TDeintrelace question: process two versions of clip same way

[redfordxx]

Imagine two clips like this o0=mpeg2source("clip1.d2v").TFM(order=-1,field=0,mode=0,pp=6,MI=20).TDecimate()
b0=mpeg2source("clip1.d2v",cpu2="xxxxoo", moderate_h=15, moderate_v=25).TFM(order=-1,field=0,mode=0,pp=6,MI=20).TDecimate()
I need both clips to be field-matched, decimated and if possible postprocessed same way, because I want them Subtract then.
Input, Output files parameters don't work, because of some crc checksum...

With script like this, it happened to me, that there was suddenly completely different first line on the deblocked clip, which was really a problem later on.

The reason why context is here (http://forum.doom9.org/showthread.php?p=825082).

Tnx much,
R.

[tritical]

Remove the "crc32 =" line from the tfm output file and it will disable the crc check. You should then be able to use the output file generated on one clip with another. Likewise, you can remove the line that starts with "crc32=" from the tdecimate output file. If you use the same input files for both tfm/tdecimate sets then both clips should be processed exactly the same way. (if they aren't let me know :))

[redfordxx]

Thanx.

Now I see, post processing must be different because deblocking can produce combing.

How I understand the situation:
1. Telecined fields are encoded separately (that's why on these deinterlaced frames we see not 8x8 blocks but 8x16 blocks
2. Deblocking on highly quantized areas can cause shift of color of the whole block differently for each field
3. That's why some 8x16 blocks appear combed.
4. Solution: PP with of 8x16 blocks with low threshold but high amount of combing.
5.Theoretical example (I am not home, so I can't try)?
cthresh=2?, blockx=8, blockx=16, MI=120?

As your filter has quite many parameters, i.e. is very complex, maybe there is other more effective solution I don't see.
In that case, thanks for suggestion.


BTW: Feedback: I use PP=6, not 7, coz ela PP produces strange dots...