VideoHelp Forum
+ Reply to Thread
Results 1 to 27 of 27
Thread
  1. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi, Video Kings!

    I am back with another crappy DVD to ask advice about. I have DVD that I think exhibits the symptoms of poor pal to ntsc conversion. Here is a clip:

    http://files.videohelp.com/u/183506/t%20dragon.demuxed.m2v

    Am I right? Is this a poor pal/ntsc issue? Would this be best fixed by the following script:

    Mpeg2Source("TATTOOED DRAGON.d2v")
    AssumeTFF().QTGMC(Preset="PLACEBO")
    SRestore()
    AssumeFPS(23.976)

    Thanks to the help of Manono, I can now adjust audio in besweet to the new frame rate.

    Thanks to all of you for your help! I hope I have the right prescription!

    Best,

    h
    Quote Quote  
  2. Yep, a standard field-blended typical crap Hong Kong PAL2NTSC DVD. I don't usually set a field order unless I notice something screwy although I don't guess it hurts, unless you get it wrong.
    Thanks to the help of Manono, I can now adjust audio in besweet to the new frame rate.
    You're welcome, and now thanks to a lot of people you're getting better and better at figuring out your video sources, too.

    By the way, although I always use BeSweet to do the audio slowdown, I sometimes notice a delay of up to -200ms when done. Be sure to check your audio against the video when all done. Find some sharp sound in various places in the movie with which to test. If you do find a constant delay, remove it using DelayCut.
    Quote Quote  
  3. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi Manono & Everyone!

    I am glad to hear that I am coming along in my abilities. This stuff is so challenging and I appreciate everyone's help and advice.

    I was taking a look at the resulting video from the script in my first post and I noticed something. Here is a clip from the m2v file that I created from the lagarith AVI made by AVIsynth:

    http://files.videohelp.com/u/183506/TATTOOED%20DRAGON%2023.976.demuxed.m2v

    Both the AVI & m2v file exhibit something I can only describe as stuttering during panning shots. I think I might have seen this in other commercial DVD's. Is this is what is called "jutter"? Is there a remedy for this?

    Also, I now have a progressive file. Do progressive files have any field order? When I played my new m2v file in the DGIndex preview window. It says frame type "progressive", field order "top". Is this normal?

    Thanks again for all of the help!!

    Best,

    h

    PS: I will take a look at delaycut!
    Quote Quote  
  4. That m2v file is progressive with 3:2 pulldown flags, bff.
    Quote Quote  
  5. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Originally Posted by jagabo View Post
    That m2v file is progressive with 3:2 pulldown flags, bff.
    Hi!

    Is that normal? Have I done something wrong?

    Thanks!
    Quote Quote  
  6. TFF is more common with pulldown flags but it should play properly in any DVD player, or software that's set up properly.
    Quote Quote  
  7. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi,

    What would I do to make it TFF? This is my script:

    Mpeg2Source("TATTOOED DRAGON.d2v")
    AssumeTFF().QTGMC(Preset="PLACEBO")
    SRestore()
    AssumeFPS(23.976)

    I would have thought it was TFF.

    I found this online:

    "Progressive has no fields, thus BFF or TFF does not exist. That's why TMPGenc does not allow it. What CCE says for TFF refers strictly to the interlaced source."

    Does this mean my file is not truly progressive?

    Thank you, Video Kings!
    Quote Quote  
  8. Both the AVI & m2v file exhibit something I can only describe as stuttering during panning shots.
    What you're seeing is either normal 24fps stutter, or the stuttering induced by 3:2 pulldown, or both. There's nothing wrong with that second sample, in the sense of it maybe missing frames.
    Originally Posted by hizzy7 View Post
    Does this mean my file is not truly progressive?
    It's progressive. Just run the second sample through DGIndex using the Preview.

    By the way, with progressive sources scan type should be set for Zigzag and not Alternate in your MPEG-2 encoder. DCT Precision should almost always be set for 10.
    Quote Quote  
  9. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi Everyone!

    Is there any remedy for stuttering induced by 3:2 pulldown or 24fps stutter?

    I was using adobe media encoder, but I will switch. There is no option for Zigzag or Alternate that I can see.

    Thanks!
    Quote Quote  
  10. Originally Posted by hizzy7 View Post
    Is there any remedy for stuttering induced by 3:2 pulldown or 24fps stutter?
    Not really. Almost everything you've watch on TV (except live sports, soap operas, and news) has had it. You can get a 120 Hz TV and watch it with 5:5 duplicates -- that's a little less jerky. Or watch with motion interpolation and get annoying artifacts now and then.

