Intel i915GM with SDVO CH7021A support?

will nodenet at
Tue Jan 2 05:46:46 PST 2007

> s-video would be very nice, but I assume there would be additional
> issues such as field parity which is hard to work with in any case in an
> X11 environment. 

Field parity is a problem (as you mentioned), afaik it's only supported
properly in a few dedicated solutions, perhaps the matrox specific code for
example. I'm not sure whether ...

a. The hardware in question will return field parity. Presumably even if it
does it would need to be passed through into the XV code correctly to work.

b. XV supports interlacing (I may have seen something that suggests it was
at least considered at some point)

c. Any video players support it using a generic interface i.e. whether any
video players would support XV interlaced output (assuming this mode exists)

d. Availability / Practicality of alternate solutions e.g. OpenGL

It is theoretically possible to attach a sync separator to the composite
sync output and determine whether it's a top or bottom field. I've also seen
USB interface kits designed for enthusiasts to build their own USB device.
So again theoretically you could design a USB device to read the sync signal
and presumably the USB interrupt can be set by sending data triggered by top
or bottom field using the v sync (it may be possible to work off one only,
i.e. if you know where the start of the frame is this should be sufficient
and TFF/BFF can be toggled appropriately). All that's before you
modify/write any software so I think it would be quite a lot of work.
Generally turning on software deinterlacing is considered sufficient I
believe the line doubling deinterlacer is supposed to be best for this
purpose. It seemed to work when I tested it (briefly) and I know others have
used it on similar setups, so it would just remain as to whether or not the
quality is considered to be adequate.

> Have you got any guess at how hard it would be to add
> s-video support at a later stage?

I think it would be a case of trial and error as the BIOS seemed to revert
back to composite even if booted under S-Video. i.e. I don't think BIOS
tracing will work in this instance. Composite and S-Video both use DACB on
the Chrontel so it may be that an additional flag or sdvo op code is used to
switch between the 2 modes.

More information about the xorg mailing list