Recording Levels

[philbo]

Just extend bit depth to 24 (if needed), normalize to -17 dBFS, then mix.

Not that big of a deal, unless all the tracks are compressed already - - if so, give it back and ask for uncompressed tracks.
________
M62

[MASSIVE Mastering]

The DAW will probably throw calculations in 24 bit or 32-bit FP.

Normalizing to -17 isn't going to solve anything. Nothing.

And it *IS* a big deal. If the tracks were recorded that hot, they've already been overdriven at the input stage. Reducing the volume isn't going to correct that - Nothing will.

[RodC]

If the tracks were recorded that hot, they've already been overdriven at the input stage. Reducing the volume isn't going to correct that - Nothing will.

So every preamp out there would be overdriven if I tried to record at -6 levels? How can you make a blanket statement like that? Can you point us to some specs to read?

Many preamps will prob provide this much gain without too much problem. With my delta 1010 set on -10 db input I'm not even touching the top of my headroom with many preamps. With it on +4 there are some that I run out of room and then normaly record at lower levels.

I'm not trying to be nasty but unless I hear a preamp at various levels I cant judge if it is breaking up at unacceptable levels at 50% or 90% of its headroom.

How can you judge a signal and say that it would have been better when you did not get to hear what the eng heard when he did the setup?

[Mark Alan Miller]

That was my point. That I can often record hitting pretty damn hot without even stressing the preamp. Some signal paths will be able to drive the inputs of a digital system cleanly to 0dbfs, some will not.
But that is only half of the issue. The other half is having too hot of output during mix, where one is forced to reduce the gain anyway. Two different, but often related, issues.
Sorry to be repetitive.

[Mark Alan Miller]

Mark, just wanted to say that I'm a born and raised Western MA. boy and I've always been impressed by your recordings for the various valley indie rock bands I knew and loved back in the old days- the Mitchells come to mind immediately. When I move back from Brooklyn, can I have a job? I'll make you coffee. I'll do all the soldering. Anything.

Wow! Thanks for the kind words! Have we met?
I don't have a staff... but, PM me anyhow.

Oh, the new Mitchells CD should be coming out soon, BTW. Lovely record, IMHO.

[MASSIVE Mastering]

So every preamp out there would be overdriven if I tried to record at -6 levels? How can you make a blanket statement like that? Can you point us to some specs to read?

Well, by definition, once the pre-gain signal of the preamp goes beyond line level, it's being overdriven. So, yeah, I suppose I can make a blanket statement like that. Unless your converters are calibrated to -6dBFS = line level...

Of course - If you like the sound, I'm not telling anyone not to do it - In 25 years of engineering, I've *never* heard a signal shoved that hard that actually sounded better than the same signal at nominal. Of course, 25 years ago, hitting tape that hard would cause all sorts of problems and print-through. I don't have "specs" but it's not a secret that the distortion levels rise dramatically and the S/N gets thrown out of whack when you drive a preamp that hard - It's no different than a Marshall stack. Some distort less than others, but 0dBVU has been the standard for over 60 years. Digital didn't change that - It was designed to work around it.

But again - If that's what you're looking for, go for it. As long as you're aware of what you're doing, it's all good. But that's not my issue:

My concern is with the people who don't understand the basics of gain - The people who think that "making the signal hot without clipping" is proper and normal. The ones that *don't* understand that they're *NOT* getting a better S/N by recording hot - They're getting worse. The ones that think they're getting *less* distortion, when they're actually getting more - Sometimes considerably more. The ones who "just can't get the mix to sound right" -- This is something that I run across so often that it blows my mind.

And every single time, I tell them to try it themselves - Record something hot and the same thing normal. Mix them both and turn the loud one down to match the normal one. And every single time, without a single exception that I can remember, they come to the same conclusion - They were tracking too hot all along. They didn't notice the distortion with a couple tracks - But when they mixed it all together, it adds up. The clarity and focus were veiled, the "punch" wasn't "punchy," the highs weren't open and airy - The usual. Again, this isn't a secret - This is why preamps are spec'd particular level. If you think the distortion at 0dBVU is the same as it is at +12, I'd argue that it isn't... Just because a preamp can handle a hot signal without *CLIPPING* doesn't mean that it's perfectly clean up to the clip point - The clip is the point of complete failure. There's an awful lot of degredation before that point. The headroom is there to handle transients - not the "weight" of the entire signal.

