Pros of gainstaging (and, especially, at -18dBFS or 0dBVU):

• leaving enough room for level-dependant processes;
• avoiding aliasing that even crazy high oversampling wouldn't fix;
• covering all the bases when it comes to diverse developers' so-called sweet spots (this is kind of snake oil, really);
• making it easier to change the order of effects in the chain (e.g. EQ first or compressor first, linearity issues aside);
• working with faders in the upper, high resolution area (more ergonomical);
• ensuring solid A/B testing for subtle effects;
• knowing where the issue is (or isn't) when problem-solving;

However...

Bear in mind that your signal will NOT sound noticeably bad if you don't gainstage, as this is VERY rare in the digital realm. Mind ACTUAL problems: say you feed a -40dBFS signal into a compressor... when it responds, your threshold will be up the roof, analog emulation or not. Another cenario: the level of your Rhodes is very quiet, so you try to compensate by raising every note velocity to 127 and with that you lose all the expressiveness of the instrument, as velocity isn't merely "level control". When you feed a quiet signal into a saturator, it won't yield the same results as a loud signal. All this could be solved by dynamic control, which really is the broad area of which the process of gainstaging is but a part, along with balancing, volume automation, compression etc.

 

Well, it depends on what the input/output/drive control does, right? In most cases, it is either entirely linear or pushes the level into a transfer curve. Either with an inverse function or without.
Sometimes, however, a little more happens there. Dynamic waveshaping, pre-/de-emphasis, interventions in the stereo correlation by inducing different harmonic patterns in the channels depending on the control setting, etc. In this case, the specification of a "sweet spot" in the manual is usually the (VU) ref level that leads to the most linear result with the Init setting. Most of these plug-ins also have an additional, clean input gain somewhere, but some do not. The gain must therefore be set correctly before the plugin. However, I haven't seen a plugin that works like this for a long time without providing a clean gain somewhere, but I don't test every new plugin either, so... I agree with you on the most. But I have a question about your workflow:

Isn't that a lot of up-and-down levelling? Not that that's a problem in the digital environment.
It just sounds like a lot of (unnecessary) work to me. Or have you created "low input gain" presets for your analog emus?

 

ssl plugins are -14, just sayng

 

Yes, absolutely, I'm aware of these things. Additionally there are plugins with no specified calibration and/or any meter. Happens with AA and Kush for instance. According to their behaviour some calibrations are not specified correctly for Waves and PA's. Furthermore I've one plugin that I'm using pretty often that comes with a wannabe VU meter which shows 0 for a digital peak of 0dBFS and I'm 'underdriving' it heavily.
In all these problematic cases and the ones you mentioned, I found only one solution, use my ears. I know, not very popular, but...

Well, all purely digital plugins are out of the equation, input-wise. If you're gainstaging, you need to adjust the output of every plugin anyways, right? So by setting all the audio files to 0VU (BTW, how are you doing this? For -1TP I can select them all, hit a keycommand, hit enter and good, but this doesn't work for VU) you probably hit the right input level more often than I do (if the plugin works with the same calibration), I'll give you that. But I'm only using analogue emu compressors, rarely EQs and these not even on every track, so the additional work isn't that much and I prefer the advantages that come with this method. And with my DAW adjusting the level before and after the plugin is extremely easy with this FX chain https://imgur.com/TaC92fr. I just place the plugin in question between these. But honestly, I don't have to use it very often.

 

...key?!

Surprise! I don't! At least not to the same reference level.
Whether I normalize tracks or not depends on the source material. If source files are not in level balance, I normalize them with LUFS so that the VU on my master shows 0VU (the VU meter on my master has reference level -18dBFS because my converter outputs +4dBu at -18dBFS). Then I insert VU Meters on tracks and busses. I have presets with different reference levels for my VU Meter. I call up these presets using a script that selects the preset based on the track name. This way I usually have 0VU (+-3dB) on all tracks with one keystroke. +-3db tolerance based on genre or song type. If an instrument is more than +-3dB off, then something is wrong. Be it dynamics, saturation, frequency response or level in general. So I can see immediately when an instrument needs "something". It doesn't replace the ears, but it's helpful. My hardware always gets the level it should have and the control paths in plugins are usually very short.

Interesting!

 

Last edited: Nov 23, 2023

Er, haven't you asked if my approach is a lot of work?
Compare your (endless ) list to mine:
[ctrl + A], [ctrl + N], [Enter], done, all levels set.

Seriously, blessed be our times, because now there're as many ways to Rome as people are travelling there.

 

Ehm... I don't want to make it a dick comparison, but...
Joking.
I see what you're doing. Your approach definitely has its merits. I've just never seen anyone work this way.

I use 2 more keycommands than you. One to insert the vu meters and one to execute the script. I level the tracks that are out of line to about 0VU. Then I have my rough mix. I'm currently working on an auto-gain for the vu meter, which does this step for me with 1 mouse click. Then I only need to make some small genre or song dependent level adjustments.

 

99.9% don't have a "sweet spot". If they do it's mostly comps or saturators with a fixed threshold. If they have a fixed threshold, they will (or should) most likely have input/output gain knobs instead, which you are supposed to use (think 1176).

If a modern plugin does not have input/output gain parameters, and the dev says that you should gain-stage to outside of the plugin to hit it's "sweet spot", then I would suggest to ditch that plugin (and all plugins from that dev).

You don't gain-stage to get a better sound out of plugins, you gain-stage to make mixing easier for you (and to some degree to not get fooled by plugins that boosts the loudness of your tracks). If you mix analog, hitting the sweet spot for your analog gear is a part of gain-staging, but not if you mix ITB.

 

gain staging as advanced concept in modern mixing is determined by micro- (=short term) and macro-dynamics (within whole song) across tracks/busses,
as an example, I highly recommend studying how TDR Kotelnikov works, and how its advanced trickery works

 

That's probably because I'm self-taught and developed this method on my own.

So you wrote a script to insert a VU meter as the first plugin on all tracks at once??? That sounds cool. I've never seen anyone work this way.

Keep at it.

 

Yeah well, not quite. I wrote a script that changes the reference setting of each inserted vu meter instance based on tags in the name of the track on which the vu meter is inserted. Got it?

 

Last edited: Nov 23, 2023

Hmm... I learnt these terms different:

• Microdynamics usually refers to the dynamic range within the temporal course of a single sound, i.e. the envelope.
• Macrodynamics describes the dynamic range of at least two individual sound events within a musical performance in relation to each other.
• Musical dynamics describes the overall dynamic range of a song from the beginning to the end.

This corresponds to:

• Peak / Momentary
• Momentary / Short Term
• Intergrated

I make no claim to The Truth ©. It's just how I learnt it.

 

Last edited: Nov 23, 2023

Sure, that wasn't that difficult. Interesting approach.