General question on dynamic range and music

[Jochen]

That's the difference between the average (RMS) and the peaks (or sometimes another measure). It's NOT the full range which on a digital recording can be from total silence (-infinity dB) to the peak.

This is what I would like to understand. Apparently there are two ways of measuring it, PMF (TT DR Offline Meter) and MAAT DROffline and its more potent version MkII. But how do these measured numbers transform into the "full dynamic range" then?

With records or tapes the format is important because even with highly-compressed (undynamic) music (or program material) you still don't want to hear noise between tracks or during the rare quiet parts or during the fade-out, etc.

This is a point, but when for example using 24 bit with 144dB, there should be ways to have the noise inaudible and still much higher DR, no?

I wouldn't want the full range but I grew-up with vinyl and when CDs were introduced I expected musicians & producers to start taking advantage of the wider dynamic range... Boy, was I WRONG!!!!

That is the point, redbook CD has about 30dB more DR then vinyl, but at least in that database the DR of the same recording in both formats is basically the same. Why not make use of the additional capacity?

 

[pablolie]

I have a general question on dynamic range (DR). Redbook CD offers 16 bit, corresponding to 96dB DR. With dither up to 120dB. High-Res 24 bit audio even 144dB. Vinyl, according to different sources up to 70dB and tape machines roughly 65dB. On the other hand, recorded music has very limited DR. I am not referring to the loudness war (https://dr.loudness-war.info/), but if one refers to this database of almost 1.7 million albums, even for very well recorded, mixed and mastered albums the DR is only between 10 and 25 dB. Recorded classical music is said to deliver up to 40 dB, although in the mentioned database this value is not found and always well below 25 dB. On the other hand, live music supposedly may have up to 80-90 dB (see https://en.wikipedia.org/wiki/Dynamic_range and https://hub.yamaha.com/audio/music/what-is-dynamic-range-and-why-does-it-matter/). This is close to the threshold of hearing, which resides between 90-115 dB. So my question would be, why is it not possible (or desirable?) to record the full dynamic range of live music and why is the capacity of the standard CD not taken advantage of? Is the 96dB - 25dB = 71dB difference of capacity minus actual content on CD actually wasted? Not talking about 3dB loudness war abominations or even High-Res. Or am I mixing things up?

DR in a recording does not necessarily correlate (or is limited) to/by the DR of the format itself.

Classical music -if you listen to an entire concert- often has huge DR changes, hence the obsession with Tchaikovsky's cannon shot in 1812. It's in the very nature of the compositions, having adagios and pianissimos and then molto allegros and fortissimos thrown in. It would be horrible to compress it because it would complete violate the composer's stated intent. And yet you will not derive an additional yota of detail if you play a 24/192 file. With other music genres it is much easier (albeit a criminal practice imo) to go for compression of the dynamics without sacrificing fidelity entirely.

That is why digital formats are so awesome for classical - with vinyl you'd hear all the needle noise during a quiet adagio passage. But hey, there are analog recordings from the 50s and 60s that were re-issued on CD and I love them, even though you can hear the original master tape hiss in the quiet passages, which clearly diminishes DR.

So... "it's complicated"...

 

[Jochen]

DR in a recording does not necessarily correlate (or is limited) to/by the DR of the format itself.

Classical music -if you listen to an entire concert- often has huge DR changes, hence the obsession with Tchaikovsky's cannon shot in 1812. It's in the very nature of the compositions, having adagios and pianissimos and then molto allegros and fortissimos thrown in. It would be horrible to compress it because it would complete violate the composer's stated intent. And yet you will not derive an additional yota of detail if you play a 24/192 file. With other music genres it is much easier (albeit a criminal practice imo) to go for compression of the dynamics without sacrificing fidelity entirely.

That is my point. Even the 96dB of CD are hardly used even for the best recordings. What one gains with more bit-depth or higher sample rate is negligible, or even completely null. Why not go for higher DR recording?

 

[Ifrit]

Lower DR have two advantages: they play louder and clearer on portable devices and in cars, and they can be more exciting to the average listener.

 

[Jochen]

Lower DR have two advantages: they play louder and clearer on portable devices and in cars, and they can be more exciting to the average listener.

I understand this and agree. But why at least aren't there some special edition high DR recordings? I mean there are all sorts of Hi-Res, DSD, SACD and whatever, that, at least in my eyes, have no real benefit. Higher DR might have (at least I would like to try it out).

 

[BrianD]

My understanding is that one of the first recordings that intentionally took advantage of the increased dynamic range of the new CD format was Dire Straits Brothers in Arms. It lists a dynamic range of 20 dB, but that is just how that measure is defined. The quiet to loud on that album is much more than 20 dB. And 20 is more than most CDs have for DR.

