IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
TT Dynamic Range meter "lies" about vinyl's dynamic rang, plus why you can't trust waveforms either
2Bdecided
post Oct 4 2013, 14:48
Post #1


ReplayGain developer


Group: Developer
Posts: 5282
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



Anyone who still clings to waveforms and DR numbers to "prove" that a particular vinyl release has a greater dynamic range (i.e. less compression) than the CD version must watch this video...

http://www.youtube.com/watch?v=n-AE9dL5FG8


Found courtesy of this thread...
http://forums.stevehoffman.tv/threads/is-t...r-vinyl.330706/


Some of the YouTube comments are what you'd expect on YouTube (=clueless!), and some of the Steve Hoffman forum comments are what you'd expect from vinyl fans. However, the video itself is very clear, and it makes a change to hear someone in the industry talk absolute sense in public.

We have had similar threads on HA where we've dissected this stuff, but this video has a very clear fact in it that we've been unable establish with 100% certainty before. I won't say what it is in this first post because that would spoil it. smile.gif

Cheers,
David.
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 15:48
Post #2





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Thanks for this.

Those who took exception to any previous comments I made on the subject can suck it note that I was correct.

BTW, the link should point to post #95. Crapple's Safari on IOS seems to linking to the next post. I haven't tried other browsers. Firefox correctly points to my post.

This post has been edited by greynol: Oct 4 2013, 17:21


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 16:05
Post #3





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Seriously, suck it!

...and by suck it, I mean buy me a beer.

EDIT: just to pile on. wink.gif

This post has been edited by greynol: Oct 5 2013, 08:10


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
bilbo
post Oct 4 2013, 16:20
Post #4





Group: Members
Posts: 191
Joined: 16-April 07
Member No.: 42593



QUOTE (greynol @ Oct 4 2013, 10:05) *
Seriously, suck it!


But how do you really feel smile.gif


--------------------
Glass half full!
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 16:21
Post #5





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Vindicated! biggrin.gif


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
krabapple
post Oct 4 2013, 16:36
Post #6





Group: Members
Posts: 2453
Joined: 18-December 03
Member No.: 10538



Bravo!
Go to the top of the page
+Quote Post
cliveb
post Oct 4 2013, 16:42
Post #7


WaveRepair developer


Group: Developer
Posts: 852
Joined: 28-July 04
Member No.: 15845



His results are certainly interesting, but I question his conclusions.

He knows for sure that the vinyl was cut from the same digital master as the CD, so he concludes that it cannot possibly have more DR. But the TT meter unambiguously shows that the vinyl has a higher peak-to-avg ratio. So in an objective sense, the DR of the vinyl *is* greater. But of course whatever that increased DR is on the vinyl, it certainly is *not* genuine signal that is missing from the CD.

Something in the process of cutting and/or playing the vinyl is artificially expanding the DR. My uninformed guess would be mechanical overshoot of an insufficiently damped cutting head. (I can't see how the playback stylus could possibly overshoot - it can't move beyond the confines of the groove).

And finally, he states the vinyl cannot possibly sound better than the CD. But in my book, it's entirely plausible that this artificially expanded DR may sound more pleasing. Or less pleasing, of course. The vinyl's extra DR is some kind of distortion, for sure, but it could be euphonic.

The question to ask is whether this characteristic is typical. Do *all* vinyl cuts tend to expand the DR of the master they are made from, or is it peculiar to this particular example?

In summary, this youtube posting proves nothing - it merely adds a further complication. The plot thickens.

EDIT: sorry - it *does* prove one thing: that if the DR values for CD and vinyl releases are different, it does not necessarily show that they were cut from different masters.

This post has been edited by cliveb: Oct 4 2013, 16:43
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 16:50
Post #8





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (cliveb @ Oct 4 2013, 08:42) *
overshoot

Nah, try running a square wave through a high-pass or all-pass filter.

QUOTE (cliveb @ Oct 4 2013, 08:42) *
is it peculiar to this particular example?

Based on the preponderance of the evidence I have seen? No!

QUOTE (cliveb @ Oct 4 2013, 08:42) *
if the DR values for CD and vinyl releases are different, it does not necessarily show that they were cut from different masters.

...and that differences in waveforms do not necessarily show that they were cut from different masters.

AFAIC, this is all that matters.

Slam dunk!

