filtering hybrid DVD footage and encoding to AVCHD or BluRay

[OvejaNegra]

Let's say i have a typical hybrid anime DVD with the usuall problems (rainbow, dothcrawl, and of course mixed frame rate)

after the usuall filtering + tivtc i have a vfr video with a time code

i cannot put that on avchd or blu ray with x264, those require constant frame rate.

what can i do?

TFM will match the frames and remove the combing but i still have duplicate frames.

Yes, i know i can produce constant frame rate with tivtc but that will introduce stutering on the pannig scenes or any scene with 29.976 fps.

Yes, i know i can encode the original stream in interlaced mode, but what about the other filters on my chain (denoisers, color correction, levels, etc) some of them will work with interlaced input, but some will require a progressive input)

Maybe i can revert the tivtc after all processing but i dont know how to do it.

Before you ask, yes, i know this post is maybe best suited for the avisynth trheads but maybe, maybe there is a way of encoding this correcly with x264 using h264 pulldown flags and the timecode produced by tivtc, maybe not.

The reason to do this is because if i can put my x264 encodes on a disc, with a standard format (BD5 / AVCHD) with menus, well why not?

thanks

[manono]

Yes, i know i can encode the original stream in interlaced mode, but what about the other filters on my chain (denoisers, color correction, levels, etc) some of them will work with interlaced input, but some will require a progressive input)

Bob it, filter it, reinterlace it.

[OvejaNegra]

i assume you are talking about bob() on avisynth
why not separate fields?
because the picture wont be large enough to some filters?
can you recommend a bob filter that does not interpolate the picture (well i think that's impossible)
i meant, a way to bob the picture without change the content of the lines (to avoid losing detail)

thank you

[nm]

can you recommend a bob filter that does not interpolate the picture (well i think that's impossible)
i meant, a way to bob the picture without change the content of the lines (to avoid losing detail)
Most simple bobbers keep the original lines unchanged and only interpolate the added ones.

[Guest]

why not separate fields?
because the picture wont be large enough to some filters? No, it's not a size issue. When you separate fields, you need to allow for the field offsets. It's a difficult calculation that most users don't understand and will mess up, destroying resolution. The correct alignments are achieved automatically when using the bobbing approach.

[OvejaNegra]

Most simple bobbers keep the original lines unchanged and only interpolate the added ones.

According to the aisynth documentation:
To strictly preserve the original fields and just fill in the missing lines.
bob(0.0, 1.0)
Bob(0.0, 1.0) preserves the original fields for RGB and YUY2 and preserves the Luma but not the Chroma for YV12.

Whats the color output of dgdecode?

Well i think i can use
converttoYUY2() (maybe)
bob(0.0, 1.0)
my_filters()
AssumeTFF()
Separatefields()
Selectevery(4,0,3)
Weave()
encode on interlaced mode with TFF

please correct me if im wrong!!

I will give a try to that
Thanks!!