the EXP LEG-patches work great, no doubt, and they provide maximum flexibility for legato- and slow- to medium-tempo lines. I loved to set up every arrangement with EXP LEG\'s for every section, but - whooops! - I am running out of voices so fast, that I only take EXP LEG\'s for the most significant lines.
As discussed before, a single legato EXP LEG-trigger causes a short peak of 16-voices followed by an 8-voice consumption for the sustain. So, if you don\'t stagger voices, 5 simultaneously \"legating\" sections hit the 80-voice mark for short (!). This would be almost agreeable, but when a line contains a passage where a bunch of LEG/masking samples follow each other closely, the resulting overlaps double the count of voices needed - to all 160 voices (you hopefully have). Such a short peak brutally cuts off any other sound helplessly trying to get through...
I know, that some of you guys have already installed a second or even a third GS-machine to solve this problem. For my part, I have spent weeks to make my 3 machines (with 8 HD\'s) run fast and silent and I hate the plan to set up another noise-and-heat-spreader - not to mention the hours I have to spend for maintaining each additional system (be honest, guys!!). Btw, I won\'t leave the chair to change CDs in some closet; besides, I like to keep an eye and eventually an ear on the machines - paranoid as I am.
Therefore I like to propose a different approach here, and I like to know, what you might think about that:
If we would destructively merge each sustained sample of a section with the corresponding masking sample in an average balance (I love to set CC80 to the max.) with an average masking-sample-length, we could set up another complete map of these merged legato-samples (using the EXP-control-layers-fashion). Then we were able switch the non-legato-sustains we already have to this new map using a plain ol\' GS-dimension-split (triggered by the sustain pedal). Maybe it was helpful to integrate an option into MaestroTools, that disabled the masking-sample-triggers for LEG-mode and lets the sustain-pedal-data pass, since our split could use the same original note-range for both non-legato and legato sustains. I know, we loose the possibility to adjust the masking-sample\'s volume and length in real-time, but look at the pro-side:
The voice-munching monster was condemned to half of his former monster-portion. No single note, that used more than 8 voices then (or 16 in times of overlaps).
I consider 80 voices/max more than enough for strings. Remember: This is all only the salza, someone might like to trigger some real instruments, too .
Go and get me! Don\'t hesitate - I am curious...
[This message has been edited by Roman Beilharz (edited 02-27-2002).]
This idea (and a couple of variations on that theme) were considered during development. We affectionately called them the “brute force” approach to legato. There is a major manufacturing drawback to doing this: Long Bow file size and the resultant disc space are approximately doubled. If I remember our calculations correctly, taking this approach would have increased the already huge 16 disc library to 24 discs. Not only that, but some of the Long Bow discs would have required a dual CD split since file size would have considerably exceeded the single CD limit (at least without compression). Gary did the math and said, “wrong choice guys.”
In a world with no practical limitations, the brute force method would have been my preference, partly because I could have programmed it in a fraction of the time. By the way, one of the variations, in effect, did not give up the adjustability of the transitions. All of them had the advantage of reduced polyphony demands. Unfortunately, it came down to giving the user a workable LEG choice using masking samples that required insignificant additional disc space (but at the cost of polyphony) or no choice at all.
To help matters I could suggest two things in the here and now: Use the LEG instruments only where the passage(s) absolutely require it. Use the standard EXP instruments elsewhere. Or, convert the completed (or even temp) LEG tracks to audio. Not the most convenient way of dealing with this, but it does free up considerable polyphony. As you’re already aware, the other solution is to add more PCs.
As DVDs inevitably become the delivery system of choice and as new features are added to GigaStudio we may revisit this subject.
I understand I came over with ancient brabraluga. Sorry. Obviously I am on this train for too short for some things. And I didn\'t do maths.
Nevertheless I am whirling the theme over and over, because I am not yet ready to accept this polyphonic nightmare...
And I am sure we will have to rethink the \"brute-force\" theme again when eventually working at a solo-library. By the way, didn\'t I tell you this story that happened in 2003? Some famous composer cried out loud against the roaring fans of his machines:
\"Why does this oboe-gigastudio always freeze in the middle of the main theme?\"
there is the option of creating a new gig your self with recorded versions of the legato files. This will be tedious tho.
Honestly tho. I understand the polyphony issues, but as computer speed increases and GigaStudio Max Poly increases this \"problem\" will become less of an issue. For me realism and flexibilty are the main issues. If it costs me 32 voices of poly for one note, then so be it. Just give me a \"lite\" version of the file that I can do my mockups with, then mix down/bounce with the \"real\" files.
As libraries grow, we of course run into more than just poly count issues, but sample count/memory issues.
It becopmes a \"you want your cake and eat it too\" type situation.
I would hate to see developers (as many do now) limit themmselves to what can be achieved because of poly count and memory issues. Putting a leash onthem and not letting them push technology to the extreme and often uncharted territory just for \"conveniences\" sake just doesn\'t agree with me.
Of course conveniece is a big issue and should be something to focus on, which is why I love \"lite\" instruments and other solutions of the same. Good things usually come at a high cost in terms of resources. Its jsut the way it goes.
thanks for your opinion. Yes, I WILL create a brute-force version of each section, I too love \"lite\" versions, especially when they include all samples .
Other than you seem to understand my post, I completely agree with you not to stop breaking frontiers just to fit average resources. Hey, today I increased my RAM to 1GB and changed mobos to get my piece of the cake, I will eat it, no doubt!It\'s just, that one arrangement I\'ve done was actually hitting the 160-voices max for short when I hadn\'t even finished the strings! Although I recorded strictly mono I had to delete any masking-samples I was able to live without and to manually reduce voice-overlaps by shortening some notes. This has been one of these moments, you know...