VideoHelp Forum
+ Reply to Thread
Results 1 to 21 of 21
Thread
  1. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    I know that telecined material SHOULD follow a pattern of 3 progressive frames and 2 blended ones, but when I step through this film, I am SUPER confused because I truly cannot tell exactly what the pattern is. It kind of LOOKS like 1 good frame and then a frame where it almost looks like a duplicate and then a frame where it doesn't exactly look 'blended' as in the example of the truck in the older post on Doom9. In that picture it is pretty clear that you have 3 clear frames of movement and then 2 blended or blurred looking frames, but in this one I'll be danged if I can figure it out (Manono said, 'Use your eyes' Well, I'm bloody trying...) It looks like a couple of good, clear frames, then it seems that there are duplicate frames SOMETIMES after the good ones, and the like I said, they don't look 'blended' like the truck, but they kind of look like the edges are catching up slightly with the rest of the frame, or something like that. Weird...

    I did load it into MeGUI and analyzed it, and it said that it was 'Hybrid film Interlaced / Mostly film' HUH...??! And the script it recommended or came up with was exactly the same as the last film I did with it, but it wasn't titled quite that way and I think that film was BFF and this one it says is TFF. Here is the script it recommended. These are basically Mpeg2 captures to DVD. I'm kind of surprised that they are not flagged as straight telecined because of the source...

    LoadPlugin("D:\EXECUTABLES\MeGUI_2624_x86\tools\ls mash\LSMASHSource.dll")
    LWLibavVideoSource("C:\x\00000.mkv")
    LoadPlugin("D:\EXECUTABLES\MeGUI_2624_x86\tools\av isynth_plugin\TIVTC.dll")
    tfm(order=1).tdecimate(hybrid=1)
    #crop
    LanczosResize(640,480) # Lanczos (Sharp)

    http://lathe-of-heaven.com/RoE.mkv
    Quote Quote  
  2. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Hmmm, I'm reading this thread here and it LOOKS like Hello_Hello was asking something similar (It still doesn't help me VISUALLY identify what type of encode it is though...)

    https://forum.videohelp.com/threads/354617-De-interlacing-and-MeGUI
    Quote Quote  
  3. Member
    Join Date
    Dec 2005
    Location
    Finland
    Search Comp PM
    There may be some confusion here regarding "interlacing" and "blending". Clean, telecined video has no "blended" frames. What it does have is alternating progressive and interlaced frames, specifically three progressive followed by two interlaced.

    You can easily see this using your eyes Look (with no filters applied) at the scene about halfway through the clip where the door swings open, following the corner of the door. Or later as the people are walking out. Step through the frames, you'll see that there are always three intact looking frames and two interlaced frames (visible combing effect) for each group of five frames. This pattern is not so pronounced in the earlier scenes with less movement, but it is there.
    Quote Quote  
  4. It's from a crummy VHS tape but, yes, it's telecined. Once again, if the source is the least bit problematic don't ever rely on an automatic analysis such as MeGUI's or AutoGK's. Whoever captured it didn't use a TBC as they should have. Notice all the vertical lines are kind of 'jagged'. TBCs correct that.

    Also, your sample kind of sucks. Most of it has no movement. You want to find 10 well chosen seconds of steady movement.

    I know that telecined material SHOULD follow a pattern of 3 progressive frames and 2 blended ones...
    Three progressive and two interlaced, not blended.
    Quote Quote  
  5. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    So, you DO see that pattern then... Good, at least then I know more about this film and how it was encoded.

    Of COURSE you are supposed to look mainly at the last 20 seconds or so of the clip. The whole point is obviously where they are walking to their car... C'mon Manono... DUH! Sorry I kept you waiting for like 15 seconds until then...
    Quote Quote  
  6. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Hmmm, I'm using Windows Classic Media player to step through it, and darned if I can't make out much difference between frames. I wonder if it is doing something to it on the fly. Try as I might, I cannot see any obvious combing or lines like you do when frames are interlaced (I've seen enough pictures of what they look like!) And YES I am looking at where they are walking outside.

