Is the DR database really accurate for vinyl?

When you say Stan is no longer with us do you mean you work at A+R in Dallas and he's no longer with your company, or that he's no longer alive? Ray cuts with whatever he gets, so it's a case by case basis. I am grateful DMB have woken up but I suspect the live releases are not handled by the same people that do the regular vinyl, 2009's 2LP was cut from a separate 24/48 master, the 2012 Lillywhite-produced album was cut from the 16/44.1 CD files. The CD master was a lot less slammed than the 2009 CD which may explain that, but still could have done with more open dynamics. Unfortunately the 2009 2LP was poorly pressed at United and I had to buy three copies (@ $13.99) before I got one with SIDE D centred properly, thankfully the upcoming RSD live releases are being done at Furnace MFG.

I have bad examples too, last Jane's Addiction LP was awful (Meller), thought the Telephantasm box set vinyl (Ludwig) actually sounded noticeably worse than the CDs. Case by case is only my point. As for the liar comment it was referring to the article stating "no matter what you've been told".

 

Thanks ClausH I wasn't aware of him. That would make more sense

 

I haven't read the entire thread, so I apologize if I'm repeating something here, but I do want to mention my own observations that differ somewhat from your statements above:

Concerning the first of your points, I'm following the vinyl ripping scene where it's common for rippers to swap decks and cartridges and then rerip albums all the time, and yet the DR values hardly ever change significantly. Admittedly, we're talking about high-end equipment here. Yes, if a cartridge sounds brigther, the DR values might be a little higher, but seldom more than 1dB.

About the second point, in the last months HDtracks has released numerous albums (e.g. from ZZ Top and Yes) that were clearly derived from LP cutting masters, thus allowing to compare their DR numbers with those of rips of the original LPs. I've done that comparison in about a dozen cases (I can elaborate on which ones if needed) and in each case the DR values matched, again to within about 1dB. Never have I experienced differences as large as 4dB. Those findings also apply to CDs mastered in the 80s where often LP cutting masters where used. Here I can give dozens of examples.

The track in the video is a different animal, however. The master was (judging by the waveform and the DR value, quite severly) limited, and vinyl cutting as well as playback of the resulting LP will act like a declipper, which increases the DR values. In such cases I agree that the latter can be quite misleading. In the classic rock examples I mentioned above, there is no such severe limiting present in the master, and in such cases CDs and carefully ripped LPs will give comparable DR values.

 

Be care

This was really unecessary. I never said TT Meter is valid or not valid for vinyl (I've never even used the tool, I couldn't care less), I just proved that whatever affected TT Meter could come from any kind of source file, not necessarily from a vinyl rip, and all I needed to prove that was simple logic, common sense and unbiased reasoning...

It seems I'm not the only one imagining things... many other members are asking exactly the same thing I was pointing out that you dismissed as being "stupid" but now you admit they are valid points. Oh well...

 

Thanks for the feedback, that's interesting. The trouble is, there's no way to tell about the ripping conditions, or the quality of the rip. So your experiences are good, but mine (and many other's) are not.

I disagree that the track in the video was heavily limited - it certainly wasn't clipped. Do you see the paradox of you making that judgement "judging by the waveform and the DR value" ? That's not a valid way to make this judgement, that's the point.

There are plenty of examples of discrepancies of 4dB - for example the recent Daft Punk album. And unfortunately, the low-DR-CD cases are exactly the ones where we need reliability, if the TT meter (and especially the database) were to be useful.

 

Sorry if I sounded snarky, but you didn't "prove" anything, you were just speculating. I certainly didn't call anyone stupid. Which points do you think I denied and then accepted from other people ? Yes, the TT meter can be "confused", regardless of the source, but no, that isn't a universal problem with lossless digital sources. With vinyl, it is.

 

If the bulk of DR increase comes from 'remastering' the source for LP so that the image is narrower -- and remember, this is done simply to accommodate the format -- then measurements of digital versions of the LP production master, and of the LP itself, could be similar. Though not identical -- variations in playback will change the DR further, perhaps (according to yourself ) as much or more than 1 dB.

