And The Winner Is …

JR_Audio · March 3, 2014

Track 10 and 11 of the 24 96 High-Res are from 44k1 MP3

Besides the high compression for this kind of music, where the bass drum kicks extremely often into hard limiting and the bass notes most times also, I have recognized, that for example the Track 10 and also Track 11 of the 24 Bit 96 kHz High Res Version are 100 % from 44k1 MP3. They show all the typical artifacts of modulating a 16 kHz low pass filter with level. So basically, here we have MP3s sold as 24/96 High Res. All other High Res Tracks are based on 44k1 sources with the same DR and Bandwidth as with the CD release.

Juergen

Skeptic · March 3, 2014

Here are two posts from another thread that comment on the DR issue and some of the possible concerns.

The "DR" tool is a measure of peak-to-loudness ratio, which is a good indicator of how "dynamic" a track is.

It is not a measure of dynamic range, which is the difference between the loudest and the quietest parts of a track.

JRiver's file analysis includes both Dynamic Range (DR) and Dynamic Range (R128) measurements. R128 is an updated spec based upon BS.1770.

While the R128 information is useful to have, the DR tool seems to be a much better indicator of how compressed a track is.

JRiver also includes R128 "True Peak" measurements (inter-sample peaks) rather than relying on sample peaks, and will use this measurement with the Volume Leveling or Peak Level Normalization features to prevent inter-sample clipping during playback.

Yes, Adobe Audition seems to be a much better tool. DR-Meter is great tool - or rather has the potential to be- but people just need to be aware of the serious limitations of "Official" DR values from DR-Meter. It could be misleading.

It gives the vinyl experience a DR rating of 9 and HD Tracks a rating of only 6 despite the fact that the HD Tracks version has a whopping 42dB more of DR.

You need to use statistically weighted values for the dynamic range. (R128/BS.1770) There are things which can make a "dumb" dynamic range meter register very high values which don't correspond to the music.

Archimago · March 3, 2014

Track 10 and 11 of the 24 96 High-Res are from 44k1 MP3

Besides the high compression for this kind of music, where the bass drum kicks extremely often into hard limiting and the bass notes most times also, I have recognized, that for example the Track 10 and also Track 11 of the 24 Bit 96 kHz High Res Version are 100 % from 44k1 MP3. They show all the typical artifacts of modulating a 16 kHz low pass filter with level. So basically, here we have MP3s sold as 24/96 High Res. All other High Res Tracks are based on 44k1 sources with the same DR and Bandwidth as with the CD release.

...

Indeed Juergen!

I put up a similar post last night on the blog after receiving some info on this. Nice to see someone else confirming it as well...

Most unfortunate.

firedog · March 3, 2014

Follow up on JRAudio's post here in this article:

Beck's <i>Morning Phase</i>: Is the 24/96 Download Worth It? | AudioStream

junker · March 3, 2014

Synfreak,

Maybe you should "read-read the informations again" and check your software. The numbers are all posted by Chris in the article. Can you please comment when you have some concrete numbers and examples provided from the article and the Audition data in his screenshots. Thanks.

Junker,
"sorry but I can not follow"

"i haven't checked with my software"

junker · March 3, 2014

Yep, there doesn't seem to be and HD content >22KHz on the posted screenshot in the article. Disappointing. Would HDTrack make a refund? If not, maybe have a talk with your CC company since the product they sold wasn't disclosed or represented as being MP3 quality.

The Computer Audiophile · March 3, 2014

Yep, there doesn't seem to be and HD content >22KHz on the posted screenshot in the article. Disappointing. Would HDTrack make a refund? If not, maybe have a talk with your CC company since the product they sold wasn't disclosed or represented as being MP3 quality.

The files are 24/96 even if the original music used to create the files was 44.1 MP3. I don't think HDtracks can guarantee sound quality of anything. In fact, I don't want any retailer to judge sound quality for me.

Arg · March 4, 2014

I love the concept behind the DR DB but it's too bad the algo is so flawed.

What is so flawed about the "DR" algorithm?

I suspect that it is this issue that is causing some to think the measuring tool is imperfect: Why the TT Meter doesn't work on vinyl - YouTube

That's right, not only is DR useless for vinyl vs digital analysis as proven by Ian Shepherd above (and let's face it, the "Vinyl Exp." is a needle drop), but it is also useless for digital vs digital comparisons. I realised this when I saw Ian Shepherd comment somewhere in a discussion of his video above, that when he simply varied the pan width of the digital edition by a small degree, the DR number jumped by four points! That's bad, really bad.

