What script would you use when resize to PAL Anamorphic16:9?

[Draven]

I have a source in 640x256 (2.50:1) PAL.

And I use
LanczosResize(704,412,0,0,640,256)
AddBorders(8,82,8,82)

I get 0,3% difference from the original aspect ratio ,in FitCD.

And I convert and resize to watch the output on a widscreen TV.

What would you use and why ??

[Draven]

Ok u don't have to explain why only what script u use when u have a source in 2.50:1 and if u resize it to anamorphic. And then play it on youre DVD player and widescreen TV.

[ronnylov]

I would use FitCD (and you already did use it) and let it generate an avisynth script. However I would enable macroblock aligned resizing to make it easier for the encoder. I would also set the destination resolution to 704x576 (XVCD) because it is a valid DVD-resolution and you have less black border to encode. I would also set 2 blocks overscan and crop-resize it to 672x416 with 16 pixels black border on left and right and
80 pixels black borders on top and bottom.

The reason of using a active picture size of mod32 pixels is that you can fill complete macroblocks with black around the actual picture area. The mpeg2 encoding is done on blocks of 8x8 pixels and macroblocks of 16x16 pixels. It requires more bitrate to encode a border edge within a block and within a macroblock compared to letting the border edges be aligned to the encoded macroblocks. This is not very important at DVD bitrates but when making SVCD it's good to make all optimizing you can. Since I started with SVCD before making DVD's I am used to always make macroblock aligned resizing when reencoding.

But all of this can be calculated with FitCD. An aspect ratio difference within 1 % or so can not be detected easily with the human eye. I would use the lanczos resize method because it gives a sharper picture than the other alternatives.

[Draven]

Yees this is what I'm talking about . Cause everybody has their own opinion and others can always learn something from them.

Thanks Ronnylov.