DAW Mixing: Lower Levels=Better Sounding Mixes?

[uptoolate]

I've been reading alot about keeping the input and channel levels lower (around -6) to get cleaner, better mixes.

Has anyone found this to be tue? I for one, am a believer.

My question is this though: When doing "do it yourself mastering" ITB on a 24 bit file, is it still desirable to keep everything lower?

I basically use EQ,mulitband compression and the L2 plug to punch it up, but do I just accept that it will be more quiet?

Or, since this is the final stage, is it ok to go ahead an take it close to zero?

I know the real answer is "dont master it yourself", but for some of my clients, pro mastering is out of their budget.

Thanks!

[Kiwiburger]

If you want quality, it's a good idea to never hit 0dB. Bear in mind that the converters in domestic CD players don't often have brilliant analog chips, and they get a bit crunchy sounding as you approach 0dB. They can always turn up the volume if they want it louder.

But in the real world - modern CD's are ridiculously loud and clipped, and nobody seems to give a toss. Also, there are so many digital players that attenuate in the digital domain, so maybe those cheap converters will never be have to approach 0dB anyway ... it's not the 70's anymore, few people would recognise hifi if it bit them on the arse.

If you want to sound like a modern CD, smash the tops off your waveform - judge how much distortion you can tolerate.

[uptoolate]

Originally Posted by Kiwiburger

If you want quality, it's a good idea to never hit 0dB. Bear in mind that the converters in domestic CD players don't often have brilliant analog chips, and they get a bit crunchy sounding as you approach 0dB. They can always turn up the volume if they want it louder.

But in the real world - modern CD's are ridiculously loud and clipped, and nobody seems to give a toss. Also, there are so many digital players that attenuate in the digital domain, so maybe those cheap converters will never be have to approach 0dB anyway ... it's not the 70's anymore, few people would recognise hifi if it bit them on the arse.

If you want to sound like a modern CD, smash the tops off your waveform - judge how much distortion you can tolerate.

Thanks. It's a trade off. Less squashed stuff shure has more punch.

[Stiff]

Sorry, misread your post

[deuc647]

Yes, i found that if u just kep it at like -3 db it gives the track more room to breath and sounds more open.

[adrianex]

My mixing level has always been around -10db (picks at -6db) and one of the reason is that I rather leave the over all volumen part to mastering and My mixes has come better sounding to me when I give the ME space to work...

Also, I've been reading Bob katz's book (good reading BTW) and I agree on one part, when he says that only high quality meters can really tells you when you get to 0db or clipping.

Jus my $0.02

[The MPCist]

For me, if it's ITB, it stays at a max of -3. If I'm mixing on the SSL, I just forget the levels (VU's on the SSL) and get it to the sweet spot. I do take care not to slam the AD's for the mix though...

[Clarence]

Interesting topic.
I was once asked to do a remix (itb) for a friend, who was not 100% satisfied with his own work.

Because I liked the overall balance, I decided not to start from a scratch but to look at his mix first: The level was pretty close to the 0dB (not clipping, but close) so I decided to bring it down a little. In the process I carefully adjusted every track, checked every plug-in for its in- and out gain. I found a few comp/eq combinations where the output gain from the first was to hot for the following plug so I simply brought the levels down—carefully adjusting comp thresholds to match the original settings.

In the end I just moved faders to keep an even balance between tracks and bus (eg track +3 but bus down to –20 became track –6 bus –6) and aimed for a more conservative level of –3dB for the stereo bus.

I expected the differences between the remix and the original to be subtle—they were not. Of course I must have messed with the original balance, but being careful and sober and what not I am pretty sure the biggest step forward was to keep all the signals at healthy levels.
I am aware of the floating-point-thingy (almost unlimited headroom itb) but it seems that plug-ins behave quite different from their analog counterparts when driven hard.

I usually keep all stuff with figures on it, manuals and college education in the “I-don’t-care-locker” but in this case I would be grateful for somebody to comment, explain and hopefully confirm this experience.

BTW: I was talking about levels that are usually considered to be okay—no audible distortion etc.— and “bringing down levels” meant 0.5 – 1dB.



Clarence

[adrianex]

Originally Posted by Clarence

.....it seems that plug-ins behave quite different from their analog counterparts when driven hard.