    Are you sure your seeing normal 3:2 pulldown judder? Maybe your player is playing the video wrong? The camerawork is a little jerky too. Is what you're seeing worse than the attached p24 file? You want it to look more like the attached p60 file?
    Attached Files
    Last edited by jagabo; 3rd Aug 2014 at 20:25.
    Quote Quote  
  11. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi!

    I don't know if I am seeing jutter... I do notice something during shots when the camera is panning. I am only guessing it is jutter. I saw it in the video clip that I posted in post # 3.

    This discussion has got me interested in encoding. Does pulldown follow any rule of thumb. Should I always use 3:2 pulldown? I am playing around with encoders with many more options than I am used to. There are so many options! Are there optimal settings?

    Thank you!

    Best,

    h
    Quote Quote  
  12. Originally Posted by hizzy7 View Post
    Should I always use 3:2 pulldown?
    No. Not for anything encoded as either progressive or interlaced 29.97fps. Not for your Hong Kong movie if you had kept it at 25fps after the unblending. Then you'd apply a 3:2:3:2:2 pulldown (set DGPulldown for 25->29.97fps). Not for silent films or 16mm films at a base framerate less than 23.976fps. 3:2 pulldown is applied only to something encoded as progressive 23.976fps.

    And no MPEG-2 encoder will be able to apply any pulldown other than 3:2. The rest you do afterwards using DGPulldown.
    Quote Quote  
  13. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi!

    When I create an m2v file with adobe media encoder, I could sweat that it applies pulldown for me, but, being a novice, I shouldn't swear on anything, I guess!

    After you mentioned scan types, Manono, I looked at my adobe encoder and could not find any setting for that. I then took a look at mainconcept totalcode, and found many more settings, among them scanning order and DC precision.

    Mainconcept also can apply pulldown, or so it appears to me. They offer the choice of 2:3, 3:2, and 2:2 pulldown. They also offer chroma formats, quantization scale, and some other options that adobe didn't. I will look at some of the free programs next.
    Quote Quote  
  14. What you should know:

    Standard definition TV was always interlaced. You saw 59.94 fields per second on the screen. Each field was half the screen. The "top" field was every other scanline starting with the first (0,2,4,6...478), the "bottom" field was every other scan line starting with the second (1,3,5,7...479). Interlaced TV always alternated between top and bottom fields: top, bottom top, bottom, top, bottom... TVs couldn't even display an entire frame at once so you never saw an entire frame. With a video camera every single field could come from a different point in time. That delivered very smooth motion for high action content like sports. At the cost of a little flicker and reduced vertical resolution.

    Film is shot at 24 frames per second. To display film on an interlaced SD TV you have to convert 24 progressive frames into 59.94 interlaced fields per second. First the film is slowed to 23.976 fps, then pulldown is applied. It might be easier to think of it this way. The film is slowed to 23.976 fps and frames are repeated in a 3:2 pattern. So six film frames:

    Code:
    A B C D E F
    are repeated with a 3:2 pattern:

    Code:
    A A A B B C C C D D E E E F F
    Then fields are pulled out of those frames alternating between top and bottom fields:

    Code:
    At Ab At Bb Bt Cb Ct Cb Dt Db Et Eb Et Fb Ft
    Note how that sequence always alternates between top and bottom fields.

    For the sake of simplicity I'm going to use 24, 30 and 60 instead of 23.976, 29.97 and 59.94 from here on out.

    In addition to the inherent jerkiness and flicker of 24 fps that adds a judder because frame A is seen on the screen for 3/60 of a second, whereas frame B is only seen for 2/60 of a second. This judder is most apparent with high contrast content and moderate motion -- like smooth panning shots. See the 24v30v60.avi video in the following post:

    http://forum.videohelp.com/threads/307004-Best-framerate-conversion-%28eg-23-97-to-30-...=1#post1888926

    The top row is 24p with a 3:2 repeat pattern to make 60 fps. The middle row is 30p with a 2:2 repeat pattern to make 60 fps. The bottom row is 60p. The middle row doesn't look very smooth and appears to flicker because 30 fps simply to slow to deliver smooth flicker free motion. You can see that 24 fps has judder in addition to being jerky/flickery. 60P looks nice and smooth (unless your computer monitor isn't running 60 fps, but even then it probably looks much smoother than the others).

    When a computer captures 60 field per second interlaced video it weaves pairs of field together into frames before storing or compressing them. This creates 30 fps interlaced video. The capture device can start with a top field then add the next field (a bottom field) to produces top field first video. Or it can start with a bottom field then add the next field (a top field) to produce bottom field first video. A playback device has to know whether the video is TFF or BFF so it can send the fields to the TV in the correct order.