[farview]

That was my point. That I can often record hitting pretty damn hot without even stressing the preamp. Some signal paths will be able to drive the inputs of a digital system cleanly to 0dbfs, some will not.

It's not just about the preamp. There is also analog circuitry in the converters all the way up to the point of conversion. If you have a compressor (for example) between the preamp and the converter, you are also running that hot. The cheaper your equipment is, the more this will make a difference.

Everything in your audio chain is spec'd at line level. That is where it was designed to have the best S/N ratio, the best distortion characteristics, this is where it can best handle transients. It's designed that way.

Depending on your calibration, somewhere around -18dbfs is line level. It is always easier to do this when you have actual VU meters, you just set the analog chain so that it hovers around 0dbVU and the digital side will take care of itself. That is the way digital was designed to work with analog gear. Oddly enough, when you use the stuff as it was intended and as it was designed, your results get better.

[jeddypoo]

Mark, just wanted to say that I'm a born and raised Western MA. boy and I've always been impressed by your recordings for the various valley indie rock bands I knew and loved back in the old days- the Mitchells come to mind immediately. When I move back from Brooklyn, can I have a job? I'll make you coffee. I'll do all the soldering. Anything.

Wow! Thanks for the kind words! Have we met?
I don't have a staff... but, PM me anyhow.

Oh, the new Mitchells CD should be coming out soon, BTW. Lovely record, IMHO.

I don't think we've met, but I was always hearing about Slaughterhouse, and how you would sometimes record drums in the old meat locker or something- anyway, I always liked everything I heard from there sound-wise! Back then, I was still 4-trackin' myself- the only proper recording I did was at Steve Wardlaw's place. Did you every do The Marshes? I imagine you know Steve. Anyway, I'll PM you. Yeah, I love the Mitchells. I saw Caleb a little while ago here when New Radian Storm King played in the city. I've always been a big advocate of his songwriting.

[joel hamilton]

I love mixing stuff that people have tracked "too quiet" in their words.
I get to wind on some gain from any number of sources, like tube comps, tube EQ's, beautiful equipment produces very nice gain and proper gain staging( for tone and the task at hand) is always more fun than something really fucking loud that always has "that" sound... The sound of ATTENUATING at every frikkin stage.
Cranking something up in the mix is SO satisfying when it wasnt plowed into the converters OR to tape when being recorded. Same deal with tape, different outcome, but still... a loss of transient information, loss of overall punch and clarity, loss of "quality" overall.

You can always make something recorded well sound beautifully F'd up, but something recorded WAY too hot just always sounds poopy. It can get less poopy, and maybe good, but not better than if it was recorded right in the first place!

use your ears.

[Mark Alan Miller]

That was my point. That I can often record hitting pretty damn hot without even stressing the preamp. Some signal paths will be able to drive the inputs of a digital system cleanly to 0dbfs, some will not.

It's not just about the preamp. There is also analog circuitry in the converters all the way up to the point of conversion. If you have a compressor (for example) between the preamp and the converter, you are also running that hot. The cheaper your equipment is, the more this will make a difference.

Everything in your audio chain is spec'd at line level. That is where it was designed to have the best S/N ratio, the best distortion characteristics, this is where it can best handle transients. It's designed that way.

Depending on your calibration, somewhere around -18dbfs is line level. It is always easier to do this when you have actual VU meters, you just set the analog chain so that it hovers around 0dbVU and the digital side will take care of itself. That is the way digital was designed to work with analog gear. Oddly enough, when you use the stuff as it was intended and as it was designed, your results get better.

I know it's not just the preamp. Sorry for the oversimplification. I've said in other posts in this thread things like "signal path" - figured that was already a given in the discussion.
Yeah, sometimes my console's pres will put out a clean, but hot, level that something between it and the converters can't handle. And sometimes that sounds cool. Sometimes not. Recognising that difference is the key.

And yes, operating within the gear's design parameters is part of proper gain staging. But sometimes some gear will yeild excellent results on the outer edges of that range of tolerance. That is what I'm talkng about. And in my particular case, I'm getting great stuff even when my converters see occasional peaks approaching full scale.
As always, YMMV.

Also, when people mix and match gear operating at different nominal levels, things can get even hairier!