So I suggest the DR scores could be removed from the article and it would be less misleading.

In terms of process, it seems that Chris did the objective analysis before the subjective analysis, which is unfortunately biased because he would have known the (misleading) DR scores and seen the waveform plots before listening.

Le Concombre Masqué · March 4, 2014

Originally Posted by InfernoSTi

I suspect that it is this issue that is causing some to think the measuring tool is imperfect: Why the TT Meter doesn't work on vinyl - YouTube

HIGHLY RECOMMENDED

firedog · March 4, 2014

Well, based on lots of experience, the DR meter results seem to confirm what I hear. So I'm pretty sure that at least for comparisons of digital files it is accurate enough to be useful. I wouldn't give a lot of weight to DR differences of "1", but when I buy a remaster that sounds heavily compressed and check the DR, I find the result is much lower than the previously remasatered CD that sounds fine to me.

Likewise, when a hi-res remaster sounds much better and more dynamic than the corresponding CD, the sound is confirmed by the DR results.

In addition, just looking at the waveforms shows that the DR results are meaningful. Perhaps not as precise as you would like, but still quite useful.

That's right, not only is DR useless for vinyl vs digital analysis as proven by Ian Shepherd above (and let's face it, the "Vinyl Exp." is a needle drop), but it is also useless for digital vs digital comparisons. I realised this when I saw Ian Shepherd comment somewhere in a discussion of his video above, that when he simply varied the pan width of the digital edition by a small degree, the DR number jumped by four points! That's bad, really bad.

So I suggest the DR scores could be removed from the article and it would be less misleading.

In terms of process, it seems that Chris did the objective analysis before the subjective analysis, which is unfortunately biased because he would have known the (misleading) DR scores and seen the waveform plots before listening.

Le Concombre Masqué · March 4, 2014

Why the TT Meter doesn't work on vinyl - YouTube initially Posted by InfernoSTi is very worth watching

junker · March 4, 2014

Good link. Thanks!

Yeah, the RMS loudness value only being derived from the top(loudest) 20 average RMS measurements is so utterly arbitrary. Is this algo R128 weighted or something ad hoc? Is the code open-source and vetted in the public domain? I don't see how a fixed value of 20 applies to long classical tracks, or quiet, dynamic recordings that have 20 loud passages, etc.

Skeptic · March 4, 2014

It seems that Junker was misinformed about what the "DR" tool does (it doesn't measure "dynamic range" at all, for one thing) and I would not agree that it is fundamentally flawed.

I do agree that you cannot use it with vinyl recordings, or at least you cannot compare those results to digital files, but that does not mean the tool is useless. Like any tool, you have to know when and how to use it.

This is part of the reason why I asked if anyone knows how the "Vinyl Exp" tracks were sourced; whether they are actually a recording, or files processed to sound like a record. Even that processing could be producing invalid results.

Using the DR tool, we can confirm that all the current digital releases of this album are terribly compressed.

We cannot definitively say that the "Vinyl Exp" MP3 files have much better dynamics using this tool, but listening tests suggest that this is the case - or at least, that they are more pleasing to listen to.

Synfreak · March 4, 2014

Exactly what I have tried to show to @junker in my last post.

There seem to be a lot of misunderstanding of the technical processes used in recording and mastering to different media (digital versus vinyl in our case).

I do agree that you cannot use it with vinyl recordings, or at least you cannot compare those results to digital files, but that does not mean the tool is useless. Like any tool, you have to know when and how to use it.

+1

The DR numbers for vinyl shouldn't be compared _directly_ to the ones of digital files, because vinyl mastering uses some (needed) processing which can alter the DR value to some extent, without actually changing the compression (or limiting) of the recording.

These DR tools should be seen as a method to put numbers in relation to what/how our ear-brain system reacts to the sound.

JR_Audio · March 4, 2014

DR Meter is not misleading and is doing exactly what it intended to do

But first some words about the music. Yes, the mood and the melody of the lead vocal is good, without doubt, so from the musical side, there are some good amount of pros.

And secondly, as a background information: I am an audio engineer, own a recording studio and do also make music, so they may be at bit more than just using Audicity (that I don't use)).

