Timing

by **billv** » Wed May 08, 2013 7:26 am

@Nubeat7
RE: ppq_asm_opt.fsm ....from: PositionSyncedStepLFO - help?

Ok, screwed my head back on and took another look at what your doing......

First test, and it's like this at every (song jump) marker...

: ScreenShot217.png (5.84 KiB) Viewed 22563 times

Great work mate

And even better is that the test unit was thrown together really really quick,
the circuit looking back on it, is rough as guts. So, yeh, a good sign for your ticker

by **Drnkhobo** » Thu May 09, 2013 12:32 pm

Hey guys, been following your discussion here. . .

I suppose as an offshoot of this topic, I have been playing around with your PPQ module Trog. While very impressive at the least I still experience lags & a delay of around 1500 samples when using it in my proj. This might be down to my error but i have tested in FL/Reaper/Cubase & get similar results.

What I am trying to do is trigger a mono2graph module on every beat.

Now using Trog's PPQ Test Rig I grab the current beat & timing. . . which then passes a trigger to get a snapshot of the audio (mono2graph) every beat.

Each time its slightly off. Like I said around 1500 samples.

Notice I calculate how many samples to grab determined by the time of each beat. . . (Changing the tempo quickly in your host will cause the plug to crash.....absolutely no idea why!)

Is this delay something to do with the PPQ detection? I modified the Ruby to exclude some midi triggering.

I also get 20-30% CPU spikes in FL!!!

In my fsm I am using PPQ 006. I tried with the latest PPQ file but its not 100% with the timing. . . ?

by **Drnkhobo** » Thu May 09, 2013 5:51 pm

Surely accurate vst synced triggers is something FUNDAMENTAL to VST dev in FS?? :roll:

There are a number of uses for this. . . take a look at how long this thread (and Trogs PPQ one) is already!

C'mon guys. . .

by **billv** » Thu May 09, 2013 8:46 pm

Drnkhobo wrote:This might be down to my error

No, like i said before, they are a little bit out....
Not sure if Trog's still working on them.

: ScreenShot204.png (8.49 KiB) Viewed 22524 times

It's kind of fustrating a bit...Trogs got the PPQ right, I've got the timing right, I just wish
we could combine the two somehow, end this subject for everybody.

Drnkhobo wrote:You can look at my fsm here:

I got issues here....couldn't open FSM.....locked up CPU, force re-start...???????

: ScreenShot219.png (15.02 KiB) Viewed 22524 times

by **Drnkhobo** » Thu May 09, 2013 9:35 pm

Old SM Wave player module creating a problem . . . :mrgreen:

Updated here, you should export to vst to check it out.
Im getting 2 problems in my schm:

1. It doesnt grab the correct amount of samples per beat (usually OVER the calculated "Samples per beat")
2.Timing is off on every beat.

Test Beat.fsm: (7.15 KiB) Downloaded 1325 times

Trogs got the PPQ right, I've got the timing right, I just wish
we could combine the two somehow, end this subject for everybody.

Hell yeah, by getting this precise it opens a lot of doors to us VST devs. Whats the biggest issue here? Is it dodgy hosts (Im looking at you FL!) or is it more on FS's side?

Damn, I wish there was a stock FS module that did this already!

by **billv** » Fri May 10, 2013 9:10 am

Drnkhobo wrote:Updated here

Cheers mate, thanks for that. File opened + saved..all good.
Give me few days to check it out, I'm sick as a dog at the moment and struggling to focus on anything

Drnkhobo wrote:Damn, I wish there was a stock FS module that did this already!

.
There is hope in that.
Trog has already talked to malc about some "missing element", that's sort of preventing Trog
from going all the way, that something is "not right"...
Also if we keep failing in this objective, DSPR might see this, in a way, as "negative publicity"
for FS, and bail us all out in a future update....
Also, Nubeat7 seems to have got one hand on the timing "holy grail', he's just got to stick in a
tempo changer that rings true, and were all home

I'm fortunate my current project(X11) dosn't require PPQ as it's provided by automation.
But if was building a regular type VST synth, I'd be very, very fustrated by now.

We will get there eventually.......it's just a patience thing i suppose....... :roll:

by **Drnkhobo** » Fri May 10, 2013 11:12 am

sick as a dog at the moment and struggling to focus on anything

Not good man, hope you get better soon

Trog has already talked to malc about some "missing element", that's sort of preventing Trog
from going all the way, that something is "not right"...

hmmm. . . . :roll:

I think there are quite a few things here & there that DSPR can work on. It would certainly keep me busy for a while!

But this issue needs to be sorted asap. Lucky for you its not needed in X11, for me its ESSENTIAL! Although im using it for a slightly different reason (not to trig midi events) its still very important.

Im glad Trog has been in contact with Malc & they know about this "missing element".

by **billv** » Fri May 10, 2013 7:35 pm

Drnkhobo wrote:Updated here, you should export to vst to check it out.
Im getting 2 problems in my schm:
1. It doesnt grab the correct amount of samples per beat (usually OVER the calculated "Samples per beat")
2.Timing is off on every beat.

Yeh, it's way out.
With Trogs PPQ versions, I tried lots of different "cowboy" attempts to re-work them, but never could
get a result on my main CPU. On my laptop, for example, PPQ version 6, is absolutly perfect.
It's really "tear your hair out" type of stuff-damn thing is perfect but dosn't perform on my main CPU

This clearly says to me that it's not Dodgy Hosts, but the type of CPU systems we run.
So Tog's work here, in my view, is failing to address some aspect of hardware behaviour.
Know what i mean,
..... if I'm using the same host, and the same plug-in, and then getting different results using
different CPU's,......it just seems the only answer is that the Plug in is not addressing the hardware
protocols properly....
I don't know......all this is Guru stuff......i can only see up to a certain point......

by **Drnkhobo** » Fri May 10, 2013 7:56 pm

..... if I'm using the same host, and the same plug-in, and then getting different results using
different CPU's,......it just seems the only answer is that the Plug in is not addressing the hardware
protocols properly....

I get you, I think you are on the money here!

Just so im at the same place as you guys, why is this not a simple thing? Why is the PPQ module not 100% accurate? Its vital to syncing VST plugs!

See things like this make me angry. Take for instance the sine module, or saw, or delay module. . . they are integral just like PPQ! Man I could go on all night. I suppose I can answer myself with your previous posts about the "missing element"

I don't know......all this is Guru stuff......i can only see up to a certain point......

bro, you are further down the rabbit hole than I am!!

by **Nubeat7** » Fri May 10, 2013 8:09 pm

hallo, ok i experimented a bit more and i found the perfect solution now to get stepposition and ticks in green, i`was using trogs ppq version 7 and modified it a bit just to read the incoming monosignal (ppq or raamp) for thisi`m using my ppq stepposition (asm optimized) this schematic is perfect for using steplfos to read from an array and when you need the array data in green or when you need a variable tick for quantizing a keybord (works perfectly ) i also was modifiing trogs ppq version 10 alone without the mono ppq which worked well too but the performance was not as good as this solution, the good thing is there are ways which are working to get accurate green ppq triggers and intchanges

Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Re: Timing

Who is online