question about tracking too hot......

[Ethan Winer]

Shit, it took 212 post for you to actually come out and say that tracking at +22 is not a good idea?

WTF? Seriously, dood, WTF? See my second post to this thread, which was the first time I talked about levels and gain staging:

I would never run anything at +20 because it's not necessary and just wastes electricity and makes gear run hot aging the components faster. I think a good approach is to set your mixer so outputting around 0 to +5 or so hits digital zero at the converter inputs. But anything within +/- 10 dB of that is fine, especially when there are no transformers in the path.

So yet again Ethan is proven 100 percent correct. Though it took 212 posts for some of you guys to accept and admit that.

--Ethan

 

[Ethan Winer]

So the whole argument is a matter of perspective.

No, it's a matter of how much distortion one gets when running various levels through preamps and converters. Since this is easy to determine by measuring, I don't understand why it's become an "argument" as you call it.

Most of the interfaces out there don't allow you to calibrate them, so recording at just under digital zero IS running the mic preamps close to +22.

Say what? I just went to Mercenary Audio's site and picked the first high-quality converter I saw, the Apogee Rosetta 800:

http://lib.store.yahoo.net/lib/mercenary-audio/apor8man.pdf

Page 13 shows how to set the internal jumpers to accommodate a wide range of input and output levels including -10 dBV = -13 dBfs.

Mid-level outboard converters like the Presonus models have input and output level controls. PCI type sound cards expect consumer levels, so they won't demand +22 either to hit digital zero.

Just curious: How did you calibrate the delta card to +4dbu = -4dbfs?

The Delta 66 has a software control panel that lets you control all sorts of stuff, including input and output levels on a per-channel basis.

--Ethan

 

[miroslav]

Good. Under 18 would be illegal in most states.

That's why I'm leaving the extra +1dBu....for added headroom.

 

[masteringhouse]

In addition to getting audio into the computer one has to get it out.

It's still best not to keep peak levels too high during recording so that when you are processing the track it will not clip. For example if you are EQing a track it doesn't make sense to constantly have to trim the input to the EQ (plug-in or hardware) to avoid clipping. There would also be a minor degradation of sound quality in digital due to performing the gain calculation over not having to do it at all.

This is a part of gain staging that I don't believe has really been addressed here yet.

 

[miroslav]

Maybe...but not if you mix back OTB....

I think this thread really underscores how there is NO one way to look at this, and it always comes back to you listening to your signals and being happy with what they are doing...and then not worrying so much if a certain *number* is or isn't being hit on the mark.

 

[masteringhouse]

Maybe...but not if you mix back OTB....

I think this thread really underscores how there is NO one way to look at this, and it always comes back to you listening to your signals and being happy with what they are doing...and then not worrying so much if a certain *number* is or isn't being hit on the mark.

Well you would still need to trim the input of the hardware and be aware of how the levels of the D/A are affecting the sound. Given proper calibration you could get this to work as a default though.

The problem here is that many home recordists have no idea on how to properly calibrate their converters and the "defaults" on prosumer gear do not support tracking at hot levels in order to keep costs low. So as a "general rule of thumb" keeping levels lower makes more sense if you want clean audio than tracking hot.

But yeah, if I were mixing or recording a Rock track through a Neve console that rule of thumb would likely go out the window.

 

[Ethan Winer]

if you are EQing a track it doesn't make sense to constantly have to trim the input to the EQ (plug-in or hardware) to avoid clipping. There would also be a minor degradation of sound quality in digital due to performing the gain calculation over not having to do it at all. This is a part of gain staging that I don't believe has really been addressed here yet.

I addressed that in my very first post to this thread, and offered proof that this is not a problem with plug-ins in a modern DAW:

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.

As for "minor degradation" during volume changes, "insignificant" is the better word. A volume change manifests as added distortion, but with 32-bit floating point math there are more zeros after the decimal point than I have time to type.

--Ethan

 

[masteringhouse]

Ethan, it still depends. If you are mixing OTB or are inserting hardware anywhere in the path then you still are restricted to 24 bit fixed due to the D/A. Also Pro Tools TDM systems are restricted to 24 bit paths.

When would tracking near 0dBFS actually be a benefit?

Unless you can propose a compelling reason it makes more sense not to given that there are more potential problems in doing it.

 

[Ethan Winer]

If you are mixing OTB or are inserting hardware anywhere in the path then you still are restricted to 24 bit fixed due to the D/A.

Yes.

When would tracking near 0dBFS actually be a benefit?

Only when recording at 16 bits, but that's never been my point in this thread. All I have objected to is the claim that it's better to record at low levels. Especially in the context that recording closer to zero somehow makes it more difficult to get tracks to blend well in a mix.

--Ethan

 

[masteringhouse]

Only when recording at 16 bits, but that's never been my point in this thread. All I have objected to is the claim that it's better to record at low levels. Especially in the context that recording closer to zero somehow makes it more difficult to get tracks to blend well in a mix.

--Ethan

If I can draw an analogy.

Suppose that someone new to an area asks for directions. There are two ways that one can get there, one goes through a very bad part of town that can be dangerous for this person unless they know the territory well, and another way that is safer. Both paths are the same distance and take the same amount of time to get to the destination.

Which directions are best to give?

For the newbie it's "better" to track at conservative levels, for someone who knows gain staging, calibration, how their DAW works, etc. it matters less.