But mainly, I try to bring a bit orientation in this discussion about the DR Meter and the Vinyl Experience mix. But next two comments first, about two points, that appeared here:

1. The comment / video that I have read / seen here, about DR Meter and Vinyl is only partly right but also partly wrong. When you give a peak limited wave file to a record plant, they will look into the file and mostly optimize the file a bit to better match the possibilities and limitations of the record because you can't simply press a LP to have the same constant peak limits as with the wave file. A record can't have this highly compressed content, that you can have on a CD (or High Res Files). In this case it doesn't matter, that you gave the same file to the Vinyl plant as to the CD plant. When the outcome after pressing the CD and Vinyl is different (for whatever reasons), then the sound is different and so the measurements are different as well.

2. When I change the stereo width pot during mastering, this has not only a great impact on the DR Meter, it has for most a great impact on the perceived sound. The Mid / Side ratio has an immense part of the perceived loudness and sound pressure.

But now back to the discussion about the difference of the Vinyl Experience Mix and the “official” CD / MP3 / High-Res Mix. Fact is, that both mixes differs in sound and so also differs in DR Meter. Next some details on that.

For better comparison, I have normalized both files to 0 dBFS max level (I could have also normalized both files to the same perceived loudness level, but I wanted to start “easy”, so normalized the peaks).

Waveform: When you comparing the waveform, then it is clearly visible, that both files differs in the overall loudness and in dynamic (taken into account that both have now the same maximum peak level (and the peaks are music, not clicks …)).

DR Meter: When looking at the DR Meter of the track 03 (to what I refer here), then the official release does have DR06 and the Vinyl Experience has DR11 with 5 dB more “Dynamic” (so there is a sort of correlation, between looking at the wave form and this number).

LU Loudness Level: Next I compare the Loudness Level of both files. The official release shows up LU09, where the Vinyl Experience LU14. This means that the official release would be 5 dB louder (after normalization), or the other way around, taking into account that they have now the same maximum peak level, the official release does have 5 dB less dynamic.

Histogram: Or when comparing the histogram of both files, to see which levels do occur how often or for the most time, I do see, that with the official release, the histogram peaks at – 12 dBFS and that of the Vinyl Experience at – 17 dBFS (both having the same normalized peak level during this comparison). So also here 5 dB difference.

Fact is, that both files do sound different and so they measure different. At the above example, I have compared both with set to the same maximum peak level, but it would be even more fair, to adjust them to have the same overall loudness level (LU). So when doing this, both tracks sound similar loud and then I can clearly (and fairly) hear, that the official release is much more compressed than the Vinyl Experience release.

I am sorry for this length of explanation, but I wanted to bring some light into these points.

But from the mixing perspective, in my opinion this mix is made and intended mainly for EarPods listening on the go. I personally (and this is my opinion, every one has his different one and I do accept others, so you should too) can't listen to the tracks on a big system, because the low bass drum and bass are by far too loud. The acoustic guitar is halve as big as my living room and the vocal is big and flat in the face.

So at the end one more time: This is not about the music, not about the taste of the music, my main reason for writing this was to explain some technical points about DR Meter numbers.

Juergen

Synfreak · March 4, 2014

Thanks Juergen!

Nice to see you contributing some of your knowledge here again!

The Computer Audiophile · March 4, 2014

So at the end one more time: This is not about the music, not about the taste of the music, my main reason for writing this was to explain some technical points about DR Meter numbers.

Juergen

Thanks Juergen!

InfernoSTi · March 4, 2014

Here is my concern about DR as a tool. Look at this wonderful song by Wilco with a DR rating DR5. It should be a flattop worse than the Beck album we are discussing. Yet it isn't.

DR5 -0.08 dB -9.09 dB 01-Misunderstood.aif

(from this discussion: http://www.computeraudiophile.com/f13-audiophile-downloads/new-wilco-hdtracks-17560/#post256024 )

Please explain how the DR5 is a meaningful measure of what the song Misunderstood sounds like from a dynamic range perspective. I believe it is entirely misleading to call this a DR5. The "good" parts should be a DR12 or greater. Truly an artistic choice and a poor tool to use to measure the quality of the song.

I am open to understanding how the loudest 20% of this song represents the entire song, the dramatic nature of the loud vs. soft passages, etc.

Please help me believe this tool is worth using as a measure of this song…

Best,

John

JR_Audio · March 4, 2014

Just read the papers about ITU-R BS.1770 and EBU R128, how the loudness meter was been developed over the years. And the AES Group and also TC and other companies and institutions have putten a lot of efforts (analyzing centuries of records) for that and they are clear reasons, why it works that way.

Juergen

InfernoSTi · March 5, 2014

Just read the papers about ITU-R BS.1770 and EBU R128, how the loudness meter was been developed over the years. And the AES Group and also TC and other companies and institutions have putten a lot of efforts (analyzing centuries of records) for that and they are clear reasons, why it works that way.

