KernelDeint documantation ?

I heard a lot of good things about this plug-in. Is there any documentation on how to use it ?

Any other deinterlacing plug-ins that you can recommend ?

TIA

Originally posted by numlock
I heard a lot of good things about this plug-in. Is there any documentation on how to use it ?

Any other deinterlacing plug-ins that you can recommend ?

Well, I don't know where you got it, but the download on the author's site contains the documentation in HTML form:

kerneldeint/kerneldeint.html

np: Adam Johnson - Autopop (Chigliak)

I don't see any documentation there. Just revision history.

I also don't see any download link. Except for the source code.

Where can I download the latest compiled dll

TIA

In the documentation you got when installing AviSynth -gt; english -gt; externalfilters -gt; kerneldeint.htm

I assume you have seen that?

btw, the download link is left to the source code (v1.4.0)!

OK, thanks I got it.
And the documentation was includede in the zip file also.

Now if somebody could explain this to me (it's from the documentation):

To use this filter as a post-processor for Telecide(), use the following script:

Telecide(...,post=1,hints=true)
KernelDeint(...)

What the heck is post-processor for Telecide() ?
All I want is to deinterlace stuff.

tia

Originally posted by numlock
All I want is to deinterlace stuff.Then you don't need to worry about the Telecide() bit, just use KernelDeint(...) on it's own.

Telecide is an field matching filter - part of the Decomb package.  You won't need it if you only have pure video (it is mainly designed for use on telecined NTSC films).

When you use the term quot;deinterlacequot;, it really means one of two things.  One is what you are probably thinking- you have images that have a comb-like pattern and you wish to use a filter to elimate the line patterns.  That's what filters like KernelDeint() and TomsMoComp() are for.  Note that since you are only filtering images, the framerate of the video won't change.

The second thing applies mainly to telecined sources (as Bogalvator mentioned), where it is possible to intelligently recreate the original frames.  The lines that you associate with interlacing comes from the fact that images are divided up and displayed together in an alternating fashion (see here for a better description).  Filters such as neuron2's Telecide() attempt to bring the original images together, this is also often referred to as deinterlacing.  If it is done well, you will have eliminated the combs from your video without any image filtering, and you're done (this also changes the framerate of the video back to the quot;originalquot; framerate).  However, it is generally difficult if not impossible to perfectly match and elimate combs this way, therefore requiring the usage of post-processing filters on the frames which are still combed, in which we are back to filters like KernelDeint() or TomsMoComp().  The Telecide() function uses FieldDeinterlace() as a postprocessor (built into decomb.dll), although you can use other filters if you so choose (hence the reference to Telecide() in the documentation of KernelDeint()).

Whew, that was a pretty long post.  Anyway, I don't consider myself to be an expert, this is all stuff from a layman's perspective, so I wouldn't be surprised if some of the stuff I said wasn't perfectly correct.  It took me a while to absorb most of this stuff so hopefully that explanation helps out a bit.