I'm back, with more questions. I've been trying to piece
together a stack of libs to gain the best leverage before laying
in my app. The goal is a system that will capture video on a
Radeon ZV port and display it, surrounded with some
computer-generated graphics in the margins. The imagery need go
no further than the video overlay and out to the display, until
the user requests a hardcopy, at which time a single frame would
be snapshotted and saved out to a file.
I've got a stack running that looks like so:
[ Demo App ]
[ DirectFB ]
[ V4L2 ]
------------------
[ radeonfb ]
[ VideoDevX ]
[ V4L_compat ]
[ V4L2Cap ]
V4L2Cap is a SW-only 'pseudo' video-capture driver that provides
animated video up to the V4L2 lib. It's a test pattern w/ running
time clock. Kinda' neat for testing. I'm using one of the demos
in DirectFB-examples for now, while getting the drivers sorted
out.
Anyway, the next step is to replace the V4L2Cap driver with
something that actually knows how to control the Radeon capture
port. In conversation on the GATOS mail list
(www.sourceforge.net/projects/gatos/), I'm coming to realize their
drivers assume X11 and are incompatible with framebuffers. I'm
trying to put this together without going so heavy-weight as to
use X11, which is what led me to DirectFB initially.
Can anyone here point me at any alternative ways of reaching my
goal? Are there other projects working on video (and video
capture in particular) on ATI boards?
Assuming nothing else available 'off the shelf', could I engage
the folks here with the best understanding of DirectFB in some
discussion as to the best way to architect such a feature? Would
it make the best sense to lay it in here in DirectFB as a new
video provider, or organize a new 'driver module effort' under the
GATOS banner?
Benton 19feb04
--
BC&G Holzwarth
[EMAIL PROTECTED]
--
Info: To unsubscribe send a mail to [EMAIL PROTECTED] with
"unsubscribe directfb-dev" as subject.
