[awolfoutwest]

Originally Posted by AAR.oner

i hope i didn't come off as saying, "either buy the best gear available or don't transfer it at all", cuz thats not what i'm saying...

Nah, I never got that from your comments. However, I do feel a certain responsibility to transfer as best I can, hopefully improving the material from the source when possible. While it is not curing cancer, some of these transfers are archival until a better version comes along. If no better version surfaces, then what I put out had better be as good as I can get it.

As far as quality goes, I like to take my processed video and subtract it from the original video. Doing this shows flat grey when the video is unaltered, and shows clearly where noise has been reduced, edges sharpened, etc. For a while I was using a technique that required 3 captures of the source, then the AVIsynth script would take the 2 closest pixels of the 3 captures and pass on that information. It was an effective method of filtering out the noise generated in playback, but I have since found a more effective all-around filter for spatio-temporal noise reduction.

My cause celebre has been cascading compression. I have tried to impress on those that will listen how essential it is to avoid re-encoding previously encoded video. I have used MPEG2 video to patch a dropout in something I am working on, but that is on the order of a couple of seconds of video. There are those who don't think twice about re-encoding an entire video to employ color correction, noise reduction or even editing. When I suggest that they find an analog source instead, or employ a GOP-level editor if all frames are not being altered, they maintain that they have "improved" the MPEG2 source and it "looks better". I remain "skeptimistic"