Is the DR database really accurate for vinyl?

As far as why the waveform changes, we don't know for sure. The RIAA filtering process almost certainly accounts for some (perhaps all) of it, maybe the EQ a little, quite possibly all the other variables I list in the video.

Let's avoid "better" or "worse". Does it sound more dynamic to anyone here ?

(Remember, in my head "more dynamic" means "more punch, snap, light and shade, subtlety, impact, variety, excitement...")

 

Here's a PDF that shows the math behind how the DR value is calculated: Measuring DR (pdf)
It's explained with math and equations. I don't have a full understanding of the algorithm to be able to give a proper explanation in natural language rather than math equations. The math does obfuscate what is really going on if you don't understand the math.

 

It's basically the difference between the peak level and the RMS, averaged over a 300ms time window, but biased to only pay attention to the top 20% of the range - so it tells you roughly how squashed the loudest moments are.

It's probably worth saying at this point that it doesn't really measure "dynamic range" at all - there's a technical definition for that which is quite different. So in hindsight it's a shame they used that as the name - on the other hand it works very well on an intuitive level, so maybe it is a good name, after all...?Regardless, it works well in digital processing at the very end of the processing chain. But other processing that causes variations in the peak level not related to the original music, can "fool" it into giving a much less useful reading - like cutting to and playing from vinyl as in this case.

 

Well, that's interesting. So it really only reflects the loudest portions of the music? I would then have to agree with you that the DR has very little to do with true dynamic range, which intuitively would seem to include all of the range.

I have to say that I personally do find it difficult to listen to "brickwalled" CDs at high volume - they just seem to irritate my ears. Well mastered non-brickwalled LPs and CDs present no such problem.

 

Yet, if you compare the DR numbers from some of the audiophile CD/LP releases where the mastering chains are more similar, they are usually about the same.

 

Sorry but I have a problem with the use of the term "brickwall" in that way. I used a "brickwall" limiter on the folk album I linked to a couple of posts ago, but I guarantee you can't hear it working. (Even though you can see it in the waveform, btw]

It's the more extreme use of limiters, compression and clipping that cause the "fatigue" you mention, not the processors in and of themselves. Sadly this is the norm, these days...

 

That's a good natural language explanation that folks can understand. I understand the math, but not enough of the details to be confident in explaining it properly in English without the math equations. I also didn't know it used a 300ms window. The PDF says 3 seconds. They must have meant 0.3 seconds. Details... details...

The DR numbers and the unofficial DR database and the dynamic range days and all of the other DR related promotion has gotten the issue of the loudness war talked about. Now the challenge is getting people to actually understand what DR is and what constitutes bad DR or good DR, or what constitutes a too squashed DR, or that all compression or use of limiting isn't evil.

And thank you for what you are doing to help explain all of this.

 

Well, I've given the video a very careful listening with my Sennheiser HD600 headphones.

Youtube is probably not helping here, but I can't really say that I can hear any real difference in Dynamic Range... at least nothing really significant. If I had no previous knowledge I might be tempted to describe the CD as being marginally more compressed or limited but nothing extreme. Then the waveform would just confirm this, at least visually it does look that way.

I can hear other differences, some of them are very obvious and you describe most of them pretty well. The added crispness and perceived increase in transient impact is there in the CD version, but I wouldn't automatically attribute that to more or less Dynamic Range, it seems to be the result of EQ changes introduced somewhere in the mastering chain or playback chain... and they don't sound very natural to me, if anything I would say the Vinyl version sounds with less impact but in a good way, it's more pleasant to hear the vinyl version and I can even turn it a little louder to make it "pop out" a little more without the side effects that would result from such a move on the CD version with the added "slightly artificial sounding" crispness that would start to become a problem above a certain SPL.

So, if you ask me... real DR might even be the same, maybe TT Meter is reading it wrong... or... TT Meter is reading it right but the name Dynamic Range is not really the best way to describe it... maybe it should be called Pleasure Range?

This is all somewhat irrelevant to me because I never rely on any tools to tell me what sounds better or worse in any way. I understand what you're trying to do and in fact I think you're mostly right about TT Meter results being "abused" in some ways when comparisons are made... but I wouldn't say that it can't be used for vinyl sourced files because maybe the TT Meter reading s telling us more a better information than real DR alone...

 

Hi Ian,
Your viewpoint (specifically, what you wrote in the article you linked to earlier) doesn't seem to be at all in agreement with the people I mentioned earlier. Although you did say things in the article that are completely agreeable, things no one would ever disagree with - such as, "A great master for CD can be a great master for vinyl, too". Who would disagree with that? Of course a great master for CD would make for a great vinyl cut. We seem to be debating a whole lot of nothing, except little details (and the devil is in the details).
But when was there ever a "myth" that a dedicated master for vinyl is always needed? No mastering engineer would ever say it's an absolute necessity. Seems like a bit of a straw man argument. But if you check the words of many well known mastering engineers on the topic, they all seem to very strongly recommend a separate mastering for vinyl.
On the Masterdisk website, you can see the following pdf file on mastering for vinyl: http://www.masterdisk.com.php53-6.dfw1-1.websitetestlink.com/PDF/Why-You-Should-Consider-Vinyl.pdf
To quote from it - "[question]Can’t I just press my CD master to vinyl later?
[answer from Scott Hull] In my experience, a lot of bands rely on their CD master to meet all future needs. But trust me, you don’t want to. If you want to release on vinyl, there are really good reasons to have your final mix mastered specifically for that format."

