record and mix levels. loud or soft?

[arpodthegreat]

using PT LE

my friend (who i have been recording and working with off and on for the past few years) and i always get into the arguments. i say don't worry about recording at a level right before it clips, you can always turn it up later. then, when we mix i get so frustrated when he pushes the (digital) faders past 0. i like to turn everything down to that no red lights go off even in the master section. you can always turn it up right? he says it doesn't matter and the other people he's worked with (who are actually a lot more experienced then me and actually get paid good money to do this stuff) have tracks going in the red all the time and that it doesn't matter and not to worry and even that it sounds good. i try to tell him it sounds good because it's louder but that's all. i tell him you get a cleaner better sound if you do things at a moderate level and then just make up gain when you bounce or burn at the end.
the only thing i can think of in his defense is that i can push my preamp a little harder and get a better sound. actually this is a problem for me. how do i push my preamp to get the best sound out of it without clipping in PT. for example, when i record snare i want to crank the pre to get a nice full sound but can't cuz it clips PT.

any thoughts and these issues is muchos appreciated.

[jgrif08]

There's a VERY big difference between cranking the output of your preamp and cranking on an output in the digital realm. I personally would shy away from doing either. Part of the problem with recording and mixing the way your friend is doing it is that a) when things clip in the digital realm it sounds like CRAP-O-LA and b) you're not leaving any headroom. If your tracks are going 'into the red' when you send them to a mastering house then there's not whole lot they're going to be able to do with 'em. You're friend is probably under the impression that, with your current setup, he can get tracks to sound as loud as the commercial records you hear on the radio. There's no quicker way to ruin your mixes than to gun for that 'loudness'. Mastering houses have very expensive, very specific peices of gear that allow them to attain that sound.

Always be mindful of your levels throughout the entire process. Record and mix at a nice healthy level, and STAY AWAY FROM THE RED! They make those lights red for a reason...

Good Luck

[cdog]

You're right buddy.

Digital clipping is bad. Always. Forever.

Analog clipping is good. Usually.

Digital clipping might not always sound bad, but..... it is.

With 24 bit digital systems the beauty is the huge dynamic range.

You can record to medium soft levels then bring the gain up digitally to an appropriate mix level, you don't add any noise/distortion unlike an analog desk.

And always keep the digital master buss from clipping, just turn down the damn tracks, don't resort to using a limiter on the master buss. If anything a fast compressor taking off only 2-3db.

[theblue1]

Well, I agree 100% on the front end... if you're recording using a 24 bit interface, you have a fair amount of headroom and can afford to record at a lower, comfortable level. As noted, clipping your converter produces the bad kind of distortion. (And I don't mean good.)


Now, overloading the channels in a 32 bit float software mixer (as I believe LE is) is a little more ambiguous, since you don't run into a digital brick wall at 0 dB and can prevent disastrous clipping of your hardware by compensating for the over 0 levels of individual tracks by lowering the level of the bus.

BUT... not all plug ins are necessarily so forgiving. Overloading some of your plugs may cause unexpected results. It's reasonable to assume that many plugs are optimized for normal, below zero levels.

So, while your DAW software may handle track overs gracefully -- some of your plugs might not.

It seems to me that the safest bet is to keep on doing what you're doing.

[quincyg]

you want to get the hottest clean signal coming in without clipping...at least according to "mixing with your mind" which i can't argue with.

[Greg Curtis]

"Use your bits!" The lower the level, the lesser the digital resolution. It may be OK to occaisionally clip a percussion transient, but digital clipping is not a good thing.

[Geert van den Berg]

Originally Posted by Greg Curtis

"Use your bits!" The lower the level, the lesser the digital resolution. It may be OK to occaisionally clip a percussion transient, but digital clipping is not a good thing.

Because your mic and analog preamps/compressor also has some selfnoise and also the converters it's no point in trying to use all your bits. You won't gain extra precision.

You should use all of the analog gear at their optimum levels, in my book that does mean you not always hit the digital ceiling, maybe only on very dynamic and loud sources.

Decide for a reference level between -18dB and -12dB with peaks occassionaly going to -6 dB.

I wouldn't even let a percussion transient clip, offcourse if it happens there's no man overboard, but better to not let it clip.

After it's digitized there's also issue of the headroom that certain plugins require.

[RCM - Ronan]

There is no reason to get close to clipping the inputs of your digital recorder, especially at 24bit. just don't do it. Its bad, its wrong. tutt

I make about half my living mixing records that other engineers have recorded, and I am stunned at how sloppy some engineers are when it comes to clipping. A clipped out signal can be unusable in a mix sometimes, but on the other extreme if something peaks at -30 dBfs, that is not really a big problem for me at all.