I don’t understand the loudness wars, you can’t have loud without the soft.

 

[Jochen]

My understanding is that one of the first recordings that intentionally took advantage of the increased dynamic range of the new CD format was Dire Straits Brothers in Arms.

Two points here.

1) "Brothers in Arms" may be the most famos example of a well recorded early CD, but it is from 1985. Talking Heads' "Speaking in Tongues" is from 1983 and has equally high DR, as well as Donald Fagen's "Nightfly" from 1982. There may be other earliers examples.

2) The corresponding vinyl publications from these recordings have only slighly less (by a few dB) DR. So it appears to me that these are simply well mixed and mastered recordings, but not necessarly because they were published on CD.

DR.It lists a dynamic range of 20 dB, but that is just how that measure is defined. The quiet to loud on that album is much more than 20 dB.

Well, how so ? How do you know?

And 20 is more than most CDs have for DR.

True.

I don’t understand the loudness wars, you can’t have loud without the soft.

100% agree.

 

[Blumlein 88]

The TT meter based DR rating is not what the OP thinks it is by comments. In simple terms they take the RMS average and compare that to peak levels.

So for instance average RMS levels could be -60 dbFS with peaks up to -40 dbFS and DR would be DR20. This is a little simplified though essentially how this works.

So you could have average RMS levels for a track of -20 dbFS and peak levels of 0 dbFS and also DR20. So it isn't at all related to the total dynamic range of a medium.
BTW in case you don't know dbFS means db relative to Full Scale.

To achieve a DR of 70, you might need to record hall silence punctuated by periodic loud drum shots at 0 dbFS. In other words not really music.

So in addition, the near silence in quieter parts of the recording are averaged in or glossed over by the averaging of RMS levels. I think the TT meter for DR does this on some sliding time scale as well. Been awhile since I looked into it in detail. Still the key point is DR ratings are the running average levels versus peaks. 20 is pretty high, 12-15 for most non-classical music quite enough. Less starts to get into loudness wars. DR4 means average levels are only 4 db below peak levels. Or a near non-varying level of loudness.

 

[Jochen]

The TT meter based DR rating is not what the OP thinks it is by comments. In simple terms they take the RMS average and compare that to peak levels.

So for instance average RMS levels could be -60 dbFS with peaks up to -40 dbFS and DR would be DR20. This is a little simplified though essentially how this works.

So you could have average RMS levels for a track of -20 dbFS and peak levels of 0 dbFS and also DR20. So it isn't at all related to the total dynamic range of a medium.
BTW in case you don't know dbFS means db relative to Full Scale.

To achieve a DR of 70, you might need to record hall silence punctuated by periodic loud drum shots at 0 dbFS. In other words not really music.

So in addition, the near silence in quieter parts of the recording are averaged in or glossed over by the averaging of RMS levels. I think the TT meter for DR does this on some sliding time scale as well. Been awhile since I looked into it in detail. Still the key point is DR ratings are the running average levels versus peaks. 20 is pretty high, 12-15 for most non-classical music quite enough. Less starts to get into loudness wars. DR4 means average levels are only 4 db below peak levels. Or a near non-varying level of loudness.

So why does the literature say that live orchestra music has DR of up to 80dB? A TT meter based measurement would not deliver this value?

 

[Blumlein 88]

So why does the literature say that live orchestra music has DR of up to 80dB? A TT meter based measurement would not deliver this value?

No, and I'm not aware of anything claiming 80 db for the DR rating as obtained with TT meter. Now DR, as in dynamic range of an orchestra, yes that can be 80 db in some of the finest recordings. Few reach that. The confusion is DR rating and DR as in short for dynamic range. Two different things.

Dynamic range has some relevance to the medium while DR rating effectively does not. Redbook CD could have dynamic range of 96 db, but self noise of microphones, microphone preamps and hall noise mean you won't get that on any actual recording of orchestral music. That is why worrying about dynamic range of the modern digital medium is not worth doing.

The other thing, the DR rating of the dynamics in a recording are about how it was mastered.

 

[RayDunzl]

And there are recordings out there that have very large dynamic range. Mostly 'classical', if I'd have to guess.

The "visually" most "dynamic" recording I think I have.

It is far from Classical, but is definitely Orchestral:

Just under an hour of music above.

Zooming in on the very beginning:

There is silence from the orchestra, the record level is brought up a little, bringing up the ambient noise level from silence so the start is not so jarring

A single instrument begins playing softly, barely heard.

