Redbook x HDtracks dynamic range comparison

[junker]

The TT Dynamic Range algorithm is flawed and that single number in and of itself is not very insightful. I would highly recommend that it only be used to compare different versions of the same recording. Other than that it is unfortunately next to useless.

[Julf]

The TT Dynamic Range algorithm is flawed and that single number in and of itself is not very insightful. I would highly recommend that it only be used to compare different versions of the same recording. Other than that it is unfortunately next to useless.

 

I am not sure what algorithm that is. The established industry standards are EBU Tech 3341 and 3342, covered by EBU recommendation R 128.

[junker]

Sorry, I don't know what any of that means, but they compare the 20 highest RMS measurements to the peak and FLOOR the difference:

 

 

You could have a recording where most of it is quiet, but has 20 very loud passages resulting in a low official DR value. It is also biased against longer recordings.

 

Here is an example with Hans Zimmer - The Man of Steel that I uploaded a few days ago. It ranged from DR 16 to 7. Very, very dynamic recording with a track of almost 30 minutes. I don't feel the metric is very helpful in this case. It is after all a single number. Now if they calculated the absolute RMS value for every sample and compared that to the peak then it would be helpful in comparing different recordings.

 

http://www.dr.loudness-war.info/details.php?id=41057

 

I am not sure what algorithm that is. The established industry standards are EBU Tech 3341 and 3342, covered by EBU recommendation R 128.

[Julf]

It is after all a single number. Now if they calculated the absolute RMS value for every sample and compared that to the peak then it would be helpful in comparing different recordings.

 

Suggested reading: tech.ebu.ch/loudness

[Jud]

The TT Dynamic Range algorithm is flawed and that single number in and of itself is not very insightful. I would highly recommend that it only be used to compare different versions of the same recording. Other than that it is unfortunately next to useless.

 

Thanks for the explanation (in a later comment) of how the DR is calculated. I agree it's flawed. Nevertheless, even flawed as it is, it does have a couple of uses.

 

One is the one you point out here, comparison of different versions of the same recording. (This is in my opinion sufficiently helpful to audiophiles who want to consider whether a particular version of a recording has been compressed relative to other versions that it would make the DR Database worthwhile if its usefulness ended there).

 

Another is to check the provenance of downloads. Look for instance at the DR values for the Beach Boys' Pet Sounds, and it becomes readily apparent which of the available versions it shares an origin with. (The HDTracks download apparently comes from Mark Linnett's excellent stereo remastering, which was used as the basis of the MFSL SACD remastering.)

 

A third is (with appropriate caution) to be warned by exceedingly low DR values - say, all tracks with a DR of 6-7 or below.

[junker]

Totally agree with you Jud.

 

I just hope that people don't read too much into it and make (even unconsciously) the wrong general, sweeping conclusions. It's really kind of a shame because you can see how much the community has embraced this database.

 

You're a perfect person for me to ask this... I saw that you have been very happy with your DSD DAC and the MF SACD Pet Sounds recording.

 

I have two copies of that album (DVD 24/96 & 40th Anniversary Stereo / Mono 16/44.1 rip). They both have a DR value of 11, where your SACD has a value of 13, so I'm assuming it sounds a lot better? On my copies I can't barely get into the first 10 seconds of "Wouldn't It Be Nice" without it shrieking my bloodied eardrums into an ice-cream headache. I'm using A+ to play DSD files and was wondering if you had an opinion as to how much that impairs the sound, if any. That's how Berkeley is recommending to play DSD with the BADA. I'm assuming that you would still get the improved DR even with the conversion. Any feeling on this? I'd like to rip this on a PS3 before too long.

 

Nevertheless, even flawed as it is, it does have a couple of uses.

 

One is the one you point out here, comparison of different versions of the same recording. (This is in my opinion sufficiently helpful to audiophiles who want to consider whether a particular version of a recording has been compressed relative to other versions that it would make the DR Database worthwhile if its usefulness ended there).

 

Another is to check the provenance of downloads. Look for instance at the DR values for the Beach Boys' Pet Sounds, and it becomes readily apparent which of the available versions it shares an origin with. (The HDTracks download apparently comes from Mark Linnett's excellent stereo remastering, which was used as the basis of the MFSL SACD remastering.)

 

A third is (with appropriate caution) to be warned by exceedingly low DR values - say, all tracks with a DR of 6-7 or below.

[Jud]

You're a perfect person for me to ask this... I saw that you have been very happy with your DSD DAC and the MF SACD Pet Sounds recording.

 

I have two copies of that album (DVD 24/96 & 40th Anniversary Stereo / Mono 16/44.1 rip). They both have a DR value of 11, where your SACD has a value of 13, so I'm assuming it sounds a lot better? On my copies I can't barely get into the first 10 seconds of "Wouldn't It Be Nice" without it shrieking my bloodied eardrums into an ice-cream headache. I'm using A+ to play DSD files and was wondering if you had an opinion as to how much that impairs the sound, if any. That's how Berkeley is recommending to play DSD with the BADA. I'm assuming that you would still get the improved DR even with the conversion. Any feeling on this? I'd like to rip this on a PS3 before too long.

 

When I had the Bifrost, I still preferred DSD converted by Audirvana Plus to even the very good HDTracks download. Now that I can play it natively, no contest.

[junker]

When I had the Bifrost, I still preferred DSD converted by Audirvana Plus to even the very good HDTracks download. Now that I can play it natively, no contest.

 

Grr... okay thanks. Not exactly what I was hoping to hear, but thank you very much.