For optimized audio production, you should make a mixdown of the main elements of your piece; one stereo track for each instrument or instrument group, depending on the build.
Here’s an example of final 22 tracks, mixed down from about 60;
01 Bass drum
06 Drum Reverb
07 Drum FX
09 Bass FX
10 Bass Reverb
12 Lead FX
13 Kead Reverb
15 TB303 FX
17 Pad FX
19 Apreggio FX
20 Drum Dub
21 Bass Dub
22 Lead Dub
This is a tune that was supposed to be dubbed by vocals, percussion, and some other stuff, but I’ve lost the original tracks, and the backup dvd’s was corrupt, so it will never happen.
On my sessionpage the track ‘B10S’ is the first downmix (raw demo mix), the track ‘B10S (edit)’ was one of the first mixes in the mastering process.
I master all tracks separately, and burn a backup-dvd with all both mastered and unmastered. Then I create a rawmix of the mastered tracks, which I dub with the original tracks if anything get lost in the downmix, and repeat until I no longer can do anything without loosing quality of the sound; then I’ll have the final master.
If some guy at the record company wants more of anything he can easilly dub it from the backup tracks, but if he want you to remove something he can go flush himself, because then you have to go far back in production. This can be done without problems if you got backup of all tracks and workfiles of your acid, protools or whatever app you use.
It’s important to never mix down effects on the tracks, but always keep them seperate, this in case you have to do som reverse enginering.
What I find most difficult to produce and master myself is stereoimaging and radiocompression. I often use dolby-tricks that really does not sound good if being played trough some automated stereomastering hardware, but it rocks the juice out of the living daylight when standing on a subwoofer with an hq headset….