In order to keep this documentation
concise finite and
interesting useful, I’ll be focusing mostly on the challenges I faced that were specific to this multitrack, and describing my solutions to these challenges.
Track alignment:
* Nudged the OH L track a bit to the right so that its snare hits align with those in the OH R track
* Nudged the Pre-Gated Rack Toms track a bit to the left so that its tom hits precede those in OH tracks (because it seems that the pre-gating process had caused some delay that got baked-in to the provided wave file)
Drum track polarities:
* All track polarities seemed to be correct as provided, however - the song had more energy once I flipped the polarity of both the Kick and the Bass together, so I went with that (this worked because the Kick sounded fine against other drum tracks at either polarity after the drum tracks had all been processed in terms of EQ and dynamics)
Audible “gremlins” baked-in to provided tracks:
* OH tracks seem to be tainted with nasty clicks occurring every time a rack tom hits, so worked around this issue by ducking the OH stereo submix aggressively, triggered by a side chain from the rack toms track (not ideal, but kinda works because the attention of the listener is mostly focused on the sound coming from the rack tom close mic during rack tom hits, and there are usually no important cymbal hits during tom hits either)
* Because the OH submix is now getting ducked heavily during rack tom hits, I’m compensating for this by additionally employing an upward gate with similar timings on the Room mic, also triggered by a side chain from the rack toms track, in order to “fill in” the occasional “holes” left by that OH ducking (I made an upward gate from a normal range-limited downward gate by adding make-up gain after it). This way the loss of cymbal tails is less noticeable during OH ducking as the Room mic expansion replaces that loss at a similar level (also helps with the tom body sound a bit)
* The tom hit transients in the pre-gated rack toms close mic track were offensively clicky and pokey, so I had to tame those with a transient processor (especially because the OH were getting ducked during rack tom hits already as part of another workaround, and so I had to bring up the rack tom close mics more than I otherwise would have had to)
* Lead guitar track contains annoying metallic clanks on some picked notes, luckily these mostly disappeared into the mix, but a couple of these clanks became distractingly obvious during the first intro phrase, so I made a micro-edit and replaced a short offensive section with an identical section (notation-wise) copied from the repetition of that intro phrase where the clanks were more subdued (a listener can’t really tell the micro-edit is even there unless they read this doc, I think)
* Once the mix was almost completed, some new audible “gremlins” emerged during various sections on a few specific kick drum hits, as the high-mid frequencies from all drum tracks and the high-mid frequencies from guitars and vocals all combined together in a weird way. The only solution I could come up with was to automate a transient processor to go briefly online on the entire Drums submix during the few offensive kick hits (not ideal, but not noticeable to a non-clinical listener)
Consistency of levels / timbre / timing:
* Lead Vocals needed extensive gain automation during the final Chorus section in order to make them consistent compared to the prior Chorus sections
* Speaking of consistency, the Lead Vocals appear to have been sung either with a different throat overdrive timbre, or by a different singer during the second Verse and second Chorus sections, I can’t really tell which is the case; luckily there were also differences in instrumentation arrangement and overall energy during these sections, partially “legitimizing” the difference in vocal timbre so it doesn’t sound totally out-of-place
* Aah - now about that sweet bridge section! Ironically, the cross-stick hits on the snare were way too consistent and thus mechanical-sounding in terms of hit velocity, to the point of masking out lead vocals occasionally, so I had to automate some variation into those cross-sticks, and additionally had to automate lead vocals in terms of gain as well as some special effects in order to make the bridge section work
* The Kick and Snare sounded ok after some EQ-ing, but in the mix they ended up a bit too “default and average sounding” to my taste, and I felt like I wanted more weight and definition to them. It then occurred to me that I could simulate a tastefully muffled drum kit by using an autofilter triggered by envelope follower on the close mic tracks - so I tried that, liked the result, and went with it
Guitar sound(s):
