A little something about gain-staging, for those who don't know

I'd need someone clever like JBroll to pop in and give his thoughts, but the more I think about it... the more I think a 64-bit floating point mix engine makes this all null and void.

I'm using Reaper for this test. But...

Put a sine-tone on a track. Say it equates to 0db, give it +24db increase on the tracks gain control.
Now on the master, give it -24db decrease on the gain control.
Render this out as a mono wav.
Add it to a new track. Reset the master volume and reset the first tracks volume. Invert the phase of the 'rendered' sine tone.

Now what I'm getting here is a peak when the file starts, then complete cancellation. Which seems to indicate that even though the track was at +24db and was clipping awfully... the master track being reduced prevented any clipping in the final output file. NOTE: You don't hear any clipping whilst it plays back either.

What does this mean in regards to the issue at hand? I'm not so sure. But it's interesting and might be relevant at some point.
 
Hey drew

What you are saying is indeed correct, but I think it's just bad practice, especialy when you want to work with analogue equipment or do live shows on non-digital consoles (maybe on digital mixers too, i don't know how the summing engine tolerates levels on the new generation of digital consoles)
 
as long as you dont clip your ad converter everything is good.
the reason behind lowering the volume is to have enough headroom for sound changes.
I for myself lower every channel post fader to -10dbfs so my master bus never clips...

if you try to get every channel to -3dbfs for example your master bus will clip and thats a bad thing because you wont want the plugins on it cliped also you will export the session in a much lower quality were clipping IS a problem
 
We'll get a compromise soon.

Slate's VCC will have a VU meter you can calibrate yourself to whatever level you'd like, and have it correspond to the overload characteristics of whatever console you've got it emulating. Not the same as having a VU meter at the top of every channel in mixer view, but as close as we'll get for now.
 
tempe said:
Another question, albeit slightly off topic. I've read that when recording DI tracks that are to be reamped you want your peaks at around -3dBFS, is that a load of shit or just something that's specific to preparing for reamping?
Click to expand...


Can someone clarify that?



Say you track your DI's at -18dBFS and then send it out to the analog domain to a clean preamp for unity gain/boost levels and then on to the reamp box?


Basically, Would it be good practice to compensate the volume difference in the analog domain just before the reamp box?
 
drew_drummer said:
I'd need someone clever like JBroll to pop in and give his thoughts, but the more I think about it... the more I think a 64-bit floating point mix engine makes this all null and void.
Click to expand...

Drew, it would actually make it the same thing. It's not about the DAW it's about the preamps and converters. They are optimized (in theory) to run at 0dbVU, obviously some sound "better" when driver harder. And a 64bit mix engine would make it even less important to drive it into the DAW hot, becuase you would have more headroom.


0dbVU = somewhere around -20 to -14dbFS (you have to look up your interface/converter spec sheet to find out which exactly).

Remember guys it's about average, not about peaking.

It makes even less sense to track hotter and then pull down all the faders for headroom to mix. I find when even using these cautious tracking levels i'm still pulling down faders for mix headroom once you start stacking up 20-30 tracks.
 
drew_drummer said:
I'd need someone clever like JBroll to pop in and give his thoughts, but the more I think about it... the more I think a 64-bit floating point mix engine makes this all null and void.

I'm using Reaper for this test. But...

Put a sine-tone on a track. Say it equates to 0db, give it +24db increase on the tracks gain control.
Now on the master, give it -24db decrease on the gain control.
Render this out as a mono wav.
Add it to a new track. Reset the master volume and reset the first tracks volume. Invert the phase of the 'rendered' sine tone.

Now what I'm getting here is a peak when the file starts, then complete cancellation. Which seems to indicate that even though the track was at +24db and was clipping awfully... the master track being reduced prevented any clipping in the final output file. NOTE: You don't hear any clipping whilst it plays back either.

