Decompressor

[cool9]

I'm still learning about Virtualdub after a year (I use it everyday) and I don't know what the "decompressor" is in "File Information". I imported a .d2v file into VDUB and I noticed the decompressor was "DivX 6.2.5 decoder". I have no idea what that means. Does it decode the input video? Does it have an effect on encoding in VDUB? Does it have something to do with FFDSHOW? The only thing I have enabled on FFDSHOW is XviD (I'm not sure exactly what FFDSHOW does, either).

[setarip_old]

Does it decode the input video?It decompresses the input video - it's 1/2 of the DivX v.6.2.5 video CODEC(compressor/decompressor)...

[neuron2]

If you loaded a D2V file via an Avisynth script, which is the only way I know to load a D2V file in VirtualDub (not VirtualDubMod), then the DivX codec is being used only for color space conversion from YV12. If you put ConvertToRGB() at the bottom of your script and then look at the File Information, you'll see.

[cool9]

Does it make a difference which decompressor I use? RGB or YUY2?

[Awatef]

It is recommended to keep everything during all the process in the YV12 color space, to make things go faster and to avoid (eventually buggy) color conversions.

[cool9]

Can you deinterlace YV12? Avisynth manual says this on the Color Conversions and Interlace page about working with interlaced YV12:
Let's assume you have interlaced video, you want to work in field-based mode (to apply some filtering for example) and you also need to do some color conversion. Do you need to do the conversion on the frame-based clip or can you do it on the field-based clip? Well, that depends on the color conversion you want to apply:
YV12<->YUY2/RGB conversions should be done on the frame-based clip (with the interlaced=true setting). Doing them on the field-based clip will yield incorrect results. The exact reason of this is outside the scope of this page, but it is a consequence of how the color format YV12 is defined. The main issue is that chroma is shared between pixels on two different lines in a frame. More information can be found here Sampling.

[Awatef]

Afaik, TomsMoComp can deinterlace in the YV12 color space if AviSynth 2.5 Alpha is used (not tested though)

[cool9]

(wrong response)

[neuron2]

Can you deinterlace YV12? Yes, of course. All the well-known deinterlacers take YV12 input (Decomb, TDeint, [Leak]KernelDeint, etc.).