FlowStone 3.0.9 Beta1

[MyCo]

well, you can't connect streams to double streams or vice versa. You'll have to use the converter primitives. And then your assembler code has to get updated as well to use instructions for double precission.

[iman]

Can't test this and it's the first time I heard about that.

Very annoying bug... This is related to "Midi to Voice" primitive.

if active voices (pressed keys) become more than the maximum number of the voice and you use sustain pedal in this situation that issue will reveal.

For example set the maximum number in "M to V" primitive to 3, then play a chord with 4 notes (in this case one voice must be stolen , ok no problem) at the same time if you activate/press the sustain pedal and release the keys then deactivate/release the sustain pedal, You expect muting all notes... but that stolen notes is wrongly played again and stuck ...

Please fix this issue..

[martinvicanek]

- I did some testing with floatarrayin and memrefin and unfortunately I am still getting random crashes.
- Sometimes when I connect a text prim to a code box to reveal the code FS crashes on me
- Sometimes the analyzer connections would turn red, sometimes followed by a crash
Nice implementations of abs, sign, int and modulo %!

[MyCo]

For example set the maximum number in "M to V" primitive to 3, then play a chord with 4 notes (in this case one voice must be stolen , ok no problem) at the same time if you activate/press the sustain pedal and release the keys then deactivate/release the sustain pedal, You expect muting all notes... but that stolen notes is wrongly played again and stuck ...

That's interresting. I might be able to track that back even though I can't test it (don't have sustain pedal). So just to make sure I understand it correctly:
1. set max voices to 3
2. play 4 notes (one of them isn't played)
3. hold susain
4. release all notes
5. release sustain -> one note starts playing?

- Sometimes when I connect a text prim to a code box to reveal the code FS crashes on me
- Sometimes the analyzer connections would turn red, sometimes followed by a crash

Yeah, those random crashs are a pain in the ass. It's nearly impossible to track them down, they might happen because some old ASM code in a primitive became incompatible with the new assembler and manipulates something that somewhere later (even outside of the streams) can cause a crash. So it's most likely a primitve (one that has quite a big chunk of memory accessing code in it) that you have somewhere in your schematic.

Nice implementations of abs, sign, int and modulo %!

Thanks

[iman]

I might be able to track that back even though I can't test it (don't have sustain pedal).

No need for sustain pedal, you can emulate it using MIDI CC#64:

capture2.png (29.78 KiB) Viewed 20067 times

So just to make sure I understand it correctly:
1. set max voices to 3
2. play 4 notes (one of them isn't played)
3. hold susain
4. release all notes
5. release sustain -> one note starts playing?

Exactly but it does not happen all the times! I don't know how to say it is somehow depends on how you release the hold notes from keyboard (in ascending or descending order). It is totally random you should check out different combinations,, play several notes and release them in different orders and see what will be going on.

Here is a test schematic:

MIDI-Sustain bug.fsm

(20.81 KiB) Downloaded 897 times

[Spogg]

Exactly but it does not happen all the times! I don't know how to say it is somehow depends on how you release the hold notes from keyboard (in ascending or descending order). It is totally random you should check out different combinations,, play several notes and release them in different orders and see what will be going on.

I'm very glad that this might be resolved. Tulamide told me that it was somehow inevitable, which was a puzzle to me.
I agree with iman that it's not so predicatble, like it's a timing issue maybe. For me this is a serious bug because my playing style is that of a (very bad) pianist, so I use the sustain pedal frequently. I found that it's much less likely to happen if you set a large number for max voices but on my high CPU projects this is not always a good idea. Consequently I provide a "Panic" button which kills the audio.
This would really bug me if I was making high quality release-grade plugins but since it's a hobby I live with it.
I've tested numerous VSTi releases not made with FS and never encountered this issue. In fact, before FS I made a basic synth with SynthEdit and the MIDI worked fine.
If just this one issue could be resolved you would have a very happy Spogg on your hands!

Cheers

Spogg

[MyCo]

Thanks for the test schematic. I found a pattern that always causes the problem now (by just using PC keyboard as midi input):
1. click pedal on
2. play the keys Q,W,E,R one after another
3. click pedal off

Result is, the note of Q-key gets stuck.

We should be able to fix that. Unfortunately I'm not really familiar with Midi/Voice implementation in FS right now, might be something for Malc. I'll ask him.

[iman]

Yes. I got the same result using PC keyboard. Cannot wait to see fix for it...

I suggest to create a dedicated section in forum for bug reports. And I expect the developers to check and response us actively like old SM forum. I don't really want to hijack this thread the sustain issue is an old bug it is not generated in 3.0.9 Beta1. maybe you expect reporting/testing new features here ...

[adamszabo]

Bug: when you hover your mouse over the poly to mono, in the popup info it says ...into onesingle channel signal. there is a space missing from one and signal. Its the only bug I found

[Walter Sommerfeld]

Warning:

Even an empty Modul created with v309b1 dragged from the toolbox let FS v306 to v3081 crashes