How to test the "Tick" without using sound??

[billv]

Arp Seq Concern.......
Quick question:
Problem is I'm worried about the accuracy of the "tick" signal generated by Arp Seq Green Output.
The midi notes are spot on, but here's the thing.
If i use that tick to test it, using a the old "note on" module, it won't be accurate, I know, because of the method.
So how do I test the "Tick" without using sound...Yeh, that's my question. And without visuals to, I don't want the
test to be compromised by some switch that needs to light up.
I think Nix told me once, but i can't remember.......

[nix]

I reckon use a stream counter.
It is not audible.
So run the counter,
and use mono to float to sample every second.
If the number is the same, ie. 44100,
the tick is sample accurate.
The stream counter is feeding back the number 1,
this will generate a stream that incrementrs at sample rate.
Simply fire the tick by whatever green counter u have.
If need help setting it up,
fire me a schematic.
Cheers mate

[billv]

Can you set this up for me mate? i can't get method right....

Timing sample rate test.fsm

(45.25 KiB) Downloaded 1043 times

[nix]

That 30-150 samples I measured was DreamSequence, with the custom ticker outputting a MIDI tick.
I can knock up that timer if you'd like it?
DreamSequence is pretty fragile(buggy) at that stage.

Here is the test->

test ticktime_0.fsm

(2.14 KiB) Downloaded 1033 times

It seems that it's possible to get sample accuracy?
--with a little mod of the Ruby ticker.
Check it out!

I suspect this will deteriorate when lots of ticks are scheduled.

[billv]

Here is the test->

Thanks heaps for that Nix. Interesting test unit.

I can knock up that timer if you'd like it?

Thanks, but no need.
Your test unit shows an example that's "sample accurate".
I hooked up the test module that shows a drift after 3min in the host, and your test unit
reads it as "sample accurate" as well.
Ran it for a few minutes, still accurate.

This indicates that maybe the test unit is not accurate for the purpose of VST testing.
Don't know for sure.... More tests......

seems that it's possible to get sample accuracy?

It's always been there, it's not the issue mate.
We have to maintain it, over time. Otherwise it's not accurate.
The discovery of drift, hints at a PPQ issue- which makes me glad that i hassled Trog a while back
that a "Guru PPQ" version may be required. It's a relief to know that the FSM's there waiting
if all these experiments turn to s**t.
Here's your test rig, with my timer/midi, and 16 track seq on a green counter.
Sample accurate, just like the User guide says-...but I don't buy it....yet.....
https://secure1.storegate.com/Shares/Home.aspx?ShareID=051b2118-58d1-4fac-8e16-f60fcc534dfc

[billv]

Houston...we have a problem... do you read.....
Created a ruby midi test synth, stripped back to nothing.
Ran it for 3min at 1/4.
First picture at the start, the other at about 3min.
Not the same.

Rubytest_0.png (21.76 KiB) Viewed 21908 times

here's the test synth.

Ruby test_0.fsm

(34.93 KiB) Downloaded 1041 times

Required next step- test for 10/20 min- verify that the drift keeps stretching...
I'm pulling up stumps for the night, and taking a hot bath.
If someone else could verify that my test synth is not faulty,
would help me in making a confident next step in the right direction.

Note Nix:
Want to have a laugh mate
Have a look at this thread, check the date, and read the conversation
http://synthmaker.co.uk/forum/viewtopic.php?f=12&t=11135&hilit=seq+timing

[nix]

I don't think streams drift at all Bill.
A simple test would be to set a kick drum in FL, using FL sequencer,
and set the kick drum with ur FS seq.
Hit play, wait 15 minutes and listen to it.
If there are train tracks,
the fs seq isn't that good.

If you can be bothered u could run it for an hour

[billv]

I don't think streams drift at all Bill.

Ok. I'm not sure the stream test concept is any point here at the moment, cause I'm back to
trying to land one note, on the money, and failing.
here's the latest one, same test synth, this time went to 5 min exact.
First picture at start, the second after 5min

ScreenShot135.png (20.03 KiB) Viewed 21875 times

ScreenShot136.png (20.59 KiB) Viewed 21875 times

it's not accurate.Full stop.
My designs have been proven now, in-accurate, more than 4 times , using 3 different systems.
Sorry mate, But I'm back in the land of 99.9999% blah blah blah.....

[billv]

But I'm back in the land of 99.9999%

Back to 100%. All good.
http://www.dsprobotics.com/support/viewtopic.php?f=3&t=1255