[PD] [polywavesynth] and [polygrainsynth] bug fix

[Phil Stone]

hard off wrote:
> hi again....some more comments
>
> 1) at first i also tried changing the ADSR sliders to a log setting, 
> and totally understand what you mean about the low values and 
> clicking.  my workaround was just to change the range from 1-5000 to 
> 5-5000.  however, i think an even better solution is not the log 
> function, but an exponential one.
>
> if you change the sliders to 0-1 range and then do:
>
> [r slider-value]
> |
> [pow 2]   (or 3)
> |
> [* 5000]
> |
> [+ 1]
>
> then that gives the most user-friendly sliders in my opinion. 

Be aware that that breaks the way state is currently saved for the ADSR; 
ADR values are stored as milliseconds, not as 0-1.  That could be 
changed, but it would break all existing patches (I'm probably the only 
one for whom that causes problems -- still I try not to do it too 
often).  The "log" setting is a pretty good compromise; thanks again for 
the idea.

> 2) i understand that you can change the pan position and speed, but 
> unless i'm mistaken, there is no way just to keep everything centre 
> panned. 

Just change the radio button from "rnd" (for random panning) to "fix" 
(for "fixed" position).  Then, move the "position" slider to the center 
(or move it live for manual panning).  Sorry that the controls are 
labeled so cryptically, but real estate is at a  premium, and the web 
page documentation spells out the details.

> i think a pan width slider would be helpful.  however, maybe that's 
> just something i want and wouldn't be that important for anyone else, 
> so i might just put that on my own modified version.

Yeah, I thought of some other panning algorithms and enhancements 
besides random panning, too, but have punted this off to OSC control.  
You can do just about anything you can conceive of through the OSC 
input, as you are completely in control of what numbers you pass in.  
Most of my latest work with these two synths involves hooking up LFOs 
and other data generators to various controls through the OSC input.

That said, I *do* plan to balance the panning a little better at some 
point.  It's a simple-minded pan right now, not a balanced-power ratio, 
so there's a "center hole" effect.

> 3) not sure what was happening with the gain problem.  perhaps svf~ 
> puts out some gnarly peaks that are well above the -1 1 threshold, and 
> that was making things sound weird when i changed volume.  the problem 
> seems fine now, because i deleted the gain section from the 
> polywavevoice~ abtraction and put the gain directly onto the ouptut of 
> polywavesynth~.  

Well, I'm glad you're not shy about customizing, but that change wipes 
out the possibility of phased-in preset changes (as when the "global" 
toggle is off).  I've gone back and forth on this during the evolution 
of the instrument, and am really liking the current set up which allows 
local-per-voice gain changes, over-rideable with the "global" control, 
which causes all voices to be controlled instantly by the GUI and OSC.

> the gain slider also had a really unusable range, so i also adjusted 
> that into a more manageable range.

It is greater than unity gain above about 70% or so.  This was done to 
match up with [polygrainsynth], which sometimes needs the extra gain.  
When you have a "rack" of these, with the two types mixed together, it's 
good to have the gain slider-range match, but otherwise, you're right 
that the gain slider could be better optimized for [polywavesynth]'s 
purposes.

> i will definitely use this synth a lot from now on.  today i was 
> running it with an SH101 emulator i am making, and a modified 
> sequenced version of frank's vosim~ abstraction, and some drum synths, 
> and it was all sounding pretty sweet.  really cool to be able to get 
> some polyphony happening and do chords and stuff! 

That's really great to hear, and I look forward to hearing some music 
made with it.  Check out [polygrainsynth] if you get a chance.

Thanks a million for the detailed feedback,

Phil
http://www.pkstonemusic.com