home brew sequencer

[billv]

So the old system with a green integer works,
but nulls any accuracy advantages that could be gained

Sounds like you've come full circle-Quote:"I would really like a tick counter in Ruby".
Welcome back to the start-
I'm still here-
havn't gone anywhere-
still can't count in ruby-
still no help

[nix]

I had to build the trigger counter before I got anywhere.
It's hard to get an integer out of time.

I'm trying to mesh the codes now,
I got the array reader code to send MIDI.

baby steps

[pall]

Hi boys!

I wonder, what influence has a green counter on the Ruby ticker accuracy?
Can someone answer to this question? Because I'm not sure yet on this issue.
However in my opinion, a green counter should not be influence a Ruby ticker.
Hello support!

[nix]

Heya Pall!
In my opinion-
Once u have ASIO running the custom ticker will be quite accurate.
The actual Ruby interval should be on the same amount of samples.
So your tick is not bang on the mark because of the green components,
but it will not drift further out in time.
The core of the timer is accurate.
hmm
You could devise a test on rendered audio to see
I have done this with the custom timer and green components,
finding 30 to 150 samples error, it's not the same as ur screeny though

[pall]

tick is not bang on the mark because of the green components,but it will not drift further out in time

Aha...
Thanks Nick!
I will try to do some tests.
Would be nice if someone would make a counter in Ruby.

Cheers!

[billv]

Hi Pall, good to see you here.

a green counter should not be influence a Ruby ticker.

Thats what I'm getting in all my tests. It's bang on the mark and no drift. Everytime.
The pictures of accuracy in hosts are great, but we are not precision watchmakers,
so yeh, it would be nice if Support or someone clarified this.

I have done this with the custom timer and green components,
finding 30 to 150 samples error, it's not the same as ur screeny though

Nix, can you upload this "test fsm" so I can verify it, and see the error?

I can't help feeling we are maybe "flogging a dead horse" here.
But if Nix thinks there's a error, I'm happy to dig it all up for another Autopsy.
The Logic that we need to do the count in ruby makes sense.

[billv]

I wonder, what influence has a green counter on the Ruby ticker accuracy?

Might not have to worry about this, we may have bigger problems.
Did a timing test on one of my seqs and I detected drift after 3min.
Ok. Looks like Nix maybe right.....So I back to the source, to verify that I have created the error.
I cut another vst- this time no counters, one note made/sent in ruby-
the drift is still there .
Notice the first line how it lines up spot on...

rubystart.png (7.91 KiB) Viewed 16687 times

And then after 3 min, it's not the same...

ruby3min.png (6.75 KiB) Viewed 16687 times

So Nix, at the moment, looks like you were right about an error, but maybe wrong about the place.
No ticker, No counter. Just ruby

So, what good is sample accuracy if we can't maintain it?

Don't know, but I'm back to square one. More tests.
I'm hoping I've made some error with the test vst....

[pall]

Yes, I see the difference between the two pictures, but you could measure the time difference between these two moments? If so, how is it?

[billv]

measure the time difference

Don't know how at the moment. can't get the s/r test happening, that Nix gave me a tip on
how to do. Not sure how to work it, Will upload it for Nix to fix the method, and maybe give
you a better answer.

difference between these two moments

Pretty safe to say that you won't hear it with your ears.
For general VST stuff, won't be an issue for most people

Still early yet. More tests when i get home in about 12 hours.

The ruby code i used had a bit of stuff in it that wasn't being used in the test.
It's a pipe dream, but I'm not hoping that the extra code is the cause.
Will strip it back to nothing and test again.
Also need to re-test the Arp method again and see if the same drift is there.
After that will try Trogs "Guru" version again, and see if he's getting the same drift after 3 min.

[nix]

I upped the sample stream counter in ur thread.
omg! green accuracy is obtainable in a simple schematic,
think not with larger ones,
coz DS was 30-150 samples out.

I won't up my pianoroll,
I think I'll hang on to it.
It's really hard to integrate my changed module into my reader/MIDI out code.

But I thought use might like 'changed' in a single code block->

Ruby changed.fsm

(1.3 KiB) Downloaded 1039 times

Cheers!