Mastering Vinyl, Myths, questions, discussion Options

[Axon]

This is a split of a discussion Hancoque and I had, which desperately needs to be elaborated on.

First of all, I've dumped my thoughts on the matter into the Wiki: http://wiki.hydrogenaudio.org/index.php?ti...Vinyl_Mastering. Please comment on it.

A couple of the points there are explicitly worth starting this topic off:

• There is no consistent, unambiguous way to identify the provenance of a vinyl master. Most methods - but most of all, the zoomed-out waveform plot - are inaccurate in one way or another. Visual examination of clipped samples is fairly telling, but not applicable or effective in all cases. How can a vinyl master be accurately observed as being the same or different as a CD master?
• Even if a difference exists, whether or not it is audible is another question entirely. IIRC, brickwall limiting of 3-4db, correctly applied, may not be audible. If so, Hancoque's Slayer waveforms might not mean anything, because his peaks are all within 4db of the CD peaks. What audibility thresholds exist for clipped signals?

That said, I may be eating a hearty dinner of crow tonight, as I just examined one of the records I thought was of the same master as the CD (Battles - Mirrored), and I was able to observe pretty strong differences in clipping between the vinyl and CD versions. I'll reply later with some visual examples.

[Axon]

"There's this idea floating around that vinyl records must have intrinsically different masterings than CDs of the same material."

I am really not familiar with this myth. While most *do* have different masterings I am completely unaware of any myth that it must be so for technical reasons. Certainly one can find certain extremes inwhich one can't cut the signal to a laquer. But they tend to be extreme and rare. For the most part the same signal can be cut on laquer and mastered to CD. Who is saying otherwise?

I think that sort of thinking is inherent with the belief that vinyl must necessarily be cut from a less hypercompressed/clipped master than was used for the CD, to avoid boogaboos like mistracking/overheating cutting head/etc - so therefore (this line of thinking goes) vinyl purchases are safer than CD purchases for the purpose of avoiding hypercompression. I think that's a very pervasive myth but I'm going to take enough time with this reply that I'd rather not go through actually digging up an example of it right now... certainly it ought to ring a bell to everybody contributing to this thread?

It would be more accurate to say the same signal goes to the ADC and the preamp that feeds the cutting lathe.

Jesus, ok ok: when I say "the same signal goes on the ADC goes on the cutting head", I meant to say "the same signal goes on the ADC goes into the inverse RIAA filter and then to the cutting head." Sorry - with these sorts of discussions, in my head, I think of the "quality" of the signal as being in some sense invariant across all linear operations, in the sense that people generally care less about linear differences between different masters than they do nonlinear differences. Obviously that's not true from any remotely formal point of view of the signal itself.

A quality LP would rarely if ever made that way, because you need to condition many if not most of the signals you would record on a CD to make them fit within the limited dynamic range of the LP. The dynamic range limits of the LP conditioned the entire music production business back in the day. Music was even sometimes arranged and played so that it would "fit".

I never said this situation was "usually" the case, only that it might be the case. My understanding of current practice, flawed though it may be, is that this is indeed almost never the case - at the very least acceleration limiting and tape delay are always present to prevent accidental cutting head blowout. But there isn't any fundamental reason why you couldn't formulate a PCM signal with a peak acceleration under required limits, formulate another PCM signal to control pitch width (if that is even necessary for the application), apply inverse RIAA digitally, and wire the DAC directly up to the cutting amplifier in passthrough mode. I have asked a mastering engineer about this possibility and he said it is definitely possible.

(I know it's a wiki, but my account isn't verified, and I'll forget!).

The wiki forum has instructions - I believe the only instruction is "PM Jan for an account".

The wiki looks to me like it was written by someone who did academic-style research heavily weighted on internet chatter, It seems to dealing with a limited set of misapprehensions. The author does not seem to have the kind of understanding that would come from a good understanding of how vinyl works and its limitations.

For example it mentions elliptic filters as an explantion for bass summing. Elliptic filters are just a kind of filters like Butterworth filters. They have steep slopes. There may have been a popular implementation of bass summing that was based on elliptic filters, but that choice was irrelevant to the basics of the process wihich was to minimize vertical modulation because vertical modulation was more distortion prone and had far more limited dynamic range.

Yes, I know what an elliptic filter is, I still have those EE textbooks (and reread them from time to time). The fact is that bass-summing filters are known as elliptic filters in vinyl mastering parlance. I don't dispute that it is confusing and archaic terminology, but I did not make it up. I can change the terminology to read "bass-summing" instead of "elliptic" if you think that is clearer.

