When do YOU mix?

[JJRoach]

Hey all,

Just curious here...I'm glad I finally started reading threads in this forum. It's very relieving (and somewhat a drag that I still need to do the work...whatever, no sympathy for the whiner here!) that it's expected that I do my own mixing with things other than levels and panning.

the question I'm wondering is...

when do YOU mix? I know people who mix as they go...but I'm really wondering what the best method of working really is. I know...this is a loaded question since everyone has a different method, but I'm still trying figure out how I can get through a piece without losing myself, especially since there are so many things we can tweak with a virtual orchestra..


example: I need more violins here!...ok do I increase midi volume, increase the musical volume of the "players" (such as, more MF than MP), increase audio volume...

I mean, I try to leave my audio levels at unity gain when I'm sequencing, and also try to leave MIDI volume alone and try first to do things with expression/breath/velocity crossfading control first...but again, I think it's really easy (especially for those of us...me...who are much less experienced than you folks) to lose oneself in a piece.

Josh

[JWL]

Josh, mixing as you go serves the purposes of the moment. As you are adding instruments to your score the tendency is to favor the instrument "du jour". This is akin to a monitor mix in a live performance-- you hear what you need to hear based upon your listening position and needs on the stage. If you are sitting in the violin section, you'll need to hear the violins a bit more than other instruments. Same with sequencing. Sonic orientation sounds "right" at that particular moment.

For orchestral final mixes, I always start from scratch. For an economy of CPU resources, I never sequence with my best CPU-hungry reverb, for example. Also, for mixing (for the sake of CPU resources) I run my tracks out as audio and dump my VI's to allow for more processing plugins. Should I need to change something in a particular track, it's easy enough for me to call up the MIDI sequence, change the part where needed, and then run it out again as an audio track. This, of course, places a lot more importance of making the x-fades and expression moves as convincingly as possible as early on as possible. I worry less about overall mixing in the early stages as I do about making sure the CC's are all as believable as I can make them. I know the general levels are fine, and with VSL-VI I have no doubts that the samples will blend well when the mix is done.

For true objectivity when wearing multiple hats, I think like a composer when it's necessary, then like a tracking engineer, then like a mixing engineer, (a music copyist when it's called for), and finally like a mastering engineer. Starting with a clean slate clarifies each phase of the process as though there were a team of people working on the project.

Funny- I can only give 100% to anything if I do just one thing at a time. There are people who can multi-task with a "do-it-all-as-you-go" approach. Even though my computer has dual processors, my brain is still running on a single chip with limited memory!!

[JJRoach]

JWL,

Thanks for your thoughts. Yeah, you address a big reason I started this thread...

For example, I've tended to start working dry...and then who knows from there...as I add more instruments, I find myself going back to ajust the ones I had hoped I was finished with...again, hard to guage how much something needs to be when working on it and then putting it in the mix...but, like I was saying...I was working with string patches...and when they were dry, they sounded terrible...and then when I put them in a reverb situation, the xfades between samples/notes in a phrase were much smoother...thinking I just need more experience...and it sounds like I just need to work on 1 performance at a time...getting a strong signal, then mixing later...you agree?

of course, I should probably, at some point soon, decide what my definition of the various dynamics should be for each section, much like I do as a player (of course, when I play with an orchestra, I'm not thinking about dynamics as much as being in the right part of the blend!)

Josh

[JWL]

Hey Josh:

Yes, there is so much more to consider with VI than we're used to.

For dynamics and timing, I always make a dummy track with just piano or piano plus some other key orchestral instruments, even if I use a "lesser" library. I ride in some expression at key points just to give myself a benchmark to work with BEFORE even dealing with VI input. I'll ride those peaks and valleys to the upper and lower limits of the dynamic range I want for the final mix-- and, of course, with all timing, tempo changes, and other pretty much in place. Using a reference track from a CD always helps define how loud the brass might be in the end or how soft the strings might be or where a solo French horn or oboe might sit within that dynamic range. I just set my velocity limits of my temp track to match the CD reference track as closely as possible.

Of course, I always run out my temp track as audio because it uses fewer resources than running them with VI's, which get dumped. This may require setting up two different sequences-- the VI sequence will only contain a stereo audio of the temp track until VI instances are loaded. You can troubleshoot a great deal by detailing your temp track up front. Saves on Advil.

While working with the samples, I may opt to edit the interpretive settings, but at least I don't have to start from scratch.

I've also found it easier on the soul to just get all the note info entered. I do *some* expressive editing, perhaps 60-70% of what I'd eventually like to do, but as you say-- when other instruments are added it's a constant battle going back and forth to fine tune these elements. It's also painful for a while to have to listen to "unfinished" expression editing, but once all the instruments are entered and are in that 60-70% ball park, the fun of sculpting really begins.

This has been the reason I've chosen to use separate tracks for some cc data. Digital Performer 5 has what they call "track folders", so a group of cc MIDI tracks along with its corresponding VI can all be opened or closed in an actual folder. It's a bit cleaner than it is with Logic, I've found. Some may beg to differ. But-- say I want the flute 1's expression to follow the violin 1's expression to some extent. Sometimes it's easier to just copy in that data from one instrument's cc track to another and edit it. DP also has custom cc control consoles which can scale such mirrored data in percentages, invert the polarity of that data, and most importantly: assign any cc source data to any cc target data-- from one instrument to another or from one cc# to another. CC#1 data can be routed to CC#7 if you wanted. This saves tons of time when entering the data from scratch each time is more trouble than it's worth. Of course, where this approach creates too mechanical a result which is harder to edit, individual track entry cannot be avoided for the sake of the believability of the results.

On a similar note-- there have been many studies done on how the brains of men and women work. Women may have less trouble switching gears and doing several things at once. Guys like me have to do one thing at a time. My approach is to get all the note entry done, given that the orchestration is done before dealing with VI. It's just too much for me to think about-- arranging, orchestrating, and mixing at the same time. I've never been one to compose from scratch in a sequencer. The details of geekdom interfere with the creative thought process way too much.

Given that the music is composed and that the orchestral concept has at least been sketched out on paper, it's much easier to delve into the sequencing phase. It's even easier when working from a completed score.

Once the matrices are set, notes entered, and proper articulations and keyswitches are assigned, I tend not to do *too much* detailing too soon. I'll check a patch for the possibiities for x-fade and expression, etc., take a few notes just as a reminder, but largely I save the tweaks for later.

You learn quickly where your workflow forces you to backtrack more than you'd like. The less backtracking, the happier the project (for me, anyway).

Going slowly doesn't bother me that much, unless a deadline is bearing down. What does bother me is mental disorganization.