This post has been edited by greynol: Oct 4 2013, 17:01


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
edwardar
post Oct 4 2013, 17:14
Post #9





Group: Members
Posts: 100
Joined: 8-July 04
Member No.: 15139



Slightly off topic, but possibly linked. I noticed this recently:

I took a mildly clipped CD (Brian Wilson - Live at the Roxy), ripped to wav and halved the volume in Audacity. I then compressed this file as a LAME MP3. The MP3 had a DR of 1 or 2 higher than the source, though I could not hear a difference when compared with the original at matched volume.

[My unfounded speculation is that it's possible the DR meter takes into account very brief peaks in amplitude which contain minimal audible information ("energy"?).]

Edit: oh, and the MP3 waveform 'looked' more dynamic.

This post has been edited by edwardar: Oct 4 2013, 17:15
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 17:16
Post #10





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



Not off-topic at all and yes, I think that's absolutely correct. You could have also decoded to 32-bit float PCM an mp3 created from the original source without reducing the level.

This post has been edited by greynol: Oct 4 2013, 17:19


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
Engelsstaub
post Oct 4 2013, 17:28
Post #11





Group: Members
Posts: 566
Joined: 16-February 10
Member No.: 78200



Since I posted that thread (one of two that was linked to earlier) I think I also learned quite a bit about the TT DR Meter. I even learned in that thread from 2Bdecided that some excessive phase-nudging can dramatically change the measurement.

I noticed that Ian Shepherd (a person I've come to respect) mentioned in that video that it was a "blunt instrument." Beyond my initial impressions I think I've always viewed it as such. The problem is is that it's the best (though imprecise) instrument I"m aware of. Obviously I can't tell anyone here that I hear or perceive more dynamic range or something because, within this forum, that is rightfully irrelevant and considered subjective.

I've understood (rightly or wrongly) that phase-nudging, EQing, etc can have a moderate effect on that DR score. I'm not totally convinced that more than 3 or 4 decibels is just chalked up to some cartridge or whatever. I could be wrong.

The only way we can be certain that a vinyl master is different is if the record company gives us such info. When I've requested it in the past some always do (and don't always give me the answer I want to hear either) but many others don't even respond to such emails.

...but then again: some brief reply from a record company saying something to the effect of "we understand that the engineer submitted a separate master for the vinyl. Thanks for your interest in our products!" could very well mean "yeah, the engineer EQed the same brickwalled recording slightly differently for the vinyl." Really all we have is subjective observations when comparing to vastly different mediums like CD and LP. (And yes: CD will obviously always win in precision against an archaic but still loved format like vinyl records.)


--------------------
The Loudness War is over. Now it's a hopeless occupation.
Go to the top of the page
+Quote Post
cliveb
post Oct 4 2013, 17:29
Post #12


WaveRepair developer


Group: Developer
Posts: 852
Joined: 28-July 04
Member No.: 15845



QUOTE (greynol @ Oct 4 2013, 16:50) *
QUOTE (cliveb @ Oct 4 2013, 08:42) *
overshoot

Nah, try running a square wave through a high-pass or all-pass filter.

It seems that you have interpreted my use of the term "overshoot" as suggesting that the cutting head might "interpolate" a clipped wavetop. That is not what I was suggesting at all.

What I am wondering is when there is a (non-clipped) peak with a fast rise-time in the signal, it's possible that the cutting head might fail to stop its lateral excursion at the correct point, and continue on for a while (due to a combination of inertia and insufficient damping). This would result in bigger peaks, and hence a greater peak-to-avg ratio. But I must stress that this is pure speculation on my part.
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 17:31
Post #13





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



I wouldn't be surprised if the math involved in characterizing the mechanical mechanism and the the filter response turned out to be the same.

With that said, strike the "Nah" in my reply.

This post has been edited by greynol: Oct 4 2013, 18:00


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
krabapple
post Oct 4 2013, 17:53
Post #14





Group: Members
Posts: 2453
Joined: 18-December 03
Member No.: 10538



QUOTE (cliveb @ Oct 4 2013, 11:42) *
His results are certainly interesting, but I question his conclusions.

He knows for sure that the vinyl was cut from the same digital master as the CD, so he concludes that it cannot possibly have more DR. But the TT meter unambiguously shows that the vinyl has a higher peak-to-avg ratio. So in an objective sense, the DR of the vinyl *is* greater. But of course whatever that increased DR is on the vinyl, it certainly is *not* genuine signal that is missing from the CD.


Right, and that's what he's saying.

QUOTE
And finally, he states the vinyl cannot possibly sound better than the CD. But in my book, it's entirely plausible that this artificially expanded DR may sound more pleasing. Or less pleasing, of course. The vinyl's extra DR is some kind of distortion, for sure, but it could be euphonic.


