But I will follow your suggestions for some tests.
#Topaz clean 2017 mac series#
My TV series is 5 hours! And so to run it all through Topaz would be an ordeal. There was a thread here where I submitted stills, but nobody provided a better solution than I can get myself, even though they recommended a very complex method. I am currently, in between working for a living, restoring an old TV series I made around 2006, shot in DVCam.
You can see that very often on 4K Netflix documentaries that incorporate DV archive. Steve Fishwick wrote:Yes Carlo, it's true, it has some very innovative tools, I should experiment more with it However as I say, it's often a time/workflow issue and I have never seen an ideal solution for SD DV interlaced footage. Because, for me, as I said the difference between this and what can be achieved directly in the NLE, is subtle, I prefer to choose the quickest workflow, consistent with some useful improvement, but obviously YMMV. Grain often helps with perceptual sharpness in images. Most of it ends up looking like severe edge enhancement from the old video camera days, which is beginning to happen, I would say, in Carlo's first two examples too.įor DV footage, the best solution I have found is de-interlacing, judicious uprez to HD (4K is not going to happen) and then a very mild filmlike grain applied. It's easy to make an image very plastic looking with these tools. But nothing is really a magic bullet, and as with de-noising less is almost always better than more, the difference between results from all methods are subtle rather then dramatic.
It's perhaps best for uprezzing BMD HD slo mo shots to 4K. I bought Topaz on an offer that was going but I hardly use it, since the workflow is cumbersome and slow. Ellory Yu wrote:I kind of expected anything that does “aI” is calculation intensive and reiterative in its algorithm.ĪI is almost a misnomer these days, at least I would say it's not nearly as intelligent as the marketing would suggest, as you can see from Carlo's third example.
0 kommentar(er)
