question about tracking too hot......

[dastrick]

I just read Massive's article on tracking too hot...

I'm going under the "no stupid question" assumption, here. Let's say that I've tracked to -12dbFS, then in my DAW I increase the gain by 10db (not moving the fader, but processing the track to increase the gain).

Is that the same as tracking at -2dbFS?

 

[Paulie Jay]

Good, you've saved me from having to ask this one. I eagerly await the answer

 

[mattr]

No. You are applying clean digital gain, which is different to running everything too hot on the way in which pushes into the headroom of your equipment and possibly might bring out unwanted characteristics in your preamps, converters, etc. On the other hand, you will still make mixing more difficult by having lots of tracks hitting -2dbFS.

 

[legionserial]

Is that the same as tracking at -2dbFS?

No but if all your tracks are at -2dbFS you're still going to have very big problems keeping the master output from clipping. -2dbFS is a little loud for even the master output. I'm really not sure why you'd want to do this. Is this question purely hypothetical?

 

[SouthSIDE Glen]

(not moving the fader, but processing the track to increase the gain).

The answers so far are basically correct. But the quoted statement itself indicates a problem.

I understand that you're probably just setting up a hypothetical here to make a point, but I can't think of any legitimate reason why anyone would want or need to add "10dB of processing gain" to any track. Even if you're purposely doing a whole lot of mangling in order to come up with this way-cool unique sound, you should be keeping your gain in check either via the output gain controls on your plugs, or by simply bringing the gain down on that track altogether.

This is something that a lot of folks don't realize. If you wind up boosting the gain on a track because of signal processing, you don't need and usually shouldn't just live with that gain, but rather often should likely adjust the gain in order to keep that in check and bring the track back to normal levels.

G.

 

[dastrick]

I was just throwing those numbers out there. I was just trying to understand the theory. I could have said -18dbFS then added 3 db of processing gain. Would that be the same as tracking at -15dbFS?

 

[SouthSIDE Glen]

I was just throwing those numbers out there. I was just trying to understand the theory. I could have said -18dbFS then added 3 db of processing gain. Would that be the same as tracking at -15dbFS?

Gotcha.

The answer, as has already basically been given is no, they are not the same. The root reason behind the "don't track too hot" idea only partly has to do with what the final track level is. What's just as, if not more important is keeping the levels in check all the way down the recording chain from mic to preamp to converter, and anything you might have in-between, in order to maximize the signal quality before it even gets to your hard drive (or other recording device.)

G.

 

[Tim Gillett]

I just read Massive's article on tracking too hot...

I'm going under the "no stupid question" assumption, here. Let's say that I've tracked to -12dbFS, then in my DAW I increase the gain by 10db (not moving the fader, but processing the track to increase the gain).

Is that the same as tracking at -2dbFS?

Nobody so far seems to have mentioned what the real difference will be. Recording 10db lower will increase the noise from your A/D converter by 10db.
That 10db of extra noise may or may not be significant in the context of the stuff you're recording but that's what will happen, whether you can hear it or not in a particular instance.

Whether the track will overload your master output is very much a secondary issue. I personally wouldnt be reducing tracking level by 10db just to avoid possibly overloading the master output later on.

The important thing is to get a track which is neither noisy nor clips. If your track file doesnt clip, it's not been recorded too hot. If it does clip it was recorded too hot.
You dont get any audio brownie points for boasting, "hey I avoided clipping by more db's than you avoided clipping."

The way to get the best headroom is to look equally at the other end, the "footroom", ie: the noise floor of your A/D converter. The idea is to let the baseline noise in your mic/pre combination just mask out the converter noise. There is no point in tracking with more gain than that, even if your track never peaks above -30dbFS. Tracking louder just robs you of potential headroom. Of course that assumes your mic/pre combination has that headroom itself.

Unfortunately discussions on this subject nearly always omit the A/D noise floor issue and concentrate on setting figures for tracking relative to 0dbFS. That's only half the story.

Read Stefan Kudelski on this in his old Nagra recorder manuals. He made the point that increasing preamp gain beyond say "8" on a particular recorder's gain dial would do nothing but rob you of headroom.

These days we're so spoiled for headroom that we can often get away with not even worrying about the bottom end (the noise end) of the equation.

I come from an older time when headroom really was in limited supply.

 

[Massive Master]

Nobody so far seems to have mentioned what the real difference will be. Recording 10db lower will increase the noise from your A/D converter by 10db.

Even though you sort of touched on it, the big question would be "What noise?"

The noise (and usually fluctuating SNR) from tracking too hot is much more significant than the non-existent noise at the bottom.

And to simply clarify to the thread - I'm not talking about "low" levels - I'm talking about normal levels. This is how the gear is designed. This is how the system works.