What does this mean in regards to the issue at hand? I'm not so sure. But it's interesting and might be relevant at some point.
Click to expand...

blah blah blah

i don't really understand why people still debate this shit...to me, it's a pretty damned simple concept, and one that was taught to me before i ever plugged in a mic or touched a knob - and that was to keep the peaks of the material being recorded around 0dbvu on your analog meters when tracking. regardless of what that equates to in the digital world with any specific DAW or converter, you're always going to have better fidelity by not pushing your front end too hard.

sure, sometimes it's fun to push your preamps into the red when tracking to get a little grit into the source...but 95% of the time, just aim for 0 on the preamp, and the rest sorts itself out later!
 
colonel kurtz said:
blah blah blah

i don't really understand why people still debate this shit...to me, it's a pretty damned simple concept, and one that was taught to me before i ever plugged in a mic or touched a knob - and that was to keep the peaks of the material being recorded around 0dbvu on your analog meters when tracking. regardless of what that equates to in the digital world with any specific DAW or converter, you're always going to have better fidelity by not pushing your front end too hard.

sure, sometimes it's fun to push your preamps into the red when tracking to get a little grit into the source...but 95% of the time, just aim for 0 on the preamp, and the rest sorts itself out later!
Click to expand...

Why are you being a prick? If you're gonna get involved in the discussion, play nicely.

Let me be clear. My peaks never approach clipping the preamp. I've gotten into the habit of making sure that I peak around -8db (I made a mistake during an earlier post, which I've just corrected)

I've tried making sure that I peaked lower in the past. It never seemed to make any difference to the mix. Except for the signal to noise ratio. Running the preamp at a lower gain level, and boost internally made the noise more apparent.

After I record all the material and before I bother to add any effects. I turn all my channel faders down to -11db. This brings my -8db peaks down to -19db. Then when I start adding effects... the peaks will slowly creep up to about -17db.

What I find in this situation is after the mix is done... each track has increased by a minor amount of decibels and the mix is pretty much an even spread - all I need to do at this point is make minor adjustments. And I end up with a decent amount of headroom on the master channel.

Now... that may be the *wrong* way to work in you guys' eyes. But it's the result of years of trial and error. I'm happy to learn and improve my mixes further.

But it's not as simple as you make out.

My assertion is.... as long as you don't clip your preamps, once it is in the digital realm... it is just numbers, and can be made to work. I do not agree that you should hit each plugin with -18db/0dbVU in order to get the best gain structure.

At first I was agreeing. But I was tired, and was being a spastic. But the bit-depth of modern day DAW's is so large that you really don't need to worry about hitting each plugin with the same level. Just worry that your final output doesn't clip.
 
http://www.massivemastering.com/blog/index_files/Proper_Audio_Recording_Levels.php

This is a good article on the subject. He sums it up as:

THE "DUMBED DOWN" VERSION: Stop recording so hot. Instead of trying to get your tracks to peak at -2dBFS, have them peak between -20 and -12dBFS and your recordings will almost undoubtedly sound better. Mixing will be easier. EQ will be more effective. Compression will be smoother, more manageable and predictable. You're in the age of 24-bit digital recording - Relax and enjoy the headroom. Even if your only concern is the volume of the finished product (which would be a shame, but it happens), recordings made with a good amount of headrom are almost undoubtedly better suited to handle the "abuse" of excessive dynamics control. QUIETER recordings have more potential to be LOUD later. It's because they're usually better sounding recordings in the first place.
Click to expand...

I think essentially what I've been doing is peaking at -8db, and using the mixers gain fader to add in the extra mix headroom. Whether this is right or wrong I guess is a matter of preference. But I learnt the lesson of not peaking at -1db .. maybe I just need to be a lot more hardcore with my peaks. It's possible and I'm completely open to the notion.
 
AudioGeekZine said:
A relevant thread on the DUC got bumped tonight.
Well worth a read.
http://duc.digidesign.com/showthread.php?t=230766
Click to expand...

That was indeed very interesting... This post sums it up very good i think:

I hate to break it to you, but you have yet to list anything that would constitute "Digital" noise. I think your ignorance is showing here. Everything from RF interference to a bad power supply are types of ANALOG noise. AND!!! turning your mic preamps up and recording with hotter signals will only AMPLIFY those noises, not mask them.