    While I was checking through the player itself I stumbled on this; does this help me identify the source at all. I notice that it looks like it says NO interlace flags, right?

    CLSID: {B98D13E7-55DB-4385-A33D-09FD1BA26338}
    Filter: LAV Splitter Source

    - Connected to:

    CLSID: {04FE9017-F873-410E-871E-AB91661A4EF7}
    Filter: ffdshow Video Decoder
    Pin: In

    - Connection media type:

    Video: MPEG2 Video 704x480 (4:3) 29.97fps

    AM_MEDIA_TYPE:
    majortype: MEDIATYPE_Video {73646976-0000-0010-8000-00AA00389B71}
    subtype: MEDIASUBTYPE_MPEG2_VIDEO {E06D8026-DB46-11CF-B4D1-00805F6CBBEA}
    formattype: FORMAT_MPEG2_VIDEO {E06D80E3-DB46-11CF-B4D1-00805F6CBBEA}
    bFixedSizeSamples: 0
    bTemporalCompression: 1
    lSampleSize: 1
    cbFormat: 282

    VIDEOINFOHEADER:
    rcSource: (0,0)-(704,480)
    rcTarget: (0,0)-(704,480)
    dwBitRate: 0
    dwBitErrorRate: 0
    AvgTimePerFrame: 333667

    VIDEOINFOHEADER2:
    dwInterlaceFlags: 0x00000000
    dwCopyProtectFlags: 0x00000000
    dwPictAspectRatioX: 4
    dwPictAspectRatioY: 3
    dwControlFlags: 0x00000000
    dwReserved2: 0x00000000

    MPEG2VIDEOINFO:
    dwStartTimeCode: 0
    cbSequenceHeader: 150
    dwProfile: 0x00000004
    dwLevel: 0x00000008
    dwFlags: 0x00000000

    BITMAPINFOHEADER:
    biSize: 40
    biWidth: 704
    biHeight: 480
    biPlanes: 1
    biBitCount: 12
    biCompression: 0
    biSizeImage: 506880
    biXPelsPerMeter: 0
    biYPelsPerMeter: 0
    biClrUsed: 0
    biClrImportant: 0

    pbFormat:
    0000: 00 00 00 00 00 00 00 00 c0 02 00 00 e0 01 00 00 ........À...à...
    0010: 00 00 00 00 00 00 00 00 c0 02 00 00 e0 01 00 00 ........À...à...
    0020: 00 00 00 00 00 00 00 00 63 17 05 00 00 00 00 00 ........c.......
    0030: 00 00 00 00 00 00 00 00 04 00 00 00 03 00 00 00 ................
    0040: 00 00 00 00 00 00 00 00 28 00 00 00 c0 02 00 00 ........(...À...
    0050: e0 01 00 00 01 00 0c 00 00 00 00 00 00 bc 07 00 à............¼..
    0060: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
    0070: 00 00 00 00 96 00 00 00 04 00 00 00 08 00 00 00 ....–...........
    0080: 00 00 00 00|00 00 01 b3 2c 01 e0 24 17 67 e3 82 .......³,.à$.gã‚
    0090: 10 10 10 12 10 12 16 16 16 16 16 16 1a 18 1a 1a ................
    00a0: 1a 1a 1a 1a 1a 1a 1a 1a 1a 1c 1c 1c 22 22 22 1c ............""".
    00b0: 1c 1c 1a 1a 1c 1c 20 20 22 22 24 26 24 22 22 22 ...... ""$&$"""
    00c0: 22 26 26 28 28 28 30 30 2e 2e 38 38 3a 44 44 53 "&&(((00..88DS
    00d0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
    00e0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
    00f0: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
    0100: 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 08 ................
    0110: 00 00 01 b5 14 82 00 01 00 00 ...µ.‚....

