To what degree DR reflects SQ?

[sphinxsix]

 

Discuss  

 

 

[John Dyson]

My opinion is that the desirable amount dynamic range in general is dependent on the recorded source AND listening environment.   If there is more compression than needed to make for a listenable experience, then it can be damaging to the sound.

 

Not all compression is bad, but excessive & unartful suppression of dynamic range, especially with artifacts that create biases in the sound (distorted stereo image, muted sibilance, overly compressed percussion), are damaging changes to a recording.

As long as an understanding/knowledgeable artist is in control of whatever audio procecessing that is applied to the recording, then there is a greater probability of maintaining quality.

 

So, DR is one quantiatitve aspect of a recording, which can resolve into a qualitative aspect.   Not all reductions of DR are all bad, but I believe that unartful or too aggressive reduction of DR for loudness reasons can also be damaging.

 

Blatent, manifest, across the board reduction in DR (both the DR meter type DR and other kinds of dynamic range) are lazy and disrespectful against the original artists and the consumer.  Sadly, some artists and many consumers don't realize the extent of profound damage against the recordings.   I do believe that many consumers 'feel' that there is something wrong, but it is only recently that there is a more precise understanding of the damage.  On the other hand, there are sometimes benefits to reduction in DR, but more often than not, the recordings are more seriously damaged than what is needed for a sometimes beneficial reduction in dynamic range.

 

John

 

[DuckToller]

Without scientific approach, just anectdotical:
- Very often, when I look into music that allow tones the air to breathe, to swing and to resonate, that gives room to a deep, high and wide soundscape / soundstage, the DR is higher than DR8. Which to the contrary shouldn't say, everything having DR9 and upwards sounds good.
- It always depends on the artist, the producer, the studio (technology), the mix and the mastering and finally on the format.
- I'd describe the effect maybe this way:
 A higher DR seems to allow the production team to articulate the music more significantly in its steps to finalize the product than a lower DR. I can't say why, as I am the consumer, not the producer.

As an example: Radical Dance Faction's "Hope" from the "Wasteland" recording, showing DR15 in the JRIVER music analysis. (here's my description from 2018: Radical Dance Faction - Hope (1991) - RDF is masterminded by Chris Bowsher, a musician from Hungerford, whose articulation could have mentored the Sleaford Mods. Meanwhile, his taste in creating spacial soundscapes including accentuation for voices and instruments, just provides a completely different mindset (than the Sleaford Mods). This song talks about the indecision after terrorists attacks. With a speedy reggae rhythm basement that create the space for the pronounced, in my ears "jazzy" guitar riffs, a female voice contrasted by a male background singer, some meandering saxophone and hand percussion, the music leads the "long long road" to freedom from fear.) 

- If you compare it to Metallica's Death Magnetic (DR3), the difference in SQ is more than audible.

- FWIW, it always depends on a system that allow you to hear the difference and on your own personal capabilities and gusto.

- High DR usually demands a lot more from your music reproduction chain.

[firedog]

The DR by itself means nothing. DR reflects the difference between the loudest and softest parts of the music. Lot of music doesn't have very much volume dynamics, so it will have a low DR. Doesn't mean it wasn't recorded/mastered well.

What you don't want is music that was originally made with a relatively high DR, and then heavily compressed. That doesn't sound good. 

[Jud]

23 minutes ago, firedog said:

The DR by itself means nothing. DR reflects the difference between the loudest and softest parts of the music. Lot of music doesn't have very much volume dynamics, so it will have a low DR. Doesn't mean it wasn't recorded/mastered well.

What you don't want is music that was originally made with a relatively high DR, and then heavily compressed. That doesn't sound good. 

 

I'm thinking (perhaps incorrectly) that DR as shown in DR Database and as defined in a standard reflects the difference between the loudest parts of a track and the *average* loudness (rather than the quietest part).

 

For me it's most useful when comparing different versions of the same track, realizing (1) vinyl will always show 2-3dB more range for a given version (this is an artifact of the test process; I forget the reason); and (2) 1-2dB difference in DR is likely not going to be very noticeable.

 

It's *one* factor to evaluate when choosing among versions of a track or album. Older versions will often have higher DR and *may* sound better.

[Allan F]

31 minutes ago, Jud said:

 

I'm thinking (perhaps incorrectly) that DR as shown in DR Database and as defined in a standard reflects the difference between the loudest parts of a track and the *average* loudness (rather than the quietest part).

 

As @firedog correctly states, dynamic range in audio is defined as the difference between the quietest and loudest volume of a piece of music, measured in dB. More generally, dynamic range is defined as the difference between the smallest and largest values that a certain quantity can assume.