Juergen

If you recall, I'm the one who brought forward the link to the ITU-R BS.1770 standards back in post #49. But I'm not sure a reading of the standards is going to answer the question I asked.

What I am asking is if this tool is used by people to judge the quality of the mastering AND it sometimes is completely misleading in that regard, how can I tell when it is a "good" tool and when it is a "bad" tool for judging the quality of the mastering?

And if the answer is "you can't, it won't tell you that information due to the standards it was designed to" then I can't rely on the DR value to be a clear indicator of a poor mastering.

Perhaps I am weary of the constant rattle of "it has a low DR so it sounds bad" or "it has a high DR so it sounds good" when either one of those situations may simply not mean a thing (as in the specific example posted above).

I also wonder if all the clicks and pops that were added to the "vinyl" experience were removed, if it would have the same DR as the regular edition?

Certainly I don't want to listen to those clicks and pops as I would find them completely gimmicky and the worst kind of marketing. I wouldn't be convinced that they were artistically intentional but were added by the marketing group to be "vinyl cool" or some nonsense.

John

JR_Audio · March 5, 2014

Three points:

In the Vinyl Experience, there are no clicks and pops that would misleading the DR values. Ready my comments about the DR value of the Vinyl Experience above.

With some decade of experience (similar to the people who have worded on the Loudness Measurement), I personally can confirm a strong relation between the numbers and how the loud sequences are compressed and locking at the above graph, I do clearly see, even without DR Meters, that the loud parts (mostly refrains) are heavily squeezed, even when in between there are much more quieter parts. If you read the papers (and I am a member of AES for nearly 30 years and attending every AES convention) you will see the reasons, why exclude the quiet parts, and why exclude the very loud parts and why pick up mainly the "refrain" parts.

But don't mix up DR values with Sound. I have said this also above. So what I was trying to explain with DR values, came from the technical side. And what I was saying about the sound, was from the listening side. Both can correlate, but you can't go from the number and judge the sound.

But want I stand for is that with lower compression, you have better chances to have an open and transparent mix and with high compression it is much more difficult, to bring all the information of all instruments into it. This doesn't mean lower compression equals better sound, but the chances are better for that.

And another important part for me is, that higher compression does stress the ear / brain much more than lower compression and I do like (and have to) listen to music for several hours and high compression makes me stressed and tired after that, where lower compression don't.

Juergen

InfernoSTi · March 5, 2014

Three points:

In the Vinyl Experience, there are no clicks and pops that would misleading the DR values. Ready my comments about the DR value of the Vinyl Experience above.

With some decade of experience (similar to the people who have worded on the Loudness Measurement), I personally can confirm a strong relation between the numbers and how the loud sequences are compressed and locking at the above graph, I do clearly see, even without DR Meters, that the loud parts (mostly refrains) are heavily squeezed, even when in between there are much more quieter parts. If you read the papers (and I am a member of AES for nearly 30 years and attending every AES convention) you will see the reasons, why exclude the quiet parts, and why exclude the very loud parts and why pick up mainly the "refrain" parts.

But don't mix up DR values with Sound. I have said this also above. So what I was trying to explain with DR values, came from the technical side. And what I was saying about the sound, was from the listening side. Both can correlate, but you can't go from the number and judge the sound.

Juergen

Thank you, Juergen,

I don't doubt your credentials and that you know more than I ever will on this topic. And that you take the time to post here is not taken lightly and I, along with many many others, appreciate you very much.

On your first point, it seems that people are reporting the addition of record surface noise, often referencing this as a needle drop. Here is one such quote from somebody describing the noise that was added as "unmistakable":

The Vinyl EQ version has unmistakable record surface noise especially audible in the beginning of the first track. I listened to the NPR stream multiple times. I listened to the beginning more than I wanted too because I lost connection a few times and the stream would start over at the beginning I don't recall hearing the surface noise in the stream.

Perhaps I don't understand your analysis of why those unmistakable record surface noises aren't meaningful. I will reread your posts a third time as I am dense.

1. The comment / video that I have read / seen here, about DR Meter and Vinyl is only partly right but also partly wrong. When you give a peak limited wave file to a record plant, they will look into the file and mostly optimize the file a bit to better match the possibilities and limitations of the record because you can't simply press a LP to have the same constant peak limits as with the wave file. A record can't have this highly compressed content, that you can have on a CD (or High Res Files). In this case it doesn't matter, that you gave the same file to the Vinyl plant as to the CD plant. When the outcome after pressing the CD and Vinyl is different (for whatever reasons), then the sound is different and so the measurements are different as well.

snip

So when doing this, both tracks sound similar loud and then I can clearly (and fairly) hear, that the official release is much more compressed than the Vinyl Experience release.

