VideoHelp Forum
+ Reply to Thread
Results 1 to 25 of 25
Thread
  1. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    2 SD sources , one 29.97i 704x480 and another 30p 640x480

    I will be doing level ajustment on both and stabilizing on the progressive one, in media fx

    i don't have access to scopes within VMS , is their any free OFX plugin that have them ?
    else i have to render portions and check them and take notes

    i have 2 output goal , one for dvd (mixed interlaced/progressive content ?)
    and another one for youtube , for my YT output i'll have to deinterlace one source

    should i apply smart resample on both ?

    if i have to do deinterlacing i would prefer to do it outside vegas movie studio , before editing

    i did my audio with audacity and using that for sync

    anything else to look out when mixing progressive and interlaced content ?
    Quote Quote  
  2. If it's a film, it shouldn't be interlaced at all after you're done, and without using a damaging deinterlacer on it. Can you post a short sample of an interlaced section? And if it's a film, the 30p version was created by an idiot.

    If it's a mix of progressive and interlace, you might be better off bobbing it for YouTube with QTGMC in AviSynth, and keeping it interlaced for the DVD.
    Quote Quote  
  3. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    the ASF was cut with Virtualdub2 in stream copy from original the other with Mpg2cut2

    the MPG is from a dvd recorder , the original version (dv-cam) is lost,
    it's my only backup
    the ASF is from a kodak cam ( i think)
    Image Attached Files
    Quote Quote  
  4. Thanks. I don't know what to do with an ASF. With it open in VDub2, I see a bunch of missing frames and a bunch of duplicate frames. It's a mess,but I believe the dupe-every-4th-frame can be interpolated with new frames to replace the missing frames. The MPG, as you already know, is pure interlace, but the whites are so blown out that you'll never be able to get any detail back from, for example, the band members' faces.

    Unless you're doing a whole bunch of back and forth interlaced and progressive sections for DVD, you can encode the interlaced parts as interlaced and the progressive parts as progressive and use the 'ADD' button in Muxman to join the pieces together at the authoring stage. Or, if the progressive parts are also 29.97fps just encode the entire thing as interlaced. For YouTube I might bob everything with AviSynth's QTGMC.
    Quote Quote  
  5. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    an ASF is really like a WMV file ( i think)

    last year i did a test for what i wanted to do .. but i think a can do better

    https://youtu.be/ivtUdjFfYKY
    Quote Quote  
  6. A modification of johnmeyer's filldrops/filldrops2 will smooth out a lot of jerks in the ASF file. His functions replace the second frame of a pair of duplicates with a motion interpolated frame -- but it needs to be modified to replace the first.

    https://forum.videohelp.com/threads/386186-Fix-duplicate-frames-on-old-digicam-video#post2504520
    Quote Quote  
  7. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    the ASF is the one that needs stabilizing...
    would MFlowfps to 60fps help stabilizing ?

    I still have difficulties building scripts *finding the right plugins that work toghetter
    ..don't think i would jump into modifying plugins

    *i usually work with simple scripts built with xvid4psp and tweak them

    xvid4psp has filldrops1 in it ..if that help
    Last edited by smartel; 1st Jul 2019 at 01:35.
    Quote Quote  
  8. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    is this the one i need to modify ?

    function filldrops2 (clip c)
    {
    even = c
    super_even=MSuper(even,pel=2)
    vfe=manalyse(super_even,truemotion=true,isb=false, delta=1)
    vbe=manalyse(super_even,truemotion=true,isb=true,d elta=1)
    filldrops_e = mflowinter(even,super_even,vbe,vfe,time=50)

    ConditionalFilter(even, filldrops_e, even, "YDifferenceFromPrevious()", "lessthan", "1.3")

    }

    where do i go from that ? is it only changing even for odd ??
    Quote Quote  
  9. Yes, that's the function. As written it replaces the second frame of a duplicate pair. It was written for captures where a frame is missing and replaced with a duplicate of the frame before. Instead of frames 1234 you have 1224. In that sequence the real frame 3 is missing and replaced with a copy of frame 2. The filldrops2 function replaces the second 2 with a frame motion interpolated half way between the frame before and the frame after. In your video most of the duplicates are in the form 1334 -- frame 2 is missing so the first of the 3s needs to be replaced with a motion interpolated frame. I thought all that needed to be done was replace the YDifferenceFromPrevious() with YDifferenceToNext(). But that's not working. I'm looking into it but won't have time to dive in til later today.
    Quote Quote  
  10. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    Well thank you for the help
    and being it in avisynth i'll change it to 704x480 and 29,97fps to better match my other file

    would another sample help ?
    Quote Quote  
  11. A hackish workaround to target the other duplicate frame would be to reverse() , apply filldrops , reverse() back to normal
    Quote Quote  
  12. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    the script seems to work...
    converting to 59,94 for my interlaced version in this script
    would this screw things up ?
    Last edited by smartel; 1st Jul 2019 at 14:25.
    Quote Quote  
  13. Originally Posted by smartel View Post
    the script seems to work...
    converting to 59,94 for my interlaced version in this script
    would this screw things up ?

    Converting to 59.94 for interlace should not "screw things up" if it's done properly
    Quote Quote  
  14. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    using MFlowfps...
    i don't know the results with the inverse process and filldrops
    Quote Quote  
  15. Originally Posted by poisondeathray View Post
    A hackish workaround to target the other duplicate frame would be to reverse() , apply filldrops , reverse() back to normal
    It's weird -- I tried that but it didn't work.