I'm pretty sure he agrees , either in the video or the comments below it. He certainly says , circa 12:25, that if you like the sound of vinyl, you may well prefer the sound of the vinyl version of this track, regardless of how its DR was generated.



QUOTE
In summary, this youtube posting proves nothing - it merely adds a further complication. The plot thickens.

EDIT: sorry - it *does* prove one thing: that if the DR values for CD and vinyl releases are different, it does not necessarily show that they were cut from different masters.



Exactly. And if it's the case that the masters were the same, then....the 'extra' DR is entirely due to either artifactual or intentional post-processes, not a natively higher DR of the source. This is logically obvious, but I think having 'proved' it this way is a pretty significant thing!
Go to the top of the page
+Quote Post
2Bdecided
post Oct 4 2013, 18:06
Post #15


ReplayGain developer


Group: Developer
Posts: 5282
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



Once you square-wave / clip something, it's really hard to do anything to it without raising the measured peak-to-average level.

Anyone remember this noise clip?
http://web.archive.org/web/20050209010438/...3.net/norm.html
(about 1/4 down the page)

Many modern CDs are just pathological examples of this. A low pass filter, a high pass filter, proper sub-sample interpolation, and an all pass filter - all these things will raise the peak-to-average ratio compared with calculating it for the raw samples. Yet the sound doesn't change at all - it's subjectively the same signal. TT DR is measuring numbers, not sounds.


I think the easiest way to compare like-for-like with an "objective" "DR" tool is to run the CD version through an all-pass filter first and use that as the baseline against which you compare the vinyl version. Or just assume, if the CD looks clipped, that there's 3-4dB more energy in the clipped peaks just waiting to burst out the moment you fiddle with them.

Cheers,
David.

P.S. Glad you like this greynol. Many of us suspected this, and pointed to identical clipping on CD and LP to suggest the same master was used, but it's nice to get it confirmed by a mastering engineer: "exact same master, 4dB higher DR measurement on vinyl, no actual improvement".

This post has been edited by 2Bdecided: Oct 4 2013, 18:08
Go to the top of the page
+Quote Post
greynol
post Oct 4 2013, 18:13
Post #16





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



http://www.hydrogenaudio.org/forums/index....showtopic=66401
http://www.hydrogenaudio.org/forums/index....showtopic=68641


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
Engelsstaub
post Oct 5 2013, 06:43
Post #17





Group: Members
Posts: 566
Joined: 16-February 10
Member No.: 78200



In general and for the sake of constructive argument I have a few earnest/non-rhetorical questions:

Did anyone really not understand before that a 3-to-4 dB difference with the TT meter could be attributed to vinyl playback itself? (That was somewhat rhetorical but seriously: discuss if possible.)

How do you think this would play out if a non-dynamic master was used on both the CD and the LP? (e.g. all of my needle drops that have really bricked masters have very similar looking waveforms and are usually within ~1 dB with that meter.) Ian Shepherd is known to engineer decent recordings in a time when most do not. His example was a pretty dynamic CD mix.

Semi-related: I was reading elsewhere where Mr. Shepherd said that Nine Inch Nails' newest album "Hesitation Marks" on vinyl sounded subjectively better than either the CD or (controversially) so-called "audiophile" release. I think I can submit examples of each and we can look at that one example and observe differences the TT meter "detects/"

Thoughts?


--------------------
The Loudness War is over. Now it's a hopeless occupation.
Go to the top of the page
+Quote Post
krabapple
post Oct 7 2013, 01:14
Post #18





Group: Members
Posts: 2453
Joined: 18-December 03
Member No.: 10538



Rather interesting experiment done by Ian Shepherd, identifying a likely cause of the DR discrepancy:

http://forums.stevehoffman.tv/threads/is-t...11#post-9481216

This post has been edited by krabapple: Oct 7 2013, 01:15
Go to the top of the page
+Quote Post
krabapple
post Oct 7 2013, 01:18
Post #19





Group: Members
Posts: 2453
Joined: 18-December 03
Member No.: 10538



QUOTE (Engelsstaub @ Oct 5 2013, 01:43) *
How do you think this would play out if a non-dynamic master was used on both the CD and the LP? (e.g. all of my needle drops that have really bricked masters have very similar looking waveforms and are usually within ~1 dB with that meter.) Ian Shepherd is known to engineer decent recordings in a time when most do not. His example was a pretty dynamic CD mix.



But that's the funny thing. To many audiophiles of the sort that populate the Hoffman board, the CD waveform on Ian's video would be considered 'brickwalled' and 'non-dynamic'.

