Vinyl Mastering

From Hydrogenaudio Knowledgebase
Jump to: navigation, search

Does vinyl intrinsically require a superior master than CD?

There's this idea floating around that vinyl records must have intrinsically different masterings than CDs of the same material. There's both a kernel of truth to this, and a few gigantic myths.

CDs have only one (extremely strong) restriction on how loud they can be cut - the digital peak level, 0dbFS - and (almost) anything that doesn't violate that restriction is permissible. Vinyl manufacturing has many different restrictions, and they are all rather loose, in that the restrictions can be sometimes relaxed.

  • The grooves can actually overlap each other if they are too loud. This can be alleviated by spacing the grooves further apart.
  • If the groove "moves" quickly enough - its velocity is high enough - some turntable cartridges will be unable to track the groove, and a skip results.
  • Just like the voice coils in a speaker can burn up if enough energy is dumped into them, the voice coils on a cutter head can burn up if the signal is of a high enough power. (The amplifiers range into the hundreds of watts and the coils themselves are liquid- or helium-cooled, depending on who you ask, so the powers involved here really are quite substantial.) The historical solution to this is a special limiterthat squashes the high frequencies in the music if the cutting head temperature exceeded some threshold. Clearly not a high-fidelity solution (and many mastering engineers do not use them nowadays).
  • Excessive stereo bass content (bass in one channel or another) can compromise tracking or even make the cutting head jump out of the groove. This is sometimes solved with an elliptical filter, which sums bass frequencies to mono. Again, not all mastering engineers use this.

None of these restrictions explicitly say "hypercompressed, distorting music cannot be cut onto vinyl". Rather, that music may be more difficult to cut and play back than other music.

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

  1. The CD and vinyl masters might just be exactly the same: the same signal that goes on the ADC goes on the cutting head.
  2. Acceleration limiting might be used on the vinyl master.
  3. Elliptic filtering (bass sums to mono) might be used on the vinyl master.
  4. The vinyl master may be sourced from a 24-bit version of the CD master. (However, the high noise content of vinyl generally makes this a meaningless distinction.)
  5. The vinyl master may be sourced from a higher-sampling-rate version of the CD master. (However, the demonstrated inaudibility of frequencies above 20khz makes this a meaningless distinction.)
  6. The vinyl master may be EQ'd differently to account for equalization differences in the cutting head, electronics, or playback devices.
  7. Finally, the vinyl master might be sourced from a master with less dynamic range compression or limiting than the CD master. This is the only distinction between a vinyl an CD master that is meaningful - in the sense that information exists on the vinyl master, in terms of reduced compression, that does not exist on the CD master.

How do you know if a vinyl master is audibly superior than the CD master?

You ask the mastering engineer what he did. Other that, that, generally, you don't know. There are certainly many wrong ways to determine this, which can lead to false positives and false negatives.

  • Many people look at large-scale waveform plots, like those available in Audacity and Audition, and compare the waveforms across the entire piece of music. This does not work. The distortions present in vinyl - everything from subsample delays in the recording process to phase errors in the analog electronics to tracking and tracing distortion - ensure that even if the vinyl is cut with the exact same master as the CD, the peaks will be considerably higher, even during regions of gross clipping. Thus this technique is generally not acceptable, even though it is by far the most popular.
  • RMS loudness estimates, such as the industry standard RMS figure and ReplayGain, are ineffective because they require a reference level to compare the vinyl and CD versions against. No such reference level exists.
  • Experimental dynamic range estimators, such as pfpf and SparkleMeter, are useful in teasing out substantial differences in dynamic range, and may be quite useful in estimating when they become audible, rather than . pfpf, in particular, is designed to be immune to moderate levels of clipping distortion, under the expectation that clipping is either going to be inaudible or going to affect the timbral character of the music, not the dynamic range.

The one consistently accepted method of showing reduced compression is to show the individual samples in a clipped waveform against the same waveform in a different master that is not clipped. But again, this method is not foolproof: Various distortions can mask the clipping so that it is not consistently at the signal peak, yet still retains its characteristic distortion. However, clipping may not exist obviously in hypercompressed music, and even if a difference exists, it very well may not be audible.

Is less compressed music always of a superior quality?

Sometimes. Low levels of clipping and hard limiting (perhaps up to 3db!) are surprisingly inaudible. Excessive dynamic range is generally contrary to many peoples' listening habits and situations; few people will tolerate the full dynamic range of a symphony orchestra in their living rooms, or the full loudness of a live rock band. Modern music listeners consistently perceive less compressed music as being drier in tone and less pleasing to the ear than modern mastering styles.

