NTSC DV, Cedocida, XVID question

[Gav]

If I wish to process NTSC DV avi's through Avisynth and then compress to XVID and I have the Cedocida codec installed, which of the three settings (DV, MPEG 2 non interlaced, MPEG 2 interlaced) should the YV12 chroma sampling be set to?

[Guest]

NTSC DV avi's Does the "DV" here give you a clue? :)

[Gav]

Does the "DV" here give you a clue? :)
I wasn't sure because if I was converting my DV avi's to DVD format then I assume I would choose the MPEG 2 interlaced setting. XVID on the other hand is based on MPEG 4, is it not? So I thought MPEG 2 and MPEG 4 may be similar so therefore...

[Guest]

OK, for decoding, if your script delivers YV12, you need to set the decoding YV12 type to the type of sampling you want to provide to the encoder. So if you encode MPEG interlaced, select that, etc. It's the same for MPEG2 and MPEG4.

My first post was hasty -- forgive me. :)

[Gav]

OK, for decoding, if your script delivers YV12, you need to set the decoding YV12 type to the type of sampling you want to provide to the encoder. So if you encode MPEG interlaced, select that, etc. It's the same for MPEG2 and MPEG4.

My first post was hasty -- forgive me. :)
Thanks for the reply. One last question, though. My DV avi's are interlaced. In Avisynth, I am deinterlacing and then feeding it to the XVID encoder. Would I set the YV12 sampling to MPEG 2 non interlaced then? This would seem the most logical to me.

[scharfis_brain]

better uses YUY2 output if you want to process you video with filters.
you can still downsample it to YV12 at the end,.

[Gav]

better uses YUY2 output if you want to process you video with filters.
you can still downsample it to YV12 at the end,.
And why is that? Even just with deinterlacing?

[scharfis_brain]

especially with deinterlacing.

NTSC DV's native colour space is YUV 4:1:1, which means that every line has a unique colour information.
While YV12 has combined two lines to one colour information, effectively throwing away image detail.
This is even more bad since this is done on the field-level.

So tell CeDoCiDa to put out YUY2.
Deinterlace, and then convert to YV12.
this will give you the most precise handling of the colour information as possible.

[Guest]

I concur with the esteemed scharfi.

[Gav]

especially with deinterlacing.

NTSC DV's native colour space is YUV 4:1:1, which means that every line has a unique colour information.
While YV12 has combined two lines to one colour information, effectively throwing away image detail.
This is even more bad since this is done on the field-level.

So tell CeDoCiDa to put out YUY2.
Deinterlace, and then convert to YV12.
this will give you the most precise handling of the colour information as possible.
Other then a possible speed increase with encoding, is there any disadvantage to not converting to YV12 before passing to the XVID encoder, i.e. just keep it in YUY2 all the way through?

[scharfis_brain]

this shouldn't even introduce a speed difference, cause xvid converts to YV12 internally anyways.

So it is up to you whether to use converttoyv12() as last line in your script, or not.

[henryho_hk]

I always use AviSource("file.avi", pixel_type="YUY2", fourCC="CEDO") in such cases to play safe. [Yup, I modified the Cedocida Inf and binary to change its fourcc.]