This post has been edited by krabapple: Oct 7 2013, 01:18
Go to the top of the page
+Quote Post
Engelsstaub
post Oct 7 2013, 01:29
Post #20





Group: Members
Posts: 566
Joined: 16-February 10
Member No.: 78200



Now this is getting pretty interesting. Thanks for the link!

I, for one, appreciate Ian's determination to get to the bottom of all of this. Does anyone think there will ever be a decent way to measure the dynamic range of a record?

I was hesitant to ascribe more than a 3-4 dB increase in vinyl measurement to the other reasons (or combinations thereof) mentioned. Yesterday I found a needledrop of mine that actually "measured" a full 5 dB greater than the CD. I'm certain they're the same master. Not only does it sound like it (the record actually sounds a little worse) but the waveform from the vinyl looks almost exactly the same as that of the CD. No real or artificial peaks.


--------------------
The Loudness War is over. Now it's a hopeless occupation.
Go to the top of the page
+Quote Post
markanini
post Oct 7 2013, 11:15
Post #21





Group: Members
Posts: 555
Joined: 22-December 03
From: Malmö, Sweden
Member No.: 10615



QUOTE (krabapple @ Oct 7 2013, 02:14) *
Rather interesting experiment done by Ian Shepherd, identifying a likely cause of the DR discrepancy:

http://forums.stevehoffman.tv/threads/is-t...11#post-9481216

So Ian was able to repoduce the higher DR measurement by narrowing stereo width of the master in his DAW. laugh.gif
Go to the top of the page
+Quote Post
Engelsstaub
post Oct 7 2013, 20:48
Post #22





Group: Members
Posts: 566
Joined: 16-February 10
Member No.: 78200



Someone in the Hoffman thread just brought up a fair question IMO and I'll paraphrase it in the following paragraphs:

Why do all (by all I'm speaking from my personal experience thus far as well) the eighties LPs measure almost exactly the same as their CD counterparts? Both are equally dynamic but for some reason always have pretty much the same DR "value" as one another.

...these records should, in theory, be "scoring" 3-4 dB higher on that meter as well. The records have the same issues of crosstalk, soundstage, etc. that we're taking Ian's word is changing the the "value" of his own recording.

(None of my musings here should be construed as me once again taking the position that the DR meter is reliable for vinyl rips.)


--------------------
The Loudness War is over. Now it's a hopeless occupation.
Go to the top of the page
+Quote Post
xnor
post Oct 7 2013, 21:09
Post #23





Group: Developer
Posts: 965
Joined: 29-April 11
From: Austria
Member No.: 90198



Why does anyone use the TT DR Meter? Where is the algorithm specified? Why isn't it based on BS-1770?

It seems like it is pretty much completely useless.


--------------------
"But you refuse to [..] back your claims of inaudibility."-amirm
Go to the top of the page
+Quote Post
greynol
post Oct 7 2013, 21:27
Post #24





Group: Super Moderator
Posts: 10257
Joined: 1-April 04
From: San Francisco
Member No.: 13167



I couldn't agree more.

In addition to that, and while this may be extraneous to your point:
ABXing vinyl from its digital master is generally assumed. When armed with the knowledge of "DR" figures, differences generally get attributed to perception of differences in dynamic range. When those specific differences don't exist, but the obvious unrelated tells do, expectation bias can (and often does) prevail. This is especially true when the imagined difference supports the point being argued.

Back to the original topic, does anyone want to add to the wiki article on vinyl myths?

This post has been edited by greynol: Oct 7 2013, 22:33


--------------------
Your eyes cannot hear.
Go to the top of the page
+Quote Post
DVDdoug
post Oct 7 2013, 23:12
Post #25





Group: Members
Posts: 2676
Joined: 24-August 07
From: Silicon Valley
Member No.: 46454



QUOTE
Why do all (by all I'm speaking from my personal experience thus far as well) the eighties LPs measure almost exactly the same as their CD counterparts? Both are equally dynamic but for some reason always have pretty much the same DR "value" as one another.

...these records should, in theory, be "scoring" 3-4 dB higher on that meter as well. The records have the same issues of crosstalk, soundstage, etc. that we're taking Ian's word is changing the the "value" of his own recording.
I believe it's the lack of excessive compression (or lack of excessive compression). If you take an "pure" uncompressed recording of an orchestra and run it through an all-pass filter, the DR "measurement" or crest factor should be unaffected, because all of the frequency-components sum-up randomly either way.

But a compressed/limited file has been limited wherever the frequency-components "randomly" sum-up to exceed the threshold/limit. When you all-pass, the timing of the various frequency components is "randomized" again...
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 28th November 2014 - 12:26