    - Enumerated media type 0:



    So, if I see that line and there are no interlace flags, does that pretty much mean then if it is coming from a 29.xx source and MediaInfo says it is interlaced, that it is in reality just straight telecined?
    Quote Quote  
  7. Member
    Join Date
    Dec 2005
    Location
    Finland
    Search Comp PM
    Here are frames 450 and 451 as viewed via VirtualDub. If you can't see that difference with your player, it must be processing the video in some manner.

    Looking at any encoding flags or automated detection results is generally speaking useless, they are so often incorrect.
    Image Attached Thumbnails Click image for larger version

Name:	450.jpg
Views:	262
Size:	106.0 KB
ID:	39248  

    Click image for larger version

Name:	451.jpg
Views:	273
Size:	130.9 KB
ID:	39249  

    Quote Quote  
  8. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by Lathe View Post
    Hmmm, I'm using Windows Classic Media player to step through it, and darned if I can't make out much difference between frames.
    Pretty sure that is automatically bob deinterlacing the video. So you wont see any of the interlaced frames as is. I'd use VLC to look for that kind of thing.

    I'd also use

    Code:
    tfm(order=-1).tdecimate()
    instead of the suggested

    Code:
    tfm(order=1).tdecimate(hybrid=1)
    Quote Quote  
  9. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Originally Posted by ajk View Post
    Here are frames 450 and 451 as viewed via VirtualDub. If you can't see that difference with your player, it must be processing the video in some manner.

    Looking at any encoding flags or automated detection results is generally speaking useless, they are so often incorrect.
    Heh..., well of COURSE I can see it now. Geez... Thanks...
    Quote Quote  
  10. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Originally Posted by KarMa View Post
    Originally Posted by Lathe View Post
    Hmmm, I'm using Windows Classic Media player to step through it, and darned if I can't make out much difference between frames.
    Pretty sure that is automatically bob deinterlacing the video. So you wont see any of the interlaced frames as is. I'd use VLC to look for that kind of thing.

    I'd also use

    Code:
    tfm(order=-1).tdecimate()
    instead of the suggested

    Code:
    tfm(order=1).tdecimate(hybrid=1)
    Yep... must be. Normally I use VLC for everything, but I guess I will have to look more closely to see how to step through the frames.

    Ah... May I ask, please, why you would use that particular code parameter () rather than (hybrid=1) Is it because it is a straight IVTC that should be done, and the 'hybrid=1' parameter is causing some kind of interpolation or something?
    Quote Quote  
  11. It is most definitely nothing more than a simple 3:2 telecine. The standard tfm/tdecimate software is all that is needed. Here's the script I used to recover the original 23.976 film and also apply a little denoising. I should point out that I did have a few problems on my computer with the way it was decoding the mkv file, so once I discovered that problem I used ffmpeg to convert it to uncompressed, and the slight weirdness (repeated frames, etc.) went away. This is the ffmpeg command I used to convert to uncompressed:
    Code:
    ffmpeg -i "RoE.mkv" -an -vcodec rawvideo -y output.avi
    Here's the script I used:
    Code:
    loadPlugin("c:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
    loadplugin("C:\Program Files\AviSynth 2.5\plugins\MVTools\mvtools2.dll")
    
    setMTMode(5,6)
    AVISource("E:\fs.avi").KillAudio().converttoYV12(interlaced=true).AssumeTFF()
    setMTMode(2)
    
    tfm() 
    tdecimate() 
    source=AssumeFPS(23.976, false)
    
    output=MDegrain2P(source,4,2,0)
    
    return output
    
    #-------------------
    function MDegrain2P(clip source, int "blksize", int "overlap", int "dct")
    {
    overlap=default(overlap,0) 
    dct=default(dct,0) # use dct=1 for clip with light flicker
    
    super = source.MSuper(pel=2, sharp=1)
    backward_vec2 = super.MAnalyse(isb = true, delta = 1, blksize=blksize, overlap=overlap, dct=dct)
    forward_vec2 = super.MAnalyse(isb = false, delta = 1, blksize=blksize, overlap=overlap, dct=dct)
    backward_vec4 = super.MAnalyse(isb = true, delta = 2, blksize=blksize, overlap=overlap, dct=dct)
    forward_vec4 = super.MAnalyse(isb = false, delta = 2, blksize=blksize, overlap=overlap, dct=dct)
    