On another note, wouldn't it make sense that the electronics in properly designed converters be designed to remain as linear as possible throughout their usable range? Why would anyone deliberately design analog front end for A/D that gets less linear as one goes above line level, be it -18 or -12 or whatever? (Aside from designs that do this on purpose, like the HEDD...) To the contrary, one would assume that the analog stage in a converter would be designed (hopefully) with a touch of headroom above what would be required to record or reproduce 0dbfs. Sure, cheaper converters, not so much, but these days even cheap converters are reaping the benefits of decades of converter design. At least we should be able to agree that decent A/D and D/A should be able to handle most of what is thrown at it in context of this discussion.

Just some more thoughts...

[Mark Alan Miller]

I don't think we've met, but I was always hearing about Slaughterhouse, and how you would sometimes record drums in the old meat locker or something- anyway, I always liked everything I heard from there sound-wise! Back then, I was still 4-trackin' myself- the only proper recording I did was at Steve Wardlaw's place. Did you every do The Marshes? I imagine you know Steve. Anyway, I'll PM you. Yeah, I love the Mitchells. I saw Caleb a little while ago here when New Radian Storm King played in the city. I've always been a big advocate of his songwriting.

Thanks again!

Yep - the old location had the old meat locker as a "drum room". Designed the new place's "drum room" to try and keep some of that sonic flavor.
Yes, I did a couple of Marshes records. Steve and I go back to High School, actually... he had a nice little studio in his house there. What band were you in?

And isn't it fun having Caleb in Storm King? Great rhythm section for them, with JJ on drums...

Sorry for the tangent folks.

[MASSIVE Mastering]

OT @ Farview: Hello Jason! Haven't seen you 'round here in... in... In a long time I'm sure.

@ Joel - Thanks for jumping in. "Too quiet..." It's actually happened once that I can think of in the last 15 years or so - A project that came in *too* quiet. And it was a malfunction of some sort - It wasn't even a recording level issue.

[RodC]

Cool info guys, makes me feel a lot better about some of the Altec pres I have. I have always recorded at the levels they sound the best at (Around -12 for the peaks) And sometimes I felt I was doing a band injustice because everthing you read says record it as hot as possible. I recorded some stuff out of a sansamp for a band that sounded killer but it was very low prob peaked at -20?? I was a bit worried but we all liked the sound.

I went back and looked at some of the tracks I did with my 1588c pres and they all produced about -12 to -18 with the lowest gain and the output level at the middle. (when I used them for drums.)

I rarely patch in a comp during recording but now Im going to give this a shot because I think some of the issues I had when doing it was that I was not happy with the gain staging. I remember a Vox soud I was trying to get through my 166a and it didnt sound right because I had to use too much makeup gain to get the peaks to -6

So what do we do when we get more bits? Are all converters that are more than 16 bits going to be calibrated for 0db = -18 ???

[joel hamilton]

So what do we do when we get more bits? Are all converters that are more than 16 bits going to be calibrated for 0db = -18 ???

More bits means better resolution near the BOTTOM, so thankfully, we could record even LOWER and still capture the sound with a great deal of fidelity. That means a nice pre, that gets "grumpy" trying to scream out its back end into your converter, can be used within its optimum operating range of levels.
This is for tone. This is not a "rule" laid out by anyone... this is basic gain staging.
With tape, if you record too little level, you get lots of hiss. With decent 24 bit converters, if you record too little level, you get... um... less level. (within reason of course). Just remember this:
The frequency response of the capture device is affected by level. tape or DAW.
Use your judgement, and your ears. Learn what you can about where YOUR system is happy, and sounds best to you.
Oh, and by the way, PT will report peak level: Just option click on the "volume window" above the fader. It will show you the peak level after showing the channel delay (if any). Shift>click on the peak level to reset it.
I almost NEVER use that function unless I KNOW I am printing a mix slightly hot, back from my two track, and I want to see if I am just too far to let it go to mastering that way... If I am getting peaks above -4 I will usually chill a liittle, and print again...
Transients make music fun to listen to!
Music IS transients!

[farview]

So what do we do when we get more bits?

Like Mr. Hamilton already explained, more bits allows you to record quieter.

24 bit gives you (theoreticaly) 144db of dynamic range. This is 40-some -odd-db more dynamic range than any of the equipment used to feed the converters can supply. In fact, you would be hard pressed to find a room quiet enough to have 144db of dynamic range available out of any instrument you are trying to record.

I'm not saying that there will never be converters with more bits, I'm just pointing out that we have already hit the point of diminishing returns.