's one that civilians may be surprised at: It took me 15 years (!) to learn how to EQ in a way I would today refer to as "correctly". Of course, everyone you ask may have a different definition of "correctly". And certainly, most engineers start EQing from day one. This is probably the subject of another thread, for another day.)
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
My impressions Barry are that many engineers hated the 1630 and its sonics? And at that time there was a lot of problems at the mastering facilities with 1630s. I remember doing a piece on an album Bernie Grundman did with a 1630 and sent what were supposed identical copies to two mastering houses and the two CDs sounded far from the same.
Hi Myles,
At the time, the 1630, especially when equipped with Apogee's retrofit filters, was the state of the digital art in terms of mastering. For a long time, it was the only game in town.
As one of the first engineers to master for CD, my opinion was always that digital was a step down from vinyl in terms of capturing the sound of the source. In fact, I said so (shyly at the time) at a very early meeting of the AES in New York when I sat on a panel of mastering engineers. Mine was the only voice in the room that said anything other than high praise for CD and digital. (It took over 20 years before I heard digital that I felt could challenge vinyl and that was only with the very few converters that could *truly* do 4x rates properly-- and only at the 4x rates, not less.)
Regarding supposedly identical CDs sounding different, there are a number of things at play. First , you mention two mastering houses. As soon as two different engineers are brought into the equation, I would withdraw any and all bets. Now, let's talk about the same engineer in all instances: I have said, since the first CD I mastered in January of 1983, that pressed CDs made at different plants (often on different lines at the same plant) all sound different from each other and that *none* sounds indistinguishable from the master from which it was made.
In my experience, this is still just as true today (though some plants make CDs that differ from the master to a much lesser degree than others do). It is true when the disc is played in any transport or player in my experience. Interestingly, when the different sounding discs are properly extracted to a computer hard drive (in a raw PCM format, such as .aif or .wav), the sonic differences go away and they now *do* sound indistinguishable from the master used to make them.
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
My apologies, Barry, I didn't mean to take a shot at your work. Thin, might be a better word than bright. It's not so much an excess high end I hear as a lack of bass. The upside is they don't assault the ears when they come on. In fact, when I have cuts from those albums on playlists, I often find myself needing to turn them up. And that's better, to me, than the opposite issue. I have enough system headroom. I can turn it up myself. Thanks for your reply. What about the 16 bit question? We're you referring to recording to 16 bit in the studio, or to 16 bit media?
Tim
Tim
Barry,
Excellent posts! It is really great to have the perspective of someone inside the industry.
Hi Tim,
If that's how you hear it, that's how you hear it. No problem.
Again, most folks complain that they sound "dull" or "too bassy" but one person's bass might be another's mid-bass... or something. ;-}
Could I do them better today? I have no doubt.
At the time, did I do them to the best of my ability given what I was presented with? I have no doubt.
With regard to the 16-bit question, if you are referring to my comments in response to the question posed by this thread, I'm referring to 16-bits anywhere -- in the studio or via media we play at home. (I think of 16-bits and CD as the "cassette" of digital.)
The low end of the dynamic range, while above the noise floor of 16-bit formats, is in my experience, inadequately represented resolution-wise, effectively being captured at considerably less than the 16-bits louder parts of the recording are captured at. (Each bit represents roughly 6 dB --- 6.02 dB actually.) If the loudest sounds in a recording peak within the top 6.02 dB, they are represented by all the bits. things like instrumental harmonics and spatial cues are well down in level from the loudest sounds - perhaps 30 or 40 dB down or more.
With a 16-bit format sounds that are 30 dB down are effectively represented by 11 or 12 bits. Sounds that are 40 dB down are represented by 9 or 10 bits. Lower level sounds are represented with even fewer bits.
Now let's look at the same sounds with a 24-bit format: The 30 dB down sounds are represented by 19 or 20 bits. Sounds at -40 dB are represented by 17 or 18 bits --- still more resolution than the loudest sounds at 16-bits.
Dither hides some of the quantizing error but dither is only at roughly the bottom 4.7 dB of the dynamic range. There's a lot of music that is well above this level and is - to my ears anyway - represented by a woefully inadequate degree of resolution in 16-bit formats. So, if I've got a really dynamic recording, the quiet parts will just not be anywhere near as well resolved as the louder parts. With that in mind, the common confusion of signal-to-noise ratio with real dynamic range makes me wonder if folks who make those claims every stop to actually listen to the musical consequences.
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
If you are an analog person the above quote =
If you are a digital person, the above quote =
Personally, I think that Barry was being too harsh on cassette tape, especially if it was recorded on a Nakamichi deck from a great analog source.
Last edited:
If you are an analog person the above quote =
If you are a digital person, the above quote =
Personally, I think that Barry was being too harsh on cassette tape, especially if it was recorded on a Nakamichi deck from a great analog source.
Hi mep,
Perhaps a bit hard on the old cassette. ;-}
However, I prefer a deck that uses a standard playback EQ curve rather than those that hype up the top end (and compensate only with their own record EQ -- fine for tapes they recorded, painful --to my ears-- for tapes recorded elsewhere).
In my early days at Atlantic, I worked in a room filled with cassette decks from all sorts of manufacturers. I aligned the decks too, to ensure top performance when we copied new masters to them for distribution to A&R and promotion personnel. Heretical as it may seem to some, I found certain $100 Sony decks to sonically wipe the floor with $1000 machines that hyped the top, when the criterion was fidelity to the source.
Again, just my perspective, as always.
Best regards,
Barry
www.soundkeeperrecordings.com
www.soundkeeperrecordings.wordpress.com (The Soundkeeper Blog)
www.barrydiamentaudio.com
As a newbie to computer audio, I must say that I too find that CDs ripped to my mac Air via AIFF sound better than if I play them with a transport through the same DAC. There is more air, spacial dimensionality, a cleaner top end and more low level detail. At the same volume levels, I hear a lower noise floor with an overall smoother sound. Of course, the better the recording and mastering job, the more the differences are noticeable. Some ripped CDs sound sensational, not as good as the best 24/96 and 24/192 tracks I have heard which are even more liquid, but sensational nonetheless.
My comment was exactly based on Barry's comment about 16 bits and CD being the "cassette" of digital. I thought it was pretty clear.
I realize my post may be a bit of a nonsequitir at this point in the conversation and that we've moved on beyond what processed audio looks like but I can't help myself.
This representation of the Fourier transformation was enlightening for me and this page describes the model used to get there. Here's the interactive version rendered in Flash.
This representation of the Fourier transformation was enlightening for me and this page describes the model used to get there. Here's the interactive version rendered in Flash.
Similar threads