That said, just like speaking in different tones and loudnesses of voice is considered more emotional and human than speaking monotonically, music with a wider dynamic range is generally perceived as being more emotional than music with a tight dynamic range. And once the dynamic range is crushed out of the music, it generally cannot be added back in at a later time. The information representing by the dynamic range is effectively destroyed. These factors, as well as the diverse other factors mentioned in discussions of the "Loudness War", ultimately reduce the value of modern-mastered records in many peoples' view.

Some known examples: Vinyl releases with a different master than the CD

Bob Katz has stated that all of his vinyl masters are sourced from a mix "prior to any of the peak limiting or any additional loudness makers other than the ones there for esthetic purposes."

Steve Hoffman's work is generally known for distinctly different masterings compared to equivalent CD releases. His mastering of ZZ Top's Tres Hombres includes diverse changes, including a much less compressed drum track. His mastering of the White Stripes's Icky Thump is also well praised for being distinctly different (and better) than the CD release.

Some known counterexamples: Vinyl releases with same/similar hypercompressed master as on CD

This is a list of vinyl/CD releases, where the CD release has been considered compressed or clipped for a popular music audience, and the vinyl master has been proven to be sourced from digital audio equal in compression/clipping content to that of the CD master. This list excludes albums whose CD releases are considered relatively uncompressed or unclipped.

It does not mean the CD master necessarily sounds poor - some CDs referenced here have been highly commended for their sound quality - but it does mean that, for any of these records, if one considers the CD to be 'hypercompressed', one ought to also consider the LP to be hypercompressed as well, so preferring LP over CD for that reason would be foolish.

When I cut the vinyl DMM masters for the first "special edition" release of the 45rpm 4-LP set for the Foo Fighters "In Your Honor" album the producers made a big deal out of wanting to have an "audiophile" release and made sure that I could work from 24bit/96kHz source (which I was able to do at Europadisk due to SAWStudio sending to a matched pair of Lavry Blues that went to both the pitch depth computer and the cutting head) and that I did as close to a flat transfer from these as possible.

I was disappointed to find that the high res files I received ... had been heavily clipped as they were the same files (just prior to SRC and dithering) that the CD master was made from.

  • Metallica, Death Magnetic. Visually confirmed, and strongly hinted at by Ted Jensen, who mentions that the mixes were clipped. No word from vinyl mastering engineer Kevin Gray. As the mix itself was clipped, it is highly anticipated that the vinyl release is of the same master as the CD. Jensen:
In this case the mixes were already brick walled before they arrived at my place. Suffice it to say I would never be pushed to overdrive things as far as they are here. Believe me I’m not proud to be associated with this one...
  1. He was specifically commended for the sound quality of the vinyl release before he pointed out that it was sourced from CD.
  2. He claims that a Stereophile reviewer once commended him in the same way about a separate record, and refused to believe that the vinyl was in fact sourced from 16/44. (emphasis added)
I once had a reviewer from Stereophile call and was raving, raving I tell you, about the sound of an LP done the same way: I sent my eq'd CD master and someone else cut the lacquer.

"The depth, the detail, the microdynamics are beyond compare, it's just more proof of the superiority of analog."
"But it was cut from a 16 bit digtial source."
"I was at the session."
"Don't you tell me what I'm hearing!"

"Uh, I gotta get back to work......."
  • Pete Lyman, Infrasonic Sound, same "Down on the Upside" thread as above, comments on how CD sources form the majority of his business in cutting records, and points out that 16-bit digital delay lines have been used since the 80s, and quite possibly for most records released since then (emphasis added):
I cut about 10 sides a week, most of them from a 16/44.1 source. If the source material sounds good and doesn't have excessive high end issues or phase issues, it should sound fine, if not better. I doubt most people could tell the difference between the same material cut at 16 bit and 24 bit. Most records (at lease since DAT/digital surpassed analog mixdown options)have gone thru some conversion during the cutting. I think it would shock quite a few people to know that those "analog" records they love so much probably passed thru a 16 bit digital delay line on their way to the cutting head. But, quite a few of us are opting to bypass that sort of delay and do it in the workstation thru a second set of converters, eliminating a stage of conversion.

Vinyl releases suspected of being of different masters than the CD

  • REM, Accelerate
  • Battles, Mirrored
    • Clips, but at a different (lower) level than the CD (!)
  • Slayer, Christ Illusion
  • Depeche Mode, All new releases since 1997.

Vinyl releases suspected of being the same master as the CD

  • The Decemberists, The Crane Wife
  • REM, Accelerate
    • Commentary is very divisive. Some people believe the vinyl is clearly superior to the CD, and some people abhor the vinyl as sounding hypercompressed.
  • Autechre: Untilted, Chiastic Slide, Gantz Graf (and probably all others)
  • Shellac: Excellent Italian Greyhound (and probably all others)
  • Of Montreal, Hissing Fauna: Are You The Destroyer?

Vinyl mastering online

Here is the list of services where you can make online masterig for vinyl:

Guideline of do's and don't for vinyl audio mastering