 

[Zeppe]

(not moving the fader, but processing the track to increase the gain).

I'm for tracking hotter to avoid processing the track to increase the gain. Even if it is a digital processing. You should let the preamp do its job. Preamp to tape or hard disk. There shouldn't be much else you need other then maybe a compressor or filter. If you mix properly you won't overload your master.

Tracking to low will always produce more noise. Then if you bring up the overall level of the track, that noise that should be on the 'floor' is hovering just below the recorded signal.

 

[Ethan Winer]

I'm not talking about "low" levels - I'm talking about normal levels. This is how the gear is designed. This is how the system works.

What DAW software are we talking about? Other than ProTools, most modern DAWs do all internal processing at 32 bits. That gives about 800 dB dynamic range (not a typo), and the distortion added by each math operation is so infinitesimal as to be irrelevant. Unlike analog tape where the distortion creeps up gradually as you approach and then exceed 0 VU, digital audio is totally clean right up to the 0 dB limit. So I don't see why it matters if you record at -30 or -0.1 dB. Further, once the audio is inside the DAW at 32 bits, you can go way over Digital Zero and still not have a problem. I proved this point using Cakewalk SONAR in my AES Audio Myths workshop video, which is now online at YouTube:

AES Audio Myths Workshop

One of the demos starts with a music track normalized near 0, then boosts that by 18 dB and sends it to an equalizer plug-in, then reduces the volume by 18 dB to compensate. Even after all that, the processed track nulls completely with an unprocessed version of the same original music track. This demo starts around 53:40 into the video.

--Ethan

 

[Chibi Nappa]

Why has the board exploded with noise floor discussion this week? Noise isn't an issue anymore. This isn't 1 inch tape or 16 bit with 1980 converters. It would probably be hard to make the noise floor audible on purpose.

How 'bout we switch the topic to "should I install a lantern-based warning system on my studio roof to warn me of Brittish soldiers" or something else simmilarly irrelevant.

Chill out about the noise. Track low to keep your headroom.

 

[DrewPeterson7]

I'm for tracking hotter to avoid processing the track to increase the gain. Even if it is a digital processing. You should let the preamp do its job. Preamp to tape or hard disk. There shouldn't be much else you need other then maybe a compressor or filter. If you mix properly you won't overload your master.

Tracking to low will always produce more noise. Then if you bring up the overall level of the track, that noise that should be on the 'floor' is hovering just below the recorded signal.

Have you read the article in question, though? The whole thing was written as an argument against that viewpoint, pointing instead to the huge amounts of usable headroom allowed by 24-bit recording, as well as nonlinearities that begin to appear in the response of audio signals hitting AD/DA converters at points below but approaching clipping.

If your noise floor is such that you can't track at -12-18dbfs and get clean results, then I think you have bigger issues in play.

 

[Zeppe]

Noise isn't an issue anymore. Track low to keep your headroom.

Noise is not an issue? I did'nt get the memo.

Drew-

Please post a link to the article in question. My beef is with telling people that are new to the game to track low. So low, that they feel they need processing to get a proper level. I agree with the two facts that you stated in summarization of the article, but I think the article in question is being misunderstood. I've seen a few post now where people are getting the wrong idea out of it.

 

[miroslav]

Yeah...I have to agree with Ethan...and maybe this discussion is focused too much on extremes…?
I mean, if you’re digital meters are hitting -1 to 0 dBFS…then OK, you’re walking a thin line and could easily cross 0dBFS on a random peak without even noticing it as you track…
…not to mention that whatever analog front-end you are using is probably struggling to give you that much level without creating its own distortion.
But most decent analog gear should allow you a bit more than *-18dBFS* without shitting the bed…and I certainly don’t see that -18dBFS is some line you should avoid crossing when going in (though no one really said that specifically) or as the *correct* average level to use.

Since I tend to track to tape and then dump to DAW, I rarely even bother looking at the digital meters on my converters, since I’ve already got my levels set at the analog/tape stage. BUT…since I do tend to record on the hot side with my analog gear to improve my S/N ratio…when I do dump to the DAW, my levels are very often hitting around -10/-8 dBFS.

And when I do track direct-to-DAW…I usually go for those same levels or even a couple dB hotter. My digital meters have their level ranges in the usual green/yellow/red colors. I just ballpark it so I’m averaging in the yellow and only hitting red on the peaks. I believe the red zone starts at -6 dBFS on my digital meters.

 

[SouthSIDE Glen]

I think we have some people talking apples and others talking oranges here, while a couple are talking about the whole mixed orchard.

