VideoHelp Forum

Our website is made possible by displaying online advertisements to our visitors. Consider supporting us by disable your adblocker or Try ConvertXtoDVD and convert all your movies to DVD. Free trial ! :)
+ Reply to Thread
Page 2 of 2
FirstFirst 1 2
Results 31 to 53 of 53
Thread
  1. If it's footage with little motion then its easy to accurately interpolate the damaged frames with svpflow so try this:
    Note that it's always better to have progressive frames

    LoadPlugin("C:\Program Files (x86)\AviSynth 2.6\plugins\svpflow-1.0.6\svpflow1.dll")
    LoadPlugin("C:\Program Files (x86)\AviSynth 2.6\plugins\svpflow-1.0.6\svpflow2.dll")

    AVISource("M:\05-montage\glitch1video.avi").assumetff().converttoyv 12(interlaced=true)
    ReplaceFramesSVPFlow(13,3) # premier = image, 2ème= durée (nbre d'images à remplacer)


    function ReplaceFramesSVPFlow(clip Source, int N, int X)
    {
    # N is number of the 1st frame in Source that needs replacing.
    # X is total number of frames to replace
    #e.g. ReplaceFramesSVPFLow(101, 5) would replace 101,102,103,104,105 , by using 100 and 106 as reference points for SVPFlow interpolation

    start=Source.trim(N-1,-1) #one good frame before, used for interpolation reference point
    end=Source.trim(N+X,-1) #one good frame after, used for interpolation reference point

    start+end
    AssumeFPS(1) #temporarily FPS=1 to use mflowfps

    super=SVSuper("{gpu:1}")
    vectors=SVAnalyse(super, "{}")
    SVSmoothFps(super, vectors, "{rate:{num:"+String(X+1)+", den:1}}", url="www.svp-team.com", mt=1).Subtitle("SVPFlow")

    AssumeFPS(FrameRate(Source)) #return back to normal source framerate for joining
    Trim(1, framecount-1) #trim ends, leaving replacement frames

    Source.trim(0,-N) ++ last ++ Source.trim(N+X+1,0)
    }


    Quote Quote  
  2. This is the same code than poisondeathray ?
    What's the difference ?

    Yes, I used that code, it's the best option for me, except for "big movement/action situation".
    Quote Quote  
  3. svpflow use the gpu (graphic card) so it's faster (and probably more accurate for the motions vectors, to be confirmed though)

    more info here: https://forum.doom9.org/showthread.php?t=164554
    Quote Quote  
  4. Is it possible to select the area we want to get smooth ?

    Example here, the top is damaged.
    But when I apply the filter, the effect is not really good cause the guy moves his hand quickly.
    Image Attached Thumbnails Click image for larger version

Name:	glitch3.png
Views:	67
Size:	901.2 KB
ID:	45093  

    Quote Quote  
  5. You can crop, filter, then overlay.

    Code:
    patch = Crop(0,48, width,96).RX(x,n) 
    Overlay(last, patch, x=0, y=48)
    Quote Quote  
  6. Member
    Join Date
    Aug 2010
    Location
    San Francisco, California
    Search PM
    Originally Posted by jagabo View Post
    If these videos are very important to you consider getting an old Panasonic ES10 or ES15 DVD recorder to use as passthrough to clean up the time base. It may fix a lot of these problems as well as reducing the horizontal jitter on the rest of the frame.
    As I fruitlessly pointed out in #8 while everyone was focused on software manipulation.
    Originally Posted by TONYFRANCE View Post
    TAnd I also heard that sometimes, watching your tape in a simple 2 heads vcr is enough.
    Watch the same tape in a high model can provoke issues.
    Is that true ?
    It's often the case that a lower-tier Panasonic or Sharp VCR will handle EP recordings better. Your glitch looks like a tape tension problem during recording or playback.
    Quote Quote  
  7. Originally Posted by jagabo View Post
    You can crop, filter, then overlay.

    Code:
    patch = Crop(0,48, width,96).RX(x,n) 
    Overlay(last, patch, x=0, y=48)
    Can you explain me how you adjust the height or if you want to move the section ?
    How can I see the selected area ? Can I take a screenshot and put it into a software like photoshop, to see what data I need to write into the crop line ?

    Crop(0,48, width,96) : what is it ? I know what is width, but width of what ?
    Overlay(last, patch, x=0, y=48) : I need to put the same number than in the crop line ?

    Thank you again guys for answering at these stupid questions...
    Quote Quote  
  8. Remember, when you don't specify a named video in AviSynth the name "last" is used. Since I posted only a short code segment you can assume a "last" video has already been opened.

    Code:
    patch = Crop(0,48, width,96)
    This creates a new video clip called patch. Since no input clip was specified for Crop() the "last" clip is assumed. "width" and "height" are a predefined variables in AviSynth -- the width and height of a clip. Again, since no video was explicitly named, the width of "last" is assumed. The first two numeric arguments to Crop() are the x and y locations (relative to the top left corner) of the of the box you want. The second two numeric arguments, when positive values, are the width and height of the box. When negative they mean that amount less than the width and height of what's being cropped. For example, a "-8" for the width argument means "width-8".

