VideoHelp Forum
+ Reply to Thread
Results 1 to 16 of 16
Thread
  1. Member
    Join Date
    Apr 2017
    Location
    United States
    Search PM
    I have a video sourced as native interlaced in PAL but was also released as NTSC VHS. I've done a capture but I can not tell if it is pure interlaced or has some non-standard telecine pattern. I seem to see 5 interlaced frames followed by 1 progressive frame. Would that make sense? How would I handle it in avisynth? Very small avi sample attached. Thanks for any assistance.
    Image Attached Files
    Quote Quote  
  2. Try:

    Code:
    AviSource("sample_edit.avi") 
    AssumeTFF()
    Yadif(mode=1) # or better, QTGMC() if you don't mind the slow processing
    SRestore(frate=23.976)
    Quote Quote  
  3. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    It was most likely originally captured on Film @24fps, and then was simply sped up to 25fps to be compatible with PAL. Then from PAL it was telecined (with blending) to 29.97fps for North American release. The blending between the fields makes detelecine not really an option. So the other option that Jagabo pointed out was using a bobing deinterlacer to make full frames out of the fields, and then throw out the blended frames with SRestore. But instead of using a frate=23.976fps, I would recommend just going back to the PAL rate of 25fps, otherwise you are potentially throwing out 1 frame every second.
    Quote Quote  
  4. Yes, definitely lots of field blending, so SRestore is needed. I didn't test it, but I assume jagabo's code does the trick. It certainly is the approach I would have taken.
    Quote Quote  
  5. Member
    Join Date
    May 2014
    Location
    Memphis TN, US
    Search PM
    Originally Posted by KarMa View Post
    But instead of using a frate=23.976fps, I would recommend just going back to the PAL rate of 25fps, otherwise you are potentially throwing out 1 frame every second.
    It's really annoying when people take film-speed originals and speed them up to sound squeaky and look panicky at 25fps. Strictly amateur, YouTube level work. If you have to have 25fps for DVD, you can always telecine properly with pulldown for 23.976->25 playback. Motion will look normal and the British guys won't sound castrated.
    - My sister Ann's brother
    Quote Quote  
  6. Originally Posted by LMotlow View Post
    It's really annoying when people take film-speed originals and speed them up to sound squeaky and look panicky at 25fps. Strictly amateur
    No, it was the industry standard for many years. Almost all movies broadcast on analog PAL TV were done that way.
    Quote Quote  
  7. Originally Posted by KarMa View Post
    But instead of using a frate=23.976fps, I would recommend just going back to the PAL rate of 25fps, otherwise you are potentially throwing out 1 frame every second.
    I was all set to contradict you, as field-blended DVD releases from film to NTSC do occur. But you're right. And I apologize for thinking otherwise before studying the sample. So, for the sample:

    Yadif(Order=1,Mode=1)#field order might not be the same for the complete video. And a better bobber can be used.
    SRestore(Frate=25)
    Quote Quote  
  8. Originally Posted by LMotlow View Post
    It's really annoying when people take film-speed originals and speed them up to sound squeaky and look panicky at 25fps. Strictly amateur, YouTube level work. If you have to have 25fps for DVD, you can always telecine properly with pulldown for 23.976->25 playback. Motion will look normal and the British guys won't sound castrated.
    That is how film has been broadcast in the UK and most of Europe since the beginning of time. Remember, until twenty years ago we didn't have digital video and all the things we do now that seem trivial had to be done with analog circuits.

    And, of all the broadcasters in the world, I think the BBC had (and has) some of the highest standards in the world, much like PBS has long been the leader in producing the highest quality video in this country.
    Quote Quote  
  9. Member
    Join Date
    May 2014
    Location
    Memphis TN, US
    Search PM
    It's still annoying. And now that we have the means to correct those issues, there's no excuse for repeating the same sloppy work.