I think John was referring to the gear on the analog side of the chain when he was talking about how the gear was designed and all that. And for those who talk about noise not being an issue and floating point bit depths and all that, they are thinking about and talking about only the digital side of the chain.

The whole answer has to involve both sides of the chain, no matter how short or long they may be, and this means talking about purposeful gain structure and gain staging from the microphone to the hard drive, inclusive. It's just not all that meaningful to talk about digital levels without also considering the analog levels at the same time, it's *one* chain and *one* system, with the master link being in the converters and their design performance and calibration.

It's easy for a lot of folks to ignore the analog side because so many home wreckers have it so (visually) short these days, with just a microphone and an integrated interface. But that does not lessen the importance or the need to look at the whole path when talking about levels.

G.

 

[DrewPeterson7]

Drew-

Please post a link to the article in question. My beef is with telling people that are new to the game to track low. So low, that they feel they need processing to get a proper level. I agree with the two facts that you stated in summarization of the article, but I think the article in question is being misunderstood. I've seen a few post now where people are getting the wrong idea out of it.

Sure thing.

http://www.massivemastering.com/blog/index_files/Proper_Audio_Recording_Levels.php

I think the point isn't that "noise isn't an issue anymore" but that knowing your signal chain is very important, you don't "gain" anything by trying to "use up all the bits" and peak just under -0db, and in fact you more often than not lose a fair amount since in doing so you're operating quite a bit above the ideal operating range of some of your (especially entry level) gear.

I've been tracking in the 12-20dbfs peak range ever since I read that, and anecdotally I think my mixes sound better for it (it's tough to say for sure because in that time I've also gotten a kickass new amp and finally managed to get a bass tone I'm happy with - of course, that second item may be related to MM's point somehow, but I've at least changed how I track bass and am happier with the results), and since I've been demoing for an instrumental rock project with drum loops, I've basically changed the source sound for every single thing I'm tracking myself. ). However, more importantly, I found I'm rarely having to cut levels while mixing, only occasionally having to boost something by more than a couple db, I haven't noticed a change in signal-to-noise, and I'm having mixdowns come in around the -6db range that mastering engineers seem to prefer anyway, so I don't really think I'm hurting myself. So, the risk-to-reward picture seems pretty good...

EDIT - in fact, may as well explicitly ask a question I hinted at earlier. Could reducing my tracking volume on a 5-string bass guitar with relatively hot 'buckers from around -6dbfs to the 12-20 range potentially be a partial reason for the fact that the low end in my mixes lately has seemed better defined and less "mushy?" or does that have to do more with changes in the way I've been recording (I experimented with a ton of different approaches, none of which really got me where I wanted, and finally hit on running through a Sansamp pre, recording an uneffected DI and a distorted effected DI, and then EQing them in a complimentary manner, high-passing the distorted DI and low passing the clean DI, adding very light compression to the distorted version and quite a bit more to the clean, and then sliding one a few samples at a time against the other until I liked what I heard, to give a sound with a lot of body to it but still plenty of punch and growl)?

 

[miroslav]

It's easy for a lot of folks to ignore the analog side because so many home wreckers have it so (visually) short these days, with just a microphone and an integrated interface.

Yeah...most don't realize that the built-in preamp on their converters is an analog front-end (not to mentoin their mic or DI).

 

[miroslav]

I found I'm rarely having to cut levels...

And this has been my case too, using the average input levels I mentioned earlier.

Mind you...I don't sum in the DAW...so quite frankly, I could probably have all my individual tracks hitting 0dBFS in the DAW after I'm done editing/comping them, and it really wouldn't matter that much...it would only cause me to lower the gain on my console channels... ...but my audio would still be clean/undistorted.

 

[Farview]

What DAW software are we talking about? Other than ProTools, most modern DAWs do all internal processing at 32 bits. That gives about 800 dB dynamic range (not a typo), and the distortion added by each math operation is so infinitesimal as to be irrelevant. Unlike analog tape where the distortion creeps up gradually as you approach and then exceed 0 VU, digital audio is totally clean right up to the 0 dB limit. So I don't see why it matters if you record at -30 or -0.1 dB. Further, once the audio is inside the DAW at 32 bits, you can go way over Digital Zero and still not have a problem. I proved this point using Cakewalk SONAR in my AES Audio Myths workshop video, which is now online at YouTube:--Ethan

But that misses the overall point about tracking at line level. The point is not to overcook the analog chain on the way into the converters.

Once a signal is digital, you can do just about anything you want with it level wise without much consequence. But in the analog world, running your preamps at +20db won't sound the same as running them at +4db.

In a perfect world, there would be no difference between tracking at -2 and tracking at -12 and normalizing to -2. In reality there is, most of that difference is caused by the analog side of the chain on the way into the converters.