logo Sign In

Post #525062

Author
msycamore
Parent topic
Info: a Smear-free '93 ?
Link to post in topic
https://originaltrilogy.com/post/id/525062/action/topic#525062
Date created
19-Aug-2011, 7:14 PM

Moth3r said:

I liked the sample, these old video transfers have an aesthetic appeal that's hard to describe.

I would definitely increase the white level (contrast), then perhaps you might want to carry out a gamma adjustment to darken the lower end of the luma curve.

So I take it the sample video is from a multi-capture median script, while the screenshots above are from the raw captures?

Yes, that is correct, those screenshots are from the raw capture. The sample video was from a median of 5, but to be honest I didn't notice any difference between 5 and a median of 3 but I wanted to try it out. Thanks for the advice, a level tweak is definitely needed to get this to look acceptable.

Moth3r said:

Yes the script above is simple to modify for five captures. There is a noticeable improvement when going from a 3-cap to a 5-cap average, but I've found that five is the sweetspot; more than this and the additional time and storage required outweigh the benefits of a small increase in quality.

Regarding the temporal soften command, I don't think the scenechange parameter (the 4th figure) is required, and I cannot tell any difference between modes 1 and 2 (the 5th figure). Obviously you would use radius=2 for a 5-cap average.

Ok will try it out, thanks for the help. :) Using his DVD recorders combfilter by using composite instead didn't do this any better unfortunately. Sure the picture was much cleaner with less dot-crawl/rainbowing but it looked less saturated and it clearly killed some detail.

Moth3r said:

What method did you use for the IVTC and resize? I ask because there's a fair amount of aliasing visible (it's actually as bad as the GOUT) and I wondered if that was from your processing or present on the source.

For the IVTC I used DoubleWeave(). Pulldown(,) and for the resize I used Spline36Resize. I can confirm that the bad aliasing is unfortunately present on the source, much seems to indicate that this LD used the same source as the GOUT. I've noticed that the JSC and the other pressings of this LD are somewhat better in this regard. This one also shares the same amount of wonderful gateweave we have on the GOUT among other things. I can post a raw sample later. Thanks for the response Moth3r, much appreciated.

EDIT: Good news is that I will be able to get some better results from a new cap, he had mistakenly a sort of dnr preset activated on his recorder when he captured it the first time. Sigh... ;)