    MDegrain2(source,super, backward_vec2,forward_vec2,backward_vec4,forward_vec4,thSAD=400) 
    }
    Finally, here is a link to the denoised result. As you can see from this script, I didn't use very high settings for denoising, so you could certainly go a little further in that direction, if that is what you're trying to do.

    IVTC'd and denoised clip

    You probably don't have to kill the audio, but I've found that it avoids problems with MVTools2 (at least in the version I have used for many years).

    Finally, isn't this simply the "Charlie's Angels" "Nips and Tucks" episode? If so, why are you bothering with some old tape. All those shows have been re-masterd, from the original film, in 16:9 HD.
    Last edited by johnmeyer; 4th Nov 2016 at 13:56. Reason: Added last paragraph
    Quote Quote  
  12. Originally Posted by Lathe View Post
    Of COURSE you are supposed to look mainly at the last 20 seconds or so of the clip. The whole point is obviously where they are walking to their car... C'mon Manono... DUH! Sorry I kept you waiting for like 15 seconds until then...
    No. You're the one asking for help. You're wasting the time of those willing to help. In the future please provide a proper sample. Even ajk mentioned it. People standing around doing nothing are not part of a proper sample. If you don't know how to cut out just the part you want, then ask.
    Hmmm, I'm using Windows Classic Media player to step through it...
    Open the clip in Virtual Dub using an AviSynth script with no filtering, something like:

    FFVideoSource("RoE.mkv")

    and stop trying to use players that may or may not be deinterlacing or bobbing it.

    May I ask, please, why you would use that particular code parameter () rather than (hybrid=1)
    Because it's not a hybrid? Although, in the sample I didn't notice any obvious damage being done.

    Because a simple:

    TFM().TDecimate()


    results in most of the frames being deinterlaced because of the funky source, and because turning off the post-processor doesn't seem to hurt, I'd probably do it like this:

    TFM(pp=0).TDecimate()
    Vinverse()


    Vinverse will pick up any residual interlacing. Or loosen up the CThresh in TFM so that fewer frames are seen as being interlaced to begin with. You can check what gets deinterlaced by putting on a temporary 'Display=True', like this:

    TFM(Display=True).TDecimate()

    For sources that can be field-matched, you don't want the deinterlacer kicking in all the time.
    Last edited by manono; 4th Nov 2016 at 13:58.
    Quote Quote  
  13. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by Lathe View Post
    Ah... May I ask, please, why you would use that particular code parameter () rather than (hybrid=1) Is it because it is a straight IVTC that should be done, and the 'hybrid=1' parameter is causing some kind of interpolation or something?
    Hybrid will allow for the blending of 2 frames together that it believes are unique frames. This is designed for videos that are mostly shot on Film (24fps-23.976fps) but also with a bit of NTSC Video included (29.97fps is all it can really deal with). So if there are only a few scenes with 29.97fps then the thought is that it's better to blend the frames and get the frame rate to 23.976fps that way, instead of simply dropping one of the two most similar frames every 5 frames (a duplicate). Unless you have a reason to believe it's anything but all Film then I would not use Hybrid. Tdecimate is not perfect, and so when using Hybrid you are increasing the chance for it to make mistakes on purely Film content.

