Gain Staging

[ghostwhistler]

I'm a bit confused by how to do this. I'm using vst synths (no live instruments).

Do i set the fader to zero and then adjust the volume in the vst to get to, eg, -18db?

I'm confused because there are several different volume sources to work with: fader, master, vst audio.

Thanks

[imrae]

If in doubt, what I do is:

• Insert a VU meter/trim plugin after the audio source and before other plugins. (I use VUMT deluxe, which is also a stereo balancer, filter, ...)
• Trim the incoming audio / soft synth to float around 0 VU in its loud sections (with a -18 dB reference). You can do this with a gain control on the meter plugin (if it has one) or with the instrument plugin output (which might interact with an output limiter/clipper?)
• For each following plugin in the chain, try to use its output gain to keep the overall level the same. This allows for better bypass/comparison testing, and ensures your gain-staging is still standard. I mostly do this by ear, but if in doubt use another VU meter. (Strictly speaking LUFS would be better.)
• This should allow you to leave the fader at 0dB and get fairly sensible level running into master when you add a bunch more channels. Then you can use the fader for mix balance adjustments. Note that most faders are least sensitive around zero so ideally this is the region you make adjustments in.

[ramseysounds]

With your gain staging set, you want the fader to start at 0db.
Solo the track you want to gain stage, trim the volume down to what you desire (you use the example of -18db) and do the same for each individual track. You can use the channel trim or the output on the vst. Use the meter on the stereo out.
Just don't use the channel fader to gain stage. And remember to recheck levels as you pile multiple effects on each channel!
Best advice I was given was 'turn the level of your channels down and your speakers up.'
Good luck.

[Burillo]

there are a lot of myths about gain staging, but depending on what you're doing, it might not even matter. gain staging is important for things that emulate analog (such as preamps), but for purely digital processing (like a clean digital EQ), no gain staging is necessary. DAW itself also doesn't care about gain staging, you can happily boost channel volume by +300dB and then dial it back down on master with no loss of signal quality.

it's good practice to keep your levels in check, but generally speaking, unless you use a lot of analog-emulated processing, it doesn't matter how you do it.

EDIT: to add, there's no one correct way to gain stage. it's a purely technical process. as long as you understand the goal, you can figure out where in the chain it does matter, and where it doesn't. don't do it just because this other guy said you should - do it because it solves a specific problem (be it not oversaturating the mix, keeping the volume knobs around zero, or avoiding level spikes when bypassing plugins). the question of "how" will solve itself if you look at it that way.

want to keep levels when bypassing? gain-stage. want to not have to lower the DAW faders down into -50dB? gain-stage. analog emulation sounds dirty and clips? gain-stage. that's all there is to it. it is a tool, and it solves a problem. if you don't have a problem to solve - there's no need to do it.

[ghostwhistler]

I seem to have managed to do it I set the faders to zero and then used the vst volume to set to desired level. Thanks for the replies

[imrae]

there's no one correct way to gain stage. it's a purely technical process. as long as you understand the goal, you can figure out where in the chain it does matter, and where it doesn't. don't do it just because this other guy said you should - do it because it solves a specific problem

This is good advice for pretty much everything in the Production Techniques forum!

[ghostwhistler]

When it comes to gain staging and the start of the mix process, should you do this part before you render to stems (assuming that is your preference, to mix from stems/audio files). Given my system, I'm thinking next time I might bounce the midi down into audio and mix from that

[rotku60]

I struggled with gain staging as well. After seeking out some direction, I found the method makes the mixing process much simpler for me. I also use all VST instruments in Cubase. While I am writing the song and putting the parts together, I just adjust the mixer faders to suit the sound I am after. Once it is complete, then I review the parts to see what effects are not necessary (delays, reverb, chorus) and can be adding during the WAV file mix. It gives you the latitude to change things during the mix process. Then I set all the faders to 0 and adjust the gain trim to -18 (plus or minus a little bit). I render each element to a WAV file. I start a new project and import the WAV files and start the mixing process. I believe the reason for gain staging to -18 is so that when multiple elements are playing together the final output stays well below -0db (preferably between -6 - -10db) to leave room for processes used in Mastering. I am sure there are a lot of ways to go about this so don't take this as the gospel. It nice to have a defined method to use for each time you want to mix a project. Cheers.

