Ableton Forum

davepermen wrote:or turn the volume down on the actual sample / instrument? should do the trick?

That is the true root of the matter. It's more tedious and time consuming but it yields the best results. I hardly ever like to touch the faders, not even the master! (i.e.-leave them at zero). I try to end up with a headroom of about -2 dB. No limiters anywhere in the mix. Export as 32 bit float and then (pre)master outside of Live (Wavelab in my case).

djsynchro wrote:This is my favourite line from the first post in that thread:

Running a Digital mix right to the top of the scale is like running your SSL mix buss where the VU meters are slammed all the way to the right and you are constantly hitting it at +25.

Total nonsense!

sorry for my english.

why is it nonsense ?

paul frindle, the designer of the SSL console and Sonnox plugz say the same ...also Bob Katz similar in his "Mastering Audio - the art and the science" Book.



---this is a reply on the gearslutz thread---
Originally Posted by Blinddot

IN ANALOGUE:

nominal level = +4dBu = 0VU

max I/O level in analogue gear is something from 18dBu to 26dBu aprox (check specs)

headroom = max I/O level - nominal level

so typical analogue headroom is something from 14dB to 22dB depending on gear.


IN DIGITAL:

Digital "nominal level" is called reference level and it is not a voltage value in dBu like in analogue domain. It is just a reference below 0dBFS (max I/O digital level).

Because we need a standard, what it has done is to look at the typical analogue headroom (14dB to 22dB) and use a similar headroom in digital.

AFAIK in the music industry there is no standard reference level. In the broadcast industry the SMPTE (NTSC countries) standar is -20dBFS , and the EBU (PAL countries) standard is -18dB, both pretty similar anyway.


METERING:

The ballistics of a VU meter are 300ms slower than a digital peak meter (0ms), so you don't see fast transients, you can only see "average level", RMS is other kind of average level pretty similar to VU ballistics, even a bit slower, so you would see even less transients on a RMS meter, but similar anyway.

If you don't use some kind of average level meter (RMS or VU) you are not able to know what is the headroom you are using neither the crest factor of your signals.

The whole point of this topic is about being aware of the crest factor of your signals while tracking or mixing. In order to do this, you need both a peak meter and an average meter (VU or RMS), if you use one of these, then you need to set a headroom for it, unlike peak meters, average meters need a bit of setting up.

When tracking or mixing on an analogue desk you don't have to choose a nominal level and install average meters in your desk, they are built in, and the nominal level is fixed and standard on all analogue gear. That's why when mixing ITB if you don't have this very basic audio engineer knowledge you can easily do it wrong.

Actually, I dont' think this is so eye-opener for properly trained AEs, I guess many people were mixing and tracking correctly with analogue gear just because it was easy, Once they moved to digital, they just didnt know the technical knowledge in order to keep on doing it right.

IMHO, as some others have stated before, like UBK, this is not the reason why analogue may be superior to digital, it's all about non-linearities that analogue summing imposes over the signals unlike digital summing.

Clean it's boring, when you "push" the drum buss when mixing up to +3 or +4VU on a nice API mixing console, you get some "nice" distortion because you're pushing a summing amp quite a bit over its nominal level, if you do this with digital summing buss you either dont get any "special" flavor (32/64 float and 48bit fixed) or you get nasty digital clipping (24bit TDM buss). Analogue electronics give some of these non-linearities even when not pushed and they are working around nominal level.


hope it helps!

if it is like that, then the digital plugins you use have that issue. a digital path by itself does not have that problem. the only problem it can have is clipping, when you go ABOVE 0db. below, it works all perfect, that's the power of digital.

but if plugs can't handle the limits well, then it's their fault.


edit: and it still then means, you have to handle it so my original statement stays: know your limits, and get to the maximum

This is pretty much what happened on the GS thread - hugely respected guys with vast amounts of knowledge to impart, gained from years of working with analogue and digital systems share this with everyone to help improve their engineering abilities, and some people take that knowledge on board, some folks reject it and think they know better, and the whole thing gets confused.

The point about digital plugs being able to handle much higher levels without clipping is debated in great detail within the thread too, and there's more facts and figures then you'll ever need (or probably be able to fully understand)...inter-sample peaks causing clipping that meters cannot detect, etc. etc.

For my part, I'm not interested in pretending to know more then they do. I'm here to learn as much as I can and to test these ideas empirically.

