digitalFAQ.com Forums [Archives] - Avisynth: best Deinterlacer script?

digitalFAQ.com Forums [Archives] (http://www.digitalfaq.com/archives/)

- Avisynth Scripting (http://www.digitalfaq.com/archives/avisynth/)

- - Avisynth: best Deinterlacer script? (http://www.digitalfaq.com/archives/avisynth/11847-avisynth-best-deinterlacer.html)

right will do, I'll do a bit of messin about downloed the two thing you suggested boulder and see what I get.

And before you ask, you can determine the field order with a simple script like this:

Code:

MPEG2Source("path\clip.d2v")

AssumeTFF()

SeparateFields()

Open it in VDub. If the movement jerks back and forth, your video is bottom field first so the KernelBob line should be KernelBob(order=0,sharp=true,threshold=7). If the movement is smooth, it's order=1. You can verify this by putting AssumeBFF() instead of AssumeTFF().

ok got it all workin finally lol just one problem it says its gonna take about 4 hrs to do an 11min music video.

It's dead slow all right :lol: but try this one, maybe it's faster and might still produce good quality. I believe it's the one Inc originally posted in the sticky. If you get an colorspace error, add ConverttoYUY2() right before the Convert60ito24p(2,0) line.

Code:

MPEG2Source("path\clip.d2v",idct=7)

KernelBob(order=x,sharp=true,threshold=7) # determine the field order!

Convert60ito24p(2,0)

#and the rest of the script as in a normal encode



function convert60ito24p (clip video, int mode, int offset) 

{ 

   work = assumefieldbased(video) 



   out = (mode==2) ? interleave( 

           \selectevery( 

             \layer(trim(work, 1, 0), 

                    \layer(work, trim(work, 2, 0), "fast"), 

                  \"fast"), 5, 0 + offset), 

           \selectevery( 

                  \layer(work, trim(work, 1, 0), "fast"), 5, 3 + offset)) : 



   \    (mode==1) ? interleave( 

          \selectevery(trim(work, 1, 0), 5, 0 + offset), 

          \selectevery(layer(work, trim(work, 1, 0), "fast"), 5, 3 + offset)) : 



   \    (mode==0) ? selectevery(work, 5, 1 + offset, 4 + offset) : work 



   assumeframebased(out) 

}