How to average best picture from multiple captures....

What I am searching out is information regarding the averaging of let's say 3 separate LD captures for the cleanest possible picture. 

How is it done and what does one look for when doing such a project (i.e. software, process)?

Thanks

Zion said:

If they're all from the exact same LD release (i.e. they're all from Star Wars Definitive Collection LDs) then look up an avisynth filter called TooT (stands for Two out of Three). It will essentially take your 3 captures and combine them into one good one that's free of any rot spots or other imperfections not on the original LD. I'm sure there are other avisynth plugins out there as well.

If they're not from the same LD release, good luck and let me know if you come up with something.

Thanks Zion, appreciated.

Here is what I'm up to.....a learning experiment of sorts.

I have the Japan Silver Box 1997 S.E. and the Japan Second Pressing of the 1997 S.E..  The movie portions themselves don't have anything different video-wise except that the 2nd pressing seems to have a bit more detail than the 1st pressing so I figured if I could average out a cleaner archive it might be worth it to figure out how to try.

If it weren't for the Jap sub-titles I would use the U.S. S.E. set to round out the average since the video itself shouldn't be an issue.

I will let you know how I fair once I find this TooT and how to use AviSynth properly enough to make a go of it.

:)

Zion said:

TooT may or may not work, depending on how different the video is. It doesn't actually average the sources together, but rather compares all three sources and outputs the information that's consistent in at least two out of three sources. So like if there was a rot spot on one of the captures, it would take the information for that area from the two sources that don't have the rot spot. If all three sources differ, then I think it defaults to the first one.

The other avisynth filter you might try is Merge, which actually does average the two sources together.

 

Thanks for that too.  I've bookmarked it.  At this point, I've been learning to use Virtualdub for capture and trimming.  There is SO much to learn and everyone will always have a different view due to what type of hardware and software they had available to them due to their means and/or knowledge or both. 

I am SOOOOO not in the ball park that some are here when it comes to understanding and implementing all the different angles.  All I ever really wanted to do was to make a BETTER "5 Star" Collection but the contrast levels alone on the S.E. can be a nightmare.....it's almost as if Rick McCallum, Dennis Muren and team just redid everything with different Popsicle sticks to make the deadline for the 20th anniversary.

So far, I've captured each side of each movie for both sets in Virtualdub using the huffyuv codec.  I've also trimmed off the excess video so I have just the movies.  Now, I've got to figure out how to IVTC without compressing the AVI's I have and at least do a conversion to mpeg2 to see what I've got to work with or what I need to do after the IVTC, if I even need that.  It's crazy for a noob.......  

Puggo - Jar Jar's Yoda said:

I've been using an avisynth script on my ESB 16mm telecine that has been working great... but your captures absolutely have to be frame for frame identical.  It's been tweaked a zillion times... I'll try to remember to post it here tomorrow.

Okay then, how do I ensure that my captures are "frame by frame" identical? 

Does anyone know if either Japan version of the S.E. are different in any way?

The only difference I have seen obviously is the picture quality....seems like 2 different transfers as the 1st pressing seems to have more DNR than the 2nd pressing....at least I believe that is what I'm seeing.

My other questions are naturally going to be....

Do I use notepad for the script and what do I save it as so AviSynth reads it properly?

Thanks 

Orinoco_Womble said:

You might like to have a look at Arnie.d's V8 thread which had a very interesting discussion about averaging here

G-Force was pretty adamant that doing a median of multiple captures will give better results than averaging.

Hope this helps.

Thanks Orinoco...I have wandered in to that thread from time to time but as a noob I was miffed by some of the techie talk.  Right now I have 1 capture of each Japan S.E. set....I guess I'll need to decide which one I want to use for the 3rd source so I can move forward.

:) 

Thanks Moth3r, I got it bookmarked and read.  Learning how write scripts specifically to what I might want to do will be fun I'm sure.  There are some good examples here so that helps.

:)