Linear scale - 1/100th of full linear scale:

Decibel scale:

You can see the ambient noise as spikes as the record level is raised before the single instrument starts.

The softest level of the program is around -50dBFS

The "quiet" sections later in the piece around -30dBFS, See the skinny flat parts of the first graphic.

 

[RayDunzl]

But why at least aren't there some special edition high DR recordings?

Assume 24bit 144dB range.

The bottom 70dB is all but silence if the playback level in your room of the loud parts is 85 average, 105 peaks, as is about all you can stand for more than a short period.

I hit 116.9dB peak on a "calibrated" drum solo and it sounded good, but also like the speakers were going to come apart if I kept it up.

 

[RayDunzl]

This is a point, but when for example using 24 bit with 144dB, there should be ways to have the noise inaudible and still much higher DR, no?

144dB SPL is about the level of sound on the deck of an aircraft carrier as jets lauch with full afterburner.

The deck crew wears earplugs.

And noise reducing ear muffs.

And a noise deadening helmet over those.

"The U. S. Navy, with U.S. Air Force and industry partners, is working to improve hearing protection and speech intelligibility for aircraft carrier flight deck crews who work up to 16 hr per day in 130-150 dBA tactical jet aircraft noise. Currently, flight deck crews are required to wear double hearing protection earplugs and earmuffs in a cranial helmet."

You can't play at that level that at home. Maybe in a stadium standing 200 feet back.

So you turn down the volume.

At most, you end up with 70~80dB SPL audible range above the ambient noise of you room, of 30dB SPL or thereabouts.

The recording is adjusted so the softest parts are brought up to where you can hear them.

That's pretty much the end of the dynamic range story at home.

 

[Sokel]

The "visually" most "dynamic" recording I think I have.

It is far from Classical, but is definitely Orchestral:

View attachment 283640

Just under an hour of music above.

Zooming in on the very beginning:

There is silence from the orchestra, the record level is brought up a little, bringing up the ambient noise level from silence so the start is not so jarring

A single instrument begins playing softly, barely heard.

Linear scale - 1/100th of full linear scale:

View attachment 283639

Decibel scale:

You can see the ambient noise as spikes as the record level is raised before the single instrument starts.

The softest level of the program is around -50dBFS

View attachment 283638

The "quiet" sections later in the piece around -30dBFS, See the skinny flat parts of the first graphic.

I zoomed out as much as I could:

 

[RayDunzl]

So why does the literature say that live orchestra music has DR of up to 80dB?

Probably becase a full orchestra might peak around 120dB SPL with what I'd consider a very optimistic hall noise level of 40dB SPL if there is an audience breathing and rustling.

They won't play so softly that you can't hear them.

120 - 40 = 80. Or thereabout.

Architectural Acoustics - Acceptable Room Sound Levels - Archtoolbox

A description of the various background noise criteria for buildings and rooms, including a table of acceptable noise levels for a number of space types.

www.archtoolbox.com

 

[RayDunzl]

As for the 96db range of CD and the 144dB range of 24bit recordings.

My Practical Opinion:

The extra dB are subdivisions of the quietest part.

A DAC playing full scale 16 or 24 bit will have the same maximum output level of 2 volts or whatever.

In practice, it is just that the quietest part that you can't hear anyway can be quieter.

24bit just subdivides each of 16 bits into another 256 possible levels, possibly more precision in the mid levels, but that may or may not make an audible difference.

 

[bodhi]

I don’t think this answers your question, but it’s an interesting data point. Listen to PJ Harvey’s “Rid of Me”. Make sure to turn it up right at the beginning loud enough that the vocal is a natural volume and the guitar is has a nice live punchy sound, then wait for the chorus to come in. Steve Albini wasn’t playing with any loudness war nonsense when he produced this album. it registers “14” on the dynamic range database; the high end of their scale (for whatever that’s worth).

Holy s***. Very nice, but almost spilled my coffee there.

 

[RayDunzl]

The TT meter based DR rating is not what the OP thinks it is by comments. In simple terms they take the RMS average and compare that to peak levels.

Yeah.

Back to the Redwood Symphony:

I'd give it a pessimistic rating of about 35dB to 40dB of "dynamic range"

But the Dynamic Range Database says:

 

[danadam]

In simple terms they take the RMS average and compare that to peak levels.

And not just that, they take into account only the 20% loudest parts of the track (or at least the older PMF (TT DR Offline Meter) algorithm does).

 

[sigbergaudio]

It's at least safe to say we don't need higher dynamic range than CD for consumer playback.