Dynamic Range

[mglobe]

Only value I find in SACD is possibly a better end product in terms of mixing/mastering, but mostly just for multich options vs 2ch generally.

Actually I think the data shows that the 2ch dynamic range is improved as well in most cases.

 

[Chrispy]

Actually I think the data shows that the 2ch dynamic range is improved as well in most cases.

That could be a product of the mixing/mastering, but inherently in the recording media, no.

 

[Keith_W]

Technically that's easy: the opposite to a compressor is an expander. It expands dynamic range instead of reducing it. Use one in your signal chain. I'd it's a digital one, you can even save presets for different styles and albums and stuff.

How well has that worked for you? I tried an expander, and it was a really finicky thing. The problem was: you need an individual setting for each track. If you're not aggressive enough, the effect is so subtle that you may as well not bother. If you are too aggressive, you hear all sorts of digital artefacts. The window between "too little" and "too much" is very small. I have never found a setting that provided satisfactory dynamic expansion without digital artefacts. So I gave up.

BTW there are two types of expanders - the ones that increase the peaks, and the ones that soften the quiet parts. I can't remember what they are called. If you soften the quiet parts, it destroys the ambience of acoustic recordings. If you increase the peaks, you get those strange digital artefacts. They are sometimes combined into the same tool.

 

[mglobe]

That could be a product of the mixing/mastering, but inherently in the recording media, no.

Absolutely. I didn’t mean to imply that it was the recording media. It’s the fact that mixing/mastering (I suspect the latter) is done with higher DR on the SACD. And that is the reason I’ve gone down the SACD route. Sorry for not being more explicit before.

 

[levimax]

How well has that worked for you? I tried an expander, and it was a really finicky thing. The problem was: you need an individual setting for each track. If you're not aggressive enough, the effect is so subtle that you may as well not bother. If you are too aggressive, you hear all sorts of digital artefacts. The window between "too little" and "too much" is very small. I have never found a setting that provided satisfactory dynamic expansion without digital artefacts. So I gave up.

BTW there are two types of expanders - the ones that increase the peaks, and the ones that soften the quiet parts. I can't remember what they are called. If you soften the quiet parts, it destroys the ambience of acoustic recordings. If you increase the peaks, you get those strange digital artefacts. They are sometimes combined into the same tool.

Expanders are not even close to lossless to the original.

 

[Chrispy]

Absolutely. I didn’t mean to imply that it was the recording media. It’s the fact that mixing/mastering (I suspect the latter) is done with higher DR on the SACD. And that is the reason I’ve gone down the SACD route. Sorry for not being more explicit before.

Maybe, generally the DR in recordings just isn't that great compared to the media potential at least. Choices of production largely but also limited consumer desirability I'd think as well. I haven't found generally an advantage in the 2ch tracks of my SACDs....I just use them for multich content.

 

[Chris A]

How well has that worked for you? I tried an expander, and it was a really finicky thing. The problem was: you need an individual setting for each track. If you're not aggressive enough, the effect is so subtle that you may as well not bother. If you are too aggressive, you hear all sorts of digital artifacts. The window between "too little" and "too much" is very small. I have never found a setting that provided satisfactory dynamic expansion without digital artifacts. So I gave up.

Well, let's just say that I think you gave up too soon.

Additionally, the exact expander used has a lot to do with how well it can be "trimmed". For instance, I first used other expanders (freeware) that didn't work at all (not multiband and not visual as well as real-time listening capable). Then I used FabFilter Pro-MB (multi-band compressor/expander) and had some real success for about 5-6 months, but it took a little tweaking (full disclosure: I have way over 10K hours demastering music tracks to help me in that task).

Now, I use Pro-Q4 (a dynamic/static EQ plugin from the same company), and the task is actually fairly trivial--and pretty stable from album to album, genre to genre. You can watch the original and resultant tracks jump around in real time, as well as listen to the results at the same time. So now the process is pretty mundane and the results are, well, pretty spectacular at times, and sometimes quite revealing of the underlying track quality (a constraint which we must live with).