    Code:
    .RX(X,N)
    The period between Crop() and RX() indicates that the output of Crop() is piped directly to RX(). So the full line

    Code:
    patch = Crop(0,48, width,96).RX(x,n)
    creates a new clip called patch which is a crop taken from "last" with some frames replaced by RX().

    Code:
    Overlay(last, patch, x=0, y=48)
    This overlays patch onto last, at position x,y. Since patch was cropped starting at line 48 it needs to be overlaid back in the same place, line 48.

    You can determine coordinates by exporting a frame and examining it in an image editor. Or you can just eyeball it. You can visualize the crop (or any other operation) in many ways. You can just return the crop clip {add return(patch) after the patch=Crop line}. But sometimes it's difficult to identify the crop, especially with small crops. In this case you could Invert() (the brightness and colors of) the crop so that it stands out:

    Code:
    patch = Crop(0,48, width,96).RX(x,n).Invert()
    The cropped portion will be obvious.

    Maybe this will make it clearer to you:

    Code:
    # Show the values we're going to use symbolically:
    PATCH_XPOS = 0
    PATCH_YPOS = 48
    PATCH_WIDTH=width  # the width of "last"
    PATCH_HEIGHT = 96
    
    patch = Crop(PATCH_XPOS, PATCH_YPOS, PATCH_WIDTH, PATCH_HEIGHT).RX(x,n) 
    Overlay(last, patch, x=PATCH_XPOS, y=PATCH_YPOS)
    Quote Quote  
  9. Some people like using avspmod, it's a script editor for avisynth. There is an interactive crop editor in avspmod . video=>crop editor . You can left click and drag the edges with the mouse, and it uses the invert trick jagabo mentioned to enhance visibility . It's a bit faster than having to go back & forth into an image editor . You can edit & preview/refresh script (f5) directly
    Quote Quote  
  10. Thank you for the tips everyone.
    Quote Quote  
  11. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Originally Posted by TONYFRANCE View Post
    And I also heard that sometimes, watching your tape in a simple 2 heads vcr is enough.
    Watch the same tape in a high model can provoke issues.
    Is that true ?
    No.
    Not true.
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  12. Originally Posted by lordsmurf View Post
    Originally Posted by TONYFRANCE View Post
    And I also heard that sometimes, watching your tape in a simple 2 heads vcr is enough.
    Watch the same tape in a high model can provoke issues.
    Is that true ?
    No.
    Not true.
    Simple answer but clear
    Quote Quote  
  13. Originally Posted by TONYFRANCE View Post
    Originally Posted by lordsmurf View Post
    Originally Posted by TONYFRANCE View Post
    And I also heard that sometimes, watching your tape in a simple 2 heads vcr is enough.
    Watch the same tape in a high model can provoke issues.
    Is that true ?
    No.
    Not true.
    Simple answer but clear
    It is also wrong. The original advice was correct: cheap VCRs often do a better job on EP (6-hour) tapes. I know because I've done it.

    However, that whole conversation is irrelevant to your problem because it looks like some sort of tape issue (stretch, drop out, etc.). You might get the problem to appear in a slightly different way if you used a different VCR -- cheap or not -- but I doubt a new capture would be totally clean.

    I do have pieces of the solution, although not the whole thing. I wrote a somewhat generalized script that does automatic detection and replacement of bad frames:

    Finding individual "bad" frames in video; save frame number; or repair

    If you were to replace the detection logic in the script I developed in that thread with detection based on StainlessS' RTStats, you might be able to do what I did in this video:

    https://www.youtube.com/watch?v=qx26T6WOZ_4

    In that Kinescope film, as you can see, I was able to both detect and then remove noise bars that were on every frame, and were in a different position on every frame. The entire process was automatic, once I got everything working. Your noise bars are actually much easier to detect because they are stronger; they persist for only a few frames, and they are very cleanly delineated. Here is the thread, over in doom9.org, where I asked for some help and eventually was able to develop an automatic solution:

    Bad 1950s Kinescope - Hopeless?

    I won't be able to do any of the work on your video, however, because I just don't have that kind of time to do the development. If you have a technical background, you should be able to adapt this work.
    Quote Quote  
  14. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    Originally Posted by johnmeyer View Post
    cheap VCRs often do a better job on EP (6-hour) tapes.
    No.

    You're ignoring the actual underlying tech, masking it with a vague generality.

    What you need to pay attention to is the transport, alignment, and heads condition. It's certainly possible for a usually-better deck to perform poorly, and a crummy deck to perform well. But this is not typical, and I've rarely seen it happen in 25+ years.

    Another issue is the tape itself. If said tape was made with poor alignment, an equally badly-aligned deck probably will play it better than a properly aligned deck.

