VideoHelp Forum
+ Reply to Thread
Results 1 to 10 of 10
Thread
  1. Is there anyway to identify a video that has been recorded with Progressive Segmented Frame?
    Canon calls it PF, ie PF24 or PF30. Apparently they wrap up their progressive video in a 60i containers with no indication that the video is actually progressive.

    All the metadata information indicates that it is interlaced.
    thanks
    Quote Quote  
  2. Originally Posted by TwoJ View Post
    Is there anyway to identify a video that has been recorded with Progressive Segmented Frame?
    Canon calls it PF, ie PF24 or PF30. Apparently they wrap up their progressive video in a 60i containers with no indication that the video is actually progressive.

    All the metadata information indicates that it is interlaced.
    thanks
    Not with metadata

    If you double rate deinterlace and step through frame by frame in a scene with motion, you will get duplicate frames. This would be 29.97p content. If you get 3-2 repeats, it's 23.976p content

    The proper way to handle 29.97p content would be to weave the segmented frames, so you don't degrade it . The proper way to handle the 23.976p content is to inverse telecine (field match and decimate) to return the orginal progressive frames. In both cases, deinterlacing will degrade the footage
    Quote Quote  
  3. aBigMeanie aedipuss's Avatar
    Join Date
    Oct 2005
    Location
    666th portal
    Search Comp PM
    i don't recall there being a "pf24" on any canon HDV cam. "pf30" HDV i shot all the time. it's recorded as 60i. 2 half frames recorded at the same time that when combined makes it 30p. doesn't make a shit's difference if you use it on a timeline as 60i or tell the editing program to "interpret" it as 30p. same outcome.
    --
    "a lot of people are better dead" - prisoner KSC2-303
    Quote Quote  
  4. Originally Posted by aedipuss View Post
    i don't recall there being a "pf24" on any canon HDV cam. "pf30" HDV i shot all the time. it's recorded as 60i. 2 half frames recorded at the same time that when combined makes it 30p. doesn't make a shit's difference if you use it on a timeline as 60i or tell the editing program to "interpret" it as 30p. same outcome.

    PF24 was featured in some AVCHD cameras

    If "29.97p" video gets treated as interlaced by a NLE or other program, it will get deinterlaced and degraded when you export a progressive format.

    If you keep it as fields on an interlaced project timeline, and export it as 29.97p in interlaced stream (like the original 29.97i or 59.94i if you want to call it, same thing), then it's ok. But then you're stuck with this same PSF format instead of native progressive
    Quote Quote  
  5. aBigMeanie aedipuss's Avatar
    Join Date
    Oct 2005
    Location
    666th portal
    Search Comp PM
    don't know about 24pf but 30pf was a dream to work with. the 2 60i fields combined perfectly into 30p and there was nothing for any de-interlacer to **** up.
    --
    "a lot of people are better dead" - prisoner KSC2-303
    Quote Quote  
  6. Originally Posted by aedipuss View Post
    don't know about 24pf but 30pf was a dream to work with. the 2 60i fields combined perfectly into 30p and there was nothing for any de-interlacer to **** up.

    30PF got messed up all the time. Many threads on this, many forums. People would just import into a NLE and not know it was getting mangled. It had to be handled correctly, interpreted correctly, setup on a progressive timeline for exporting progressive - hence the early warning in my first reply in this thread.

    I still have some PF30 footage on some old HDD's if anyone is interested in seeing the difference
    Quote Quote  
  7. i was really looking for a way to automate a scan to check files - there are just too many to go one by one.
    One of the other threads on the subject mentioned converting to mkv - i tried an initial attempt and i believe the tbr value correctly reflects that 60i is 59.94 and PF30 is 29.97. I can probably script it to convert the file, grab that value, but it just seems crazy that there isn't a better way.
    Quote Quote  
  8. Originally Posted by TwoJ View Post
    i was really looking for a way to automate a scan to check files - there are just too many to go one by one.
    One of the other threads on the subject mentioned converting to mkv - i tried an initial attempt and i believe the tbr value correctly reflects that 60i is 59.94 and PF30 is 29.97. I can probably script it to convert the file, grab that value, but it just seems crazy that there isn't a better way.
    Unfortunately, tbr is not necessarily accurate or reliable indicator, and it can vary on how it's converted to mkv

    Here is an example, of 30PF content re-wrapped into a MKV with mkvmerge
    Code:
        Stream #0:0: Video: h264 (High), yuv420p(top first), 1920x1080 [SAR 1:1 DAR
    16:9], 59.94 fps, 59.94 tbr, 1k tbn, 59.94 tbc (default)
    If you re-wrap true interlaced "60i" content with ffmpeg, it will say 29.97 tbr, even though it's interlaced

    You need your eyes, or some sort of motion detection script that detection motion in every field for a motion shot to distinguish between interlaced content vs. PF30 vs. PF24
    Quote Quote  
  9. I'm not trying to belabor the issue, i just want to cover all the possibilities. I did the process of checking the true 60i vs the PsF30 and your right, tbr is the same.
    As i said, i just have too many videos to check each one, and i am still shaking my head that Canon didn't put in any way to differentiate a real 60i from a PF30!

    I did come accross one comment who had said/speculated that the timecode / timestamp of the individual fields might show that in the real 60i the timestamp of the individual fields would be different since there would be a time difference between each field, but in the 30PsF the timestamp of the frame would be the same even if it was broken into 2 sequential fields, ie 60i would have different timecode and timestamp for each frame but the 30PsF, each frame would likewise have a different timecode but 2 sequential fields could have the same timestamp?

    I admit that i am way out of my depth on understanding how the video is stored, whether there are timecodes/timestamps for individual fields, and how one could read this information.
    but there always seems people here that understand this far more. Thanks
    Quote Quote  
  10. Originally Posted by TwoJ View Post
    I'm not trying to belabor the issue, i just want to cover all the possibilities. I did the process of checking the true 60i vs the PsF30 and your right, tbr is the same.
    As i said, i just have too many videos to check each one, and i am still shaking my head that Canon didn't put in any way to differentiate a real 60i from a PF30!

    I did come accross one comment who had said/speculated that the timecode / timestamp of the individual fields might show that in the real 60i the timestamp of the individual fields would be different since there would be a time difference between each field, but in the 30PsF the timestamp of the frame would be the same even if it was broken into 2 sequential fields, ie 60i would have different timecode and timestamp for each frame but the 30PsF, each frame would likewise have a different timecode but 2 sequential fields could have the same timestamp?

    I admit that i am way out of my depth on understanding how the video is stored, whether there are timecodes/timestamps for individual fields, and how one could read this information.
    but there always seems people here that understand this far more. Thanks
    Nope - timestamps are the same for all 3 cases (and for PF24) , that's the whole reason why canon (and others) used this format, to make it compatible with legacy NTSC standards. Instead of native 23.976 pN , or 29.97 pN.

    They are all encoded as fields, and all are transmitted as 59.94 fields per second interlaced. All 3 need to be treated differently because, obviously, all 3 are different
    Quote Quote  



Similar Threads

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