The point , though , is that higher DR values of the LP (a typical finding) does not necessarily mean that it came from a naturally more dynamic, less compressed source, or from going back to the 'original master tape'. The LP production master -- which records the cutting room moves as well as the 'pre-mastering' moves -- encodes any moves that artificially raised the "DR" of the source (such as narrowing the image). And in this case , the move wasn't even 'meant' to do that -- the higher DR reading is a by-product.

There are certainly semantic issues here though. What is the 'natural' DR of a pop/rock recording made using overdubs, separate tracking sessions, direct signal to board, etc. There might have been no 'acoustic' event in a live space at all. So is further manipulating the DR at the cutting stage (whether intended or a byproduct of some other process), less 'natural' ?

 

If you have no information about the ripping conditions, I agree. The rippers I'm following are very forthcoming with such information though, down to which cables were used.

Yes and no. I cannot really judge sound quality from a Youtube clip, so the waveform and DR value are all I could go by. And you yourself state in the video that the track was limited, saying that it looks like a brickwall, don't you? (Sorry, I can't watch it again where I'm typing this at the moment - hope I'm not misquoting you.)

The Daft Punk album appears to be a case similar to the track you used. All CD tracks peaking at 0dB makes it clear that a limiter was applied. As I said, I agree with you that the micro-dynamics may be artificially increased on the vinyl rips in such cases. All I want to caution against is that people start believing that vinyl (and ripping the same) will always yield artificially increased dynamics. That's absolutely not the case. Check out older albums and CD editions of the same on the DR database and you will find very few - if any - such big discrepancies with vinyl rips.

Agreed.

 

I have noticed this as well. For most CDs and vinyl from the 80's I have checked out, the DR is almost identicle.

 

This is true and I've seen this myself - I can up the rating by 1 or even 2 points just by changing my cart / TT / ADC. This is also audible, I might add - so again it backs up what I'm hearing.

Now was the "better" TT setup the legitimate rating, or was it just colorizing more than the previous one? There's really no way for me to know.

This has generally been my experience as well, the numbers are usally very similar.

But they don't sound the same, at least not all the time. So a different situation where, again, the numbers don't really tell you about what's "better". Only that the dynamics seem to be similar.

 

Surely this has already been stated in this thread, but here goes - the TT meter reads digital files. It just reads digital data. It has no idea if the source was a CD or a vinyl record. If it's a file from a CD, the meter has no idea if it's a CD that was mastered from an analog tape, or a "purely digital" Pro-tools recording. If you're saying that the TT meter gives an inflated reading (one which doesn't correspond to what you actually hear) when measuring a digital file taken from a vinyl record, and one reason being that the stereo image is slightly narrower, then the same must hold true for a CD that has a narrow stereo image - that is, the TT meter reading would be inflated, and not correspond to what you hear. Thus the "suitability" of the meter for giving a meaningful reading for anything, including CDs, is questionable.
Obviously the reading for files taken from vinyl rips will vary quite a lot, depending on various factors (as stated here repeatedly).

I don't really see a need for debate on the issue... if people are going to go to the online DR database and look up some anonymously submitted info, and compare results from a Cd rip and a vinyl rip of the same record, and then come to conclusions about the mastering, and then make purchasing decisions based on that... that's just foolish. (Perhaps that point should be made over & over.) We should only pay attention to honest reviews (that's what this site is good for), people who have heard both versions (or several versions), and the more reviews the better (especially if there is a consensus on how things sound), and then maybe decide whether or not something is worth buying.

 

Agree completely.

 

You are - I said I used "what some people call" a brick-wall limiter. We discussed this earlier in the thread - if you watch the section where I zoom in on the waveform it's clear it hasn't been "brick walled".

My opinion is that it's the relatively high levels of sub-bass, phase smeared by the RIAA filtering, plus some degree of cross-talk - but either way we can agree its a common feature of modern records that doesn't necessarily mean poor sound.

Yes, and such cases are very common - the norm, in my opinion.

Agreed !

