Using master fader for more headroom?

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

09 Feb 2016

odarmonix wrote:
ScuzzyEye wrote:Doesn't happen with floating point. You always have 24-bits of precision in the mantissa, the exponent changes as the scale of the values change. Floating point is very well suited for PCM encoding. The size of the steps between values get smaller as the values themselves get smaller.

A crude example would be values from 0.1 to 1 get 10 steps:
0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0

When you go up a level magnitude from 1 to 10, you still get 10 steps:
1, 2, 3, 4, 5, 6, 7, 8, 9, 10

And one more, from 10 to 100:
10, 20, 30, 40, 50, 60, 70, 80, 90, 100.

Except you actually get 16 million steps at each scale level, and there are 256 scales.
I'm well aware of the purpose of floating point audio and that it basically offers virtually unlimited headroom. ;) But I'm not sure this is truly relevant here, how many musicians use floating point audio when sending their tracks to mastering studios ? Us Reason users have floating point audio processing for certain things like the SSL, but we're still limited to a maximum of fixed 24 bit for import & export.

That being said, does anyone REALLY need the insane dynamic range allowed by 32 bit float (let alone 64) for PCM encoding ? While the theory sounds good on paper it sounds like complete overkill to me, I'm not even sure it's all that useful for DAWs' internal processing and I don't remember having ever needed more than fixed 24 bit for any kind of task, although I can definitely hear the difference between 16 and 24 in some very specific situations.

Regardless of the above, in my previous post I only wanted to point out that working at extremely weak levels can be just as a bad idea as working with too hot levels, even though irreversible signal degradation is less likely with the former. Music definitely has been WAY more often butchered by excessively hot levels during the last two decades. :twisted:
First of all, it's not "excessively hot levels" that has destroyed music, it's the extreme limiting (or more specifically, the distortion caused by the limiting) of the dynamic range - the levels are exactly as "hot" as they've ever been, that is to say, no digital file can exceed 0 dBFS, so no file can be at a hotter level than another. As for extreme low levels, you CAN have a level down at - 200 dBFS in a floating point system - you'll have to raise it significantly before you could hear it, but you'll have the complete original dynamic range preserved when you do so. Seems maybe folks are mixing floating and fixed point examples here, hence the confusion?

Next: the dynamic range of a floating point system (with regards to internal processing) has one huge advantage IMO, and that is the almost impossibility of clipping internally. Anyone that remembers working with Pro Tools (or any fixed point system) when it was 48 bit fixed will know what I'm talking about. :(
The advantage of 32 bit floating point's additional dynamic range is that there is not only the additional dynamic range BELOW 0 dBFS, but also ABOVE IT.

For me, it's not at all about being able to "hear a difference" with increased bit depths (I still export mastered mixes at 16 bits), it's about the way you can work with multitrack audio when you have additional dynamic range available. Looking to the past again, anyone who remembers 16 bit audio recording will appreciate 24 bits when it comes to recording audio. Talking recording live humans here, and being able to leave plenty of headroom when doing so - I haven't clipped an audio recording in YEARS because of this additional headroom.
Selig Audio, LLC

User avatar
ScuzzyEye
Moderator
Posts: 1402
Joined: 15 Jan 2015
Contact:

09 Feb 2016

selig wrote:As for extreme low levels, you CAN have a level down at - 200 dBFS in a floating point system - you'll have to raise it significantly before you could hear it, but you'll have the complete original dynamic range preserved when you do so.
I wrote the code to test this, and yeah -200 dB does cause a slight rounding error in the 8th decimal place, just as I predicted. But funnily enough -204.7 dB (34 bits) is OK. The problem is that -200 dB introduces an additional significant digit (being 33.219 bits). But -20 dB (3.322 bits) is OK, because 3.322 is less than 24. The loss of the 8th decimal place with non-integer bit depths starts above -144.49 dB (24 bits).

Still, not bad at all.

siln
Posts: 349
Joined: 11 May 2015
Location: france

10 Feb 2016

that means you can lower your mix as much as you want , and mix at -20 db you will have no problem
Was saying -20 per tracks, instead of -12 , some says -15 is a good spot , that s 8db less per tracks (12 to 20) i was assuming this at an maximum of lowering, this is not an advice but it was to point out that lowering is less destructive than maxing levels , in the range of -20 to 0 db .

Thanks also for all these clarifications about bits/lvls/headrooms :thumbs_up: ,

this might be relevant to post here the link for the short introduction manual of the SSL X ISM meter because it explains a few things about intersamples and BIT meters .
http://www.solidstatelogic.com/docs/X-ISM_Manual.pdf intersamples/bit meter, how much do you think it s useful ? at least it 's free :)

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

10 Feb 2016

siln wrote:
that means you can lower your mix as much as you want , and mix at -20 db you will have no problem
Was saying -20 per tracks, instead of -12 , some says -15 is a good spot , that s 8db less per tracks (12 to 20) i was assuming this at an maximum of lowering, this is not an advice but it was to point out that lowering is less destructive than maxing levels , in the range of -20 to 0 db .

Thanks also for all these clarifications about bits/lvls/headrooms :thumbs_up: ,

this might be relevant to post here the link for the short introduction manual of the SSL X ISM meter because it explains a few things about intersamples and BIT meters .
http://www.solidstatelogic.com/docs/X-ISM_Manual.pdf intersamples/bit meter, how much do you think it s useful ? at least it 's free :)
It should be noted that the more tracks you are prone to use, the lower your peak reference level should be. Consider an average of 3 dB gain for each doubling of tracks, and consider that not all tracks play at the same time, and the -12 dBFS peak level is a good starting place. If going into more detail, I would advise users to start at -12 dBFS peaks for all tracks, and adjust with their workflow if they are constantly still clipping the mix bus.

