[LAD] Tests directly routing pc's midi-in to midi-out

[Ralf Mardorf]

On Thu, 2010-07-15 at 11:48 +0200, Arnout Engelen wrote:
> > Who knows? Perhaps 2ms shown by Audacity are 20ms, because Audacity has
> > a bug? 
> 
> Sounds unlikely

Indeed.

> > Perhaps 2ms shown by Audacity are 2ms, but when playing JACK, the
> > sound card or what ever will add additional jitter?
> 
> You claimed the latency/timing was obviously noticably broken without needing
> 'golden ears' even without JACK, right?

Correct.

> Let's first try to get this reliable without JACK, just ALSA

So I'll do

- instead of dev.hpet.max-user-freq=64 I'll try 1024 or 2048 as Robin
mentioned

- connect the DX7 port to the D4 port with aconnect(gui) as Clemens
recommended

If nobody should demand to it, I won't use

- aplaymidi, because there might be a problem. All instruments at the
same time by the machine is missing the reference of recorded
instruments that might have had another fluctuation for the timing or
the alternative, it's missing the feeling of a musician playing live

How should I do

- I should be able to get my Atari ST working. It's MIDI is good enough
for my needs. So, how should I use the Atari ST to make a
'timestamp'-test as somebody recommended?
The Atari e.g. could play 4 on the floor at 120 BPM. I could watch the
received events by a Linux MIDI monitor or record it by a Linux MIDI
sequencer. But, should I sync by MTC or anything else etc.?

Cheers!

Ralf