I'm happy to take your word for this - unfortunately it doesn't make the vinyl entries in the database any more useful. :-/

 

No. The reading made by the TT meter on a digital file correlates closely to what we hear - ie. what comes out of the D to A converter. But the reading from a rip of vinyl may or may not correlate - and so the two are not compatible.

Yes ! And that was why I made the video - I don't think enough people realise that.

On this, we completely agree

 

But, when you narrowed the image slightly of the CD file, and gained 2db from the meter reading, you said that it did not yield a "real" increase in dynamics.

So it logically follows from your experiment, that any Cd master with a narrow stereo image, will yield a TT meter reading that does not correlate to what you hear - that is, the stereo image of the file has a direct bearing on the TT meter measurement, but not the "real" dynamics. Based on this, I still see the meter readings as dubious, even for measuring files from CDs.

Also, you have not made an argument at all, that explains why you don't hear the increased dynamics that the meter gives for a file from a vinyl rip - we've covered what could account for the increased reading (phonographic playback equipment, things done during cutting, stereo image differences, etc), but not why exactly it doesn't correlate to what you hear - ie. what comes out of the D/A converter.

 

That point has been made over and over by me in this thread and many others in this forum. However I have been attacked by other members for making these statements. That is why I felt the need to create this post.

 

I think there's some confusion among three related, but separate, points being discussed here.

First there's the issue of repeatability: A given CD release will register identically on the TT DR meter every time. But, as Ian points out, a given vinyl LP will not register identically every time, primarily because of variations in different people's playback equipment (including, probably to a lesser degree, their A/D interfaces). As someone above pointed out, a vinyl rip is a "snapshot," and different equipment produces a different snapshot, with potentially different sound and DR values. CD rips are, by contrast, identical snapshots (putting aside certain truly micro, bit-level differences that don't impact DR and usually don't impact sound in any detectable way).

Second there's the question of comparability: Can you compare a vinyl rip's DR with a CD rip's DR? Ian and others say no, primarily because of the repeatability problem.

Third there's the question of explaining differences: Why do vinyl rips sometimes (often?) have higher DR ratings than similar CDs/digital sources? Ian suggests it might be a combination of crosstalk, stereo-image narrowing, RIAA EQ, and phase-shifting, all of which are part of the vinyl production and/or playback environment. But he's clear that he's speculating there. Others are saying they don't find his speculations convincing, because they don't always seem to hold true. But no one's trying to say that vinyl always has higher DR than digital, and no one's trying to say vinyl will always differ for certain clear or specific reasons. It's just guesswork.

Finally, to address one specific bone of contention here: crosstalk. Ian isn't saying crosstalk is the reason (or at least the only, reliable reason) LPs register high on the DR meter. He's saying crosstalk is a major reason that multiple rips of the same LP on different equipment don't sound the same or register the same DR.

If you narrow the stereo image on a CD rip, you might increase the DR. But then that DR isn't valid for that CD anymore, and you won't get the same change in DR on every single CD by narrowing the stereo image - it depends on the music and mastering.

Put it another way: by narrowing a CD rip's stereo image, you're not magically making the DR meter reliable for vinyl rips. You're just making it unreliable for the manipulated CD rip.

 

I have had a similar experience. I remember comparing my LP rips of R.E.M.'s IRS albums to the original IRS CDs, and in terms of dynamic range they were nearly identical (within +/-0.5 dB).

I think it is different for highly compressed material. For example, I did a needledrop of R.E.M.'s Accelerate and compared it to the highly compressed CD. The waveform would give you the impression of slightly greater dynamic range for the LP, but to my ears the sound was pretty similar. The LP and CD for this release actually sounded much more alike than the older IRS albums and CD whose dynamic ranges were a close match. Was Accelerate mastered from the same (compressed) source for both CD and LP? I think it is impossible to know for sure without inside info. They sure sounded similar.

In the end, I concluded that these kinds of comparisons couldn't really tell me anything I wanted to know. Absent really gross differences in dynamic range, I don't think you can conclude very much from these analog/digital DR comparisons.

 

Well I purchased the Daft Punk album on vinyl rather than digital - of course after consulting the DR database. Not sure what those numbers mean but let me tell you, this is one of the best sounding records I own!