The example you site, quoted above, will get louder as you record louder. The cell phone interference is analog, and as such happens either before you get to your DAW audio interface or after your interface on your way to the speakers. Turning up the preamp to record hotter also increases these artifacts. Recording with LOWER levels will actually help to get rid of them since you are pushing them down closer to the noise floor of the system.

Anything at the same level or lower than the noise floor of the system gets "swallowed up" by the noise in the system. This is true for analog tape and for digital recorders. The difference is, the noise in digital recorders is 30dB to 40dB lower than that of any analog counterpart. So, recording your levels lower puts your analog noise down into that realm (and if it isn't swallowed up by the system noise it might very well get swallowed up by dither when you master to 16bits). Recording lower also gives you more headroom for digital processing when mixing.

Let me ask you this... if you record as hot as possible, what happens when you EQ something? If you are recording a signal up to -1dB, and then you boost any frequency band using EQ by more than +1dB you are clipping the signal inside the EQ.

Your "sterile digital" recordings are "sterile" because you are recording too hot. No tube preamp is going to help that, no matter how much marketing hype you read. You are getting all this clipping and harmonic distortion within your plugins and throughout your DAW because your level to disk is too hot. That is what creates a "digital" sounding mix.

If you are recording into your DAW trying to get your peaks as close to 0dBFS as possible. That is the EXACT same thing as recording to an analog tape and trying to get your levels to tape up to +20dBVU.

In the end it all comes down to voltage. Equipment designed to operate optimally at +4dBu do just that; they sound "best" when the signal is +4dBu or as close to it as possible. +4dBu on a Protols HD system (192 IO) is at -18dBFS (or -20dBFS if you use the B trim). If you don't believe me, but a volt meter. +4dBu is equal to 1.228 Volts. Test all of your "analog tube" equipment. You'll see that when a signal is coming out at 1.228 Volts, it is showing 0dB on the VU meter of the tube gear. When you send a 1.228 Volt signal into protools HD (192 IO) it shows up at -18dBFS.

When you are recording as hot as possible you are trying to push upwards of 9 volts (with peaks way above that) out of a mic preamp that is designed to operate down around 1.228 Volts. And you are pushing that 9 volt signal into a line input and A/D that is designed to be linear at or near 1.228 Volts. EVERYTHING about the mic preamps, line amps and converter go "out of spec" when driving the signal that hard. The frequency response is no longer linear and the harmonic distortion increases dramatically, yielding a "sterile" and unpleasant sound. It's compounding. The preamp adds harmonic distortions trying to output such a hot signal...the line input adds harmonic distortion trying to receive that hot a signal, your plugins will potentially add distortion (boosting an EQ freq for example) and then your D/A will add distortion to it because of Intersample peaks... and your line output will add yet more harmonic distortion trying to output a signal that hot and most likely the input of your power amp or speaker system will also add harmonic distortion since the signal is so hot as well.
Click to expand...
 
drew_drummer said:
Why are you being a prick? If you're gonna get involved in the discussion, play nicely.

Let me be clear. My peaks never approach clipping the preamp. I've gotten into the habit of making sure that I peak around -8db (I made a mistake during an earlier post, which I've just corrected)

I've tried making sure that I peaked lower in the past. It never seemed to make any difference to the mix. Except for the signal to noise ratio. Running the preamp at a lower gain level, and boost internally made the noise more apparent.

After I record all the material and before I bother to add any effects. I turn all my channel faders down to -11db. This brings my -8db peaks down to -19db. Then when I start adding effects... the peaks will slowly creep up to about -17db.

What I find in this situation is after the mix is done... each track has increased by a minor amount of decibels and the mix is pretty much an even spread - all I need to do at this point is make minor adjustments. And I end up with a decent amount of headroom on the master channel.

Now... that may be the *wrong* way to work in you guys' eyes. But it's the result of years of trial and error. I'm happy to learn and improve my mixes further.

But it's not as simple as you make out.

My assertion is.... as long as you don't clip your preamps, once it is in the digital realm... it is just numbers, and can be made to work. I do not agree that you should hit each plugin with -18db/0dbVU in order to get the best gain structure.