    DVD was designed in the era of interlaced TV. Early DVD players weren't expected to be smart enough to perform frame/field rate conversions themselves. So the DVDs themselves tell the player how to produce fields from frames -- 3:2 pulldown flags. That is, 24 fps material is stored as 24 fps progressive frames with instructions that tell the player how to produce 60 fields per second from those frames. It's also possible to perform hard pulldown where the pulldown and field weaving is done before encoding as 30 interlaced frames per second. If you're making DVDs you are stuck with using one of those methods.
    Last edited by jagabo; 3rd Aug 2014 at 21:41.
    Quote Quote  
  15. Originally Posted by hizzy7 View Post
    When I create an m2v file with adobe media encoder, I could sweat that it applies pulldown for me, but, being a novice, I shouldn't swear on anything, I guess!
    It does apply soft 3:2 pulldown. One of many ways to confirm that is to run the Preview in DGIndex. If it has field repeats, it has soft pulldown of some type. And since, as I mentioned, the only kind of pulldown any MPEG-2 encoder will apply for progressive 23.976fps sources is 3:2 (or 2:3) pulldown, then your second sample has it. GSpot can confirm it, too.
    Quote Quote  
  16. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi!

    I hope this doesn't sound stupid, but is 3:2 pulldown & 2:3 pulldown the same thing? I have seen people say that online. My mpeg encoder offers 3:2, 2:3, and 2:2. I have now set the scanning order for zig zag

    Thanks!

    h
    Quote Quote  
  17. Originally Posted by hizzy7 View Post
    Hi!

    I hope this doesn't sound stupid, but is 3:2 pulldown & 2:3 pulldown the same thing?
    Yes, it's just a matter of whether you start with a 2 or 3. Remember, the pattern repeats over and over.
    Code:
    3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2...
    Quote Quote  
  18. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi!

    Is there a general preference? 3:2 or 2:3? Is one of them the industry standard?

    Thanks!
    Quote Quote  
  19. About the only time you might use one rather than the other is if you are encoding in sections that you will append later. If one segment ends on a 3 you want to start the next segment with a 2. And vice versa.
    Quote Quote  
  20. Member Cornucopia's Avatar
    Join Date: Oct 2001
    Location: Deep in the Heart of Texas
    Search Comp PM
    There is a standard: 2:3, however, it is only loosely followed, so you'll see just as many 3:2 as 2:3.

    Pick which one works with your system/workflow.

    Scott
    "When will the rhetorical questions end?!" - George Carlin
    Quote Quote  
  21. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Hi Jagabo & all,

    I was wondering - In post 10, Jagabo posted a clip of my video at 60p. Is it my eye or does 60p look way different than film? It looks so strange!!

    Thanks!!!
    Quote Quote  
  22. Originally Posted by hizzy7 View Post
    In post 10, Jagabo posted a clip of my video at 60p. Is it my eye or does 60p look way different than film?
    Yes, it looks more live video -- the "soap opera look: that many film aficionados complain about.
    Quote Quote  
  23. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Originally Posted by jagabo View Post
    Originally Posted by hizzy7 View Post
    Hi!

    I hope this doesn't sound stupid, but is 3:2 pulldown & 2:3 pulldown the same thing?
    Yes, it's just a matter of whether you start with a 2 or 3. Remember, the pattern repeats over and over.
    Code:
    3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2...
    Hi!

    It's funny that they are considered the same, despite the numbers being different. One is 3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2..., and the other is 2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3:2:3... They are kind of different, even if the effect is the same. What a strange thing!

    Thanks!
    Quote Quote  
  24. As I told you, the only difference is whether they start with a 2 or a 3. The end result is the same, 59.94 fields per second, the same running time, the same amount of pulldown judder.
    Quote Quote  
  25. Or to look at it another way, if you create a 2:3 video and trim just 1 frame from the start, you've got a 3:2 video. (I'm already regretting posting this.)
    Quote Quote  
  26. Member
    Join Date: Feb 2010
    Location: canada
    Search PM
    Originally Posted by manono View Post
    By the way, with progressive sources scan type should be set for Zigzag and not Alternate in your MPEG-2 encoder. DCT Precision should almost always be set for 10.
    I was just wondering, what program told you which scan type I used? Mediainfo doesn't have a field for Zigzag or Alternate.

    Thanks!!!
    Quote Quote  
  27. Member
    Join Date: Sep 2007
    Location: Canada
    Search Comp PM
    Originally Posted by hizzy7 View Post
    Originally Posted by manono View Post
    By the way, with progressive sources scan type should be set for Zigzag and not Alternate in your MPEG-2 encoder. DCT Precision should almost always be set for 10.
    I was just wondering, what program told you which scan type I used? Mediainfo doesn't have a field for Zigzag or Alternate.

    Thanks!!!

    You can use various stream analyzers.

    Restream can also tell you that information, look under the picture coding extension
    Quote Quote