Realistic multi-instrument context with MIDI alone: share your tips.

Hello.

I can play the keyboard and that’s it. No side(wo)men around. To emulate a multi-instrument context (as posted in the Music showcase section or here:

(David: if this double posting is abusive, suppress this link thank you), I have to go along the lines of what I consider a satisfactory method.

Here are my principles:

-I don’t consider that all music parts should be improvised: string sections, horn sections in particular should most of the time be duly written, if you have the technical possibility to do it. If not, probably give those instruments as little harmonic importance as possible. Leave it melodic.

-I imagine the basic rythmic pattern that fits my Melodic/Harmonic phrases the best over 2 bars (say: bass drum and snare drum). Then I tinker with Drummer tracks to come up with something that works and let it find solutions regarding Hi-Hat, and fillups. I use quantize for bass and snare drums, probably 100% but I know some of you will disagree. The advantage of exact quantization at this stage is to both provide a rock-solid on-the-beat backing for the other recordings and get you more in the mood than a click. I don’t use clicks. If you do, tell me what advantages you think it has.

Then I record the instruments. I always train for a few passes before going “live”.

-If you can: don’t record at a lower tempo. It really can mess up the feeling once you play the whole thing at the initial tempo.

-Order of recording/sequencing:

1) Bass. Here, I use the following: quantize strength 75% and exclude quantization within 12% of the quantize value (if 1/16 note: 12% of 240 ticks is 29. So I type Q-Range= -29 though I sometimes remain skeptical about how Logic works in this regard).

2) Piano (or keyboard) or guitars, whatever the harmonic pillar is.

3) At this point, I stop all recordings and focus on the instrument choices. This basic setup (what we may call the rhythmic section) is key to the whole music. This means reviewing here all possible drums, bass, and keyboard samples in my library. I’m not kidding. You never know what you may come up with if you apply this radical but often fruitful step. This is a reason why I resent growing my library. I’ve noticed how people with evergrowing libraries know less and less what they have or don’t have, hence don’t bother anymore with trying to find a better solution than the one, two or three that come to their mind. They can't reasonably test the whole of their sound bank, and I think this is a danger.

4) Then recording resumes: Melodic counterpoint, whatever it is: a moog, a string section, a 2nd kbd, a guitar… Quantize may be of no need here. If this part is both melodic and harmonic and has parts itself (for example a string section written in V1, V2, Vla, Cellos and Ctbs): I always record from top to bottom for a matter of volume balance of these parts. Harmonic balance of secundary groups has to be probably more top-down than bottom-up for EQ and spectral reasons.

4) Rythmic add-ons: horn riffs, or the kind. I find quantization often detrimental here. Short riffs tend to compress rythmic lengths in real life. Longer ones, probably not so much for obvious reasons.

5) I never assume that what I planned was wrong if it doesn’t sound well from the get-go. In particular, a MIDI part played alone may seem to not cut it and become realistic with the rest. Anyway, at this stage, balancing the volumes is SO key to a well-balanced “MIDI band”. Way more than EQ… which I’ll still have to fight with when the singing comes in.

6) “Less is more” is generally a better rule if I’m not satisfied with the result or feel something is amiss.

OK, that’s about it. All this is certainly not perfect but I find it generally a good method.

I’d like to hear see some of you add or subtract to/from what I lay out here. I don’t think these general principles are often discussed.

Thank you for reading/listening/watching.

Thanx for sharing!

Many interesting and informative insights in there!

For the sake of preserving creativity, if productivity time limit isn't in the equation I try to avoid using the same methodology. I am a strong believer that unusuality is mother of creativity, and sister of serendipity.

Well, I think it may all depend on the kind of music you're in. Out of experience, I know that I probably miss something everytime I neglect to loop-play a chorus or a verse, trying to add whatever part on the fly. There's always good chops to extract from this process. But, I know too that I'll miss something else and at least as much if I neglect the think-it-over step. A harmonic string session can hardly be improvised, it really has to be writen=thought over. Varying an accompaniement formula on a piano, kbd, or guitar is indeed another matter (not always though) and I think we're better off with a dose of trial and error. All this remains linked to the cultural nature of the instruments: in a "band", most pianists, kbdists, guitarists don't read what they're actually playing. But we'd never see a string session (and most of the time a brass session) play without reading, or at least when they don't, they know their parts by heart, or simply fake the game as is soooooo often the case on TV.

That being said, what exactly do you mean with unusuality? What are these different approaches you have ?