Just for fun (and to heat up your CPU)
47 views
Skip to first unread message
Keith Blackwell
Jan 30, 2016, 6:10:49 PM1/30/16
to Analog Box
I've played around a lot with piano sounds recently, and wasn't
really all that into doing it "right", but, you know, just playing
around. So here are a few modules I stuck together into a demo bot
that you might find interesting or even useful.
The reverb has multiple all-pass chains but also allows varying coloration with comb-filter-like delay segments as well as throwing on John's EON (modified) to help lengthen it, not to mention variable room size. It isn't perfect but it does alright for me.
The instrument voice stack tries to make some instrument body noise and feed that back in to the delay oscillators that play a small role in each "string", to get a bit of piano-like sound, but it takes up considerable CPU as is. The stack tries to re-use existing outputs if they are the same note, but I changed a 4-output stack to be used only for bass notes so that it would reuse anything within a minor third. At least, that's what I intended it to do (not sure if it really works). The idea being that if the sequencer throws out a sequence of closely-spaced notes, I really don't want them all to sustain in the bass register. So counting the middle notes 8 stack, the upper notes 8 stack with no damper, and the lower notes 4 stack, that gives 20 voice polyphony. Each one of those voices can also keep ringing it's previous note long enough after a switch to give a bit of natural decay (from a delay oscillator).
Yes, I know, it does NOT sound like a piano. Too harsh. You can go into the stack and adjust some controls (c1 through c3), make it sound like a thumb-piano, or sound pretty dull and simple ramp+sine-wavey. But the basic algorithm in each of those 20 blocks would have to be changed to get it to sound more like a piano. I'll leave that to an expert.
The sequencer can be kind of fun at times, going through a pattern (basically) 4 times before changing it, and one out of every dozen sounds really cool. :-)
Oh yeah, there's a multi-voice chorus I created recently, even though it's use here is questionable. You might (or might not) like it for use in other circuits somewhere.
I have older versions of all these modules that are much, much simpler. But what's the fun of that? So here you go. You'll need ample CPU to run it, I expect. Or bypass things you don't want in order to reduce that requirement.
--
Keith W. Blackwell
The reverb has multiple all-pass chains but also allows varying coloration with comb-filter-like delay segments as well as throwing on John's EON (modified) to help lengthen it, not to mention variable room size. It isn't perfect but it does alright for me.
The instrument voice stack tries to make some instrument body noise and feed that back in to the delay oscillators that play a small role in each "string", to get a bit of piano-like sound, but it takes up considerable CPU as is. The stack tries to re-use existing outputs if they are the same note, but I changed a 4-output stack to be used only for bass notes so that it would reuse anything within a minor third. At least, that's what I intended it to do (not sure if it really works). The idea being that if the sequencer throws out a sequence of closely-spaced notes, I really don't want them all to sustain in the bass register. So counting the middle notes 8 stack, the upper notes 8 stack with no damper, and the lower notes 4 stack, that gives 20 voice polyphony. Each one of those voices can also keep ringing it's previous note long enough after a switch to give a bit of natural decay (from a delay oscillator).
Yes, I know, it does NOT sound like a piano. Too harsh. You can go into the stack and adjust some controls (c1 through c3), make it sound like a thumb-piano, or sound pretty dull and simple ramp+sine-wavey. But the basic algorithm in each of those 20 blocks would have to be changed to get it to sound more like a piano. I'll leave that to an expert.
The sequencer can be kind of fun at times, going through a pattern (basically) 4 times before changing it, and one out of every dozen sounds really cool. :-)
Oh yeah, there's a multi-voice chorus I created recently, even though it's use here is questionable. You might (or might not) like it for use in other circuits somewhere.
I have older versions of all these modules that are much, much simpler. But what's the fun of that? So here you go. You'll need ample CPU to run it, I expect. Or bypass things you don't want in order to reduce that requirement.
--
Keith W. Blackwell
Reply all
Reply to author
Forward
0 new messages