    This is also a little hackish but it works:

    Code:
    function filldrops2first (clip c)
    {
      prev = c.Loop(2,0,0)
      super = MSuper(prev, pel=2)
      vf = manalyse(super, truemotion=true, isb=false, delta=1)
      vb = manalyse(super, truemotion=true, isb=true, delta=1)
      fill = mflowinter(prev, super, vb, vf, time=50).Subtitle("interpolated")
      ConditionalFilter(c, fill, c, "YDifferenceToNext()", "lessthan", "2.5") # you may need to adjust the threshold value
    } 
    
    ffVideoSource("FNM_Jam06-01.ASF") 
    filldrops2first()
    As with the other filldrops functions it only works correctly when there are two near identical frames together. If there are more it doesn't interpolate across all of them. You can call the filter more than once but the interpolation won't be right. For example, say you have an object that moves across the screen. In the first frame it's a x=100, at the next x=110, then x=120, x=130, x=140...

    Code:
    100 110 120 130 140
    Say second and third frames are replaced by copies of the fourth due to a capture problem:

    Code:
    100 130 130 130 140
    The first call to filldrops2first() will interpolate the first 130 frame, halfway between the 100 and the second 130:

    Code:
    100 115 130 130 140
    Instead of the object appearing at x=110 it appears at x=115. The next call to filldrops2first() will then interpolate the next 130, halfway between the 115 and third 130:

    Code:
    100 115 122.5 130 140
    So the the object has moved from 100 to 130 but not in equal increments.

    And, as usual, motion interpolation like this doesn't always work well. It will often just blur two frames together or create gross distortions.
    Quote Quote  
  16. Originally Posted by jagabo View Post
    Originally Posted by poisondeathray View Post
    A hackish workaround to target the other duplicate frame would be to reverse() , apply filldrops , reverse() back to normal
    It's weird -- I tried that but it didn't work.
    At all ?

    Seems to work ok here. The OP reported it working too

    (or at least as well can be expected for interpolation)



    As with the other filldrops functions it only works correctly when there are two near identical frames together. If there are more it doesn't interpolate across all of them. You can call the filter more than once but the interpolation won't be right.
    Yes, pairs only
    Quote Quote  
  17. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    i reported it like "working" maybe just because i surely don't have the eyes that you have

    but for safety i'll call it twice ..maybe in combinaison with the reverse() fonction !?
    Quote Quote  
  18. Originally Posted by poisondeathray View Post
    Originally Posted by jagabo View Post
    Originally Posted by poisondeathray View Post
    A hackish workaround to target the other duplicate frame would be to reverse() , apply filldrops , reverse() back to normal
    It's weird -- I tried that but it didn't work.
    At all ?
    My mistake. Originally I was using LWlibavVideoSource() to read the ASF file -- that was delivering corrupted frames when the video was reversed. Later (after your suggestion) I tested again with ffVideoSource() which alleviated the corrupt frames problem, but I had forgotten I had modified filldrops2 and it was no longer working correctly.
    Quote Quote  
  19. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    So... where should i go ... FFvideosource..ok, reverse with original filldrops2?
    Quote Quote  
  20. Maybe something like:

    Code:
    # using the original filldrops2()
    
    ffVideoSource("FNM_Jam06-01.ASF")
     
    ColorYUV(gain_y=-15, off_y=6, gamma_y=100) # bring out dark detail, legalize levels
    
    reverse().filldrops2().filldrops2().Reverse()
    
    maxstabH = 40
    maxstabV = 30
    mdata=DePanEstimate(last, trust=1.0, dxmax=maxstabH, dymax=maxstabV)
    DePanStabilize(last, data=mdata, dxmax=maxstabH, dymax=maxstabV, rotmax=0.0, method=1, mirror=15, cutoff=0.2, damping=3.0, prev=10, next=10, blur=0)
    Last edited by jagabo; 2nd Jul 2019 at 18:02.
    Quote Quote  
  21. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    i was going to use the mercalli plugin inside vegas ,
    'cause on some parts theirs a whole lotta shakin' goin' on

    but i'm gonna try your suggestion
    Quote Quote  
  22. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    Looks like i'm having banding issues (note- did't used the depan plugin, only Coloryuv)
    and i went back to LWLibavVideoSource , FFmpegSource2 gave big blocks on occasions
    Image Attached Thumbnails Click image for larger version

Name:	Sans titre.png
Views:	60
Size:	470.0 KB
ID:	49501  

    Quote Quote  
  23. Originally Posted by smartel View Post
    Looks like i'm having banding issues (...Coloryuv)
    Yes, the strong gamma adjustment brings out banding in dark areas. You don't have to use ColorYUV if you don't want. Of you can use less gamma adjustment. Or use SmoothLevels() or SmoothTweak().

    Originally Posted by smartel View Post
    and i went back to LWLibavVideoSource , FFmpegSource2 gave big blocks on occasions
    It's the exact opposite for me.
    Quote Quote  
  24. Member
    Join Date
    Dec 2010
    Location
    quebec
    Search Comp PM
    Should i use Vegas movie studio's mainconcept mpeg encoder for dvd* or output it lossless and use another encoder ?

    *it's a 25 minutes project so bitrate will be the max for dvd

    same question for mp4 output mainconcept avc or sony avc or lossless and amother encoder with youtube's suggested settings ?

    and what about vegas's direct to youtube option ? anybody tried it ?
    Quote Quote  
  25. I don't use Vegas so I can't answer from my own experience. But if you can use max bitrate for DVD you'll probably be fine -- especially if you reduce the noise in your video (noise eats up bitrate).

    I've read that the included AVC encoder isn't very good. x264 is better. I believe you can frame serve to the x264 encoder from Vegas so it's probably not necessary to export to lossless first. On the other hand, you can probably make up for the inferior encoder by using more bitrate (1.5x, 2x?). It's up to you.
    Last edited by jagabo; 2nd Aug 2019 at 21:32.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!