[Softdevice-devel] viafb + cle266mpegdec + TVout : field parity status and patches [update]
stefan at lucke.in-berlin.de
Sun Oct 8 22:29:10 CEST 2006
On Sonntag 08 Oktober 2006 19:45, Nicolas Huillard wrote:
> Hi all,
> Here's my update on the matter after upgrading quite everything. I would
> like to thank those who replied to my questions.
> First : I still have field parity problems with interlaced material on
> interlaced TV.
> Second : there are a lot things there, from the kernel, viafb,
> libcle266mpegdec, DirectFB, DFB++, up to softdevice...
> DFB++ CVS 20061008 (today)
> * applied the GetFBOffsetDFB++.patch from libcle266mpegdec
> * this patch should make it into CVS, to keep in sync with DirectFB
> * there is nothing about field parity there, since Mark's patch was
> about the unichrome driver, not about DirectFB API (which DFB++ only
> deals with)
I think the DFB++ patch is still on Denis' radar, see:
So I don't think we need to push him further.
The missing DFB++ part, is one reason not to publish a new softdevice release now.
> softdevice CVS 20061008 (today)
> * no patch to apply ?
> * uses libclempegdec, viafb, etc.
> * to attached log shows field parity is enabled ([softdevice] enabling
> field parity / [softdevice] enabling CLE266 HW decoding)
> I don't know what I miss to get rid of interlacing artifacts on TV out...
If the interlacing problem is a permanent issue, you may try to change
line 294 in video-dfb.c:
fieldParity = 0; // 0 - top field first, 1 - bottom field first
If that issue is a changing one, we need to set this from
stream information and/or give the user a chance to invert field
parity derived from stream info.
More information about the Softdevice-devel