Preparing DP Mixes for Mastering

[Prime Mover]

But they still have to work beneath digital zero, just like you do. They can't put back any dynamic range that you have removed by overcompressing the material; but they can compress it as hot as you want (and more transparently), if that's what you tell them to do. So give them room to do it, or not do it.

I wasn't actually refering to master compression or limiting. Obviously you don't want to put anything on the master track that the mastering engeineer can already do. I'm simply refering to peak volume on the master track. As long as you don't hit 0, you're not going to get any artifacts, right? There's no loss in dynamic level or quality.

In any case, if mastering engineers DO like 4db below 0, it's just a simple matter of taking the master slider down to -4, if you've been mixing near 0. It's a 32bit fp operation anyway, so it shouldn't impact the signal at all. I still don't understand why giving them less bits makes it any better for the mastering engineers. Bob Katz hasn't sugggested anything like that, as far as I've read.

Only reason I'm a bit defensive is I've just been mixing an album and using "just below 0" as a reference point, and this kinda complicates things. And further more, I just don't see any logic behind it. Leave dynamic range, yes... but why let your top 3 bits go to waste? It's basically like giving them a 21bit file with a bunch of added 0s. I've never heard of a mastering engineer promoting that.

[Dan Worley]

But they still have to work beneath digital zero, just like you do. They can't put back any dynamic range that you have removed by overcompressing the material; but they can compress it as hot as you want (and more transparently), if that's what you tell them to do. So give them room to do it, or not do it.

I wasn't actually refering to master compression or limiting. Obviously you don't want to put anything on the master track that the mastering engeineer can already do. I'm simply refering to peak volume on the master track. As long as you don't hit 0, you're not going to get any artifacts, right? There's no loss in dynamic level or quality.

In any case, if mastering engineers DO like 4db below 0, it's just a simple matter of taking the master slider down to -4, if you've been mixing near 0. It's a 32bit fp operation anyway, so it shouldn't impact the signal at all. I still don't understand why giving them less bits makes it any better for the mastering engineers. Bob Katz hasn't sugggested anything like that, as far as I've read.

Only reason I'm a bit defensive is I've just been mixing an album and using "just below 0" as a reference point, and this kinda complicates things. And further more, I just don't see any logic behind it. Leave dynamic range, yes... but why let your top 3 bits go to waste? It's basically like giving them a 21bit file with a bunch of added 0s. I've never heard of a mastering engineer promoting that.

Bob Katz does recommend mixing peaks down to below -3, and even as low as -10dBFS. This helps keep distortion from happening without the need for an extra stage of attenuation before applying various mastering processes.

Many knowledgeable engineers will advise you to bring your mix levels down and turn your monitors up.

Remember this, a peak meter is quite useless when it comes to mixing the body of the mix where the impact and power is.

c-ya,

Dan Worley

[jloeb]

David's right that it can be as little as -3 dBFS.

Katz, pp. 71-72, after a discussion of intersample peaks:

"What this means is that if you are mixing with a standard digital meter, keep peaks below -3dBFS, especially if you are using aggressive bus processing. The more severely processed, equalized or compressed a master, the more problems it can cause when it leaves the mastering studio... Even though 24-bit recording is now the norm, some engineers retain the habit of trying to hit the top of the meters, which is totally unnecessary as illustrated at left. Note that a 16-bit recording fits entirely in the bottom 91 dB of the 24-bit. You would have to lower the peak level of a 24-bit recording by 48 dB to yield an effective 16-bit recording! There is a lot of room at the bottom, so you won't lose any dynamic range if you peak to -3dBFS or even as low as -10dBFS, and you'll end up with a cleaner recording...A digital mix that peaks to -3dBFS or lower makes it easier to equalize and otherwise process without needing an extra stage of attenuation in the mastering."

Leave dynamic range, yes... but why let your top 3 bits go to waste? It's basically like giving them a 21bit file with a bunch of added 0s. I've never heard of a mastering engineer promoting that.

Bits don't map 1-to-1 to dBFS. A 24-bit recording has 139 dB of dynamic range.

[Dan Worley]

David's right that it can be as little as -3 dBFS.

Katz, pp. 71-72, after a discussion of intersample peaks:

"...Even though 24-bit recording is not the norm,..."

Should read, "is the norm."

c-ya,

Dan Worley

[jloeb]

Should read, "is the norm."

It's actually "is now the norm." Corrected above.

[intrepritra]

Bits don't map 1-to-1 to dBFS. A 24-bit recording has 139 dB of dynamic range.

If you UTILIZE it. It takes a GREAT converter and QUIET electronics to achieve that great a dynamic range and most (but ME's) don't have that level of quality so...
It boils down to a great inspirational performance and good techniques all the way through these days.

[jloeb]

The point was that you can turn down the volume 3 to 10 dBFS on your mix and still be utilizing it, very very much so. I.e., a recording with peak at -3 dbFS is not equivalent to a 21-bit recording, because the math is wrong.

It should be obvious that following technical best practices is not in any way opposed to things like artistry and musicianship. The whole point of having a Motunation and reference books like Katz etc. is to help us quickly figure out what the answers are, so we can just use them and forget about it, with max effort left over to put into getting a (well-recorded) good piece of musical art.