Clarence

That so true!!!

[uptoolate]

Originally Posted by Clarence

Interesting topic.
I was once asked to do a remix (itb) for a friend, who was not 100% satisfied with his own work.

Because I liked the overall balance, I decided not to start from a scratch but to look at his mix first: The level was pretty close to the 0dB (not clipping, but close) so I decided to bring it down a little. In the process I carefully adjusted every track, checked every plug-in for its in- and out gain. I found a few comp/eq combinations where the output gain from the first was to hot for the following plug so I simply brought the levels down—carefully adjusting comp thresholds to match the original settings.

In the end I just moved faders to keep an even balance between tracks and bus (eg track +3 but bus down to –20 became track –6 bus –6) and aimed for a more conservative level of –3dB for the stereo bus.

I expected the differences between the remix and the original to be subtle—they were not. Of course I must have messed with the original balance, but being careful and sober and what not I am pretty sure the biggest step forward was to keep all the signals at healthy levels.
I am aware of the floating-point-thingy (almost unlimited headroom itb) but it seems that plug-ins behave quite different from their analog counterparts when driven hard.

I usually keep all stuff with figures on it, manuals and college education in the “I-don’t-care-locker” but in this case I would be grateful for somebody to comment, explain and hopefully confirm this experience.

BTW: I was talking about levels that are usually considered to be okay—no audible distortion etc.— and “bringing down levels” meant 0.5 – 1dB.



Clarence

That is really interesting. I have never thought about gain structure within plugin chains. I certainly would in the analog domain. To do this, I guess you would just look at any meters that the plugins provide? How would you check i/o levels on plugins that don't have meters?

Thanks!

[AMIEL]

Originally Posted by Clarence

but it seems that plug-ins behave quite different from their analog counterparts when driven hard.
Clarence

Yes...but is nothing wrong in pushing certain plug ins as you would do it with analog gear as long it sounds good...the important thing if you do that is to use the Trim utility in Protools o Trim plug in in others to have a clean input in the overall track!!! that makes a difference!!

[not_so_new]

Here is what Paul Frindle has to say about this over at R/E/P... for what it is worth I agree 100% with this.... especially the last line....

"Every time this comes up I am left with exactly the same totally frustrated feeling.
The reason people do not get best results from ITB mixes and digital processing in general is that the whole cultural environment of metering, level control, overload and production styles within the digital domain is based on SAMPLE VALUE and not SIGNAL
However many times I re-itterate this very important fact it seems impossible for people to grasp exactly what it means and what the gravity of ignoring it actually is in repect of their audio results. And this is NOT even the user's fault, they cannot be expected to grasp it because they are totally buried in systems that are wholly based on sample value misconceptions and always display values which are NOT signal
For a really fair analysis, this is not primarily a user problem - it is an equipment problem that the user must make himself aware of if he is to avoid it.
People who hear differences are not wrong - the equipment is lying to you - it is ecouraging you to produce illegal results that you are not made aware of.
IMHO & LE this is the sole reason underpinning ALL the arguments about ITB mixing, sample rates, 'resolution' - you name it."

[AllBread]

Originally Posted by not_so_new

Here is what Paul Frindle has to say about this over at R/E/P... for what it is worth I agree 100% with this.... especially the last line....

Great, but what the hell is he talking about? What does he propose is the cure to the problem?

[over-man]

Originally Posted by AllBread

Great, but what the hell is he talking about? What does he propose is the cure to the problem?

I was a little confused with that quote too.

[not_so_new]

Sorry guys... I guess I should have posted the link as well

Let's just say this. What Paul and others are eluding to are a few different things. Overall one of the biggest keys to good recording if not the biggest is following proper gain staging.

24 bit resolution is much bigger than the dynamic range of most of your hardware. Whey trying to squeeze out that last bit of resolution what you are doing is pushing your preamps and other external gear way over it's operating voltage in most cases.

Also while in the computer using plugins you need headroom for processing. On top of that the mix buss needs headroom as well.

So to your questions about that to do...

1) don't ever worry about getting your input levels up to 0dbfs, keep your levels low.

2) In the computer back down all your channel inputs to clean up your plug ins and mix buss.

