the piece is called SymTest because it's a test-bed for Gigastudio, Sonar, and Wavelab. I've been trying to iron out that old learning curve and this is where I am so far. What I'm mainly interested in is critique on the more technical aspects of creating an orchestral composition using samples and a sequencer. I've heard some of your pieces and believe that simply from experience you can point out my beginners errors so that I can make my second composition a much better one. Thank you.
First of all, really nice job. I have 2 technical suggestions:
1) Go easy on the volume-controlled crescedos/Diminuendos. They can sound like you're just moving a fader. See if you can't create a patch that has crossfading between provided dynamic levels using the giga-editor. A lot of messin' around, but worth it. If you must use a channel volume for dynamics, use it as sparingly as possible.
2) Pan your instruments like a real orchestra (garritans site has a nice section about this, complete with diagram), and use as many instances of gigapulse as needed to give each section the right palcement. This make a HUGE difference. See
Thanks for the tips. That thread has some usefull info. I wanted to use Fairhurst's GigaPulse set-up, but was previously prohibited because it seemed I was running too many instances of GigaPulse . . . it might have been because of a bad hard drive that I recently replaced. I'll have to try it again.
I'll also check out the Garritan site for the orchestra panning . . . although I specifically downloaded and printed a layout, I noticed that instrument placement can differ among various orchestras.
You can also run a more interesting GigaPulse setup in multiple passes. Just capture half of the orchestra at a time and mix in the sequencer.
I like the composition and style. One technical note is that the dynamics at the front end are a bit extreme. Some parts are loud, while I can barely hear other parts. This would be fine on a large stereo in a quiet room. Here on my cheap heaphones at work with PC and A/C noise in the background I coulnd't really find a good balance. The second half was more consistent in volume, but that was mostly to do with the extended theme, rather than technical production.
You might consider mixing or orchestrating up front for more consistent dynamics. Or you could use a volume maximizer on the back end. It really depends on your target. Your current mix would work in the concert hall, but not on radio or TV.
JonFairhurst, your points are also well taken. I will stay with this piece until it is right. I was also a bit concerned about the dynamics in terms of presentability if I want to interest potential customers out there. I guess I was trying to emulate a CD I have of Holst's The Planets in terms of dynamics. Like you said, works for a concert hall, but not in the business of mass media.
I will revisit the GigaPulse set-up with 3 instances and go from there.
I incorporated some of your advice into the new version of SymTest. I am still experimenting with different aspects of GigaPulse and the DSP (and WaveLab), so it will take some time to get everything right.
1) I brought up the overall level, but not to your specs. For artistic reasons, this piece will retain it's overall dynamics . . . but I will NOT naiively present a piece like this to a potential customer without first considering the eventual target medium. That part of your advice I will not forget. Thanks.
2) I made changes in the GigaPulse set up (mic placement) that made a big difference in individual instrument definition (vis a vis "things getting lost in the soup" if I may quote you).
3) I decreased the overall wideness of my extreme left and right pannings.
A note: I thought about the comments concerning panning according to realistic orchestra layouts and was reminded of an article by an author recommended here (forgot his name). Anyway, he mentioned putting some violins in the right of the mix to even things out against the heavy left violin presence (even though that instrument is traditionally to the left of the listener's psychoaccoustic perspective). In a way, his idea tied in with your comments about adjusting dynamic according to the target medium.
The thing is that we take liberties in order to make things work for the intended medium. The discerning listener will notice the discrepancies between a doctored mix and a concert recording, but until I find that Giga-Nirvana of ultra-realism I will make do with best results. In other words, as I ask for advice in the future don't think I am disregarding it by not incorporating all of it into my work. Your advice has helped me greatly and I look forward to reading your posts here.