DSP Robotics Support

billv wrote:"silly" idea of bringing the Ruby accuracy down to green

Wasn't so silly after all. i managed to do it using the Arp Module.

Here it is, 100% Accurate,
NOTES CREATED AND MODULATED IN GREEN
NOTES THEN GENERATED BY RUBY, along with TICK % PPQ.
Sky is the limit now. Pump whatever notes you want in on the fly.

Note: This is the first prototype. Still some work to be done. Will be updated.
Thanks Trog. Maybe in a few days i can finish the "Seq" section. Then address the synth timing.

Thanks Nix: i pulled this off using modifiyng YOUR "on/off Ruby switch".
viewtopic.php?f=2&t=1260
You say,

nix wrote:DreamSequence is actually seriousy broken.

We have the tools now to "seriously Fix it.

Here's an updated version, that uses the "Modded Custom ticker" for the tick.
Tick is locked to Host time.
This Seq is Easy to modify. Get rid of wave player, and just feed your green notes in.
Tick/PPq and Clock value already there.

billv wrote:Here's an updated version, that uses the "Modded Custom ticker" for the tick.

So does that mean the last version was 100% accurate, & this one is 100.00% accurate?

infuzion wrote:So does that mean the last version was 100% accurate, & this one is 100.00% accurate?

Ha ha
No, it just means that

billv wrote:Here's an updated version, that uses the "Modded Custom ticker" for the tick.
Tick is locked to Host time.

The tick from the arp gets modulated by tempo changes. Different system.
Also, it's an example to show how the "Stock custom ticker" with that little "Output 0,t" mod, is 100%.
And i thought it would be a better example for Nix, 'cause he's working on the "stop/start" issue as well.
But

infuzion wrote:So does that mean the last version was 100% accurate, & this one is 100.00% accurate?

I know your just winding me up.
Nice to hear from you mate.

I'm just about done for the weekend. got to crash.
Still a couple of weeks needed to finish X11 update.
X11 re-build running great as VST and standalone
Rebuilt 90% of the whole timing system.
All 6 Seqs are 100%... Ruby Rocks!

billv wrote:This Seq is Easy to modify

That set-up is not easy to modify.
Spent two days tearing my hair out trying to modify it, and got failure after failure.
It might be accurate, but it's as dodgy as hell, design wise.
The main problem is the "@note" midi generator. It's wrong code/method for this type of circuit.
I have since come up with the best solution, will post in a day or two after more tests.
in the meantime, this is the best example of the "Arp module, dodgy seq, backdoor method",
that I came up with. You can see clearly whats going on, as there's 16 notes on a selector.
Traditional type that should make sense to everybody. Accurate as hell, just hard to modify.
It is locked to certain behaviour. There is a good side to this. I will be adding this exact Seq
to the X11 as it has a certain behaviour, that some users might prefer.

Behaviour definition.
16 Notes Seq
If 1/4 is selected, it will not play the first 4 notes, but
note-off-off-off-note-off-off-off-note-off-off-off-note-off-off-off

Sorry guys, but all of my designs have been proven in-accurate.
They start drifting after a few minutes and continue to drift with time.
before and after 5 min
Look at the black line(host marker) in the background

And look again


So the timing issue is not dead- the dead horse is not only alive, but has bolted
from the scene.
So the headline "Ruby delivers ", for me, is FALSE.
Ruby has not delivered. Why?...
Because it is up to the "programmer" now, who can make ruby VST sample Accurate.
Yeh, so for me, ruby has lost its grip on the"holy grail" of timing accuracy.
The Ghost of the Holy Grail was last seen loitering around this man...
viewtopic.php?f=3&t=1267

Ok. So I've gone back to initial theory that PPQ is required.
(Still having issues in FL with Trogs set-up, so.....)
Did a quick version, where the PPQ when changed triggers the note s&hold into the midi/timer
: still getting drift at about 5min

Poking around my toolbox, i found this old attempt made in SM that's interesting
There's two modules. One is a 99% carbon copy of the FLSM Clock, that has been
redone to take a blue input from the other module, which was lifted from something
that Trog made(can't remember where) .
It's a coded version with PPQ/Bar Start,Sample rate...ect, all the good stuff...
Still drifting at about 5 min

Please change the topic title too.

Tronic wrote:Please change the topic title too.

Did that, but it looks like a false alarm.

Got it with my first attempt for the weekend.
It's just a "tiny fix" to the standard design I've been using.

Trog's work in his VST sample Accuracy post pointed the way.

here's the first test result:
At start

At about 5min


Give me a day or two, to test in other hosts and verify.
Will upload then.
Sorry if i gave anyone a heartattack. thanks for your patience.

Not sure weather to blame ruby now, or blame trog for the success, as the "tiny fix"
is sort of halftrog/halfruby.