    Originally Posted by manono View Post
    Originally Posted by Lathe View Post
    Of COURSE you are supposed to look mainly at the last 20 seconds or so of the clip. The whole point is obviously where they are walking to their car... C'mon Manono... DUH! Sorry I kept you waiting for like 15 seconds until then...
    No. You're the one asking for help. You're wasting the time of those willing to help. In the future please provide a proper sample. Even ajk mentioned it. People standing around doing nothing are not part of a proper sample. If you don't know how to cut out just the part you want, then ask.
    His sample was fine, it had ~10 seconds a little movement and ~10 seconds of movement.
    Last edited by KarMa; 4th Nov 2016 at 15:14.
    Quote Quote  
  14. His clip is not hybrid.
    Quote Quote  
  15. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by johnmeyer View Post
    His clip is not hybrid.
    Agreed.
    Quote Quote  
  16. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Originally Posted by KarMa View Post
    Originally Posted by Lathe View Post
    Ah... May I ask, please, why you would use that particular code parameter () rather than (hybrid=1) Is it because it is a straight IVTC that should be done, and the 'hybrid=1' parameter is causing some kind of interpolation or something?
    Hybrid will allow for the blending of 2 frames together that it believes are unique frames. This is designed for videos that are mostly shot on Film (24fps-23.976fps) but also with a bit of NTSC Video included (29.97fps is all it can really deal with). So if there are only a few scenes with 29.97fps then the thought is that it's better to blend the frames and get the frame rate to 23.976fps that way, instead of simply dropping one of the two most similar frames every 5 frames (a duplicate). Unless you have a reason to believe it's anything but all Film then I would not use Hybrid. Tdecimate is not perfect, and so when using Hybrid you are increasing the chance for it to make mistakes on purely Film content.

    Originally Posted by manono View Post
    Originally Posted by Lathe View Post
    Of COURSE you are supposed to look mainly at the last 20 seconds or so of the clip. The whole point is obviously where they are walking to their car... C'mon Manono... DUH! Sorry I kept you waiting for like 15 seconds until then...
    No. You're the one asking for help. You're wasting the time of those willing to help. In the future please provide a proper sample. Even ajk mentioned it. People standing around doing nothing are not part of a proper sample. If you don't know how to cut out just the part you want, then ask.
    His sample was fine, it had ~10 seconds a little movement and ~10 seconds of movement.
    Thank you kindly for the excellent information. And the comments from you others too...

    So, that is good to know that although MeGUI read it as 'hybrid', it actually is not. Pretty much all of these older films are indeed 'films', so I GUESS that I will not have to use the hybrid setting. I was thinking that that was there to correct something wrong or not done properly when the film was encoded. But, I'm getting from what you say that it is going by the content primarily.
    Quote Quote  
  17. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Originally Posted by johnmeyer View Post
    It is most definitely nothing more than a simple 3:2 telecine. The standard tfm/tdecimate software is all that is needed. Here's the script I used to recover the original 23.976 film and also apply a little denoising. I should point out that I did have a few problems on my computer with the way it was decoding the mkv file, so once I discovered that problem I used ffmpeg to convert it to uncompressed, and the slight weirdness (repeated frames, etc.) went away. This is the ffmpeg command I used to convert to uncompressed:
    Code:
    ffmpeg -i "RoE.mkv" -an -vcodec rawvideo -y output.avi
    Here's the script I used:
    Code:
    loadPlugin("c:\Program Files\AviSynth 2.5\plugins\TIVTC.dll")
    loadplugin("C:\Program Files\AviSynth 2.5\plugins\MVTools\mvtools2.dll")
    
    setMTMode(5,6)
    AVISource("E:\fs.avi").KillAudio().converttoYV12(interlaced=true).AssumeTFF()
    setMTMode(2)
    
    tfm() 
    tdecimate() 
    source=AssumeFPS(23.976, false)
    
    output=MDegrain2P(source,4,2,0)
    
    return output
    
    #-------------------
    function MDegrain2P(clip source, int "blksize", int "overlap", int "dct")
    {
    overlap=default(overlap,0) 
    dct=default(dct,0) # use dct=1 for clip with light flicker
    
    super = source.MSuper(pel=2, sharp=1)
    backward_vec2 = super.MAnalyse(isb = true, delta = 1, blksize=blksize, overlap=overlap, dct=dct)
    forward_vec2 = super.MAnalyse(isb = false, delta = 1, blksize=blksize, overlap=overlap, dct=dct)
    backward_vec4 = super.MAnalyse(isb = true, delta = 2, blksize=blksize, overlap=overlap, dct=dct)
    forward_vec4 = super.MAnalyse(isb = false, delta = 2, blksize=blksize, overlap=overlap, dct=dct)
    