At first I was agreeing. But I was tired, and was being a spastic. But the bit-depth of modern day DAW's is so large that you really don't need to worry about hitting each plugin with the same level. Just worry that your final output doesn't clip.
Click to expand...

didn't mean to be a dick

it's just that this is a topic i've seen beaten to death multiple times, all over the internet, and i dunno why so many people don't understand that they need to just pay no mind to digital meters when tracking...keep the analog end outta the red, and then on mixdown, keep the tracks in your DAW from clipping. such a simple concept(at least to me), but with so many countless hours spent debating it...
 
I think I agree..


The amount of absolute bollox thats discussed on forums about crap that makes fuck all difference is absurd.

Ok yes keep out the red on the output and input and quite frankly your fine, I mean on poor convertors there might be a tiny insignificant change but it wont effect the end result.

Sound on Sound (UK mag I am sure you all read) did a review of convertors and found that the only thing would be a slight change in the sound they impart due to characteristics (ie slightly warmer Apogee convertors) but they all handle things in a similar way - we can hear a difference but this is not to do with the convertors. I wonder if pushing the input is just picking up more of these characteristics.
Ultimatly it does not effect you enough to care.

But I digress, we spend a lot of out time talking about this stuff and the Noobies out there begin to believe (as I used to) that this is the shit that makes a difference.

Give a pro a dictaphone and he could make a better recording than a noobie with 10 grands worth of kit and all the fucking headroom he will ever need.


My view with digital is stay out of danger (red)and push it a much as you fucking want. If you record everything in brickwalled at 0db thats your mess and probably you need better hearing!

These are confusing discussions at the best of times because it appears like there might be some truth to it but I am becoming more convinced that noone around here designs convertors!

I only bring stuff in now around -10db cause it makes life easier.
 
good read, and sorry to bump and old thread but quick question: would this mean that I should be adjusting the gain for "quieter" say vocal parts to get them to peak at -18, then when a louder vocal part needs to be recorded, readjusting the levels in order to get the peaks to stay at -18 by turning down my preamp?
 
Mjespo125 said:
good read, and sorry to bump and old thread but quick question: would this mean that I should be adjusting the gain for "quieter" say vocal parts to get them to peak at -18, then when a louder vocal part needs to be recorded, readjusting the levels in order to get the peaks to stay at -18 by turning down my preamp?
Click to expand...

The way I would look at it is this: you have a lot of headroom and with vocals they usually get slammed by a compressor which will reduce a lot of volume. So track to where the average volume peaks around -18 to -12 or so just as long as your loudest peak doesn't clip. As long as you don't clip everything is ok, and since the compressor will hit things harder, the end result if you did track a bit hot would be the fact that you wouldn't be using so much makeup gain on the compressor.
 
yeah. I'm gonna go ahead and ask if I'm missing anything in summery of this one.
I felt like I understood gain staging until this thread popped back up.
so basically:

step 1: track relatively hot at the preamp, but make sure meters don't clip at the preamp.
step 2: don't push faders too hot. leave lots of headroom. no peaking.
step 3: make sure each fader's gain does not exceed their buss's gain.
step 4: make sure each buss doesn't clip.
step 5: make sure each buss doesn't exceed the master buss
step 6: make sure master buss doesn't clip
step 7: leave a generous amount of headroom for the master buss

without getting into any specific numbers, is this correct?
 
You must log in or register to reply here.

Similar threads

AudioPhile777
A little something about Phase Relationships, for those who don't know
Replies
0
Views
601
F.O.H.
AudioPhile777
A
X
noob gain-staging question
Replies
23
Views
6K
F.O.H.
r3c0rd_3m_411
R
LeSedna
Gain staging / trim plug / printing to wav across board
Replies
29
Views
3K
F.O.H.
LeSedna
L
K
Gain Staging with Slate VCC...im confused.
Replies
5
Views
3K
F.O.H.
mstone564
M
Ice Man
Ugh, I can't get the loudness I want! How skewed is my gain staging?
Replies
11
Views
1K
F.O.H.
smy1
S
Top Bottom
Back