This is quite a contrast to what you say in your article. You seem to whitewash the use of a typical Cd master for cutting vinyl. Forgive me for not coming up with a better word... but you do minimize the importance of a higher quality master for vinyl. You also say "There’s no reason not to put “loudness war” style music on vinyl". Yet the engineers I mentioned earlier would all say there are plenty of reasons not to put "loudness war" style music on vinyl. But again, the devil is in the details (ie, what is "loudness war" style music?) Maybe we're all in complete agreement afterall.
Here's a link to an audio recording of a vinyl mastering panel with Doug Sax, Bob Ludwig and Michael Fremer from SXSW 2011. http://schedule.sxsw.com/2011/events/event_MP5680 At the beginning Bob talks about how the frequency response on a vinyl disc is wider than a compact disc, and how the quality of higher resolution files (24bit, 96khz) can be resolved on a vinyl disc. Towards the end they talk disparigingly about using compressed CD masters for cutting vinyl. They mention the 2010 reissue of "Exile on Main St.", cut from a CD master (actually I believe it was 24/44.1) with the same compression as the CD - which scores an overall DR7 on the TT meter ( a few tracks hit 10 & 11) -- one less than what you describe as optimal "DR8 or more overall". Everyone who's heard that LP thinks it sounds terrible. Doug Sax actually cut that, and said "garbage in, garbage out". Bob Ludwig was quick to point out that he did not master the digital files for vinyl cutting. At the very end of the talk, Scott Hull from Masterdisk gives some technical reasons for not using compressed CD masters for cutting vinyl. In particular, he says that for every db "louder" you're able to get the music signal, you'd have to cut the vinyl 1.5 to 2 db lower (which of course raises the vinyl noise floor). He also says that in cases where a client may not be able to afford two completely separate masters, they prioritize the vinyl mastering, and then tweak a little further for the CD mastering.

I listened to that whole album, and although the quality over the net leaves a lot to be desired, the music was really nice. I wish I could hear it with zero compression used in mastering.
I heard back from one of the "heavy hitter" vinyl mastering engineers I mentioned earlier in the thread. I won't identify him since he didn't say I could quote him, but regarding cutting really dynamic music to disc, he simply said - If a vinyl engineer hasn't figured out how to fit a wide dynamic range on LP, then I guess he has a lot to learn.

 

I agree completely.

Another record we transcribed the other day is by Lee Ritenour called Friends and cut live by Stan Ricker at the JVC cutting center in May 78. I don't remember buying this record, so perhaps Stan gave me this copy, as I also did a few 1/2 speed mastering sessions with him about that same time. I can't find much about it on the internet so it must have been a very limited release.

It's also got Steve Gadd and Abe Laboriel playing on it. It sounds a bit familiar to me, but Lee and Abe also played on other projects at our studio at about the same time. It's got a jazz fusion type of vibe and has some very nice moments.

Anyhow, this is a direct to disc record with incredible dynamics and it's on very quiet vinyl.

 

That's the case. EQ, noise and going through tube amps produce sonic changes to the audio. That noise now is part of the sounds itself. It doesn't matter where the extra dynamics are coming from, the TT meter will pick it up anyway.

What's "better" anyway? Some people like that sound, but it was "added", not originally there.
Some digital masterings are very tough on the ears. The whole coloring the vinyl introduces "softens" the audio, compensating for the harsh mastering. Death Magnetic example comes to mind. The LP sounds "better" because the clipping is not that apparent as in the CD.

Ultimately, vinyl rips can't be compared. Because if the result is DR12, you don't know wether it was originally DR8 or DR11. It could be real dynamic range, or it could be just coloring. And given modern-day mastering, it's quite probably the latter.

 

I have the Death Magnetic vinyl and the issues I had with the CD were still prevalent in the vinyl release. It may be more "dynamic" but it was still overly compressed and limited on the vinyl release. I was very disappointed because I spend a lot of money on that vinyl only to hear the same problems I heard on the CD.

 

That is true to a point but if you have a master where the DR rating (for a lack of a better measurement) is 16 lets say and you bring that down to say 13 I have a hard time believing that will automatically make the record 4.5 to 6 db quieter. I can see that if you are going from lets say 10 to 7 but once you hit a certain level of dynamics I don't think that rule necessarily applies because if the music is really dynamic they need to start to lower the levels also so they don't force the needle out of the groove during playback. I also think that is a simplified way to explain this because obviously there many other factors that determine the volume level the vinyl will ultimately be cut at, things like stereo panning and how much information you are trying to fit onto a side of a record etc etc.

 

Because Shepherd admitted brickwalling the example he used.

 

From what I've read here, the only decent-souding version of Death Mag is the file ripped from the PS/3.

 

Sorry, but you're mis-quotiong me.