3) Never let any of your plug ins clip at all.

4) Don't rely on peak meetings for overs, just never let your levels go above -6 or -8.

This is a different thread but it covers almost the same thing.... good stuff.

PSW Recording Forums: Whatever Works => Digital tracking with low levels = better...is this new???

[Thebassist]

There's a fine paper about this that Nica from Swetwater wrote years ago. Anyone know the paper, or have a link? If I recall correctly, he was advocating -6db at 24 bit.

[RainbowStorm]

Originally Posted by uptoolate

I've been reading alot about keeping the input and channel levels lower (around -6) to get cleaner, better mixes.

Has anyone found this to be tue? I for one, am a believer.

My question is this though: When doing "do it yourself mastering" ITB on a 24 bit file, is it still desirable to keep everything lower?

I basically use EQ,mulitband compression and the L2 plug to punch it up, but do I just accept that it will be more quiet?

Or, since this is the final stage, is it ok to go ahead an take it close to zero?

I know the real answer is "dont master it yourself", but for some of my clients, pro mastering is out of their budget.

Thanks!

Ok, or not, my view is that in the digital domain the converted signal is already so messed up in terms of harmonics and frequency response that a little too low signal will become a problem sooner or later depending on the context. Since a track can be recorded in so many ways focusing too much on the peak level can make the song pretty harsh sounding in the end if you forget using your ears and taking care of other more important sound parameters. I focus more on attack and duration (track and mix based) to make sure these parameters are optimized in the context. I also want to make sure there are no peaks on individual tracks damaging the headroom of the whole mix, often found on kick and vocals. So I think these three parameters are more important than knowing whether the ceiling should be set on -0.5dB or -1.4dB... I also find it more important to use a good peak meter and calibrate it correctly.

[Bob Olhsson]

Originally Posted by RainbowStorm

...my view is that in the digital domain the converted signal is already so messed up in terms of harmonics and frequency response that a little too low signal will become a problem sooner or later depending on the context. ...

Not nearly as messed up as a clipped signal. But rather than theorize, just try it and listen!

[amost]

Originally Posted by Kiwiburger

If you want to sound like a modern CD, smash the tops off your waveform - judge how much distortion you can tolerate.

That just made me laugh...excellent!

[RainbowStorm]

Originally Posted by Bob Olhsson

Not nearly as messed up as a clipped signal. But rather than theorize, just try it and listen!

I agree about that, but in my opinion you don't have to choose either way, you can use the headroom efficiently and still not have a final mix that clips. Use a good peak meter configured correctly, use your ears rather than your eyes, don't limit yourself to fake levels, practises and rules. Do what works for you, for me it works to target -0,3dB.

[Mondo]

i mix very low. the only time this has been a problem is with drumagog... i think that plugin WANTS you to clip its input, or it'll make your kick sample sound like a feather.

[bobwarren]

Back when I recorded analog, my goal was to push the input gain going to tape to keep the signal away from the noise floor and balance it with puching too hard to avoid tape distortion/compression (although sometimes desirable). When I started recording digital, I was encouraged to get as close to zero without going over to "fillout the bit map". When tracking, I put up the DAW input faders all the way and judge the input gain on the preamp by listening for distortion and monitoring the waveform. The idea is to provide enough gain to keep the audio away from the noise floor, "fill out the bit map", and avoid distortion. Am I doing something incorrect?

[not_so_new]

Originally Posted by bobwarren

Back when I recorded analog, my goal was to push the input gain going to tape to keep the signal away from the noise floor and balance it with puching too hard to avoid tape distortion/compression (although sometimes desirable). When I started recording digital, I was encouraged to get as close to zero without going over to "fillout the bit map". When tracking, I put up the DAW input faders all the way and judge the input gain on the preamp by listening for distortion and monitoring the waveform. The idea is to provide enough gain to keep the audio away from the noise floor, "fill out the bit map", and avoid distortion. Am I doing something incorrect?

Yes, absolutely 100% yes.. yes yes yes you are doing this wrong at least in my book. Others will disagree but this is one time where the "math" and the "ears" are on the same page. Try much MUCH more conservative levels and watch your ITB mixes open way way up.

