No jagged DV to HD possible?

[chamo]

Hello all,

Basically what I'm trying to do is to convert the video from my DV camcorder to HD format for watching on YouTube. I've ran into interlace problems though. When I watch the captured AVI file on my computer does Not show the interlaced lines. However if I edit the AVI video on Sony Vegas 8 Pro and then render it in to MP4 at 1280x720, I get a lot of jagged edges and lines.

My camcorder is a Panasonic GS250. I use the supplied Quick Movie Magic software to capture the video. The captured video is a AVI 720 x 480.

I've been doing a bit of research on the issue and thinking on following this work flow:

1. Capture video from DV tape to computer. The capture file is AVI 720 x 480.
2. Edit in Sony Vegas 8 Pro in 720 X 480.
3. Render edited video in AVI.
4. Resize and encode edited AVI to MP4 and resize to 1280x720 using VirtualDub.

Does the above seem correct? Any input would be appreciated.

Thanks!!!

[minidv2dvd]

vegas pro has the mainconcept mp4 encoder. it works fine, no need to use virtualdub. DVavi to 1280x720progressive looks fine with it.

[chamo]

Thanks for the reply minidv2dvd. Any ideas as why my video gets jagged lines after I encode on Vegas? Thanks.

[minidv2dvd]

DVavi is interlaced. what are you rendering to?

[chamo]

I'm trying to render to mp4 h.264.

[JohnnyMalaria]

If you are using Windows Media Player to view the DV files, it will deinterlace it for display (depending on the DV decoder in use). This isn't always true for other media players and/or formats.

[guns1inger]

Because you are doing a straight resize of interlaced material. You need to look as something like VideoEnchancer, or MagicBullet HD, or at the very least, try resizing correctly in avisynth.

The basic issue is that in order to do upscale to HD you have to create a lot of information that doesn't exist. Simple upscalers do this badly when asked to resize by such a large amount. Video Enhancer and Magicbullet HD use complex mathematics and motion analysis to (slowly) attempt to fill in these gaps and therefore reduce the jaggies and other issues.

Will you get true HD out of them ? No. The details don't exist, so anything they do will be guess work and damage control.

Will it look better than a straight resize ? In most cases, yes.

Will it take hours to process ? Certainly. You haven't filled in your computer specs, but on anything but a fast quad core expect this processing to be reasonably long, and on older machines to take half a day or more.

[chamo]

guns1inger thanks so much for the reply. At what point should I do the resize? after I'm all done editing and have an edited AVI file?

The video is about 8 mins long so hopefully won't take too long. I have a duo core 2.0 cpu + 3 gb ram.

After resizing in avisynth do I still need to deinterlace using vdub?

Is resizing in avisynth pretty straight forward or do I need some special settings? I'll do some googling but any input would be appreciated.

Thanks!!!

[guns1inger]

Avisynth is script based. There are plenty of examples in the forums of scripts that correctly separate the field before resizing, then put them back together again. There are also plenty of discussions regarding deinterlacing methods.

All of this should be done after you have finished working in Vegas. Output as DV avi, then process up to HD.

I would seriously look at Video Enhancer as a low cost alternative unless you really want to research motion resize techniques in Avisynth. If that is the case, Doom9 is the place to search.