Woithe<jwoithe(a)physics.adelaide.edu.au> wrote:
AFAIK FFADO got started through the cooperation
of one or two vendors
and as momentum has built additional vendors have come on board to a
greater or lesser extent. I expect something similar will have to
happen for professional USB2 audio devices to be generally supported
under Linux.
This misses out on the most important step in the evolution of FFADO.
Before FFADO was Freebob, an open source implementation of software to
support the BeBob API. Freebob was a project that was started by
someone who worked for the company that developed the Bebob chipsets
(Bridge Co.) ...
Ok, I should have included a little more detail. By "FFADO" in the above
paragraph I was referring to the FFADO/FreeBoB combination. BridgeCo was
indeed one of the cooperative companies I was referring to. I gave the
brief three line version whereas Paul has filled in a few more details.
:-) It wasn't that I missed the most important step but rather glossed over
the details in the name of brevity. Perhaps I was too brief.
TC Electronic ... the existence of Freebob
convinced them to seek out the
two people involved in that project and pursue drivers for their new
chipsets. I was happy to play a tiny role in helping this meeting to take
place (very tiny).
Ah - I didn't know that.
The problem with USB2 is that it doesn't seem
as though there is any
similar embedded systems interest in linux support for audio over
USB2, and nor is there a single company (or 2 companies) with a choke
hold on the chipsets used for this. This makes the problem a lot
harder to solve. Maybe the companies that were thinking about firewire
audio are now thinking about USB2 audio, but if they are I don't have
my ear close enough to the ground to hear.
For devices with higher channel counts USB2 hasn't really been the
interface of choice thus far. This is mainly due to the limitations of the
USB bus and the fact that it makes it hard to do low latency audio well.
Things like it being a half-duplex bus with a host-slave architecture, and
the fact that the real-world throughput is *much* less than the 480 Mbps
*signalling* rate. The picture is a little different with USB3 since this
version adds full duplex capability along with a number of other
enhancements (beyond the obvious speed increase). How well it will go with
audio given the latency requirements remains to be seen.
As a side note, it seems RME are posed to release a USB2 version of the
Fireface. I don't know much about it except that the device runs different
firmware based on the operating system it's being used with. This suggests
to me that RME work at a very low level both in the operating system driver
stack and at the bus hardware levels in an attempt to circumvent the
biggest audio-related USB showstopper issues.
Maybe one of the things to note as well is that some of the USB2 audio
interfaces are not even really keeping to the USB standard, and certainly not
to the the USB audio standard. I believe some of them aren't even allowed to
carry the logos for these protocols on the boxes.
sincerely,
Marije