I usually shoot for keeping my peaks around -6 dBfs.

The red lights at the top of the meter = bad

[tamasdragon]

Don't clip. I agree with the post above. It's interesting to see how many engineers doesn't understand the digital domain.

Tamas Dragon

[luctellier]

arpodthegreat,

Clipping is no good! tutt

Take a look at this article:

http://www.digido.com/modules.php?na...=article&sid=8

I personnally keep my tracks peaking at -6db...

I also started a thread regarding "Plugin Internal Clipping" not too long ago:

http://gearslutz.com/board/showthread.php?t=97159

Hope this help!

[theblue1]

I'm with Geert on one thing: each stage of your gear should be set to an optimal level for its own operation if at all possible. Analog gear has self-noise, of course, but even more importantly is often far from linear in its response to gain-staging issues. And -- if one stage is lower than its optimal level you might end up compensating by taking the next stage higher than its optimal level... compounding the problem.


That said, for me, trying to avoid clipping a converter isn't so much optimal as mandatory. I'll always try to sin on the side of caution, as a rule.

(But, yeah, if it happens on a transient, there's a fair chance you can mitigate the damage. I don't automatically stop a take just 'cause the 0 dBfs indicator goes on for a moment. But it is something I will carefully address if I'm going to try to keep the take. Well... as carefully as I do anything. And, since I'm my only client, these days, that ain't much. )

[George Necola]

I had to mix a blackmetal band last week (actually a pagan blackmetal band). the material was recorded by themself.. the vocals were totally clipped (convertorclipping), then I asked em why, and they answered: "we did the same on our last album.. the guy used something to get rid of the earkilling crackling":.

so I used some waves declicker and voilà:
great distortet raw tracks

the material is released as a majorlabel (blackmetal) release.. so..

besides that very special opportunity to use digitalclipping.. forget it and record at lower levels

cheers George

[Greg Curtis]

Originally Posted by Geert van den Berg

Because your mic and analog preamps/compressor also has some selfnoise and also the converters it's no point in trying to use all your bits. You won't gain extra precision.

You should use all of the analog gear at their optimum levels, in my book that does mean you not always hit the digital ceiling, maybe only on very dynamic and loud sources.

Decide for a reference level between -18dB and -12dB with peaks occassionaly going to -6 dB.

I wouldn't even let a percussion transient clip, offcourse if it happens there's no man overboard, but better to not let it clip.

After it's digitized there's also issue of the headroom that certain plugins require.

It's all relative, so I think that we are saying the same thing! The balance of input levels/digital headroom is always predicated with proper gain staging; can't have one without the other. But the mic pre's had better be so quiet that any self-noise is negligable at healthy digital levels, or I wouldn't use them. Since I prefer to record via a very simple chain without compression or anything else besides the mic and pre-amp, I usually find myself at your recommended levels on the first take, nudging them up or down depending.


Greg

.

[majellin]

I read that if you're making a 24-bit recording and if your level is below -6db, that you are really only making a 23-bit recording. Plus your transients come through on your 24th bit. Any digital signal that clips is really unusable, unless you are going for that nasty distortion on purpose. You can Always attenuate your signals if needed after you capture them with a really hot level. Reference monitoring should be around 85db. as that is where the frequency spectrum flattens out naturally, according to the Fletchor-Munson curve right?

[steveschizoid]

Originally Posted by majellin

I read that if you're making a 24-bit recording and if your level is below -6db, that you are really only making a 23-bit recording. Plus your transients come through on your 24th bit. Any digital signal that clips is really unusable, unless you are going for that nasty distortion on purpose. You can Always attenuate your signals if needed after you capture them with a really hot level. Reference monitoring should be around 85db. as that is where the frequency spectrum flattens out naturally, according to the Fletchor-Munson curve right?

No, no, yes, no, perhaps, and no.
Think about this. If you have a given dynamic range in the sounds you are recording, it will be converted into binary numbers that are representative of that given dynamic range, irregardless (given some common sense) of what your input dBfs meter shows. The important thing is to operate your mic pre's at their optimal gain settings. Of course, if you want to use your mic pre to generate distortion, that's different and is an artistic descision. If you are unwittingly distorting and adding uneccessary noise (due to exceeding the level where optimal signal to noise ratio exists) you cannot attentuate that sound. Read this excellent article for more detail: Proper Audio Recording Levels | Rants, Articles | My MASSIVE Blog
85 dB monitoring is the convention for what you can listen to for 8 hours without damaging your hearing, but the equal loudness contour is by no means flat at that level - more flat than at lower levels, yes.