DSP Robotics Support

My Flowstone crashes kinda often. I bet it's because my obscure wiring is far from ok, though it seems kinda random. Now finally i have a case where i'm able to reproduce the meltdown

The attached schematic is my very first go at a simple Delay module. The module works, producing delay according to parameters and chosen delay type. To select the delay type i used a Selector dropdown module, and this is the heart of my booboo If i change the delay type a few times in a row (using the mouse clicking the dropdown) my Flowstone crashes. Sometimes it takes a good handful of changes, sometimes only a few. It happens even if there is no sound going through the delay.

Remeber that i am very new to Flowstone and you may find my schem to be kinda childish So, besides kicking me in the right direction towards a solution to my Selector problem, i am also very open to general input about good and bad practice in Flowstone.

Cheers

hmm, thats a good question, your schematic should be fine, it seems that this is related to this bug

viewtopic.php?f=2&t=2674&start=40

this looks like that the bug is related to selectors and asio, because the schematic works if you choose not an asio audiodriver!

i tried to just put 2 selectors in series and the same happens!

edit: it seems that it also happens with generic lowlatency driver but not that often!

Aw, bugger anyways, thanks for checking my schem out. I'm glad it looks normal.

Too bad about the bug, but it seems from the other thread that it already has been reported.. og should i do something active to let the grownups know that we found a 'dead squirel in the sandbox'??

Thank you very much for your feedback, Nubeat7.

I actually like this forum a lot. I'm proud to have become a 'stoner'

You may use code block to "mimic" a selector/multiplexer. Usually Selectors also bypass the sections that are turned off (So the don't process = don't waste CPU), but in this particular schematic it makes no difference, so there shouldn't be any significant CPU waste at all.

Flemming wrote:Aw, bugger anyways, thanks for checking my schem out. I'm glad it looks normal.

Too bad about the bug, but it seems from the other thread that it already has been reported.. og should i do something active to let the grownups know that we found a 'dead squirel in the sandbox'??

Thank you very much for your feedback, Nubeat7.

I actually like this forum a lot. I'm proud to have become a 'stoner'

I would still pass it on, just email support.

Exo wrote:I would still pass it on, just email support.

You're right. I sent them an email with links to here and the other thread.

KG_is_back wrote:You may use code block to "mimic" a selector/multiplexer. Usually Selectors also bypass the sections that are turned off (So the don't process = don't waste CPU), but in this particular schematic it makes no difference, so there shouldn't be any significant CPU waste at all.

Code: Select all

//Multiplexer
streamin Select;
streamin in;
streamout out0;
streamout out1;
streamout out2;

out0=in&(Select==0);
out1=in&(Select==1);
out2=in&(Select==2);

I copied your code and used it to replace the Multiplexer primitives. Then i started clicking that Selector dropdown like i was all out of coffee - no crashing yet. I'll have to catch some sleep now, but i'll make sure to click some more tomorrow..... just to be sure

Thank you very much for this solution!

Nubeat7 wrote:i tried to just put 2 selectors in series and the same happens!

edit: it seems that it also happens with generic lowlatency driver but not that often!

Your Selector example crashes here for me also. I play a few notes and then click the Selector a few times, i get a crash.

So i guess that both the Multiplex and Selector primitives are affected by this bug

I tried both Flemming's and Nubeat7's example schematic and they work completely flawlessly in Flowstone 3.0.2 Firefly free version - never had a single crash even after several dozen tries. So the bug must've been introduced after that update.