    MDegrain2(source,super, backward_vec2,forward_vec2,backward_vec4,forward_vec4,thSAD=400) 
    }
    Finally, here is a link to the denoised result. As you can see from this script, I didn't use very high settings for denoising, so you could certainly go a little further in that direction, if that is what you're trying to do.

    IVTC'd and denoised clip

    You probably don't have to kill the audio, but I've found that it avoids problems with MVTools2 (at least in the version I have used for many years).

    Finally, isn't this simply the "Charlie's Angels" "Nips and Tucks" episode? If so, why are you bothering with some old tape. All those shows have been re-masterd, from the original film, in 16:9 HD.
    Thank you for the script and help JM! It's probably a little beyond me, but it WILL give me something to study

    And, heh... NO... it is NOT a Charlie's Angels episode... It is a rare 1970 Telefilm called 'RITUAL OF EVIL' that I recorded and transferred long, LONG ago, like many of my other films like this. Most of my stuff is Blu-ray and DVD where necessary in my collection (over 7500 films so far) BUT... I have these from the grand old BETA days! Thus the new interest in Avisynth and learning about it hopefully to improve these lovely, very rare, but rough looking old favourites...
    Quote Quote  
  18. Originally Posted by KarMa View Post
    His sample was fine, it had ~10 seconds a little movement and ~10 seconds of movement.
    Then he should have provided only the final 10 seconds and not the first 10. If he didn't know before that static video is absolutely useless for the purpose, he does now. It's the responsibility of the helpee to make the job of the helper as easy as possible. The good ten seconds would have been more than enough to tell what was going on. Is it all that important? Not really, but rookies should learn to do it correctly.

    This is the same fellow that claimed over on Doom9 to have spent four hours trying to get SRestore to work. Yet he had to ask what the hybrid setting did when 30 seconds with the TDecimate doc would have explained it:

    hybrid -

    Controls how or if video sections (30p) should be dealt with. Possible
    settings:

    0 - no handling
    1 - blend decimation of 30p sections into 24p and leave 24p untouched


    In addition, your own advice was 'flawed':

    Originally Posted by KarMa View Post
    I'd also use

    Code:
    tfm(order=-1).tdecimate()


    Doing that means something like half the video gets deinterlaced and not field matched. If you insist on doing that, you should probably use a better post-processor than TDeint, something like QTGMC. I might even do it myself that way, if not the way I outlined in my previous post.
    Quote Quote  
  19. Member Lathe's Avatar
    Join Date
    Nov 2007
    Location
    Sunny Southern California
    Search Comp PM
    Very good, thanks. I think we are done here...

    BTW... THIS tells me what hybrid does:

    (KarMa)

    "Hybrid will allow for the blending of 2 frames together that it believes are unique frames. This is designed for videos that are mostly shot on Film (24fps-23.976fps) but also with a bit of NTSC Video included (29.97fps is all it can really deal with). So if there are only a few scenes with 29.97fps then the thought is that it's better to blend the frames and get the frame rate to 23.976fps that way, instead of simply dropping one of the two most similar frames every 5 frames (a duplicate). Unless you have a reason to believe it's anything but all Film then I would not use Hybrid. Tdecimate is not perfect, and so when using Hybrid you are increasing the chance for it to make mistakes on purely Film content."

    THIS:

    "hybrid -

    Controls how or if video sections (30p) should be dealt with. Possible
    settings:

    0 - no handling
    1 - blend decimation of 30p sections into 24p and leave 24p untouched "

    Doesn't tell me $hit...

