DSP Robotics Support

Hi,
Over at the SynthMaker forum we have uncovered a little problem with changes in green data values occasionally being missed when fed into a stream input. Very likely that this will affect FlowStone schematics too.
You can find the details HERE - the demo file there should load OK into FlowStone if you select the 'All Files' filter in the load dialogue.

PS) Notice there is no dedicated 'Bugs' forum over here - seems like rather an oversight, the one over at SM has been very useful over the years for keeping tabs on known issues.

Just tested in FSP and SM and I get no errors?

Judging from my own results - approx.2% error on an Atom powered system - I would be very surprised to see any errors on a decent DualCore (or better) system when tested in isolation.
So, as I suggested in my SM forum post, a mostly obscure 'theoretical' problem on a decent system run at a reasonable load - especially as a skipped parameter change will most often not be critical (the most recent set-point does always arrive after a little latency).

The issue is heavily dependent on CPU load - I'm waiting to hear back from the developers, but it seems related to the way that 'Green' parts of a schematic are heavily dependent on Windows resources, particularly timers.
A similar issue affects the Tick25 and Tick100 components -the accuracy and jitter in their timing also increases as CPU load rises (tick100 can sometimes be less than half that rate on a stressed system) - the timing routines just don't get serviced often enough if the CPU has other, more urgent, threads to process.

One reason for reporting what might seem such a minor problem is that over at the SM forum, we are very keen to keep up the pressure on the developers to remove the reliance on Windows systems for 'green' timing - it can often make the timing of MIDI events too unreliable to be useful musically, and makes 'sequencer' type schematics difficult to implement without a good knowledge of code/assembly (to be able to use the much greater timing accuracy of sample-rate streams).

Unless, of course, the problem is SM specific - though I would be very surprised if that were the case (I may have to try it out with FlowStone FREE, just out of curiosity!).