Re: [FFADO-devel] stanton test-scs (alsamidibridge)

[Jonathan W. <jw...@ju...>]

Hi Rob

On Sat, Apr 28, 2012 at 11:02:15AM +0200, Rob Bothof wrote:
> On 4/28/2012 10:00 AM, Jonathan Woithe wrote:
> > The way MIDI is communicated to the device at the lower level depends 
> > on the device. As I mentioned in ffado-users, the problem could be due 
> > to what the interface expects regarding message contents (that is, it 
> > may fundamentally expect all 7 bytes in a single message) or timing 
> > issues between the separate 1-byte messages.
> 
> yes, but when a 7byte (or more) single message is send, the message is 
> split up in packs of 4 bytes (quadlets?)
> and this works, maybe it has to with the first byte (quadlet label) or 
> message type as it is called in test-scs that is in front of all 
> messages? except when a long message is split up into multiple packs of 4

I don't know.  I think the use of single-byte messages are the problem
more-so than any of the lower level processing done by ffado.  However, I
have no way of proving that except to say that I wouldn't be surprised if
the device itself saw a difference between 7 1-byte messages and a single 7
byte message, irrespective of how that stream of messages got delivered to
the wire.

Anyway, we have a solution now. :-)

Regards
  jonathan