Another example:

Under "How many different ways can a CD master differ from a vinyl master?"

"The vinyl master may be sourced from a 24-bit version of the CD master."

If a 24 bit master exists, then the CD was almost certainly sourced from it as well. There would rarely if ever be any difference in this regard.

Duh - but that's the right answer to the wrong question. In terms of strict generation trees or signal flows, there is a quite unambiguous difference - the vinyl mastering is sourced from the parent of the CD master, rather than the CD master itself.

I don't dispute that the audible difference ought to be pretty close to (if not precisely) negligible, but I don't believe that is pertinent to this discussion.

I guess my real point is that that particular wiki IMO does not need just some edits. It needs to be re-architected. I don't think the author has any real audio production chops.

For openers, IMO any discussion of mastering needs to start out with a discussion of well, mastering. This one dives into a current controversy about mastering that is very narrow.

Ummm.... guilty as charged Arny, I'm about as ivory tower as they come in terms of never actually producing audio in any meaninful sense, but you'll have to forgive me for writing this wiki entry to only deal with answering the questions that people actually care about on the subject, and in such a way as to appeal to the broadest possible audience - ie, leaving questions of audibility on separate pages, and focusing primarily on intrinsic qualities of the masters in question. The changes you seem to be proposing would alienate anti-TOS8 people for no good reason and go into minutiae of mastering which are not at all necessary for explaining the important topics. I don't dispute that audibility discussions need to be on the wiki in one form or another but I think that they ought to be separated from discussions about format/processing intrinsics, because so much of why people care about vinyl is due to intrinsics rather than sound quality, and there isn't necessarily anything wrong with that.

This is not an encyclopedia entry on the static topic ("Vinyl Mastering"). Perhaps that means it needs a new title, such as "Differences between vinyl and CD mastering" or whatnot.

[analog scott]

"There's this idea floating around that vinyl records must have intrinsically different masterings than CDs of the same material."

I am really not familiar with this myth. While most *do* have different masterings I am completely unaware of any myth that it must be so for technical reasons. Certainly one can find certain extremes inwhich one can't cut the signal to a laquer. But they tend to be extreme and rare. For the most part the same signal can be cut on laquer and mastered to CD. Who is saying otherwise?

I think that sort of thinking is inherent with the belief that vinyl must necessarily be cut from a less hypercompressed/clipped master than was used for the CD, to avoid boogaboos like mistracking/overheating cutting head/etc - so therefore (this line of thinking goes) vinyl purchases are safer than CD purchases for the purpose of avoiding hypercompression. I think that's a very pervasive myth but I'm going to take enough time with this reply that I'd rather not go through actually digging up an example of it right now... certainly it ought to ring a bell to everybody contributing to this thread?

.

You "think this line of thinking is inherent...?" That is a prime ingredient for a strawman. If people are actually pushing the idea that "vinyl *must* have intrinsically different masterings than CD of the same material" Then there should be some actual evidence supporting that *this myth* is being spread by people who actually believe it. Not by folks who suspect this is the underlying line of thinking of other (kinda dumb) people.


There are sprinklings of individual facts throughout your post, it is possible to hypercompress a CD to the point that the same signal could not be cut onto vinyl. But your allusions to these particular "lines of thinking" do not ring a bell. Maybe there are a few people who have come up with such convoluted lines of reasoning and made the gross error in making some universal "rule of thumb" that mistakenly extends to all CDs even ones with no compression. But I have not seen this aleged myth anywhere but in the Wiki aticle. If it is a "very pervasive myth" as you claim we should find it spewed all over the place in audio forums, no? So once again I would like to see this pervasive myth expressed somewhere other than this thread and the wiki article by someone who actually believes it if I'm going to buy into your claims of pervasiveness.

[cliveb]

If it is a "very pervasive myth" as you claim we should find it spewed all over the place in audio forums, no? So once again I would like to see this pervasive myth expressed somewhere other than this thread and the wiki article by someone who actually believes it if I'm going to buy into your claims of pervasiveness.

I have posted comments in the past which could be construed as supporting this myth. The gist of my argument is this:

*IF* you cut a vinyl LP from a hypercompressed master, *THEN* you will either get very little playing time per side *OR* you have to seriously limit the levels which compromises the (already woeful) S/N ratio available from vinyl.

On reflection, it occurs to me that when the source material is hypercompressed there is little need for any kind of decent S/N ratio. So perhaps turning down the levels to achieve sensible playing time isn't the big problem I had previously believed it to be.