I have, and what PF and Skip, etc. suggest as good operating benchmarks work for me. Sure, there are other ways to get to roughly the same results (like tracking cautiously, or lowering the levels of samples/instruments as suggested), but why not set aside what you think you may know and take some advice?

If you ultimately still disagree, and you've tested the ideas on offer, AND your mixes sound as good as any of the best engineered records ever made, then I salute you.

davepermen wrote:if it is like that, then the digital plugins you use have that issue. a digital path by itself does not have that problem. the only problem it can have is clipping, when you go ABOVE 0db. below, it works all perfect, that's the power of digital.

but if plugs can't handle the limits well, then it's their fault.


edit: and it still then means, you have to handle it so my original statement stays: know your limits, and get to the maximum

because my english is bad...i answer this one more time with two replys from the gearslutz thread...

Originally Posted by DSM Interactive

Pushing the peaks as close as possible to 0dB digital is exactly what you don't need to do.
It's a recipe for overs (intersample or otherwise) and for a hard, cold, grainy and two-dimensional result.
Leaving some headroom allows you to take advantage of the depth, space and clarity that the medium supports,
without scraping your knuckles on the ceiling.

------------------------------------------------------------------------------------------------------------------------------
Originally Posted by Skip Burrows

Without high end limiters protecting everything all the time better to keep your levels
referenced at 1.23 volts = -20 to give you headroom much like an analog console.
Stay away from 0DBFS unless you are in a mastering phase with the correct metering and tools.
TRUST ME, your mixes will open up at every level. Remember, when you work on an analog desk this is what it forces you to do,
with out you even knowing it. Enough said?

------------------------------------------------------------------------------------------------------------------------------
...and weedywet from thewombforums.com wrote

you care because the whole system will sound better with more HEADROOM.

those little bits of digital clipping that don't sound awful on each track, ADD UP to sounding brittle and thin and all the things you don't want when they're all mixed TOGETHER

-------------------------------------------------------------------------------------------------------------------------------

and i don't care as you don't need to compare analog and digital as they behave wastly different.
the only fact you have get is, the lower your maximum is, the less actual digital resolution you have. that doesn't mean much, as you have much too much resolution on a 24bit audio path anyways.

the thing that matters is what i said before, and evernaut said it, is, clipping can occur in places where you can't measure it, a.k.a. in the middle of a plugin. it might result in some aliasing artefacts, and other stuff.

so it results in the same i said before.. know your limits. the lower you go, the less signal quality you have, but the more chance to not hit a limit in the audio signal path somewhere.

but as long as you don't hit a limit, and this is different to analogue, you won't have any signal degradation 'just because it's more loud'. people with years of experience in analogue and analogue/digital environments have a hard time to learn that. but the digital world is different, vastly different. it everywhere is.

For myself, I'm not contesting anybody's knowledge, I'm just trying to understand what is being said. First their talking about ITB mixes, then talking about using hardware too. Then they're talking about standards in digital mixing, but say that the music industry doesn't have a standard. They're talking about VU meters and clipping, but then they mention that digital meters have 0ms response time and so they avoid the problem of missing overs.

From what I can tell, they are simply saying that digital clipping is bad, and we should all make sure we are not clipping at any point in the signal path. They suggest the way to do this is to mix towards a -20 dbFS standard. I'm I understanding this right?

Maths and physics aside for a moment - I think the main thing to take away from the points raised is that working in the mix with individual tracks down somewhere in the region of -20dB as a reference point allows you to get the best from any subsequent compression/EQ/reverb plugins or hardware outs you may wish to add.
This is as much for artistic reasons as scientific ones. Things just sound better. And if your levels creep up a bit with added DSP, then at least you've started from a sensible point and you probably won't run into trouble.

And accurate metering helps us know what we're doing, rather than relying on DAW faders which only adjust the level feeding the mix buss or master fader summing.

Also, it practically guarantees you more open and dynamic mixes...more akin to how great recordings (and great performances) actually sound. And it's true that some plugs are designed to take higher levels ( and some lovely saturation effects do come from pushing things a bit harder for flavour), but...for example... just because your fancy modern sports car can do 220 mph doesn't mean you have to drive it that fast all the time

evernaut wrote:Also, it practically guarantees you more open and dynamic mixes...more akin to how great recordings (and great performances) actually sound.

I've definitely found that to be true for me.

Use ears.

Robert