#20 Modulated (progresive) post-procesing


I have a request that may sound a bit strange, cause it's
not what's made today with FFDShow, but I think this
could be quite usefull to improve decoding visual quality.

I have seen that post-processing uses to blur the
images, actually, when it's at level 6, a lot (ok, maybe
I'm paranoid, but it seems to be the same at level 2) of
details (e.g. faces, walls or roads textures) are lost in
some blurred colors (especially on sharp movies). For
the moment, the idea is to allow max PP on quiet
scene, when CPU load is quite low, and to decrease it
on hi-motion scene.
But when do we need highly detailled scenes ? On low-
motion scenes, where we have time to see sharpness,
whereas high PP can make macroblocs dissappear on
hi-mo scenes.

So my idea (tell me if it's stupid or already implemented,
but I haven't seen this with FFDShow or any PP-filter -
Nic's or DivX's ones) would be to have a 'special quality
mode' which would only activate PP for hi-motion
scenes, to kill Mblocks, and allow max details on static
scenes without adding PP. Of course, this could be
mixed with a '100% cpu load' limit, for weak processors.

I hope what I have told is clearly understandable (it's
late, I'm preety tired right now, and english isn't my
mother's tongue)

Best regards


  • Anonymous - 2002-12-04

    Logged In: YES

    Postprocessing strength is controled by how much the source image is
    compressed and the compression should be higher in high motion
    scenes, so ffdshow postprocessing should work as you wish
    But as you notices, there is problem is slight blurring. It
    could be possible to activate/deactivate postprocessing based on
    motion (MPEG video stores motion vectors and this decision could be
    based on their lengths), but there would be problem with visible image
    changes caused by this switches. The best would be to set
    postprocessing strength more precisely, but it seems that even with
    lightest processing still blurs the image.
    Two solutions comes in my
    1. rewrite or modify postprocessing routines (deblocking,
    deringing) so would not blur so much: this would be hard, because their
    aren't written by me and I don't understand how they work
    2. blend
    processed and original image with blending strength controlled by
    motion. Something similar is implemented already for Soften filter and
    it's quite fast and gives good results. I can try to implement this to see
    whether this could help or whether it's just nonsense.

  • Inoe

    Inoe - 2002-12-04

    Logged In: YES

    Well, what can I say ? If your solutiob n2 works (I'm not a
    coder at all, so i can't even think to modify deblocking or
    deringing), then you're my saver, and I'll pray your name all
    over the forums :)
    Cause it's a shame to work hard to have the best quality
    possible, and see details disappear because of PP (which is
    very usefull most of the time). But don't think this means I
    can't wait, I have great respect for your work, FFDSow is my
    favorite decoder.
    One last question (if out of subject, just do not answer it): is
    asharp (by MarcFD) on the way ?

    Thx for everything.



Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks