A question - NTSC DV to Divx5 using AVISynth

[Byungsoo]

Hi.

I apologize for my elementary questions, first of all.

I want to compress my dv source (from sony camcoder) to divx CDs.(family movie?^^)
I will do it with divx5 one pass quality based(100%) option.
But do I need what kind of filters ?

Is it better to use Avisynth filters than Virtualdub filters ?
If it's better to use avisynth , which filters and orders ?
is it right the following filters and order ?

DirectShowSource
Tomsmocomp --> for deinterlace
Convolution3D --> for denoise
BicubicResize --> 720x480 -> 640x480 for viewing in PC

is more filters or another one needed ?

I am sorry . English is not my first language.

Thanks in advance.

[bb]

Welcome to the board!

As a starting point please note that there is a DV forum available on this board. You will find most of your questions answered there (read the sticky threads first).

AviSynth filters are faster than VirtualDub filters most of the time, because they can usually stay in YUV colorspace instead of converting to RGB, which saves a lot of time.

The filters you suggest are good and sufficient for most purposes. The order is correct (always deinterlace first), but some people prefer Convolution3D after resize because of the speed improvement. If you ask me, quality will be a tad better if you use Convolution3D before resize.

You did not mention neither the file format nor your preferred audio codec. I recommend to use OGM file format and Ogg Vorbis audio codec. You'll find a lot of information about that topic in the "New A/V Formats" forum.

bb

[High Speed Dubb]

bb wrote
The order is correct (always deinterlace first)...

I don’t think that’s always right. If a temporal noise reducer is written to correctly handle interlaced material, it makes sense to run it before deinterlacing. That should improve the accuracy of the deinterlacing step.

[Wilbert]

AviSynth filters are faster than VirtualDub filters most of the time, because they can usually stay in YUV colorspace instead of converting to RGB, which saves a lot of time.
True, but it could be that his DV is RGB (RGB24) and not YUV.

You did not mention neither the file format nor your preferred audio codec. I recommend to use OGM file format and Ogg Vorbis audio codec. You'll find a lot of information about that topic in the "New A/V Formats" forum.
I wouldn't do this if you are new to all this.

[bb]

Originally posted by High Speed Dubb
I don’t think that’s always right. If a temporal noise reducer is written to correctly handle interlaced material, it makes sense to run it before deinterlacing. That should improve the accuracy of the deinterlacing step.
Well, to always deinterlace first is a general statement, which is meant to simplify work for beginners, so that they don't resize etc. before deinterlacing. I use VirtualDub's noise reduction during capture, and I deinterlace later with good results.

bb