I said the waveform was characteristic of something that had been limited - what some people call "brick wall limiting". My point is, it may look like that, it doesn't sound like that. The whole video is intended to demonstrate that appearances can be deceptive.

Listen with your ears, not your eyes.

 

SuperFuzz, the irony of Masterdisc lecturing us about dynamics makes my teeth hurt. Regardless:

Many people believe that a dedicated vinyl master is essential, this is what I mean when I refer to the "myth". It's not essential, in fact often it's unnecessary. I agree though, we may be quibbling over semantic details. Certainly mastering engineers make more money when they do two mastering sessions instead of one... but I'm sure that's not a factor in the comments of any of the engineers you quoted !

Seriously though, if you want the absolute best master for vinyl from an analogue source then Yes, avoiding unnecessary A/D and D/A conversion is desirable (as with the new In Utero re-master) - but that scenario is rare, these days.

Certainly using a heavily compressed CD master for vinyl isn't the best idea - but I don't advocate heavily compressed CD masters at all ! (You seem to think that all compression and limiting is undesirable, so I'm sure you'll say that DR10 is still too compressed - we'll have to agree to differ on that point.)

When I said "there's no reason not to put "loudness war" music on vinyl", I meant more precisely "there's no technical reason why you can't put..." - of course there's a good reason not to, it will sound crappy.

I've never said DR8 is "optimal", I've said DR8 is a good rough guideline for the minimum you can hit without damaging the music, hence DR8 or more is optimal. Personally I aim for DR10, maxing out at DR8 in the loud sections - I've amended the text of that post to be clearer. I haven't listened to "Exile On Main Street" but there are many other reasons it could sound crappy as well as poor dynamics...

I'm glad you liked the album link I posted - where do you hear undesirable compression artefacts that make you wish you'd heard the uncompressed one ? Which are your favourite entirely un-compressed recordings ?

Finally:

Actually - I agree. My clients went to a large vinyl broker and had a nightmare with the cutting process - in common with very many people who decide to have vinyl pressed, these days. I dread to think how the record would have sounded if that company had been responsible for the mastering, too...

None of this stuff is clear-cut, or easy - I'm just trying to offer good advice to a wide range of people. If that upsets some purists, so be it - it's still sound advice.

 

Sorry but I am NOT misquoting you, I'm giving new information. This has nothing to do with the video. If you want a decent sounding version of the Metallica release, you have to rip it from the Sony Playstation from the Rockband version.

 

This news is a bummer, but one I was fully expecting. Not so much the DR database, I suspected this to be the case for a long time, which is why I don't even bother using it as a tool. I'm referring to the fact that the vinyl and CD masters are probably the same more often than not. Most bands aren't going to make two separate masters. Why would they when most people, including themselves, probably won't notice any difference? And I'm not against limiting, I understand how the tool is supposed to work and realize it's not a modern invention. How do you think that Raw Power got to be dubbed the loudest album of all time pre digital limiting? I have a problem with the Death Magnetics and Stadium Arcadiums, albums that have no dynamics and do truly sound like s**t. At the same time, when I buy vinyl I am hoping that because of the fact I like a pain in the ass format, I do get better mastering than its digital counterpart. I knew most times I probably wasn't and this confirmed that.

Thanks Mr. Shepherd for making this knowledge available. Very interesting stuff.

 

Correction: Not Rockband, Guitar Hero 3.

 

Hmmm - OK, here is something that may be a difference in definitions used. I use "limiting" and "brickwalling" differently, not as synonymous. I recognize that many (or most) recordings have some limiting applied, and I accept that as something that may in some cases be desirable. I use "brickwalling" to mean when that limiting is taken to the max, causing significant damage to the inherent dynamics of the music; this I think is never really desirable (at least to me). My former comment about unpleasant listening experiences with brickwalled recordings was meant to refer to this latter definition (extensive, damaging limiting) not simply to any limiting.

I may be wrong in my separate uses of these two words, but I suspect that it is a common understanding (or perhaps misunderstanding) of the two.

 

Hopefully you bought the Stadium Arcadium Limited Edition vinyl that was mastered by Steve Hoffman. That sounds much better than the CD/digital versions. I have no doubt that if there was a true digital version of that mastering it would be much more dynamic than the original.

But I agree. It would be nice since you are paying all that extra money for the vinyl that they would take the extra time and money to make a special master for it. Alas they often don't.

 

That is how I define it myself. There are many people here that say if you use a brick walling limiter even the right way as to not damage the inherent dynamics to the point of obvious distortion (when played not viewed in waveform) that is still brick walling, but I don't agree.

 

Thank you for explaining it in an easy to understand way.

I have been saying this for years. It does not measure Dynamic range, and many recordings are being given a bum rap for not being dynamic, but it is the DR meter that is to blame.

Since it only captures the top 20%, music with sustained fairly low passages, with lots of other loud parts, will not measure dynamic, but it is most likely the most dynamic music made. Really is confusing.


DSOTM, on "Speak to me" is the perfect example. It starts out barely audible and builds to a very very loud cresendo, but on the DR scale it looks like nothing special.