Most preamps and other outboard gear were never designed to run this hot. You might not notice flat top wave forms but all gear has a sweet spot and usually it is not at the very top of it's operational range. Your hardware will speak much better at lower levels 99 times out of 100.

As I said above...

"24 bit resolution is much bigger than the dynamic range of most of your hardware. Whey trying to squeeze out that last bit of resolution what you are doing is pushing your preamps and other external gear way over it's operating voltage in most cases.

Also while in the computer using plugins you need headroom for processing. On top of that the mix buss needs headroom as well."

I really think it is helpful to check out the R/E/P thread I posted above.... VERY eye opening.

READ HERE

[not_so_new]

From that thread, Terry Manning says

Originally Posted by Terry THE MAN Manning

The stuff about "not using all the bits" approaches untrue urban myth status.

The old enemy in the analogue world was noise. We all learned recording techniques based upon that premise, whether we worked in the analogue world, or just learned from those who did so.

Noise IS NOT A REAL FACTOR in digiworld.

Keeping levels low helps almost every aspect of what people thought was bad about digital (harshness, tinniness, crunchiness, distortion).

You could even record -40 to -50 dB down, and be better! But -12 to -18 is great.

[uptoolate]

This is all really helpful info. Thanks.

A few have mentioned that the meters are not acurate in software. As far as getting a high quality meter, and calibrating it. I'm afraid I'm a little in the dark about how to do that.

Are we talking about a dual channel meter?
Where would it go in the chain for someone who mixes entirely ITB?
How would you calibrate it?
What would a good one cost? (not the high end, but one that would be acurate still)

Thanks!

[Hepha Luemp]

Originally Posted by uptoolate

This is all really helpful info. Thanks.

A few have mentioned that the meters are not acurate in software. As far as getting a high quality meter, and calibrating it. I'm afraid I'm a little in the dark about how to do that.

Are we talking about a dual channel meter?
Where would it go in the chain for someone who mixes entirely ITB?
How would you calibrate it?
What would a good one cost? (not the high end, but one that would be acurate still)

Thanks!

The question to ask here is: What is intersample peaks?
intersample peaks - Google Search

Hepha Luemp

[not_so_new]

Originally Posted by uptoolate

This is all really helpful info. Thanks.

A few have mentioned that the meters are not acurate in software. As far as getting a high quality meter, and calibrating it. I'm afraid I'm a little in the dark about how to do that.

Are we talking about a dual channel meter?
Where would it go in the chain for someone who mixes entirely ITB?
How would you calibrate it?
What would a good one cost? (not the high end, but one that would be acurate still)

Thanks!

Check out the thread I posted, Terry talks a little about metering in the digital world as well but in the end the problem is NOT just levels in the box, it starts with level TO the box. As I said, there are problems with the rail voltage of your preamps and other external gear.

The only reason why we used to push things in the analog days were beause of the high noise floor of the tape. Digital noise floor is way way lower so don't use analog thinking in the digital world!

If you want to minimize your issues as Terry says above just keep your PEEK LEVELS around -12 to -18 and that should be enough to keep your gain stage from preamp to converter to plug in to digital mix buss clean.

Good luck.

[bobwarren]

Hey, thanks for the thread. I read through most of it and I understand the concept. I have a question: If I set the input DAW fader at 0 and adjust the gain on the pre-amp to where the sweet spot is, sometimes I get the square topped waveforms. To me it's kind of like an amp that has both gain and master volume in that you push the gain into distortion and keep the master fader relitavely low. Is there a difference if I record the way I've been doing it and just pull the DAW faders down a bit when mixing?

[charles maynes]

I cannot agree with Terry Manning's contention that there is no noise in the digital domain...

Whatever noise is bring heard by the mic will be amplified if that channel is raised in volume, Unless you are using gates on everything, I would suggest trying to get peaks that are no less than -12 dbfs... that should allow for plenty of headroom in most cases.

charles maynes

[Spyrow]

I understand that with 16-24 bits tracks clipping can occur and it's not good to be near of 0dBs in the DAW but... almost every DAW is working now with inside 32bits files which gives lots of headrooms so, why care? Plugins are working the same, aren't they? Just care about the master out if you are going down to 24 or 16 and that's all. What do you think?

Anyway I'll test it tomorrow with one mix I'm working right now..