[ubuntu-studio-devel] Elementary OS

[Ralf Mardorf]

On Wed, 2 Sep 2015 10:09:12 +0200, ttoine wrote:
>Ralph,
>I agree with Len, you don't need a pcie sound card, pci is fast enough
>for audio (actually, usb2 is fast enough...) RME did new pcie version
>of their card likely for the mac pro. And the RME driver has been
>developed by an independant guy, not by RME officially. This guy may
>not have the time anymore to work on the driver. Or he still have a
>pci RME soundcard, and he might find that this is good enough.
>About midi jitter with usb sound cards, I don't know. I use an Akai
>pmd18 without any issue. If there are some tests I can do to check,
>let me know. It might also depend on the midi chipset, no ?

I was in contact with RME and with the Linux guy, I'm aware about all
this. You missed the point. At that time Linux developers recommended
the PCIe RME card. With one thing they were right, it nowadays is not
easy to get mobos with PCI slots, so regarding this PCIe was the right
choice, assumed there should be the need to replace my mobo. However,
they were mistaken regarding the support, even the Linux driver
developer is mistaken regarding the quality of his work.

My point is, that it's even nearly impossible for an experienced user
like me to buy Linux compatible hardware. I often got other hardware
that was mentioned as Linux compatible, but some revisions were not and
other hardware didn't work in combination with some other hardware.

Windows and Mac user seldom run into such issues, while a lot of
inexperienced Linux users can't avoid it.

The best test regarding MIDI jitter is listening, another test is

https://github.com/koppi/alsa-midi-latency-test , unfortunately it
measures a loop, the measured system measures itself, so the validity
of this test is limited, but not completely useless.

I didn't read the links, at least I made tests years later that are
available somewhere in the LAU, LAD, Qtractor, Ubuntu Studio, 64 Studio
archives. To make a long story short, my PCI TerraTec cards and my RME
PCIe card have much better test results than my swissonic USB interface.

I'm aware how to unbind evil USB ports, IOW I'm also aware what USB
port to use for MIDI and how to make it head of the USB ports. To do
this you need to launch a terminal emulation. I doubt that there's any
chance that optimising this could be automated. This not only is an
issue for Linux workstations. USB isn't a good choice. I suspect that
the issue with galvanic isolation for expensive USB audio devices is
solved, since the missing galvanic isolation when using USB MIDI was
another argument against USB MIDI.

Years ago "koppi", Jakob Flierl from Augsburg send me a
circuit board with MIDI in to MIDI throu (out). By a tap the MIDI
signal is connected by a resistor to cinch, so it's possible to
record the signal parallel by a MIDI and audio track. For several
reasons I didn't use the circuit board. I might to it in the future.

I didn't remember his email address and wasn't aware that he's the
coder of the ALSA latency test, so I'll try to drop him a note within
the next days.

Regards,
Ralf