[firedog]

Here's what the Roon Knowledge base says about computing dynamic range:

Quote

 

How is Dynamic Range computed?

As with Volume Leveling Roon's dynamic range calculation is done based on R128 standards. In technical terms, Roon's "Dynamic Range" is the same as R128's "Loudness Range".

That there are older methods of computing Dynamic range out there--most commonly by measuring the "Crest Factor". 

Crest Factor measurements reflect the difference between the average volume and the peak volume--so they are easily swayed by periods of silence or near-silence (which distorts the average), and by short-duration peaks--which may not represent the volume of the loudest parts of the track accurately.

The R128 method is more resillient. It begins by computing the statistical distribution of loudness values present at different points in the track, ignoring periods of silence. The computed dynamic range represents the difference between the 10th percentile and the 95th percentile of that distribution. In other words, the "top" of the range is the volume level that 95% of the track sits below, and the "bottom" is the volume level that 10% of the track sits above.

Though both methods portray roughly the same information, Crest Factor values aren't directly comparable to values produced using the R128 method

 

The R128 method makes more sense to me, but I find in practice that the DR number more accurately reflects what I hear.

[Summit]

1 hour ago, Jud said:

 

I'm thinking (perhaps incorrectly) that DR as shown in DR Database and as defined in a standard reflects the difference between the loudest parts of a track and the *average* loudness (rather than the quietest part).

 

For me it's most useful when comparing different versions of the same track, realizing (1) vinyl will always show 2-3dB more range for a given version (this is an artifact of the test process; I forget the reason); and (2) 1-2dB difference in DR is likely not going to be very noticeable.

 

It's *one* factor to evaluate when choosing among versions of a track or album. Older versions will often have higher DR and *may* sound better.

 

Yep, that's right.

 

"The “Unofficial” Dynamic Range Database uses a sliding scale from 1 to 20 (1 being the worst, 20 being the best) to rank the dynamic quality of each of the recordings they list. This number represents the difference between the peak decibel level on a recording and the recording’s average loudness. DRD applies the following descriptors to these ranges: 1-7=bad; 8-13=transition; and 14-20=good. Evaluation for each album on the website includes the album’s average dynamic range, the track with the weakest dynamic range, and the track with the greatest. The Unofficial DRD also provides individual dynamic range measurements for each track on the album."

https://www.stereophile.com/content/unofficial-dynamic-range-database

[Summit]

3 hours ago, sphinxsix said:

 

Discuss  

 

 

 

I believe that DR correlate with SQ especially if you compare different versions of the same record. In all other cases the correlation is less consistent, but, t's still better than nothing and can grossly indicate if a record is very compressed or not.   

[sphinxsix]

51 minutes ago, Summit said:

 

I believe that DR correlate with SQ especially if you compare different versions of the same record. In all other cases the correlation is less consistent, but, t's still better than nothing and can grossly indicate if a record is very compressed or not.   

I wouldn't agree with it in 100%. I have personally in many cases preferred a well done mastering with lower DR, and I know it's not only me (scroll down if you are interested):

 

 

Don't want to engage too much into this discussion but I believe the DR factor had gained its popularity and almost fetish status first of all due to loudness war of course. Secondly it's easy to just measure DR (not many even know about the existance of DR R128 which takes into account slightly different factors than DR) which takes a second and not to engage into time and attention consuming listening comparisons - the higher the better - end of story 

IMO the DR is widely and highly overrated as a tool to judge and even compare the SQ of the same music material and in many cases a better mastering (e.g. smoother, more detailed with more 'air' around the instruments) can sound better than a worse mastering with higher DR. 

Of course the optimum would be a good mastering without or with minimal dynamic compression.

 

4 hours ago, sphinxsix said:

 

Discuss  

 

 

 

[John Dyson]

3 hours ago, sphinxsix said:

I wouldn't agree with it in 100%. I have personally in many cases preferred a well done mastering with lower DR, and I know it's not only me (scroll down if you are interested):

 

 

Don't want to engage too much into this discussion but I believe the DR factor had gained its popularity and almost fetish status first of all due to loudness war of course. Secondly it's easy to just measure DR (not many even know about the existance of DR R128 which takes into account slightly different factors than DR) which takes a second and not to engage into time and attention consuming listening comparisons - the higher the better - end of story 

IMO the DR is widely and highly overrated as a tool to judge and even compare the SQ of the same music material and in many cases a better mastering (e.g. smoother, more detailed with more 'air' around the instruments) can sound better than a worse mastering with higher DR. 

Of course the optimum would be a good mastering without or with minimal dynamic compression.

