Capturing Using Virtualdub

[Tony Yip]

I've read that there are 2 ways to capture off of a TV
source using Virtualdub at 352x480 NTSC.

The first method is as follows:

1) Capture at 352x480
2) Then use

a) smart de-interlace filter
b) 2:1 vertical reduction filter
c) dynamic noise reduction

to get your 352x240 avi file that captures both fields

The second method I've read about is as follows:

1) Capture at 352x480 using the
video/horizontal size/2:1 bilinear resize


Which method is better? I've heard that to capture off
of a TV source, it is best to capture both fields and
to de-interlace afterwards. This would make Method 1
better. Or is Method 2 just as good?

[ironwood321]

Try here for some help http://www.vcdhelp.com/phpBB/viewtopic.php?topic=52258&forum=2&4

http://www.digital-digest.com/nickyguides/ for some guides

I personally capture from VHS 352x576 PAL with avi-io and HUFFYUV. Then I load my video into virtual dub and use the smart deinterlace / smart resize filter. Link provided in first post.

Video must be deinterlace first before any further processing is done. Deinterlacing reorders the fields so it is progressive. Progressive video looks better and compresses better. Next I resize with the smart resize filter bilinear. You notice that bilinear files are smaller and not as blocky when reducing in size, plus things like Hanover bars are not emphasized.

b) 2:1 vertical reduction filter
c) dynamic noise reduction

The 2:1 reduction filter does not deinterlace it merges or strips one field to reduce the size of the video. You could have double images because your video not progressive or not sharp images. Smart Resize is a lot more better when it merges the fields.

Assuming you selected the correct deinterlaced option for your video first.

Dynamic noise reduction just reduce noise and you'll have to make a judgement if you want to apply this after you've seen the output from the previous filters. Remember the more filters you use the worse the video gets. Better filter you can use but I have not because I think the video I have ok are various VHS ones these take account of the reduced chroma etc and compensate for it but you have to experiment.

You don't have to use 2:1 in the smart resize filter you can specify the resolution you want 352x288 if you use PAL or 352x240 if you use NTSC.

If you have any black boarders you want to remove consider using crop or 0 transform. But remember they can be a nightmare if anyone wants to reconvert video to another format.

I'm thinking of putting a little website together with this sort of info. Does this sound a good idea or can I submit some of my findings including images to this one?

<font size=-1>[ This Message was edited by: ironwood321 on 2001-08-01 02:56:51 ]</font>