I posted one of the dynamic demastering EQ curves that I often use, above. Only relatively mild tweaking of the static EQ, dynamic EQ and parametric band centers and widths ("Q") are needed, and it's immediately obvious in the real-time listening of the track if you're going in the right direction.

One other point: these dynamic EQ settings are sensitive to playback level, so if you play the track much louder or much softer than the nominal 75-83 dB ave, playback level, you're going to need to adjust the filters used for the same track.

You can try FabFilter Pro-Q4 for free for 30 days. I recommend trying it. You won't be disappointed--it's the most powerful user and intuitive user interface that I think that I've used.

Again, JMTC.

Chris

 

[Chris A]

Expanders are not even close to lossless to the original.

My comment here: don't get lost in the metaphor...use your ears and your technical judgment. Better is better.

Chris

 

[Chris A]

I can't remember what they are called.

Upwards expander, and downwards expander. I recommend the upwards expander for this work. You can use the plugin to produce downward expansion if you need it. I have found that--well--you need both approaches, depending on the music tracks themselves and what's been done to them. Additionally, the Pro-MB plugin can be used both ways.

Chris

 

[levimax]

My comment here: don't get lost in the metaphor...use your ears and your technical judgment. Better is better.

Chris

I understand your perspective. Mine is a little different. If originally released compressed that is how I listen to it, if originally released dynamic and later compressed I listen to the original. No right or wrong way to enjoy listening to music.

 

[solderdude]

The problem here is that when compression is used in a studio (and it nearly always is) it is applied to individual channels and each has their own channel.
On top of that there are limiters used also on individual channels (mic, drum mics, instrument channels, not meaning L and R channel).
You can't undo that 'correctly' with an expander afterwards because that expander looks at the added signal (so all channels combined) and not the individual channels that once were before they were mixed.
Sure... in numbers you get a higher dynamic range (even when using just a de-clipper) and maybe you can make it sound a bit 'better' you can't bring back the original dynamics.

When the recording is f'ed up you might be able to polish that turd a little and believe you rescued it (because it might sound a bit better) but it still won't be anything like how it should have sounded before the sound-engineer got their 'magic' on it to comply to the wishes of whoever wants the recording squashed.

 

[Keith_W]

Then I used FabFilter Pro-MB (multi-band compressor/expander) and had some real success for about 5-6 months, but it took a little tweaking (full disclosure: I have way over 10K hours demastering music tracks to help me in that task).

Thank you for your recommendation.

It's not exactly cheap. So I went around looking for a review, and I found one here: Sound on Sound. They were impressed by it, as you were, but at the moment this tool is above my skill level. Maybe a 30 day trial might be enough for me to learn how to use it, but it very much looks like tools a professional will use. That has its pros and cons - it will be powerful, but it also looks difficult to learn. I'll give it a go some day - thanks for that!

I can't remember which expander I used, it was a free one.

 

[Scytales]

Absolutely. I didn’t mean to imply that it was the recording media. It’s the fact that mixing/mastering (I suspect the latter) is done with higher DR on the SACD. And that is the reason I’ve gone down the SACD route. Sorry for not being more explicit before.

There may be a technical explanation for this.

Levels on DSD masters intended for SACD production have to comply with mandated thresholds in two separated bands under penalty of being rejected by optical disc pressing plants: the audio band and a high frequency-band defined as a 40 kHz to 100 kHz interval with 5th order roll-off rate at corner frequencies. According to the latest revision of the SACD Audio specifications, maximum allowable levels in the audio band have to comply to a relatively complex set of rules which defines a maximum absolute DSD modulation level and what is called a "maximum long term average modulation level" obtained from peak holding the DSD modulation level with a slow down time of 258 ms [1]. Although there is no mandate to comply with it, there is also a recommendation not to exceed a certain level of DC offset that might be a by-product of sustained clipping [2].

