On Sat, 2010-07-03 at 22:53 -0600, gus...@echostar.ca wrote:
> ### First the RME results ###
> $ alsa-midi-latency-test -i 16:0 -o 16:0 -w 20 -r
> > SUCCESS
>
> best latency was 0.96 ms
> worst latency was 4.04 ms, which is great.
> ### Next are the midisport results ###
> $ alsa-midi-latency-test -i 28:1 -o 28:1 -w 20 -r
> > FAIL
>
> best latency was 2.95 ms
> worst latency was 34.96 ms, which is too much. Please check
Theoretically it could be, that the order of the tests had influenced
the tests. I will run my tests 3 or 4 times and in addition I will run
glxgears or a real benchmark test while running the ALSA MIDI latency
test, e.g.
"Real Benchmarks
All in one
* Phoronix Test Suite offers Unigine, Gtkperf, and a ton of other
tests all in one - [1]
OpenGL
* Unigine Sanctuary & Tropics Demos - [2]
GTK
* Gtkperf - offered by any major distribution's
repository."
(http://wiki.cchtml.com/index.php/Glxgears_is_not_a_Benchmark)
When we make music, we not only send and receive by the MIDI port while
just a little program is running. We also will have a sequencer GUI,
access to the hard disk drives, perhaps we run JAMin, it stress the
computer because of math.
OTOH I bet the the RME always would be ok, while the Midisport always
would fail, anyway the results might vary in value.
A drummer forum on German
http://www.drummerforum.de/forum/46746-midi-vorteile-bei-audiointerface-oder-pci-soundkarte-gegen%C3%BCber-midi-usb-interface.html
... one asked if the other one noticed the difference between audio latency
and measured MIDI latency, while nobody asked for jitter. Anyway, most people
seems to be non-gifted musicians and they don't hear jitter, while musicians
seems to tend to say that PCI devices are more stable.
I guess it's the industry making propaganda for USB, when we read USB
should be the better choice, while musicians seem to experience that PCI
is the better choice, anyway, even PCI might not be good enough.
Cheers!
Ralf
_______________________________________________
64studio-devel mailing list
64studio-devel@lists.64studio.com
http://lists.64studio.com/mailman/listinfo/64studio-devel
