Julien Bayle

https://cycling74.com/forums/1-js-parsing-8-data-sources-vs-8-js-parsing-each-1-data-source-which-one-win,

My "time critical" stuff is "just" a m4l machine taking a sync (Ableton Live transport) and which fires midi notes.

I "feel" my current JS lagging a very bit (but a bit...)

I sampled the result, aligning it on a grid.. it is a bit perceptible.

I choose JS because it was the less complex method for me ; I mean, it simplifies a lot the patch architecture.

Would a correctly done patch would be faster ?

I mean, maybe JS is: JS >> js wrapper to max >> max

hi there,
considering this https://cycling74.com/forums/1-js-parsing-8-data-sources-vs-8-js-parsing-each-1-data-source-which-one-win, I'm still hesitating...

My "time critical" stuff is "just" a m4l machine taking a sync (Ableton Live transport) and which fires midi notes.
It is a kind of sequencer.

I need a good accuracy.
I "feel" my current JS lagging a very bit (but a bit...)
I sampled the result, aligning it on a grid.. it is a bit perceptible.

big-hesitation-between-js-vs-patch-for-time-critical-stuff

http://soundcloud.com/protofuse/demo-of-the-uncanny-sequencer

I guess I would do that instead of making a week of tests :p

yep.
http://soundcloud.com/protofuse/demo-of-the-uncanny-sequencer is the demo I made with it.

port JS to C won't be hard..
I guess I would do that instead of making a week of tests :p


Yes, c would be fastest, but Java is much faster, for some things as fast as C,

and the syntax is also very JS like (all of them "C-like", I know) , and

almost the exact same set of functional wrappers on max msp are implemented in the Java scripting: it would not be as much a 'rewrite' as a simple translation of syntax, and might be faster to code the 'port' , and not as perceptibly 'sluggish': I have had a java and a c port of my 'reactive/synchronized/sequencer/arpeggiator' extension for a while. The Java has not seemed to be any different in performance than the C.

But since the scheduler is the scheduler , any difference would be due to the 'handover' twixt the original language calls and the max world, and i strongly suspect the clear order of speed of calling here is

since i mentioned said extension ('iso' -first c version was available in the orig Ircam public max object lib), I will point at some sound i did where most of the notes were generated with said code: i add only that all of my recent works are remastered recordings of live improvisations, since that's what i have time for, i guess. this work uses the java version.

http://soundcloud.com/josquin2000/shinencrackle

not my fav recent work, but a good example of how I waste(d) my time with two keyboards,a fader box, some key weights, 4 vsts, and max5.

Yes, c would be fastest, but Java is much faster, for some things as fast as C,
and the syntax is also very JS like (all of them "C-like", I know) , and
almost the exact same set of functional wrappers on max msp are implemented in the Java scripting: it would not be as much a 'rewrite' as a simple translation of syntax, and might be faster to code the 'port' , and not as perceptibly 'sluggish': I have had a java and a c port of my 'reactive/synchronized/sequencer/arpeggiator' extension for a while. The Java has not seemed to be any different in performance than the C. 

But since the scheduler is the scheduler , any difference would be due to the 'handover' twixt the original language calls and the max world, and i strongly suspect the clear order of speed of calling here is
a) C
x) Java
y) Javascript

M. JBayle Cool stuff!...
 since i mentioned said extension ('iso' -first c version was available in the orig Ircam public max object lib), I will point at some sound i did where most of the notes were generated with said code: i add only that all  of my recent works are remastered recordings of live improvisations, since that's what i have time for, i guess. this work uses the java version.
http://soundcloud.com/josquin2000/shinencrackle
not my fav recent work, but a good example of how I waste(d) my time with two keyboards,a fader box, some key weights, 4 vsts, and max5.


Charles, thanks for your opinion/feelings about that.

I'd really need a way to use standalone machines in that DAW centric world (especially because some users want to buy my sequencers but to use them with their daw)

The only answer of cycling is: "if you want it to work fine, use max for live / Ableton" ; I can really deal with that (btw, I'm an Ableton Certified Trainer :p) BUT I'd need more.

I'd really like to have your opinion, Charles, about that too.

Checking your soundcloud stuff right now!

Charles, thanks for your opinion/feelings about that.
I'd really need a way to use standalone machines in that DAW centric world (especially because some users want to buy my sequencers but to use them with their daw)
The only answer of cycling is: "if you want it to work fine, use max for live / Ableton" ; I can really deal with that (btw, I'm an Ableton Certified Trainer :p) BUT I'd need more.

Checking your soundcloud stuff right now!


BIG hesitation between JS vs patch for time critical stuff