 

[Ethan Winer]

For the newbie it's "better" to track at conservative levels, for someone who knows gain staging, calibration, how their DAW works, etc. it matters less.

I don't mean to be a jerk Tom, but I still disagree. Most entry level people use entry level gear similar to the setup I describe in my article Using a Mixer with a DAW. They are not using high-end pres that put out old-school large signal levels. A lot of people use all-in-one converters with built-in mic pres, like the Presonus and M-Audio externals, where all of this is handled automatically. You plug in a microphone, then adjust the single gain knob for a decent level in the DAW. That's the equivalent of the simple setup I use that's shown in the above article.

Tell you what - I'll be glad to change my mind if you can describe a typical newbie type setup where just plugging the pieces together with no real knowledge will lead to distortion when recording close to full scale. Of course, I'll still argue that the newbies need knowledge and real explanations, rather than dumbed down cookie cutter advice. But I'm still curious to learn how a typical newbie setup has problems recording near full scale.

--Ethan

 

[masteringhouse]

Tell you what - I'll be glad to change my mind if you can describe a typical newbie type setup where just plugging the pieces together with no real knowledge will lead to distortion when recording close to full scale.

Nearly ANY setup from a newbie might cause this even a newbie using pro gear. I see this in the mixes of some of the less seasoned engineers that I work with and in student mixes at the college where I teach using high-end gear. Most often there is clipping in the kick or snare tracks but may go unnoticed since it's short duration clipping and less audible than say on a bass track. It can happen even when recording at low levels but the tendency for less experienced engineers is that the hotter they try to push the recording level the closer they get to clipping any part of the chain.

I'm not trying to change your mind Ethan I don't really disagree with many of your points in this thread. I'm simply saying that for the inexperienced engineer it's usually better to be conservative with levels rather than trying to play the "use as many bits as possible" game, at least until they know all of the rules of the game.

 

[Ethan Winer]

Okay, fair enough Tom.

This reminds me of a conversation I had with an EE friend years ago. He was complaining that newbie circuit designers slam way more current than is needed into transistors. Even if a circuit works with times times more current than needed, it's wasteful and accomplishes nothing useful.

--Ethan

 

[miroslav]

For the newbie it's "better" to track at conservative levels, for someone who knows gain staging, calibration, how their DAW works, etc. it matters less.

Yes...I agreed to this awhile back in this thread...though initially, there was no specific focus on newbies...rather is was being argued by some as a better SOP to follow on all occasions...no matter your skill or gear.

I mean...if you lay out the conditions that IF you are a newbie and/or IF you want nothing but clean/pristine signals at all times on all tracks and IF you want to feel 100% safe about your level choices...
...then yeah, aim for about -18dBFS as your average level.

 

[Farview]

I'm sure that there was an assumption on a lot of peoples part that if he had to ask the question, he might be fairly new...

 

[miroslav]

The OP's question was answered to HIS satisfaction early on...
...the rest of the thread had nothing to do with him.

 

[noisewreck]

Yes...I agreed to this awhile back in this thread...though initially, there was no specific focus on newbies...rather is was being argued by some as a better SOP to follow on all occasions...no matter your skill or gear.

Funny you take this approach now.

So, do you pan vocals in the middle as an SOP or allow the song dictate it's position as an SOP?

 

[miroslav]

Funny you take this approach now.

So, do you pan vocals in the middle as an SOP or allow the song dictate it's position as an SOP?

I never argued that...rather just that center-panned vocals are a good starting postion, seeing how you have to start with them somehere!
Then you work the mix toward some final settings.
So really...that "SOP" has you end up doing whatever you think sounds best.

Are the -18dBFS choices a "starting position"...and then you turn up or down to taste from there?

 

[MCI2424]

And thus we come back around to the problem that DAW meters do not give you any indication of what the voltages are. And the only way to know since it will vary from gear to gear is to test your own gear and find out. This is the step that most of us have missed.

Also you have to be aware if you are using transparent or color gear as this will have a big impact where you need to run too. As soon as you introduce any gear with designed in coloration, it doesn't matter how clean the pre is, if you put a piece of coloration gear after it you may/will find the more you push into it the more the signal is going to be altered maybe in a linear fashion maybe not.

If I run a signal thru clean pre->converter->DAW so long as I don't clip I can go as hot as I like +10 will just make it louder so long as +10 is within tolerance

If I run the same signal through a coloration pre->coloration EQ->coloration Compressor at +10 the result could be pleasing euphonic distortion, could be horrible distortion, could be clipping or could be fine depending on how each piece is calibrated and if my own experince of testing my own small gear selection is in any way indicative it would suggest there will a good amount of variation of how much headroom is available.

So you have to know your own gear and each pieces limitations that way you can gain stage effectively to get the result you want. Knowing what someone else is doing is academically of interest perhaps but irrelevant unless they are running precisely the same set up that you are

Having said all of that if you are unable, unwilling or not interested in finding out where you're gear runs if you run at RMS -12 to -18 dBFS you are pretty safe in assuming you won't clip anything. (although tracking RMS level on a DAW meter isn't easy either) and unless your gear has an inherently high noise floor shouldn't make a big impact. Using all the bits meh, Ethans video and the subsequent bit crusher testing of my own that it inspired, suggests to me that I won't notice any problem untill I get way below 16 bits anyway

Second best post ever. This is, again, the answer.

 

[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?

So the answer to my question is PURPLE TURTLES. Thanks guys. I would've never guessed.