    And, OF COURSE I read that like 30 seconds after seeing the expression 'Hybrid'...

    Please STOP bloody insulting me every dang chance you get and ASS-UMING that I haven't read this or that.

    FWIW, you have been DEAD WRONG every time you have scathingly and impulsively stated that I have not read this or that.

    Please stop doing that...

    (unsubscribing now)
    Last edited by Lathe; 4th Nov 2016 at 22:23.
    Quote Quote  
  20. Originally Posted by Lathe View Post

    Doesn't tell me $hit...
    It tells you exactly the same thing as KarMa did, except using fewer words.
    FWIW, you have been DEAD WRONG every time you have scathingly and impulsively stated that I have not read this or that.
    If you have read something, then stop writing as if you haven't. You could have quoted the TDecimate doc, said you didn't understand what you read and then asked for clarification. It seems pretty self-explanatory to me, though. And, since your source is a movie (the 24p mentioned), it also tells you the MeGUI analysis shouldn't be followed as there are no 30p sections.
    Quote Quote  
  21. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by manono View Post
    In addition, your own advice was 'flawed':
    Originally Posted by KarMa View Post
    I'd also use
    Code:
    tfm(order=-1).tdecimate()
    Doing that means something like half the video gets deinterlaced and not field matched. If you insist on doing that, you should probably use a better post-processor than TDeint, something like QTGMC. I might even do it myself that way, if not the way I outlined in my previous post.
    Any of the combed frames still get field matched back together because any deinterlacing happens after the field matching is done. It seems the source is so bad that it's detecting combing present after the field matching. The default PP=6, is not only deinterlacing the once combed frame, but it's also picking up combing on the originally progressive frames so it's also deinterlacing those too. Which goes to tell you how bad the source is, with an old noisy VHS being transferred to Interlaced MPEG2 @~4500kbps.

    My sources are normally much better than this, especially if it's HD and usually H.264. So I don't normally think about that PP settings, as there should not be that much combing left over to trigger it. A TBC and a lossless/transparent transfer to digital would certainly help OP.

    I did a simple test to see how much PP deinterlacing was happening between PP=0 and the default PP=6. Interestly, the first 10s you considered a waste had a low % of triggering the deinterlacer. While the last 10s with the high motion scene triggered the deinterlacer hard. Probably having to do with the low bitrate MPEG2 video unable to maintain decent interlacing.

    Code:
    LoadPlugin("O:\Megui\Megui\tools\lsmash\LSMASHSource.dll")
    LoadPlugin("O:\Megui\Megui\tools\avisynth_plugin\TIVTC.dll")
    A=LWLibavVideoSource("C:\Users\...\Desktop\RoE.mkv").tfm(order=-1).tdecimate()
    B=LWLibavVideoSource("C:\Users\...\Desktop\RoE.mkv").tfm(order=-1, pp=0).tdecimate()
    Subtract(A,B)
    I used the above script to make a comparison video, provided below. Basically when you see a grey screen, the two videos are the same. When you see something, then deinterlacing has been triggered in the PP=6 version in the areas changed.

    OP could do something like this with QTGMC in progressive mode to try and get rid of the bad blocking/bad interlacing in the MPEG2 source.
    Code:
    LoadPlugin("O:\Megui\Megui\tools\lsmash\LSMASHSource.dll")
    LoadPlugin("O:\Megui\Megui\tools\avisynth_plugin\TIVTC.dll")
    LWLibavVideoSource("C:\Users\...\Desktop\RoE.mkv")
    tfm(order=-1,pp=0).tdecimate()
    QTGMC(Preset="Slower", InputType=1)
    And then there is the problem with OP's source and the heavily clipped luma at the upper end. But that would need to be fixed during the capture stage.
    Image Attached Files
    Last edited by KarMa; 7th Nov 2016 at 01:58.
    Quote Quote  



Similar Threads

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