It will be interesting to find out what Peter can do with this because I get the same thing with piano RT's. The reason is it's an audio RT and not midi.

Here's the thing with RT's. They are static audio recordings done by a real player in a studio. When you're talking about sustaining chords over a chord change, the player who did the recordings is not telepathic. He doesn't know what chords a user is going to come up with. Pick any chord pattern on a keyboard and play it. Now invert the second chord. Different sound and movement, right? Now, put the second chord back and invert the last chord so it transitions smoothly back to the first chord. Again a totally different sound and since you know where you're going with this you will hold the common notes from one chord to the next and only change what notes are necessary to change the chord. That's what makes the chord transitions sound smooth.

So....what should PG tell the player who did these RT's to play? What voicings should he use? There are only so many chord variations and voicings that can fit on one RT. And not just variations but the chords themselves.

I suspect the players have a chart of common chord moves but no way can they record all the different chords plus all the inversions. Without all the possible inversions if you throw a chord change that was not exactly recorded what is the program to do with your instructions? That chord movement was not originally recorded in the proper order. Now it has to find those three chords in different places in the RT file, stitch them together and the inversions probably won't match. That's where the high/low chord sound comes from. When you play that live you know what you're going to do so your hand position is in the right place to do the correct inversions to make it sound smooth without a 4 or 5 note jump in hand position. The program has no idea what chords you're going to come up with so it has to instantly search the RT file for the chords you want and put them together.

I think you can see the issue here. There's no way there's room in an RT for all the different chord movements plus all the inversions required to make all the possible chord patterns sound smooth. I have no idea of the actual number but I suspect it's in the thousands if not tens of thousands of possible combinations. It would take a player a week in the studio to lay down all that and the file size would be huge just for one Real Track.

Then there's the sustaining between chords. My example above I talked about you holding common notes with your fingers and only changing the notes to make the chord change. Those are called cluster chords because they're right next to each other. On a piano the sustain pedal is used for that and a skilled player can use the pedal to smooth out a two octave jump or sustain an arppegio.

If a RT was recorded using the sustain pedal to smooth out chord changes that is only good for those exact recorded changes. Throw different changes onto the chord grid and there's no pedal sustain so the chord change sounds choppy. This is one reason there's no octave command on the chord grid. You put in the chord but the program decides on the voicing and octave. There is absolutely no correction for that because RT's are audio files. Think about octaves for a second. Everything I just talked about would have to be recorded in duplicate to add a second octave to an RT.

Same as my example above, the player would have to record all possible chord combos and inversions with the pedal sustain on each and every one of them. Short of that what could PG do, maybe have the program somehow create an envelope to try to sustain it instantly on the fly? That would be some amazing real time programming to create a realistic piano sustain where there was none in the recorded RT.

Up top I said the RT's are audio files and not midi tracks. The difference is midi can be programmed to play whatever exact notes you want so those chord transitions are now correct because they're written into the file, not prerecorded in a studio and can't be changed. That's why your midi file sounds good but the RT does not. Same with pedal sustain. All that can be written into a midi file but there's nothing you can do with a prerecorded audio file.

This is a long and convoluted explanation of what I think the problem is.

Bob


Biab/RB latest build, Win 11 Pro, Ryzen 5 5600 G, 512 Gig SSD, 16 Gigs Ram, Steinberg UR22 MkII, Roland Sonic Cell, Kurzweil PC3, Hammond SK1, Korg PA3XPro, Garritan JABB, Hypercanvas, Sampletank 3, more.