Hardware/software digital audio workstations (DAW) such as Sony's Sonoma and SADiE's DSD8 included what their respective manuals called "Annex Metering" to check that those prescriptions are observed.

Moreover, I have noticed that the Sony Sonoma and early Merging Pyramix DAW manuals also talk about even more refined metering options that divides the high frequency band in two distinct bands : 20 kHz to 40 or 50 kHz and 50 kHz to 100 kHz, with recommendation not to go above some levels that the Merging manual says are recommended by the Audio Engineering Society (I am not aware of those).

All in all, these set of rules and recommendations may have the fortunate consequence to keep excessively compressed recordings at bay.

Here is what Nika Aldrich (Trillium Lane Labs) wrote in July 2003 in a white paper about clipping in over-sampled PCM digital files [3], at a time when the SACD audio specifications had not been completed yet and were a little more permissive (no long term average modulation level rule [4]):

"It is worth noting that Sony’s [and Philips's, may I add] new SACD format includes measures
that prevent the music from ever clipping in the way described. Mastering engineers who work
on SACD releases have observed the notion that heavily compressing the audio inevitably results
in the need to ‘turn down’ the overall level on the disk. Left with the choice of compressing
the disk and turning it down, or simply leaving it the opportunity to ‘breathe’ with some
headroom, most mastering engineers are mastering to SACD disks differently than they
have been to DVD’s and CDs. Many professionals in the audio industry are claiming that
audio that has been remastered for the SACD format ‘breathes more,’ ‘has more life,’ and
‘doesn’t have the digital harshness’ of the CD counterparts."

Reference:
[1] Super Audio CD System Description, Part 2: Audio Specification, Version 2.0, Annex D: Audio Signal Requirements (Normative), March 2004.
[2] Ibidem, Annex E: Audio Signal Requirements (Informative).
[3] Digital Distortion in CD’s and DVD’s: The Consequences of Traditional Digital Peak Meters. Nika Aldrich, Trillium Lane Labs whitepaper, July 2003.
[4] Super Audio CD System Description, Part 2: Audio Specification, Draft Version 1.3, Annex D: Audio Signal Requirements (Normative), June 2002.

 

[Multicore]

Yes, but note that R128 is a loudness scale--not really a dynamics scale

True but in the circumstances it's not a terrible thing. The circumstances being that the market rewards loudness by level maxing at the expense of dynamics.

There are niche markets in which this pressure is less but niche markets don't usually drive trends.

 

[Multicore]

This excellent presentation on dynamic range, limiting, mastering by Alan Silverman was discussed here before and largely dismissed. I otoh think it is very valuable and if you're actually interested in this topic it's worth a half an hour of your time. It takes seriously why mastering engineers do what they do and explains the characteristic parameters of the target environments. The examples of level boosting with limiters and the destructive distortion they introduce are good. The demonstration of the effects of loudness normalization in the streaming platforms on three mastered recordings with no, middling and high level boost is very impressive. Or so I thought.

The Future of Mastering: Loudness in the Age of Music Streaming​

 

[mglobe]

There may be a technical explanation for this.

Levels on DSD masters intended for SACD production have to comply with mandated thresholds in two separated bands under penalty of being rejected by optical disc pressing plants: the audio band and a high frequency-band defined as a 40 kHz to 100 kHz interval with 5th order roll-off rate at corner frequencies. According to the latest revision of the SACD Audio specifications, maximum allowable levels in the audio band have to comply to a relatively complex set of rules which defines a maximum absolute DSD modulation level and what is called a "maximum long term average modulation level" obtained from peak holding the DSD modulation level with a slow down time of 258 ms [1]. Although there is no mandate to comply with it, there is also a recommendation not to exceed a certain level of DC offset that might be a by-product of sustained clipping [2].

