Warping in Arriscan

Perry Paolantonio Feb 2, 2016

  1. We're doing some restoration work on a 35mm feature right now that was scanned on an Arriscan (at another facility), at 2k. The source was an IP, so no physical splices in the film (or at least if there are, they're rare).

    At every scene break, we're seeing warping along the lines of what you'd see in a Spirit or other continuous motion line sensor, only it's on the whole frame, not just the top/bottom of the frames surrounding the cut. Sometimes it's on the first frame after a cut, sometimes it's on the second. But it's there every time, and is really noticeable on shots where the camera is locked down.

    Here's the really weird bit - our client went back and looked at other scans they've done at the same facility, from prints, IPs and cut neg, and all of them have the same problem.

    I don't really know a lot about the Arriscan, not having used one personally, but how would this happen in a pin-registered scanner with an area sensor? Could it be related to the microscan sensor repositioning they use? If so, why would it only happen at scene breaks (and again, no physical splice in the scanned element, since it's an IP)

    Any ideas?
  2. Is it possible that the negative used to create the IP had a buckle on the splices when the film went through an optical printer? In other words, maybe it's a printed-in problem.

    I have never seen an IP with a flaw like this, but I have seen them with 1-perf mislights and things like that. There is a de-warping tool available in DRS and Pixel Farm and so on. I would try a test on a different non-pin-reg scanner and see if it does a better job. I know there are capstan systems that use software to read sprockets and do the registration after the fact, rather than using a mechanical sprocket (like the Kinetta). It may be that this is one of those projects for which that approach is a better idea.

    I just finished a project done on an Arriscanner yesterday, and had zero problems with any kind of warpage or stability. There were a handful of jittery shots, but I think this was due to camera problems and not the stability of the negative itself.
  3. Weird can you post a pic? As far as I know the Arriscan doesn't use the "microscan" piezo re-positioning unless it is in 4K mode.

    Could be broken? There is an Arriscan on ebay with a bad sensor right now.....

  4. That was my thought as well, since we're actually dealing with this at the moment on a scan done on our Northlight. That other film was shot on 16mm, but the only viable element is a 35mm CRI blowup. Splice bumps are baked into that CRI, even though it's a single strip of film, so it's clearly from the blowup process. Cleanup of that in PFClean is a relatively trivial (if time consuming) matter.

    I just wrapped up scanning on a 35mm IP on the Northlight, and spent a good deal of time yesterday afternoon looking for similar splice bumps, but couldn't find anything.

    I don't think they're going to bother fixing this, but the client's concern here is more about why this is happening since that they've had dozens of films transferred on that lab's arriscan, and now they're starting to notice the problem in all those films - IPs, Negs, Prints.



  5. I'll ask, but probably not. they're kind of sensitive about this sort of thing.
  6. I've seen bumps and weird stuff from 16mm blow-ups in the best of circumstances, even with A/B-roll checkerboarded elements. What might be the definitive answer would be to see if you could locate a theatrical print and check that and see if it, too, has the same vertical bumps printed in. Maybe it's always been like this and nobody noticed.
  7. I think your bumps are from the printing process. As the splices go through the printer, it just jumps. No one noticed this when the film was projected, as the projection shakes a little bit even when good. Put it through a frame registered scanner and view digitally and you will see a lot of splices.

    We did a scan of a print last year, and found the same issue as yours... Good luck!
  8. Jumps at splices were a big problem in early 16 and S16 blowups. I was just discussing this with David Leitner -- who ran DuArt's optical department, known for the best S16 blowups in the early days.

    One of the problems was due to a wet gate that had the register pin displaced from the aperture by two frames, so at imperfect splices you'd get a 2 frame jump. The blowup of Claudia Weill's GIRLFRIENDS had this problem. Switching to a different gate helped.

    DuArt later pioneered a system that let you cut your original single strand with handles, and "conform" it in the printer.

    Another issue is that contact printed splices often will have 1/2 of the frame at the splice being a bit soft because the splice pushes the film away from the print stock. Usually it's not noticed at 24fps because it happens at a cut, but it sure is annoying once you see it.

Share This Page