new ffdshow (with xvid devapi4 inclusions )

[yaz]

@all
just a short cut from the changelog:
2004-02-25 06:44 milan_cutka
working again... :) just few fixes and new features for now:

fixes in subtitles reader, fixed mono audio processing, SPP
deblocker from mplayer, volume normalization from mplayer, updated
installer to NSIS 2.0 final, updated xvid to latest RC, output
colorspace is preset option, updated libavcodec (now plays packed
xvid b frame streams, some corruptions of picture after seeking),
fixes of few uninitialized variables found by DevStudio


alpha compiles available from : athos (athos.leffe.dnsalias.com) & ligh (http://www.ligh.de/software/mirrors.phtml)
[EDIT]see also the comments in the 'ffdshow development' thread !

[buzzqw]

Great !

Thanks for info (and web pages)

BHH

[LigH]

You mean this thread (http://forum.doom9.org/showthread.php?s=&threadid=48511)... I sometimes wonder how far this thread is away from the top of the XviD forum of this board, may it be useful to make it sticky -- or even create a new forum like "other (MPEG4 or even more different) codecs and filters", apart from XviD and DivX?!
__

Ah, hell - forget it: it finally happened, I just looked in the wrong line of the search results somehow...

[The Link]

To give you all a reason to flame me: Is XviD (still?) using simple IDCT or is ffdshow intelligent enough to let it decide via "auto select"?

Regards,
The Link

[PowerMacG4]

Originally posted by The Link
To give you all a reason to flame me: Is XviD (still?) using simple IDCT or is ffdshow intelligent enough to let it decide via "auto select"?

Regards,
The Link

XviD has been using Walken for quite a while...

[sysKin]

Originally posted by The Link
To give you all a reason to flame me: Is XviD (still?) using simple IDCT or is ffdshow intelligent enough to let it decide via "auto select"?Just remember that Athos' builds do not have any assembler optimizations, and as a result they're not using either 'simple' or 'walken' but rather some yet-another pure-c version. The switch or auto-detection do nothing.

Radek

PS. I still can't make it use "xvid 4" - if I select that, decoding goes to 3vix decoder (which is next in line if ffdshow fails).

Radek

[yaz]

@ligh
maybe u know, ffdshow development started here in the xvid forum. iirc, the original thread was called dffshow (just to make it more confusing:-)

@syskin
yep. sg's strange around here. when i switch to xvid4, ffdshow seems to kick off. must be chained back via yv12. however, using libavcodec seems to be less risky, meaning, some weird artifacting seems to be debugged. more tests (?reading:-) needed. btw, i used your last dsf for testing. great stuff! (maybe u should switch too :-)))
afaik, assembler optimization is also under way. andy2222 & others make heroic efforts on it. i hope they will merge it to the cvs soon.

the bests
y

[Leak]

Originally posted by sysKin
PS. I still can't make it use "xvid 4" - if I select that, decoding goes to 3vix decoder (which is next in line if ffdshow fails).[/B]

With the current ffdshow-20040227.exe it's definitely working, as can be seen by the "B-Frame lag" message flashing by when you start playing a file... *eg*

(Couldn't there at least be a registry setting to simply replace this message with a totally black frame? ;))

np: Repeat - Studio 6122 (Miss Kittin - Radio Caroline Volume 1 comp.)

[athos]

first, i think ligh's is just a mirror of my build, at least it has exactly the same filesize (to the byte, 1_055_003 bytes).

second, my compiles does contain assembler optimizations, including mmx. andy2222 is making some more assembler stuff, for example for dscaler, which was very slow before. he is also converting some assembler stuff from ms vc style to gcc.

(after reading syskins post again): oh you mean assembler stuff in the IDCT specifically? i wouldnt know about that, but it should if the IDCT is in libavcodec and it has assembler stuff.

please correct me if im wrong.

[Satlover]

I have allways had problems with ffdshow which is a shame becuase it has some very powerfull features. Sometimes the frames would not flow at a constant rate - probably on high bitrate scenes I fixed this problem by removing all other codecs except XVID RC2 and installing: URL=http://koepi.roeder.goe.net/XviD-Dec-1.0-Beta3.exe]installing Latest standalone decoder DirectShow filter (installer)[/URL]

I also tried the latest ffdshow (ffdshow-20040225.exe) and that still has the same problem.

[Koepi]

the leech protection is working against direct linking.

satlover, please don't ever do that again.

Thank you.

Koepi

[Satlover]

@Koepi, maybe its time I read the Forum rules. :rolleyes: Sorry

[yaz]

ffdshow_20040304 (short cut)

... support for quantizer array exported by xvid
... removed xvid's "b frame decoding lag" message, custom quant tables selection moved to quantization type combobox
...ACM part is working
...by request: horizontal subtitles alignment (just first try)
...disabled SSE2 code in libavcodec, disabled quantizer noise shaping for codecs which don't support it and crash
...b frames seeking fix
...first successful VFW encoding
...third column in Codecs page: supported FOURCCs
...enabled mp4v FOURCC

try & enjoy :-)))
y

[kilg0r3]

Does anybody have an idea why it is impossible to use ffdshow as preprocessing filter for capture apps like iuvcr and virtual vcr. Both apps say that they cannot build the filter chain when I insert ffdshow.

"Could not render Smart Tee -> Compression Codec -> Flow Meter"

[Nic]

Could we continue the ffdshow discussion in the thread in the new codecs forum:
http://forum.doom9.org/showthread.php?s=&postid=454296#post454296

bond suggested it, and is probably right as talking about the same thing in two different places is probably not a good idea (even though the xvid forum was the original home of the ffdshow thread)

Cheers,
-Nic