[FFmpeg-user] Handy field & frame notation
Mark Filipak
markfilipak.windows+ffmpeg at gmail.com
Wed Oct 23 23:11:38 EEST 2019
ffmpeg related (I hope)...
The Problem.
Traditional notation (example: 2-3 pull-down telecine) denotes sets of 4
film frames & 8 fields thus:
...frames... ....fields.....
[1][2][3][4] --> [1][2][2][3][4]
[1][2][3][4][4]
Better, including time sequence, would be:
[1][2][3][4] --> [1] [2] [2] [3] [4]
[1] [2] [3] [4] [4]
But, for obvious reasons, this doesn't work:
[1][2][3][4] --> [1][1][2][2][2][3][3][4][4][4]
The Solution.
The new notation denotes sets of 4 film frames & 8 fields thus:
[A/a][B/b][C/c][D/d] --> [A][a][B][b][B][c][C][d][D][d]
It shows both the mechanics of 2-3 telecine and TFF='1' on a single line.
Details of field notation.
This: [A], and this: [a], are fields.
This: [A] (capitalized), is an odd (aka "top") field.
This: [a] (lower case), is an even (aka "bottom") field.
Examples of field notation:
720x480i24 can be unambiguously characterized by
[A][a][B][b][C][c][D][d] (which is 1/6th second of video).
720x576i25-telecine can be unambiguously characterized by
[A][a][B][b][C][c][D][d][E][e] (which is 1/5th second of video).
720x576i25-telecast can be unambiguously characterized by
[A][b][C][d][E][f][G][h][I][j] (which is 1/5th second of video).
720x480i30-2323telecine can be unambiguously characterized by
[A][a][B][b][B][c][C][d][D][d] (which is 1/6th second of video).
720x480i30-2332telecine can be unambiguously characterized by
[A][a][B][b][B][c][C][c][D][d] (which is 1/6th second of video).
720x480i30-telecast can be unambiguously characterized by
[A][b][C][d][E][f][G][h][I][j][K][l] (which is 1/5th second of video).
Rule: Field characterizations retain their picture identities relative
to their origins, not relative to the sources (or targets) of any
transcodes. Thus, because 1/6th second of a movie is sampled from film
pictures 'A' 'B' 'C' & 'D' (or their equivalent digital camera
captures), a conversion to 720x480i24 retains the origin's 'A' 'B' 'C' &
'D' picture identities and therefore becomes [A][a][B][b][C][c][D][d].
Similarly, a subsequent conversion to 720x480i30-2323telecine carries
the origin's picture identities to the next step as
[A][a][B][b][B][c][C][d][D][d]. In that way, a video's format can be
unambiguously characterized and traced back to its origin no matter what
(or how many) transcodes are done.
Details of frame notation.
This: [A/a], is a frame conveying both the odd & even picture lines.
Example of frame notation:
720x480p24 can be unambiguously characterized by
[A/a][B/b][C/c][D/d] (which is 1/6th second of video).
The Rule for field characterizations also applies to frame
characterizations. Thus, if 720x480p24 is converted to 720x480i24, the
resulting fields are characterized by
[A][a][B][b][C][c][D][d].
Comments are welcome & encouraged.
More information about the ffmpeg-user
mailing list