I am sorry for this length of explanation, but I wanted to bring some light into these points.

But from the mixing perspective, in my opinion this mix is made and intended mainly for EarPods listening on the go. I personally (and this is my opinion, every one has his different one and I do accept others, so you should too) can't listen to the tracks on a big system, because the low bass drum and bass are by far too loud. The acoustic guitar is halve as big as my living room and the vocal is big and flat in the face.

So at the end one more time: This is not about the music, not about the taste of the music, my main reason for writing this was to explain some technical points about DR Meter numbers.

OK, I did and you are saying that with the same master, that a record can't be "as compressed" as a CD due to the manufacturing/playback of that. You then say that the analysis shows, clicks and pops notwithstanding, the digital CD mix is more compressed than the "vinyl" mix. Even though they are both digital releases, really. Meaning that either they are different masters or that the vinyl release is really a needle drop and they may have started from the SAME master but ended up with different DRs.

Your second point is exactly why I am asking, "why is a tool that fails to show something meaningful on a track like Misunderstood so good?" I got a lot more information from the spectrograph than I did from a DR5 rating. In fact, the DR5 was really misleading as about 50% of the song is DR11 and the compressed parts are absolutely artistic choice (the music is a wave of sound that crushes you in sequence with the lyrics of a person feeling crushed).

What I am really asking here is why is the 20% applicable to high quality music files where we may want to know the softest to loudest range more than the 20% loudest to get a real idea of the compressed nature of the whole song. This isn't a TV commercial we are trying to regulate that is compressing to be heard while you are in the kitchen making a snack while watching a show.

I agree with you on the third point, you have to listen to know if it is "good" or "bad" sound.

Thank you again, I do appreciate your expertise.

Best,

John

JR_Audio · March 5, 2014

Just a speculation about the Vinyl Experience Mix from my side:

In the Vinyl Experience Mix, the kick drum and the bass guitar stands much more out of the mix and the mix as total has more space for the instruments, so I am speculating, that this was done (maybe) from a different mastering stage as the final CD was made.

But this is speculation, because as you can read above is, that for vinyl pressing, you have to most times optimize (= change) the mix to fit better to the limitations of the vinyl (and these needed changes are higher with higher compression and lower, for lower compression).

PS: Please read also my edited comment above about the last two edit paragraphs. Thank you.

Juergen

InfernoSTi · March 5, 2014

Thank you. I just read your edited comments.

I want to say that I am in no way saying I prefer overly compressed mastering. Perhaps it has come across that way. I do not want to listen for hours to overly compressed music.

My concern is that sometimes DR doesn't correspond to my actual listening experience for a variety of reasons. Sometimes it is a high number that still sounds bad or a low number that still sounds good, sometimes a mixed number such as above where you are picking between two equally valid extremes within the same track, and if it is a vinyl track for a number of reasons (clicks/pops, revised mastering, or something in the cutting/playback for unchanged masters).

It isn't just one case that it is a poor tool. I suppose it is like a really good 10 lb sledge hammer. Perfect for the right job and very useful. Terrible for tapping in finishing nails…and it isn't the finishing nails fault.

Thanks a bundle…

John

P.S. Here is what might be a better tool to use…thanks to jlohl for posting it: http://www.computeraudiophile.com/f14-music-analysis-objective-and-subjective/easy-analysis-tool-trackalyzer-19674/#post303728

JR_Audio · March 5, 2014

When recording, mixing and mastering, I have always SpectraFoo running (on a second Mac). You can see an example of an image of SpetraFoo at my post (in this blog number 51). I can have several meters running with different ballistic settings and what is also very nice, the phase torch, where I see the phase relation between left and right (this is easy), but versus frequency range (and this is unique).

And as for the dynamic range meter, for example for the file (song) Misunderstood, you can have the online version (plug-in) of that DR meter running, so you can see the dynamic range separately, for each section of the song. But I do not use the dynamic range meter for mixing or mastering. I just use the K14 (or K12) scale and setting of the "Katz-Meter" (Bob Katz) and this works great.

Juergen

And The Winner Is …

User Feedback

Recommended Comments

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites

Share this comment

Link to comment

Share on other sites