VideoHelp Forum

Our website is made possible by displaying online advertisements to our visitors. Consider supporting us by disable your adblocker or try DVDFab DRM and remove iTunes movie & music protection! :)
+ Reply to Thread
Results 1 to 6 of 6
Thread
  1. Member
    Join Date
    Aug 2017
    Location
    United States
    Search PM
    I'm almost to the conclusion that this whole series on DVD is jacked. Running the below clip through AVISynth with the following TFM line results in frames that look like low-res crap. If I create an override file and force the bad frames to be matched to their "c" combination, then the frames return to their true full-res glory. But why is this the case? Were these DVD's just authored incorrectly? Is there ANY way to automatically compensate for this behavior in a script or am I stuck hand-patching frames?

    TFM command:
    Code:
    TFM(slow=2,pp=7,mode=3,mi=70,cthresh=8, ovr="E:\Old Computer Files\Video Editing Tools\MeGUI_2624_x86\avisynth scripts\TMNT2k3\TMNT_xmas.txt")
    example override file command for bad frame:
    Code:
    13953 c
    EXAMPLE CLIP
    Quote Quote  
  2. There are problems with the source ; aliasing and some field blending

    Some sections don't have a clean match manually ; and many frames have aliasing to begin with - hand patching won't work alone unless you mean photoshop or extraordinary measures . overrides will not fix this alone

    Also there are some 29.97 sections - you'd have to make it vfr to make it perfect
    Quote Quote  
  3. Member
    Join Date
    Aug 2017
    Location
    United States
    Search PM
    I think you're totally right. This series is a stellar example of creating a cartoon specifically for interlaced viewing with no thought of the progressive future.

    Many of the problems can be hand-patched, however. And by that I mean the use of overrides to find the best (not necessarily perfect) field match and/or replacing the current "bad" frame with prior or next frames if the neighbor is otherwise identical, but "good". For those field blends and orphaned fields I've just been replacing those bad frames or simply blending them(not ideal).

    I was hoping beyond hope that there was some hidden key to automate the processing that I was missing. Looks like there's only the hard way unless someone else has an insight.
    Quote Quote  
  4. I didn't look too closely but try:

    Code:
    Mpeg2Source("TMNT2k3clip.demuxed.d2v", CPU2="ooooxx", Info=3) 
    QTGMC()
    Dup(threshold=6, blend=true)
    TDecimate(mode=1,  Cycle=10, CycleR=6)
    Maybe follow up with some noise reduction.
    Last edited by jagabo; 8th May 2018 at 13:12.
    Quote Quote  
  5. Member
    Join Date
    Aug 2017
    Location
    United States
    Search PM
    When it comes to the frame pattern, have you (either of you) seen anything like it before? When the sequence starts going bad it's very often of the pattern 1-2-3-3. So every fourth frame is a duplicate of the prior, but bad. It doesn't seem to conform to a normal telecine pattern, but like poisondeathray said many sequences are also truly 29.97fps. It's a crazy mishmash.

    EDIT: Sorry, I meant that the first of the frame 3's is usually the bad one. It's then followed by a good frame 3.
    Last edited by SaurusX; 8th May 2018 at 15:57.
    Quote Quote  
  6. It's quite common for cartoons to be "shot" on film, telecined, then edited as video. When shots are sped up or slowed down you get problems like those you are seeing. Then credits and titles are often overlaid as 30i.
    Quote Quote  



Similar Threads