avoiding colourspace conversion

[coordz]

I've now got my DV (type 1) AVI and want to convert to MPEG2 to go onto a DVD. My usual method (and the one the DV -> DVD guide suggests) seems to involve an implicit conversion from YUV (or YCbCr or whatever) -> RGB -> YUV where the DV stream is decompressed by a windows codec into a DIB before being passed to the MPEG2 encoder using VfW. I'd like to avoid this and keep my NTSC MPEG2 as close to the original NTSC DV as possible. I'm also thinking the chroma subsampling is the same???? :confused: ??? so should be able to pump the decompressed YUV from the DV stream directly to the MPEG2 encoder.

Am I worrying too much and should I just stick with my usual DVD creation path? Or can I easily (using VDub, AVISynth) keep all my YUV samples together as they should be?

Many TIA

[2Bdecided]

There has to be one colour conversion - NTSC DV is 4:1:1 YUV, MPEG-2 is 4:2:0 YV12.

You can avoid doing two conversions (i.e. RGB in the middle) if you use a DV codec that outputs the native format (e.g. cedocida).

Alternatively, AVIsynth has features to get back from RGB to YUV with less loss, e.g. convertbacktoYUY2(). There's also Reinterpolate411 - a google will tell you when it's useful.

Hope this helps. I'm not speaking from experience here!

Cheers,
David.