Hardware/software digital audio workstations (DAW) such as Sony's Sonoma and SADiE's DSD8 included what their respective manuals called "Annex Metering" to check that those prescriptions are observed.

Moreover, I have noticed that the Sony Sonoma and early Merging Pyramix DAW manuals also talk about even more refined metering options that divides the high frequency band in two distinct bands : 20 kHz to 40 or 50 kHz and 50 kHz to 100 kHz, with recommendation not to go above some levels that the Merging manual says are recommended by the Audio Engineering Society (I am not aware of those).

All in all, these set of rules and recommendations may have the fortunate consequence to keep excessively compressed recordings at bay.

Here is what Nika Aldrich (Trillium Lane Labs) wrote in July 2003 in a white paper about clipping in over-sampled PCM digital files [3], at a time when the SACD audio specifications had not been completed yet and were a little more permissive (no long term average modulation level rule [4]):

"It is worth noting that Sony’s [and Philips's, may I add] new SACD format includes measures
that prevent the music from ever clipping in the way described. Mastering engineers who work
on SACD releases have observed the notion that heavily compressing the audio inevitably results
in the need to ‘turn down’ the overall level on the disk. Left with the choice of compressing
the disk and turning it down, or simply leaving it the opportunity to ‘breathe’ with some
headroom, most mastering engineers are mastering to SACD disks differently than they
have been to DVD’s and CDs. Many professionals in the audio industry are claiming that
audio that has been remastered for the SACD format ‘breathes more,’ ‘has more life,’ and
‘doesn’t have the digital harshness’ of the CD counterparts."

Reference:
[1] Super Audio CD System Description, Part 2: Audio Specification, Version 2.0, Annex D: Audio Signal Requirements (Normative), March 2004.
[2] Ibidem, Annex E: Audio Signal Requirements (Informative).
[3] Digital Distortion in CD’s and DVD’s: The Consequences of Traditional Digital Peak Meters. Nika Aldrich, Trillium Lane Labs whitepaper, July 2003.
[4] Super Audio CD System Description, Part 2: Audio Specification, Draft Version 1.3, Annex D: Audio Signal Requirements (Normative), June 2002.

Thank you for this.

 

[mhardy6647]

unintended consequences:
Design an audio data storage system with vastly increased dynamic range capacity... and we get... minuscule dynamic range (on the average).

 

[manisandher]

What can be done to bring back dynamic range in recordings?

It's exactly the purpose of my AIiven project. Not trivial though...

Mani.

 

[goat76]

What the streaming platforms have already done. If the loudness of the recording (EBU R 128 dB LUFS measured over the entire track) is greater than their platform-wide max, e.g. -14 dB LUFS, then apply negative gain in the amount of the difference. The result makes the recordings that are pushed through brick-wall limiters to high levels sound quiet relative to tracks that retain their dynamics.

About 10 years ago, when Tidal and other streaming platforms started using volume normalization, I was quite sure most record companies would start following that new standard when realizing their loud album releases didn't sound any louder anymore. I waited, and I'm still waiting for that to happen, but unfortunately, 10 years later, I have not seen any such effect. Not even a trend pointing in that direction when looking at the DR database or when analyzing tracks on my own.

 

[goat76]

The problem here is that when compression is used in a studio (and it nearly always is) it is applied to individual channels and each has their own channel.
On top of that there are limiters used also on individual channels (mic, drum mics, instrument channels, not meaning L and R channel).

I think most people are okay with a bit of dynamic control on individual tracks within the mix, which is, in most cases, done to make them work together with the rest of the sound objects within the mix. I see that type of dynamic control as part of the artistic choices, even if it's done by the mixing engineer, for the reason of making the mix work as a whole.

The main problem with extensive limiting and compression is the ones applied to the final 2-channel mix, and that is likely what most people here are complaining about. At least I hope that's the thing most people here have a problem with.