    Agreed about general PBS quality, even when it comes to recent British and Australian imports and conversions, where the results that low-profit PBS can get visually put much high-profit cable broadcast garbage to shame.
    - My sister Ann's brother
    Quote Quote  
  10. It's hard to tell if the short clip should be 23.976 of 25.0 fps. In any case, it needs SRestore() to reduce the blending.
    Quote Quote  
  11. Originally Posted by LMotlow View Post
    It's still annoying. And now that we have the means to correct those issues, there's no excuse for repeating the same sloppy work.

    Agreed about general PBS quality, even when it comes to recent British and Australian imports and conversions, where the results that low-profit PBS can get visually put much high-profit cable broadcast garbage to shame.
    Actually, PBS is a non-profit, by design. They have no profits. However, historically (not sure about right now) they had huge amounts of money available, thanks to fund-raising and government support. The situation, therefore, was actually almost the opposite of what you describe: it was the commercial ventures, who had to make a profit, who had less time and money to produce the best stuff.

    As for "no excuse for repeating the same sloppy work," you assume that your method will produce results which look better than the 24 fps --> 25 fps speed up. Speeding up from 24 to 25 fps is actually quite subtle and may possibly be less visually upsetting than repeating one field every half a second second. I haven't tried it, but I've done a lot of similar work, and I expect that duplicating a field twice a second is such a low duty cycle that it would be more noticeable than the 24 fps --> 30 fps film-to-NTSC interlaced video pulldown pattern which duplicates every fourth field. Also, I am reasonably certain that, with modern pitch control algorithms, most film that is sped up by 4% has its pitch corrected.

    Standards conversion is amazingly difficult to do, and no matter what approach you take, you WILL see artifacts.
    Last edited by johnmeyer; 9th Apr 2018 at 11:04. Reason: grammar
    Quote Quote  
  12. Originally Posted by jagabo View Post
    It's hard to tell if the short clip should be 23.976 of 25.0 fps. In any case, it needs SRestore() to reduce the blending.
    One way to tell which it should be is to put on:

    Yadif(Order=1,Mode=1)
    SRestore(Frate=29.97)


    and then note if the dupe frames (or residual blends) come every 5 or 6 frames. Also, the OP mentioned:

    I seem to see 5 interlaced frames followed by 1 progressive frame.
    which, if true, (it is) is a dead giveaway of it being from a PAL source.
    Quote Quote  
  13. Member
    Join Date
    May 2014
    Location
    Memphis TN, US
    Search PM
    Originally Posted by johnmeyer View Post
    Originally Posted by LMotlow View Post
    It's still annoying. And now that we have the means to correct those issues, there's no excuse for repeating the same sloppy work.

    Agreed about general PBS quality, even when it comes to recent British and Australian imports and conversions, where the results that low-profit PBS can get visually put much high-profit cable broadcast garbage to shame.
    Actually, PBS is a non-profit, by design. They have no profits. However, historically (not sure about right now) they had huge amounts of money available, thanks to fund-raising and government support. The situation, therefore, was actually almost the opposite of what you describe: it was the commercial ventures, who had to make a profit, who had less time and money to produce the best stuff.

    As for "no excuse for repeating the same sloppy work," you assume that your method will produce results which look better than the 24 fps --> 25 fps speed up. Speeding up from 24 to 25 fps is actually quite subtle and may possibly be less visually upsetting than repeating one field every half a second second. I haven't tried it, but I've done a lot of similar work, and I expect that duplicating a field twice a second is such a low duty cycle that it would be more noticeable than the 24 fps --> 30 fps film-to-NTSC interlaced video pulldown pattern which duplicates every fourth field. Also, I am reasonably certain that, with modern pitch control algorithms, most film that is sped up by 4% has its pitch corrected.