Also remember that when recording via microphones, using a lower level tends to run the preamps in their sweet spot (nominal level) whereas trying to "kiss the reds" tends to run the preamps hotter than they are designed to run (which in some cases may be a good thing, if you like that sound!).

Another reason I advocate for a reference level of any sort is to keep consistent processing levels when using compressors, gates, expanders, saturators, and distortion devices. The main advantage of this is a quicker more efficient (and therefor more creative IMO) workflow, nothing more - it's not going to make your mixes sound "better" as some have claimed… ;)
Selig Audio, LLC

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

10 Feb 2016

siln wrote: this might be relevant to post here the link for the short introduction manual of the SSL X ISM meter because it explains a few things about intersamples and BIT meters .
http://www.solidstatelogic.com/docs/X-ISM_Manual.pdf intersamples/bit meter, how much do you think it s useful ? at least it 's free :)
Inter sample peaks is definitely something you should look for when producing for CD or going to a compressed format (mp3, aac etc.). Fwiw, Apple supplies an AU plugin that lets you compare original vs compressed audio and also shows you inter sample peaks before / after decompression. Interestingly there can be higher inter sample peaks after decompression.

User avatar
Kategra
Posts: 327
Joined: 18 Jan 2015

10 Feb 2016

For me it's simple:
A)If I have a project with 4-5 tracks and I see the Audio Out Clip indicator on Transport Panel light up, then I move the mix channels fader down.
B)If the project has more channels, then I move the Master Fader down and leave the mix channels as they are.
Can't see anything going wrong with this; except for the SSL EQ + Send FX (reverb, delay), I don't use other sections, not even the SSL master comp.
I believe Reason has so much headroom for audio levels (dynamic range), our monitor chain +ears+brain can never use or enjoy in full.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

10 Feb 2016

Kategra wrote:For me it's simple:
A)If I have a project with 4-5 tracks and I see the Audio Out Clip indicator on Transport Panel light up, then I move the mix channels fader down.
B)If the project has more channels, then I move the Master Fader down and leave the mix channels as they are.
Can't see anything going wrong with this; except for the SSL EQ + Send FX (reverb, delay), I don't use other sections, not even the SSL master comp.
I believe Reason has so much headroom for audio levels (dynamic range), our monitor chain +ears+brain can never use or enjoy in full.
The only difference is that if you start with lower levels, you spend all your time mixing and not watching (and adjusting for) clipping. You also move more quickly through your compressor etc. setup since if you use consistent levels (whatever they may be) you will spend less time adjusting threshold/makeup etc. Finally, through there is tons of headroom, it IS possible to have levels so low or so high as to move beyond the nominal level of non-linear devices. For example, you can have levels so hot you can't have subtle compression/distortion, or levels so low you can't have ANY compression/distortion. While these are extremes, they illustrate that there are limits to what levels will work for you, and there are advantages to being aware of these limits and working within them. Also, when adopting a peak reference level, you can A/B easier knowing that if you add EQ or compression, and keep the results within the same peak reference level, you will not be fooled by simply making a track 'louder' (which happens more often than one would think once you start paying attention to these things).

The main advantage I keep advocating adopting a peak reference level for all your audio signals for is the simplification of your work flow, and the joy that comes from simply making music and not having to chase levels and be paranoid about clipping! I feel I can be more creative working this way, as the main "technical" work I do is at the beginning whereas if you ignore levels you are having to remain in "tech mode" more at the end of your workflow - and I prefer to reserve the end of the production for creative decisions if possible - makes me happier at least, and keeps the process more "fun" for lack of a better word. :)
Selig Audio, LLC

Stranger.
Posts: 329
Joined: 25 Sep 2015

10 Feb 2016

ΣΣΣ
Last edited by Stranger. on 06 Jun 2016, edited 1 time in total.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

10 Feb 2016

Stranger. wrote:Selig seemed to give a clear indication of the events and performance internally,but your good self seemed to also carefully swerve that topic,when you could have added to it imo.Cheerz m8.
I guess it again comes down to how much of what I said in that thread you actually understood.
Stranger. wrote: :re:ason is being played and recorded right?
There's a heap of converting going on from rec in <> output to speakers,and every step/sample/phase postion changes with each process.
Theres exactly two points where actual conversion happens in which timing of a quartz is involved. At the input of the sound card and at the output. Between that there is no jitter or timing issues because there is no time syncing happening or needed. If you store a bunch of samples on a hard disk and later read them again there is no jitter added based on the time they stay on the harddisk - I guess that should be obvious. Same for when they are stored in memory, same for when they are changed in memory. If you have a row of numbers like 0.1, 1, -0.1, -0.2, 0.4, 1 how would you change the "distance" between those numbers? You can't because that distance doesn't exist in the numbers, only when these numbers are converted to electrical signals again you can do that.

User avatar
ScuzzyEye
Moderator
Posts: 1402
Joined: 15 Jan 2015
Contact:

10 Feb 2016

Stranger. wrote: :re:ason is being played and recorded right?
Nope.

Do you understand what abstract means? I'm not talking about Reason. I'm not talking about speakers. I'm talking about how Pulse Code Modulation works. Not wave files either. Just PCM data. What happens when you multiply 1.0 by 0.0000000002328306437 (that's -192 dB, for the curios who are still following along), and then divide it by the same number again.

I wasn't even really talking about sampling rate, because this thread is about bit depth.

Stranger.
Posts: 329
Joined: 25 Sep 2015

13 Feb 2016

ΣΣΣ

Post Reply
  • Information
  • Who is online

    Users browsing this forum: No registered users and 32 guests