Of course I didn't also purchase the CD or HDTracks so I can't compare.

 

One thing I don't think anyone has mentioned: if a vinyl rip has the same (or similar) DR as a CD rip of the same album, that might not mean the vinyl rip was unaffected by "vinyl issues."

Instead, it could mean that the vinyl actually has less dynamic range than the CD, and that the "vinyl issues" masked that fact by artificially pumping up the DR value. After all, if both a CD and an LP are made from the same master, the digital medium has a built-in dynamic range advantage over the vinyl medium.

 

Yes, I did. The TT meter was designed with two purposes - realtime measurements when mastering, and offline measurements to assess existing recordings.

In other words, it is intended to be used at the end of the chain. You're talking about extra post-processing on the file which can create apparent DR increases which don't correlate with audibility, but this doesn't happen when people read values directly from CD.

No, because the image narrowing is part of the process - it happens before the CD is written and any final limiting.

I'm not sure I understand. If the vinyl is genuinely more dynamic, and the meter reflects that, you will hear it. But in an example like the video, where the apparent increase in DR doesn't correlate to audibility, you don't hear it. And again, that's the problem - there's no way to tell the two apart, from the raw DR measurement.

Ian

 

Let me try this one more time - the effect I'm talking about is a limitation of the TT meter.

It doesn't measure true dynamic range, just the peak-to-RMS ratio. So it can't distinguish between real musical information, and spurious additional peak variations caused by extra processing.

Reading from a CD, or at the end of a mastering chain, it's not a big problem, but on vinyl it is, because this kind of unpredictable peak behaviour is "built in" to the format and often unavoidable, but always unpredictable.

 

You're not following the argument - probably my fault, I guess I'm not good at explaining things.
I'm not talking about any "extra post-processing". I'm talking about how the TT meter reads a digital file. The meter doesn't know whether a file has just been processed, from another previous file version which had a wider stereo image.
Imagine then, that the version you "slightly narrowed", which gave an increase in 2db on the meter, was the FINAL version, the version which was used to make the commercial CD. Now everyone out there who rips the track from the CD, runs it through the meter, will get a meter reading of DR12. But that's 2db of "fake" dynamics, according to your experiment. See? And every recording out there has different degrees of stereo imaging, due to how it was mixed and/or mastered.
And we can extrapolate from this, that the stereo image on a digital file has a direct bearing on the number that the TT meter spits out. Thus my point that the meter can give dubious number readings, for every single file (whether from CD or something else), based upon the stereo imaging.

Aha, but you're expecting us to accept that "what correlates to audibility" for you, should to the same for us. But nobody should ever assume that other people's listening experiences must equate with their own. Different people hear things differently, it's a phenomenon. Perhaps you can offer us lossless clips of the two files used in your video - from the CD, and your vinyl rip which had an apparent 4db increase in dynamics, then the rest of us can decide for ourselves if we hear it or not. Obviously, we shouldn't be expected to make critical judgements like this based on YouTube audio (not to mention the brief durations of switching back and forth between the files).

 

You're not following my argument - probably my fault, I guess I'm not good at explaining things, either !

But it wouldn't be. Stereo width processing is part of mastering, and would always be carried out before the final limiter. Otherwise, it would cause clipping. Remember, I've turned down both files to have headroom for the vinyl. So in a real situation, the apparent extra DR would be removed by the limiter.

I'm not. How is narrowing the stereo width slightly supposed to create extra dynamics ? They weren't there to being with, where are they supposed to have magically appeared from ?

Ah, in that case I have some clothes you might be interested in ! They're made of the most fabulous material ever known to man, but the catch is - you can only see them if you're really really clever...

Seriously though, you're right, different people hear different things, but in this case there's nothing to hear. And I'm not taking that on faith, I really have listened in some detail to the vinyl versus the CD - it's as a result of that listening that I decided this video was worth making.

Not sure - I'll ask the band.

 

Surely, narrowing the stereo image has (to a lesser extent) the same effect as summing to mono - sounds in the centre become disproportionately louder. That's where you get you extra dynamics from. They may have been introduced fraudulently and they may be very misleading but they're there, none the less.