    Standards conversion is amazingly difficult to do, and no matter what approach you take, you WILL see artifacts.
    The PBS shows I've seen from UK and Australia are 23.976 or 24fps originals using various forms of pulldown to get NTSC. I've converted a few of them to PAL/pulldown for other users and they report that the playback is superior to speeded-up PAL. of course if you can't tell the difference between the original film rate and speeded-up PAL (a video professional should easily be able to) then it wouldn't matter.

    As for PBS I'd consider low-profit and no-profit to be equivalent, mostly a use of terms. Government support of PBS in the USA is worse than in some third world countries and far less than in the UK. Profit-based broadcasters spend most of their huge income on executive perks. They couldn't care less what you see on the screen.
    - My sister Ann's brother
    Quote Quote  
  14. Originally Posted by LMotlow View Post
    I've converted a few of them to PAL/pulldown for other users and they report that the playback is superior to speeded-up PAL.
    I'll assume you used DGPulldown to apply soft pulldown. I have yet to see soft pulldown used for this in a retail release. I've seen a few with hard pulldown, but only a few. The vast majority have either been the progressive speeded up variety or field-blended. Maybe the companies doing these standards conversions have their reasons, reasons such a compatibilities with DVD players if soft pulldown is used. I don't know, but I've never seen one that was soft telecined. I've also read of complaints from PAL user that pulldown in any form creates playback that is, to their eyes, jerky. And 3:2:3:2:2 pulldown would be more jerky than 3:2 pulldown.
    Quote Quote  
  15. Dinosaur Supervisor KarMa's Avatar
    Join Date
    Jul 2015
    Location
    US
    Search Comp PM
    Originally Posted by LMotlow View Post
    The PBS shows I've seen from UK and Australia are 23.976 or 24fps originals using various forms of pulldown to get NTSC. I've converted a few of them to PAL/pulldown for other users and they report that the playback is superior to speeded-up PAL. of course if you can't tell the difference between the original film rate and speeded-up PAL (a video professional should easily be able to) then it wouldn't matter.
    If it was solely made for the UK, other Euro nations, or Australia; it's mostly certain that it's going to be produced for PAL rates. If it's a collaboration between a studio in one of those countries and PBS (or a affiliated PBS station), then it's still going to be most likely produced at PAL rates. Things like NOVA, Nature, Secrets of the Dead, and other PBS series where episodes were either licensed from foreign companies (BBC, Ch4, SBS) or collaborated with foreign production companies, it's just about always going to be PAL. It's certainly possible that the cameras used when capturing video for these programs are set to 24fps, but the copies PBS (or PBS stations) get are just about always 25fps which is then telecied for 29.97fps broadcast. There are exceptions to this like with Big Blue Live and Wild Alaska Live which was a co-production between BBC and PBS and I'm assuming heavily reliant on PBS infrastructure. It aired in the US mostly at 60i with some additional telecined 25fps content in the same program, but in the UK I'm going to assume it was all 25fps.

    Capturing 24fps for 25fps broadcasts is not in the interests of those foreign markets so I can't really see 24fps video captures being much of a thing. Outside of maybe freelancers who think 24fps will appeal to all markets or movie makers who are trying to make a documentary at Film frame rates first with international broadcasting being the second concern. From what I've seen/heard about BBC/ITV/Ch4 broadcasts, is that 24fps-->25fps telecine is extremely rare (assuming it has ever happened in the HDTV age in the UK). Instead the old Film to PAL speed up hack is still the norm today with those broadcasters. So they tend to just want to avoid 24fps as much possible, especially if they have any control over the production.
    Quote Quote  
  16. Originally Posted by manono View Post
    Originally Posted by jagabo View Post
    It's hard to tell if the short clip should be 23.976 of 25.0 fps. In any case, it needs SRestore() to reduce the blending.
    One way to tell which it should be is to put on:

    Yadif(Order=1,Mode=1)
    SRestore(Frate=29.97)


    and then note if the dupe frames (or residual blends) come every 5 or 6 frames.
    Nice trick.
    Quote Quote  



Similar Threads

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