 

 

bigger DR is definitely not necessarily better.  I just decoded a bunch of recordings for some friends to check for quality in a PM group -- they have TOO MUCH DR to be listenable.  However, the recordings are VERY GOOD test material.  The average peak-RMS one one of the selections is about 46dB (that is crazy high), when typical good recordings are in the 17dB to 24dB range.   More is NOT necessarily better.

 

John

 

[sphinxsix]

1 hour ago, John Dyson said:

bigger DR is definitely not necessarily better.  I just decoded a bunch of recordings for some friends to check for quality in a PM group -- they have TOO MUCH DR to be listenable.  However, the recordings are VERY GOOD test material.  The average peak-RMS one one of the selections is about 46dB (that is crazy high), when typical good recordings are in the 17dB to 24dB range.   More is NOT necessarily better.

 

John

 

It reminded me of the 24bit pre-master of a certain jazz-rock band (it's actually downsampled 32bit material), which to me has been almost unlistenable due to its huge dynamics.

 

[danadam]

6 hours ago, firedog said:

That there are older methods of computing Dynamic range out there--most commonly by measuring the "Crest Factor". 

Crest Factor measurements reflect the difference between the average volume and the peak volume--so they are easily swayed by periods of silence or near-silence (which distorts the average), and by short-duration peaks--which may not represent the volume of the loudest parts of the track accurately.

If by "most commonly" they mean Dynamic Range DB then this description is wrong. Dynamic Range DB algorithm uses the average volume of the laudest 20% of the track, i.e. the track is split into blocks of 3 seconds, average volume of each block is computed and then the top 20% is used to calculate the "crest factor". So it is not "that" easily swayed by silence.

[Audiophile Neuroscience]

10 hours ago, sphinxsix said:

 

Discuss  

 

 

 

There are obviously different measures of DR offered and it would be good to get some better handle on correlates to what we hear. I suspect no one measure fits all due to the wide variety of dynamic content in different music genres or individual compositions/songs.

 

I think DR in playback is important to the extent that it reflects the dynamics of the original music. Full stop.

[danadam]

What I noticed is that when I have an album A with higher DR and an album B with lower DR than I can comfortably listen album A with higher volume setting than album B. That's not surprising. What is maybe somewhat surprising is that even though now they are both at comfortable volume level, subjectively album A sounds louder. So to me higher DR allows for louder but still comfortable listening, which is nice.

[fas42]

1 hour ago, sphinxsix said:

It reminded me of the 24bit pre-master of a certain jazz-rock band (it's actually downsampled 32bit material), which to me has been almost unlistenable due to its huge dynamics.

 

 

Care to share the name?

[fas42]

A rule of thumb is that if a recording has natural DR encoded, and there are inherently huge level, sound intensity swings, then the playback has to have high integrity - otherwise, when it goes soft, the sound will be boring, drab; and when it goes loud, the system starts to fall apart - the levels of obvious distortion are too great, and you'll run to turn it down.

 

So, clever compression is an artful dodge to get good sounding playback on sub-par setups - the bad, current over-compression is just a fashion, and will start to really die out when enough audio playback out there works well enough, for more people to be downright annoyed at silly mastering decisions.

[John Dyson]

1 minute ago, fas42 said:

A rule of thumb is that if a recording has natural DR encoded, and there are inherently huge level, sound intensity swings, then the playback has to have high integrity - otherwise, when it goes soft, the sound will be boring, drab; and when it goes loud, the system starts to fall apart - the levels of obvious distortion are too great, and you'll run to turn it down.

 

So, clever compression is an artful dodge to get good sounding playback on sub-par setups - the bad, current over-compression is just a fashion, and will start to really die out when enough audio playback out there works well enough, for more people to be downright annoyed at silly mastering decisions.

Maybe a possible future project might be clever, gentle compressor that avoids the various distortions, but also avoids profound reduction in natrual sound.  A small amount of compression can reduce the dynamic range for normal living areas, none or minimal compression for studio/specialty listening environments, and rather more compression for automobiles.

 

The specific compressor design spec isn't important here, but an artful variable compressor, given wide dynamic range source material can provide the end-listener with the best of all worlds.   The problem is that a simple feedback or DBX style HW compressor cannot really do the quality that a GP compressor needs to do, but a full-on, maximum quality SW compressor can easily chew up a multi-core Core2 or Ryzen computer.

 

Perhaps the best answer is that the distributors make better, more reasonable choices for the procesing -- forget about persistent loudness, and just give good dynamics.   Another example of extreme dynamics is that I have a copy of love-over-gold that is a bit too dynamic, but a decode for a friend, that I just did, seems to have a better balance.  I have a 'Brothers In Arms' that is definitely a 2 pass encoding, but the dynamics after decoding are also VERY (too) extreme for me.  (The problem with doing a single pass FA decode for a recording that has been damaged by two passes will still result in somewhat 'woody' and 'artificial' sound.)