[PieBerger]

I struggled with gain staging as well. After seeking out some direction, I found the method makes the mixing process much simpler for me. I also use all VST instruments in Cubase. While I am writing the song and putting the parts together, I just adjust the mixer faders to suit the sound I am after. Once it is complete, then I review the parts to see what effects are not necessary (delays, reverb, chorus) and can be adding during the WAV file mix. It gives you the latitude to change things during the mix process. Then I set all the faders to 0 and adjust the gain trim to -18 (plus or minus a little bit). I render each element to a WAV file. I start a new project and import the WAV files and start the mixing process. I believe the reason for gain staging to -18 is so that when multiple elements are playing together the final output stays well below -0db (preferably between -6 - -10db) to leave room for processes used in Mastering. I am sure there are a lot of ways to go about this so don't take this as the gospel. It nice to have a defined method to use for each time you want to mix a project. Cheers.

You should be gainstaging to -18dbfs (= 0VU) not -18db peak. Rather than arbitrarily turning everything down by 18db, you should use a VU meter, there are many available for free/cheap, I like Hornet VU Meter because it has a handy autogain feature where it listens to the program material and adjusts the gain up or down to hit 0VU or a peak value that you can set (whichever ceiling is hit first).

[Burillo]

When it comes to gain staging and the start of the mix process, should you do this part before you render to stems (assuming that is your preference, to mix from stems/audio files). Given my system, I'm thinking next time I might bounce the midi down into audio and mix from that

that depends on what format you're bouncing into. if you're using floating point WAV, then clipping is okay and won't actually clip. if you're using integer format, then any clipping will degrade the stem.

I believe the reason for gain staging to -18 is so that when multiple elements are playing together the final output stays well below -0db (preferably between -6 - -10db) to leave room for processes used in Mastering.

this myth has to die already. what mastering cares about is 1) no clipping (using floating point mixdown with clipping will not matter - mastering engineer can just turn it down without any loss), and 2) dynamic range. if your mixdown is peaking at -6dB but is squashed to shit, you haven't "left room for mastering", you've destroyed your mix. dynamic range is what matters, not peaking levels, because mastering engineer can just make things louder/quieter to adjust the peaking levels to their liking. digital gain is a lossless process.

You should be gainstaging to -18dbfs (= 0VU) not -18db peak. Rather than arbitrarily turning everything down by 18db, you should use a VU meter, there are many available for free/cheap, I like Hornet VU Meter because it has a handy autogain feature where it listens to the program material and adjusts the gain up or down to hit 0VU or a peak value that you can set (whichever ceiling is hit first).

yep, gain staging to -18dBFS isn't wrong, just unnecessary and way too quiet. as long as you're roughly hitting 0VU everything should be fine, so this is sound advice.

[PieBerger]

You should be gainstaging to -18dbfs (= 0VU) not -18db peak. Rather than arbitrarily turning everything down by 18db, you should use a VU meter, there are many available for free/cheap, I like Hornet VU Meter because it has a handy autogain feature where it listens to the program material and adjusts the gain up or down to hit 0VU or a peak value that you can set (whichever ceiling is hit first).

yep, gain staging to -18dBFS isn't wrong, just unnecessary and way too quiet. as long as you're roughly hitting 0VU everything should be fine, so this is sound advice.

Yes, I should have been more clear, thanks for tidying things up for me

[ghostwhistler]

When it comes to gain staging and the start of the mix process, should you do this part before you render to stems (assuming that is your preference, to mix from stems/audio files). Given my system, I'm thinking next time I might bounce the midi down into audio and mix from that

that depends on what format you're bouncing into. if you're using floating point WAV, then clipping is okay and won't actually clip. if you're using integer format, then any clipping will degrade the stem.

I don't know the format, wav I assume. I'm using Reaper so it's a question of jsut rendering a midi file from the likes of DIva and then mixing from there. Easier on the cpu

[rotku60]

Burillo and PieBurger, thanks for clarifying some of things I posted. Especially, setting each channel to a -dbfs value, whatever that value might be and "digital gain is a lossless process.". I still like to separate the creation process from the mixing process and gained staged WAV files give me a consistent starting point. There are probably better methods but this is the way I could wrap my head around. Also, mixing the WAV files gives me a bit more CPU to utilize effects freely. Some vsti's get pretty hungry if you are mixing while still in the midi state.

[jonalyn]

if i understand correctly , we need to check after every plugin on a track if the signal is not louder than 0 dbvu ?
And what about group , does it need to be adujsted to 0dbvu too ?

[BertKoor]

if i understand correctly [...]

No you did not understand correctly.

Adjust the gain only at points where it matters.