You can have a look at the specs here:
http://www.math.tu-berlin.de/~sbartels/unitor/

German only and I don't speak german :-(( (if anybody wants to translate the
document I'm willing to help maintaining the driver...

Sebastien

----- Original Message -----
From: "martijn sipkema" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, December 21, 2001 4:07 PM
Subject: Re: [linux-audio-dev] Introducing DMIDI


> > The AMT8 have a similar system to reduce the timing errors. In fact the
> > sequencer device in the midi windows API does have a similar scheme: you
> > send blocks of data in which every midi message is time stamped. I
imagine
> > the driver does the clock translation for you.
>
> Emagic would not give me the details on their AMT protocol, but as far as
I
> know it does not work with timestamps. MIDI events are send ahead of time
> to the interface. When such an event(s) buffered in the interface need to
be
> actually transmitted, a message is sent to the interface. This protocol
> enables
> events to be transmitted on all ports simultaneously, given they are known
> ahead of time.
>
> The windows API using timestamps doesn't allow the use of LTB or AMT
> using this API, since the driver interface does not use timestamps. That
is
> why I would like to have a driver interface for Linux that does use
> timestamps.
>
> --martijn
>
>

Reply via email to