DSP Robotics Support

trogluddite wrote:something different about the engine of the FL sequencer

Something different alright.
Trog, I think I have stumbled across a breakthrough concerning FL behaviour.

I have set up neat test enviroment now, having installed FL 11 on crappy old laptop as well.

Subject: PPQ v6
Trog I've done some tests in a better "set" enviroment, and got back some wicked results....
Tests were done on both CPU's, running FL 11.
Firstly, the Laptop.
PPQ v6, is pretty much perfect for timing, landing like this almost everytime.
When I say almost, it's because every other time it's "perfect".

Now have a look at the same plug in, same host, but different machine, my main CPU
Sometimes it's on the money, and lots of times, it's like this

CONTINUED IN NEXT POST.........

CONTINUED FROM PREVIOUS POST

Now here are the setting's on my main CPU...
(Bear in mind here it's not my selections I want to highlight,
but the options that are available and how they differ)

Now Look at the settings on the laptop.

Clearly there not the same.
Can you use this information Trog, the technicalities are beyond me, but I hope it
triggers something in that monster brain of yours
CONTINUED IN NEXT POST

Now the problem I mentioned with PPQv6, was a slight error in the PPQ, or
sample rate maybe....don't know for sure...
To test, while running I toggled tempo's 1/4,1/2,whole,1/8 and not 1/3.
Error is like this,


Otherwise, on the laptop, v6 it's just perfect.
Note: v10 could be too, havn't tried it on laptop.
Hope this helps.

Thanks a lot, billy - will take a little while to digest, but I think that will be very useful information.

That second wave screenshot a few posts back is particularly intriguing - it is so far early in time that it looks as if there may well be PPQ delay compensation going on. I would guess that it isn't early for every single beat because the <= v6 code is assuming that there isn't any compensation - so when there is, it all goes a bit screwy!

The glitch when you switch time divisions is one of the things that I need to sort out - that bug happens in other hosts too - but I really wanted to get everything working the same in every host first so that I only have one "code base" to work out a fix for.

trogluddite wrote:I think that will be very useful information.

Thats' what I was hoping. Was good to get a result in FL. But the way I got that result highlight's
what a incredidbly tough assignment you have here, negotiating around different host's
behaviour, and now, it seems, different systems as well..!!!!....nasty job......