onerob wrote: ↑Sat May 20, 2017 14:49 CEST
Yeah, I do like to have fun on the master
Looping over the loudest segment of the song, with the brass and organ, I get the following values with my master chain engaged. (The final gain plugin is actually boosting by 2.5dB).
Hm... values look fine to me (although remember that something wasn't quite right with TBProAudio's dpMeter 2 - I think it had something to do with the AES-17 compensation / the RMS mode). So it remains as warning to maybe look out for that in the future.
Small Ocean wrote: ↑Sat May 20, 2017 16:33 CEST
I'm assuming these requirements are because you want to prevent people from "cheating" by increasing the apparent loudness of a track by crushing the dynamics etc... May I suggest it would be better, rather than discouraging people from using the mix bus (which I believe is fairly common for mixing engineers these days, thanks to increased CPU power), to simply require certain maximum RMS, or minimum Dynamic Range (as measured by something like TT-DR-Meter), and disqualify mixes that go beyond these levels? You could of course adjust the requirements depending on the submission, to account for tracks which are going to be naturally louder/quieter and less/more dynamic.
Actually, that is already covered in the "Rule Book" (incl. recommendations for loudness and maximum dBFS values)
The idea behind all of that, is that you should focus on
properly mixing individual channels and use as less as possible modules/plugins on the summing bus, not "premaster" the track. There is of course the occasional (allowed! I've posted about this here:) Mix Console simulation, a "gel compressor" (see SSL compressor "mixed through" techniques), an additional tape machine, a character EQ that fixes one frequency "globally". I've talked about this
in this post
But not an additional reverb, M/S tools, 3-4 different EQ's that do drastic sound shaping, loudness adjustments, etc.
This is a mixing challenge for a reason, not a "mix and mastering" one. You're challenged to work different, think outside the box, move away from your usual (maybe wrong learned) workflow, get access to genres you wouldn't normally touch, learn something in the process.
And learning how to break out of that habit, isn't easy. Trust me, I've been there - and these days, I'm only having a VU (safety check), a "Console Simulation" and maybe a tape machine or EQ (that fixes one band!) on the sum. Metering tools not counting. Everything else, I handle with individual channels and sub-groups.
As mentioned - ultimately the client decides who goes into the next round. I can only point out possible issues, what to look out for, etc.