    If you really want good EP playback, use a Panasonic AG-1980P. Not a cheap VCR.
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  15. Member Cornucopia's Avatar
    Join Date
    Oct 2001
    Location
    Deep in the Heart of Texas
    Search PM
    But there is some truth to the matter:
    Some decks have 1 pairs of heads, some 2, some 3, and some 4.
    One pair is for video, one pair is for HiFi audio, supplemental pairs are (optimized) for LP and/or EP/SLP (using different gap size & magnetic properties).

    So if it has only 4 heads and it is supposed to be a hifi deck, it can only be optimized for 1 speed. Which speed depends on the deck model, family, & manufacturer.
    If it has 8 heads (rare), then ALL speeds are optimal.

    IIRC, most decks had 4 or 6. Which often meant that with hifi, either SP or EP/SLP would be optimal for 4heads, and SP AND EP/SLP would be optimal for 6heads (though a few regions made LP optimal instead of EP/SLP).

    This all assumes well kept up, clean & operational machines and other things being equal.

    Scott
    Quote Quote  
  16. Video Restorer lordsmurf's Avatar
    Join Date
    Jun 2003
    Location
    dFAQ.us/lordsmurf
    Search Comp PM
    But it's never equal.

    There were some 2-head decks, but most were made in the 80s or early 90s. At this late date, those machines are probably severely out of spec. Gravity alone will tank the alignment after decades without maintenance. So an already-crappy VCR will be even crappier.

    The idea of "more heads = better" was mostly a myth, and one that was marketed heavily in the mid 90s. I remember Toshiba and GoVideo tried this, but it was only rich gullibles that fell for it. The machines were priced too high for the average consumer, and both pros and hobbyists tended to see the machines as lower quality. What always got me as that the Toshiba 6-head VHS was about the same price as a lower-end JVC S-VHS deck, and the JVC easily outperformed it. (And higher end easily outperform the lowers.) There was also this cult of GoVideo that really pissed me off at the time, because it did some major damage to the cartoon collecting/trading hobby of the era. It was all about "brand names" and "specs" and not the actual quality of the image.

    While more heads may have been better in theory, it just never materialized. Other parts of those VCRs simply sucked.

    Toshiba VHS VCRs did track better than average, but it was true for the 4-head units.

    I've had guides on VCRs for years: http://www.digitalFAQ.com/guides/video/capture-playback-hardware.htm

    Nothing about the heads count has anything to do with the "streaks" issue in the thread. At most, the heads are shot. But I'd more likely believe the tape had the issue here. This can be corrected without too much effort in Avisynth, but it needs time and power to run.
    Want my help? Ask here! (not via PM!)
    FAQs: Best Blank DiscsBest TBCsBest VCRs for captureRestore VHS
    Quote Quote  
  17. Something I just discovered.
    After I apply the corrections, I lose on brightness, why ?

    Image
    [Attachment 46137 - Click to enlarge]

    Image
    [Attachment 46138 - Click to enlarge]


    If there is no explanation, what code can I use to recover the loss ?

    NOTE : This is after the trick used in this topic 'Erase red/purple mark (VHS capture)'
    Quote Quote  
  18. Originally Posted by TONYFRANCE View Post
    This is after the trick used in this topic 'Erase red/purple mark (VHS capture)'
    What's the script/process?
    Quote Quote  
  19. Code:
    Function GetBadChroma(clip c)
    {
        c
        Merge(SelectEven(), SelectOdd())
        Merge(SelectEven(), SelectOdd())
        Merge(SelectEven(), SelectOdd())
        Trim(0,length=1)
    }
    
    AviSource("xxx.avi") 
    
    BadChroma = GetBadChroma()
    Overlay(last, BadChroma, mode="subtract")
    
    Crop(0,4,-14,-6)
    AddBorders(0,4,14,6)
    Quote Quote  
  20. You forgot the ColorYUV(off_y=-255) part.
    Quote Quote  
  21. Oh ok... Thanks!
    Quote Quote  
  22. And keep in mind that the GetBadChroma function gets the bad chroma from the first 8 frames (averaged together) of the clip it's given. If those frames don't have the chroma problem it won't work. And if those frames aren't otherwise black-and-white it won't work. So you must pass it a portion of the video that is black-and-white and has the chroma problem.
    Last edited by jagabo; 21st Jul 2018 at 08:23.
    Quote Quote  
  23. Is the TBC on Panasonic on? From the OP samples it seems that ag1980 TBC is having problem. I had something similar on some bad tapes (probably recorded on not aligned VCR), the "scrambling" was the same, it reminds me of early Sat Tv analog encryption. Turning TBC off stopped the garbage (two tapes) at least in my case. Probably with this kind of tapes TBC (in ag1980) is having problems, using DVD recorders TBC (es10/15) pass trough,its not getting the same effect (probably because its outside)
    Try recording the same tape without TBC on

    Sorry answer on wrong post, can the admins delete it
    Last edited by mammo1789; 16th Aug 2018 at 16:39.
    Quote Quote  



Similar Threads