View Full Version : Legato by MIDI controllers

08-11-2004, 10:23 AM
Is there anyone else who uses GOS without MaestroTools? I typically program everything into midi streams, adding controller events by hand where necessary. Now I'm trying to improve use of Legato samples.

So far, it seems that midi controller #17 adjusts release-note sustain -- set it to a high enough value and there's a blur of notes way past legato.

Midi controller #80 adjusts dynamic value of the attack note in a legato passage, I think.

What else is used in starting and stopping a legato phrase?

Here is a sample 'cello legato -- but I'm not fully happy with it. Comments please!

http://mysite.verizon.net/vze8z9ci/vcl_legato_01.mp3 (very short mp3, about 200k on a crude temporary web site)

Please don't just tell me 'use Maestro tools' unless you can't hold it in -- I'm happy working directly in the midi stream (use Sonar and GigaStudio, if it matters). Thanks!

08-11-2004, 10:27 AM
This may be temporary -- I couldn't download my own example, except by going first to the temporary site, then selecting the 'cello mp3 for saving:


Tom Hopkins
08-11-2004, 12:16 PM

Certain features in GOS are intrinsically tied to Maestrotools. The legato feature is one of them. There is no really practical way to use the LEG instruments’ legato feature without using Maestrotools. The legato mode involves the use of “masking” samples which are mapped to the upper part of the keyboard (with a “breakback” to the bottom of the keyboard). These samples are activated in Maestrotools with sustain pedal data. Maestrotools automatically directs the MIDI information to the proper keys to add the corresponding masking sample contribution to each note in legato (masking samples work to smooth the transitions between notes). So, if you choose not to use Maestrotools you lose the legato feature. You also lose the 2 flavors of mono mode that are very useful in legato playing. You lose a lot by deciding not to use Maestrotools. Without it the library simply does not function as designed.


08-11-2004, 02:07 PM
Surely you exaggerate! I don't mean to denigrate the MaestroTools, but since I must do so much at the midi event level anyway I'd rather add in the legato at that level -- I already handle the up/down alt stuff with relative ease.

It's possible, I think ... but if no one else has experience I'll report back later!

Anyone else delved into midi controllers, Gigastudio, and the Legato patches?

08-11-2004, 07:21 PM
Seems to be a case of Emporer's new clothes, here. I don't intend to criticize, since I'm a registered, happy, heavy user of GOS. But the legato patches have this: an attack sample, put in place 4 octaves higher than the nominal pitch, and controlled in intensity by the midi controller #80 plus the sustain of the main sample, with release controlled in duration by midi controller #17. Nothing else, unless I'm missing something in the GigaStudio editor.

So, to use legato without Maestro Tools, one doubles the initial note of every slur with a note 4 octaves higher (adjusted to tasted with CC#80), one uses a note duration very close to the full nominal note value, and one adjusts midi controller #17 to put a little extra tail on every note of the legato phrase. About 35 seems to work for my moderate eighths at mm=104. For the last note of the legato phrase, set controller #17 to 0.

Is there more, Tom?

Tom Hopkins
08-12-2004, 06:05 AM

Well, it isn’t quite as you’ve described it. Depending upon the range of the specific instrument, the masking samples are often not mapped entirely 4 octaves above. For many instruments the mapping breaks back to the bottom of the keyboard when the masking samples run out of room at the top. You will need to keep track of the disjointed mapping when you copy the needed notes. Now, why you would want to go to the trouble of manually copying notes whenever you wished to activate legato mode (rather than just stepping on the sustain pedal or entering cc64 data to use Maestrotools automatic functions) is beyond me. It appears that you are intent upon making things way more difficult than they need to be. Maestrotools simplifies so much and is so easy to use. Not to mention the valuable mono modes which I wouldn’t want to be without. But enough of my opinions. If you wish to take the long way around, that’s your privilege, and it certainly can be done if you want to wrestle with it like that. One other small detail: The masking samples are not attack samples. Each one has been extracted from a point between 500 and 750 ms into the corresponding sustain wave to match the timbre of the sustain while successfully “filling in” the dip at the transition between notes. It took me literally weeks to get the best extractions for this function during development. Boy, don’t I wish they were just simple attack samples!


08-12-2004, 06:16 AM
Thank you! That's more helpful information. I was working with the 'cello, where no upper range problems appear. Things like the pedal only work if you try to "play" the parts, which imo just is not feasible for the best realism -- so adding a cc64 is no easier or harder than a cc17, with the advantage that cc17 can be adjusted for the length of the overlapping sustain. One can 'step record' an orchestral part far, far faster than playing it with a midi keyboard to a metronome -- so anything geared to live performance doesn't help the work flow (again, imo).

Tom Hopkins
08-12-2004, 06:40 AM

Except that cc64 and cc17 are not functionally equivalent. But if it works for you, fine.

"Things like the pedal only work if you try to "play" the parts, which imo just is not feasible for the best realism."

Correct me if I misunderstood the above statement but: The greatest writing convenience (and printable result) is often achieved through using step entry of data or notation, but if you are after the most realistic MIDI mockups - that is best accomplished by actually playing in the parts one at a time with expression (followed by editing to fine-tune things). Step entry (even when taking steps to "humanize" the data) rarely approaches the results of actually playing in the parts. The most convincing demos for GOS and GPO have been consistently created using the "play it in" approach. Human timing and phrasing is a very complex thing that is best accomplished this way.


08-12-2004, 08:22 AM
Well, you understood me -- but we disagree! I am both a good enough keyboard player and a good enough string player (not claiming to be actually "GOOD" at either) to feel that a keyboard with a joystick or mod wheel can't begin to approximate a string player's efforts. Maybe with tons of other physical controller sources and a lot of practice? .... I bought a Kurzweill 2500 keyboard in order to use the neat ribbon controllers for this kind of thing -- but it still wasn't realistic.

My 'cello legato example is now much improved with your help -- also put up a score snippet to show some of the details I'm trying to accomplish, such as: a fingered, not open, D natural; up bows and down bows, legato starts and stops, crescendo and diminuendo within a forte dynamic level, etc. All things that a real 'cello section would handle at first sight.

Still on the temporary web site at
http://mysite.verizon.net/vze8z9ci/ with the new recording as vcl_legato_02.mp3

And thanks for your time and help, I DO appreciate it.