25fps 1080i mkv convert to 24fps 1080p or 720p mkv?

[brent1a]

I need to get a 25fps 1080i mkv to 24fps 1080p or 720p mkv.
So far, everything I've been reading has pushed me in the direction of deinterlacing. But I don't think that sounds right. I believe I need to do an Inverse Telecine to get what I want.....correct?
So far my inital plan has gotten as far as changing the audio/video file from 25fps to 24fps:
1. Use TSMuxer to change video framerate to 24fps
2. Use Clown-BD to convert DTS to 640 kb/s AC3
3. Use eac3togui to convert pal timed audio to 24fps timed audio
..........or will that screw it up before I even get started?

But I'm a n00b when it comes to IVTC to 1080p or 720p.

Any help would be greatly appreciated.

[jagabo]

Does your 25 fps source really contain 50 different fields per second? Or is it film sped up to 25 fps?

[brent1a]

I don't have access to it right now but I will try to let you know as soon as I can get back to it.

I can tell you it's an HDTV broadcast of a video/film I know is originally 24 fps. It has just been converted to 25 fps for viewing on PAL and converted to mkv.

[brent1a]

General / Container Stream # 1
Total Video Streams for this File -> 1
Total Audio Streams for this File -> 4
Video Codecs Used -> AVC
Audio Codecs Used -> AC3 / AC3 / DTS / DTS
File Format -> Matroska
Play Time -> 3h 7mn
Total File Size -> 17.0 GiB
Total Stream BitRate -> 13.0 Mbps
Encoded with -> mkvmerge v2.6.0 ('Kelly watch the Stars') built on Mar 24 2009 15:23:17
Encoding Library -> libebml v0.7.7 + libmatroska v0.8.1
Video Stream # 1
Codec (Human Name) -> AVC
Codec (FourCC) -> V_MPEG4/ISO/AVC
Codec Profile -> Main@L4.0
Frame Width -> 1 920 pixels
Frame Height -> 1 080 pixels
Frame Rate -> 25.000 fps
Total Frames -> 280700
Display Aspect Ratio -> 16/9
Scan Type -> Interlaced
Scan Order -> TFF
Colorimetry -> 4:2:0
Codec Settings (Summary) -> CABAC / 4 Ref Frames
QF (like Gordian Knot) -> 0.205
Codec Settings (CABAC) -> Yes
Video Stream Length -> 3h 7mn 8s 0ms
Video Stream BitRate -> 10.6 Mbps
Bit Depth -> 24 bits
Video Stream Size -> 13.9 GiB (81%)
Audio Stream # 1
Codec -> DTS
Codec (FourCC) -> A_DTS
Audio Stream Length -> 3h 7mn 7s 980ms
Audio Stream BitRate -> 754 Kbps
Audio Stream BitRate Mode -> CBR
Number of Audio Channels -> 6
Audio Channel's Positions -> Front: L C R, Surround: L R, LFE
Sampling Rate -> 48.0 KHz
Bit Depth -> 16 bits
Audio Stream Delay -> 6ms
Audio Stream Size -> 1 010 MiB (6%)
Audio Stream Language -> English

[jagabo]

PAL broadcasts usually just speed the film up to 25 fps. Although it is broadcast interlaced each pair of fields usually contains one progressive frame. If that is the case with your video (no visible comb artifacts) you can just slow the frame rate down to 23.976 fps. It's also possible that a capture will pair together fields from two separate film frames:

film frames: A B C D
brodcast TFF: A a B b C c D d (cap = top field, lower case = bottom field)
captured TFF: Aa Bb Cc Dd (frames look progressive, no need to deinterlace)
captured BFF: (first A missing) aB bC cD dE (frame look interlaced)

The fields in the latter can simply be recombined to make progressive frames:

BFF cap: aB bC cD dE
separated into fields: a B b C c D d E
throw away first field, recombine the rest: Bb Cc Dd Ee

You now have progressive frames and you can slow the frame rate down to 23.976...

[brent1a]

PAL broadcasts usually just speed the film up to 25 fps. Although it is broadcast interlaced each pair of fields usually contains one progressive frame. If that is the case with your video (no visible comb artifacts) you can just slow the frame rate down to 23.976 fps.

I will (hopefully) be able to do that. I will look into that first. If that doesn't work.....

The fields in the latter can simply be recombined to make progressive frames:

BFF cap: aB bC cD dE
separated into fields: a B b C c D d E
throw away first field, recombine the rest: Bb Cc Dd Ee

How would I go about doing the above?

[jagabo]

I usually do it in AviSynth:

SeparateFields()
Trim(1,0) #throw away first field
Weave() #weave the remaining fields back together

You can also do this in VirtualDub with it's Deinterlace (discard, double frame rate), cutting out the first field, and Interlace again. A warning though: VirtualDub does not handle interlaced chroma channels of YV12 video properly (most MPEG 1/2/4 sources).