 

It is all about a balance of all of the desirable characteristics.  It seems like the distributors only consider 'loudness' as being the singular desirable characteristic...

 

 

John

[Foggie]

7 hours ago, sphinxsix said:

Don't want to engage too much into this discussion

Wait, didn't you start the thread? 🤔😉

[R1200CL]

Why does Roon only report DR for local files ?

[sphinxsix]

10 hours ago, fas42 said:

 

Care to share the name?

Will PM you.

 

9 hours ago, Foggie said:

Wait, didn't you start the thread? 🤔😉

I apologize - I have other things on my mind now, In essence I have expressed my views in the post above but I'm really curious what other people have to say.

 

[sphinxsix]

22 hours ago, DuckToller said:

If you compare it to Metallica's Death Magnetic (DR3)

Isn't it ironic that gamers got less compressed PS 'Guitar Hero' version of 'Death (by Compression) Magnetic'.? 

 

Please don't misunderstand me, I'm not advocating dynamic compression. The loudness war is IMO one of humanity's most stupid inventions. I'm only trying to point at the limited possibility of judging SQ by DR number (alone), which IMO has gained a fetish-like status in the audiophile circles, which BTW is to some degree understandable when we look at it as a reaction to the loudness war. IMO, just like in real life situations we simply shouldn't overreact. 

 

 

[Summit]

22 hours ago, sphinxsix said:

I wouldn't agree with it in 100%. I have personally in many cases preferred a well done mastering with lower DR, and I know it's not only me (scroll down if you are interested):

 

 

Don't want to engage too much into this discussion but I believe the DR factor had gained its popularity and almost fetish status first of all due to loudness war of course. Secondly it's easy to just measure DR (not many even know about the existance of DR R128 which takes into account slightly different factors than DR) which takes a second and not to engage into time and attention consuming listening comparisons - the higher the better - end of story 

IMO the DR is widely and highly overrated as a tool to judge and even compare the SQ of the same music material and in many cases a better mastering (e.g. smoother, more detailed with more 'air' around the instruments) can sound better than a worse mastering with higher DR. 

Of course the optimum would be a good mastering without or with minimal dynamic compression.

 

 

 

DR is one SQ factor for people that believe that dynamic range is important. I’m one of them and do dislike the very compressed and lifeless sound of many records with low DR. I can’t say that I have fund that DR has gained a fetish-like status. Anyway like I said before I do consider that DR is most useful then comparing different versions of the same record. I use this one: http://dr.loudness-war.info/

[CatManDo]

DR is only relevant to detect highly compressed productions. It was a reaction to the loudness war, to make it easier to check an album release, as excessive compression is by far the most frequent mastering flaw. The times of extreme EQ choices (boosted treble especially) are fortunately over.

 

So it doesn't matter if a release has DR13 or 15, both are fine and you can forget about it.

 

In my experience, DR7 and lower seriously impacts the sound quality. It will sound lifeless and fatiguing. It depends on the type of music too. And it's important whether compression has been applied smoothly or the release has been "brick walled" by using a limiter, which will add harshness to the sound.

 

https://en.wikipedia.org/wiki/Loudness_war

 

I also use the DR measurement tool to quickly find out if two releases of the same album are identical (look at the dB values, not just the overall DR value).

[John Dyson]

On 7/21/2020 at 11:05 AM, firedog said:

The DR by itself means nothing. DR reflects the difference between the loudest and softest parts of the music. Lot of music doesn't have very much volume dynamics, so it will have a low DR. Doesn't mean it wasn't recorded/mastered well.

What you don't want is music that was originally made with a relatively high DR, and then heavily compressed. That doesn't sound good. 

I don't know if I mentioned this earlier, but I have some recordings that I have been testing that have so much REAL dynamic range that some people have complained.   Accurate gunshots really do cause troubles if not prepared for them

By the time some very strong percussive gunshots are accurately rendered, then the rest of the material is too weak to easily hear.

 

DR is one of the measures that can be indicative of one aspect of recordings.

 

Since most recent recordings are compressed to the extent that they are approx 6dB louder than natural, DR can be one of the reasonable indicators.  And yes, almost all pop recordings have at least several (probably about 5-6dB) more loudness because of fast compression.  Reasonable *improvements* in DR are a good thing.  DR for DR sake -- not very interesting.

 

Dynamic range is a real problem on consumer recordings, but is not the only indicator of quality.

 

I certainly would not even imply that DR by itself means nothing -- think about a DR of 9...   That says something about the character of the sound in a recording.

 

John