Sunday, May 11, 2008

better living through irresponsible editing

this week i am at the tail end of a month and a half of hard pushing to finish a record with an artist from LA. he's been staying here at the studio and i've averaged roughly 4 hours of sleep a night over the 6 weeks (i keep a log). i have this thing that keeps me up during the day - this job at nasa that i love a whole lot and care about - so i've been tracking, editing, producing, mixing through the nights and working during the days.

originally, when devon first showed up in early april, we were going to finish the EP (4 songs) that we'd almost finished tracking. that turned into re-recording most the parts and all the vocals. then, as i wrote about a couple weeks ago, one of our friends died, and we suddenly had two more songs.

my studio is the sort of place where musicians come to record and end up writing more new stuff than we can deal with. once, we set everything up to record vocals on a sweet little ballad and this crazy new song came out of him and a floppy tom. i was rolling tape, so i have the birth of the song - from nothingness to catchy little tune - in the middle of this other song's session.

so we jumped from 4 songs to 7, and now there's another song he's written that he wants still to record when he gets back to LA.

this makes my job as co-producer, editor, and mix engineer an entropic mess, but i don't make things easy on myself. jumbling up the writing, production, editing, and mixing processes like this is not for the weak. so you'd think with all the fancy tools i have at my disposal here, i'd look for shortcuts wherever i could. i sort of do, but then i sort of don't.

for example, one of the oldest songs on this record is one that we started tracking more than 2 years ago. some of my friends from the cassettes came over and we recorded a bunch of improvisational parts. that night i recorded two rhodes parts that together became one of the central hooks of the song. for 2 years, our rough mixes of this song have suffered these improvisational rhodes tracks unedited and mostly unmixed.

[background aside (i should learn how to do footnotes)]: i'm not being falsely modest or self-deprecating when i say that my ability to play things in time is severely challenged. i offset this serious weakness with boundless creativity, a willingness to edit, and some unusually keen ears and sense of pitch, harmony, and timber. combine these strenghts and weaknesses with a love for the physics of sound, an affinity for complex layering, and a purist attitude towards capturing special moments even if they're not perfect, and i set myself up for endless hours comping and editing.

ok so back to these rhodes parts. because it was kind of a big jam session that night, i wasn't taking the time to make sure the technical aspects of the recordings were of the highest quality. in addition to my naturally crappy timing (such that almost every note needs to be moved in time (thank the lord we at least used a click)), the rhodes was acting up that evening and would pop loudly when i hit hard. for the rough mix, i had thrown my waves xcrackle plugin on the track, but now that i'm editing for real, i just don't feel right about using that.

this is where the irresponsibility sets in. if i were reasonable and rational about this, i'd use recycle or beat detective or ableton live or some similar time compression/expansion software (i just got melodyne, YAAAAY!) to automatically quantize these parts. unfortunately, on a rhodes part like this, i don't think it sounds good enough to do that. so, over the entire length of the 5 minute song, i go through and manually chop up the part, manually move things, and manually find the best crossfade lengths. this means a several hundred edits per track, just to get the part in time. this can go faster than you think because i can almost hear what a waveform is gonna sound like just by looking at it, so most of the edits on first pass can be done visually. but still - it can take hours.

then there are the pops. this is what one looks like:



i really should just use xcrackle, but the purist in me won't allow it. i already have issues with pops and clicks getting introduced digitally into my audio by protools when rendering anything new. i don't need to encourage problems by starting with bad waveforms that popped on the way in because of faulty old analog electronics (gotta love them). so here's my 100% manual approach:

step 1: locate a pop




step 2: figure out where offending waveform begins and ends and find one adjacent to it that's healthy



step 3: replace bad waveform with healthy waveform



step 4: move rest of audio over slightly to match zero crossings





step 5: manually crossfade edges



step 6: listen

sometimes if i'm feeling extra sleepy and irresponsible i might skip this step, but not very often. it's really the most important part. the satisfaction of having eliminated the problem usually makes it worth it.

i know what you're thinking (if you're still awake). this is ridiculous. this 5 minute part has about 80 of these pops. i have no defense.

but that's what i'm up to tonight. i will be doing this until i change clothes and go to work at 8am. thankfully this push is over tomorrow. i will sleep tomorrow night, then get up and drive to st. louis to see radiohead and visit my dad. devon went to see them tonight in the pouring rain at the nissan pavilion and said it was worth freezing his butt off for. i figure it should be worth the 13 hour drive with a great friend.

4 Comments:

Blogger dd said...

Ah, editing. I focus more on picture than sound, but I totally geek out on this kind of thing regardless. Although I'm much less of a purist than you, maybe because my hearing is crap. (Although I can apparently hear MP3 compression much better than most.)

A maybe sideways question. Obviously you are passionate about your work and your avocation. How do you think your life would change if you had the ability to make your avocation your vocation? Say if this record you were working on sold a million copies and you became an in-demand engineer.

(I am struggling with the tension of having your passion be your vocation right now, why I ask. It's probably a silly question, but I need my brain jostled.)

May 12, 2008 6:52:00 AM EDT  
Blogger Carlos Anaconda said...

It would be good, for comparison purposes, to have links to the track before and after the corrections, or even just part of the track.

May 12, 2008 9:20:00 AM EDT  
Blogger cherry blossom said...

i thought about posting the audio too, but that would have involved bouncing the track and i was low on energy. i'll see if i can do it tonight.

dd, it's a very good question. i am slowly making the transition. in my own little niche market, there is already more demand for my time than i can sanely meet. it will probably just take a while for the money to catch up. this record isn't intended for a million sales - it is a devotional album of Baha'i writings and quotations, and there are only 6 million Baha'is worldwide. But that's kind of beside your point.

I'm not counting on making a living as an engineer. I'm instead working more towards the science/technology/art side of things that i wrote about in my first post here. i may even be able to get into this field of data sonification and stay at nasa, which would really be the best of all worlds.

May 12, 2008 10:03:00 AM EDT  
Blogger Wednesday said...

i may even be able to get into this field of data sonification and stay at nasa, which would really be the best of all worlds.


I have a friend who designs virtual reality for NASA down in Houston - I wonder if you'd want his contact info.

As far as the studio fixes, I think a better use of time would be to correct the input problems. The time spent there would have benefits beyond this project (i.e. learning how to play in time, fixing/replacing the rhodes). Whereas the work you're doing now will just have to be repeated next project.

May 12, 2008 10:26:00 AM EDT  

Post a Comment

<< Home