* I kinda liked the character of the provided printed ampsim sounds, and how they worked together, so with a bit of EQ I was able to make them sit in the mix quite well
* As for the DI-only tracks, I had to come up with my own ampsim settings, which was a real challenge since I’m not a real guitar player by any stretch of imagination. I ended up choosing different amp models to each of the double-tracked “siblings” in order to make the stereo spread sound less obviously double-tracked when panned, and then messed with EQ settings (including some mid-side EQ) to make the mix sound non-lopsided again, and to make it blend in with the rest of the song
* I especially like how I was able to choose my ampsim static gain settings such that the guitars actually naturally clean up a bit during the less busy bridge section - simply by the player picking a bit softer, which I think sounds fitting
Bass:
* FET-style compression alone wasn’t quite enough to make the bass audible on small speakers, so I made a frequency crossover using gentle 6dB/Oct low-pass and high-pass filters, let the high-passed half go into the mix directly, but put a bass amp sim on the low-passed half with some gain on it, and then blended between the two halves to taste - the idea being that the harmonics added by the bass amp sim being fed from low-passed bass signal will generate a synthetic impression of bass in the human auditory brain even if only the harmonics without fundamental come through on small speakers (I tried a simpler approach first without a frequency crossover - just a bass amp sim in series - but couldn’t make my simpler solution work with the mix well enough, so switched to the more complex crossover method instead)
My mix setup:
* Mixing entirely in-the-box (Logic)
* Using stock plugins only (Logic)
* Mixing into a main bus compressor from the start
* Mono button on main bus for checking mono compatibility
* L/R swap button on main bus for checking against lopsided imaging
* Pan law set to +3dB boost at extremes (the default pan law in Logic)
* Using BCF2000 (motorized faders) in Logic Control emulation mode to quickly set levels, and to quickly set EQ curves; also to “ride in” some automation curves in real-time
Strategy:
* Start out with mono button engaged and work on levels, EQ and dynamics in mono
* Once a decent mono mix has been achieved, disengage mono button and begin work on panning and imaging while occasionally toggling L/R swap button in order to make sure the mix doesn’t become noticeably lopsided
* Add specific send effect buses on-the-go as needed
* For this mix I added a send effect bus for the main glue ambience when nearing the end of the mixing process (although sometimes I add it near the beginning of the process, before doing any EQ work); I then engaged the mono button and messed around with instrument-group send levels going into the main ambience bus until I got distance impressions (depth-wise, aka. z-axis-wise) across instruments the way I felt made sense for this song, but also compared against some commercial reference mixes in order to make sure I’m in the ballpark
* While recognizing the fact that the mix should not be “mastered”, it nevertheless at some point became necessary to put low-cut and high-cut filters (along with slight corrective EQ) before and after the main bus compressor, in order to prevent “cone flappers” in the low-end, and in order to stop myself from attempting to “throw darts in the dark” by trying to mix very high frequencies that I cannot even hear, and of course, in order to make the bus compressor sound better as it is an important part of the sound I was going for; I hope that filtering out the extreme bottom and top frequencies together with slight corrective EQ is allowed on the main bus since mixing into a main bus compressor seems to be allowed. This was all done with just regular stereo EQ and regular stereo bus compressor, I’m not doing any mid-side or multi-band processing, nor any saturation or limiting on the main bus, just for the record
Monitoring:
* HS80M pair (but no subwoofer, sadly)
* MDR-7506 cans (these do reveal subs that I would miss otherwise)
* A device for checking midrange in isolation: literally a small kitchen radio / cassette / cd player combo with an “aux in” socket on it - if my mix sounds noticeably wonkier than commercial mixes on this thing then I know I’m in trouble
* Scarlett 18i20 1st Gen
* Last but not least - and this is a good one - since I don’t own a car, I’ve been replacing “the infamous car check” with walking into a corner of the room while playback is in progress on speakers in order to listen to any boominess and/or muddiness from there where the low end gets boosted by the geometry of the room
I hope this documentation was useful to someone.
)
