[PD] midi latency test

[PT147 at mdx.ac.uk]

> I can't just take a MIDI cable and connect the input to the output!
> 
> So jmf3 and PT147,  are you really talking about hardware MIDI, and if so, I 
> take it you both have a MIDI interface card other than the soundblaster 
> adaptor?  If not, the how did you do it? 

I'm using an old school MIDI to Gameport interface (originally from a
Gravis Ultrasound), to my SantaCruz, under windows XP.  So it is going
directly, through a real midi cable, but no synths, from the out to the
in.

> If you _are_ really talking about MIDI, and not software emulation of it 
> within the PC, then even so, 35 to 40 ms seems a bit long (these are ms 
> values, right?).
> 
> PT147, you say that it's going to be more than 10. If this is milliseconds, 
> I'm not so sure why that would be normal, because a complete MIDI message 
> should go through the interface in about 1 ms. So any delay much greater than 
> this should be due to the operating system and it's drivers. Am I missing 
> something? Unoftunately, neither PT147 nor jmf3 said what OS they are running 
> on. I have the feeling that this is where the problem lies. jmf3, what do you 
> mean by "very high"?

It's going to be more than ten because of the 10 ms delay between the
timer being started and the message being sent.  I don't really know
what the latency should be, although I guess one could reasonably assume
it to be less than 30 ms.  It isn't bothering me too much at the minute,
anyway - if I connect notein and noteout objects directly to an outboard
synth (exciting), the response is OK for me.  Maybe that's because I'm
used to Reaktor, which I could run with ~10/15ms latency with the
original version and setup I had, which seems to have grown to 30-35. 
I'd be interested to hear what jmf3 means by very high, too.

> 
> Regards
> 
> Larry Troxler
> 
Cheers,

Peter Todd