[Prime Mover]

For the record, I do know that dB:bits is not a direct ratio, I was just pulling out estimates to make a point. However, you have made your point as well. I still don't understand the math behind lowering mix volume, however it is a little safer, so I might as well practice it from now on.

[mhschmieder]

I leave a minimum of 12 dB of headroom at the mixing stage, and usually as much as 20 dB (genre-specific, as well as context-informed, such as film vs. stage backing vs. media distro).

You should also check out the T-RackS book from Bobby O., which has the most concise and informed/useful hints of anything I've ever read on mixing and mastering -- so much so, that I'm probably going to sell almost every other book I have on the subject (except Bob Katz' Art of Mastering, which is a different sort of book).

[melenko]

Funny, I've had the opposite problem, and if you think about it, it makes sense. Play/record is trying to do everything in realtime. Theoretically (and I can tell from speed decreases during the count), if you are bouncing to disk, DP doesn't have to do it in real time, if there's processing headroom, it can go faster, if it gets bogged down, it can go slower. Whereas playthrough-record is the computer trying to give you an accurate playback in time, and it can hickup in order to try and keep up with itself.

For that reason, I ALWAYS BTD. Maybe it's unwarranted in smaller projects, but I spent too long on an old TiBook with a slow G4 processor, and it became a habit that I have no particular need to break.

What you wrote makes total sense. That's what I would expect as well. But, un-intuitively enough, what I described is what I've experienced.

Since you are recording in real time, what you hear is what you get. If there are glitches you can immediately know.
I must say that I use different HDs for audio and samples, and I always close every window when I record like this (to relieve the CPU a little), so perhaps this helps.

When bouncing to disk, and the computer is very stressed, DP (in my G5) seems to "forget" or simplify automation. Sometimes the reverb or delays sound softer. And sometimes VIs are not even included in the final file. This doesn't make sense, as you said, because DP is doing all the processing off-line, so it could take as long as it needed to process everything, but it just doesn't seem to be the case every time, IME.

The worst thing is that this happened only SOMETIMES and under certain circumstances. Since I never knew for sure what would happen, I just stopped using BTD and just created a clipping for recording the final mix (since DP 4).

I confess I haven't really tested BTD with DP7. Maybe it works great now and I need to test it fully to trust it again... I hope it's reliable now! BTD is much faster and much easier to set up...

But since our experiences regarding this are opposite, I guess the best he can do is try both and see what works best. I suppose one way to be sure everything is perfect is to confirm that the recorded or bounced file cancels out with the session's full mix.

If you can, afford you an Intel Mac.

[FMiguelez]

If you can, afford you an Intel Mac.

I know! I've been putting it off for ages now.
Hopefully my summer royalties will be enough to get one of them newest babies. Oh, and lots of RAM. And all the apps I want to get..

[allemande]

A big vote for mixing in real time from my end.

There is no comparison in quality. I am actually going through extensive troubles (and joys) to mix in a more old school fashion to get the most transparency out of a mix:
- I have many MOTU converter channels (60 channels in my case, even more if I really need it).
- I Use a TASCAM 4800 digital console to dial in a mix, use its on-board EQ and dynamics, in addition to plug ins in DP; and mix everything into analog!
- I use the following analog stereo devices to mix through: SPL Vitalizer into API 2500, into an Avalon 747, (Yes, all at once). The 2500 does the main dynamics control. It is extremely versatile, from very subtle to very aggressive. Love it. The 747 boosts the sound quality immensely with its analog circuitry (hardly any compression from this box). The SPL allows me to stretch the stereo image a bit.
- Record the final mix onto a Masterlink.
- All D/A and A/D conversion is handled by a MOTU 192 HD (Black Lion Audio customized), with a Black Lion Audio Master Clock.
- Most reverbs are handled by external devices:Kurzweil KSP8, TC M2000s, and others.Occasionally I use Pro Verb plug-in.

This setup allows for breathing room in a mix. I could never get that with a B2D process.

[David Polich]

One thing I don't rely on is DP's master fader metering. I just try to
make sure I don't have levels clipping the top of the master fader. I
export the 2-track stereo mixdown to Peak and take a look at it there.
DP has a "find Peaks" process that will give you the highest peak and
the RMS average. If the highest peak exceeds -3 from zero DBFS in Peak,
I use "Change Gain" to reduce the gain of the file back down to -3db.
If the stereo track's highest peak is something like -8 db, I'll use
the same operation in Peak to bring it back up towards -3db from 0dbfs.

Frankly, I've never been able to get a mix in DP to average and peak
exactly where I would like it. Some mixes will be sounding great with
peaks and averages low, and some will be a little hot and sound great.
I don't care, I'll address this in Peak, not in DP. I don't care about
theorietical loss of bits or anything like that. It only has to sound
good and not distort. Maybe in theory I am losing some bits but if I
can't tell, then it doesn't matter.

Don't mix with your eyes. Mix with your ears. Seriously, it's a waste of
time to make sure all your math is correct. If it sounds good, it is good.
If it sounds crap, it is crap. Simple.