Announcement

Collapse
No announcement yet.

Frame by frame vs. real time video capture for 8mm telecine

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Frame by frame vs. real time video capture for 8mm telecine


    About the shortest, and least intimidating, account of frame by frame capture I've come across is by Pol Fieldman https://www.youtube.com/watch?v=eZniFhJR17M. This, and his generous answers to the many question below, seem almost a textbook on how to do it. His "wet gate" method (another is in the lower of the two videos at https://filmfabriek.nl/filmfabriek-hds-scanner/) and post production treatment seem to account for a large part of the stunning results demonstrated.

    Skilful original film camerawork, combined with rock steady, frame by frame "wet gate" transport and extensive post production, result in almost unbelievably serene transfers that hardly look as though they originated on film at all.

    It takes an absence of only one of these (fine camerawork in the filmfabriek transfer example, or less than rock steady transfer in the video example at https://vintage-home-movies.co.uk/20...e-early-1960s/) to suddenly be reminded that we're looking at film.

    In consideration that my family were nowhere near as skilled as Pol's dad in hefting an 8mm camera, I'll probably go the route originally intended (http://8mmforum.film-tech.com/vbb/fo...-video-capture). But I wonder if others are struck by the slightly unreal feel of Pol's fabulous transfers. We may have simply grown used to the feel of film, as so many professional film makers who insist on shooting on film have. After all, remarkably good transfers have been achieved with the simplest of set-ups, as James Miller and John Yapp have already demonstrated in the above thread. So just how necessary is frame by frame transport for successful transfers?

    What do other readers feel?

  • #2
    Hi Marcel. The wet gate idea is nice and by looking at the video it would not be hard to implement it. But I just used the cleaner and lint free cloth and just ran the film through it. And as you can see it is pretty good.
    https://drive.google.com/file/d/1ytl...ChqysORT4/view
    This scan was done with the modified Wolverine and WideoFred's post. Obviously frame by frame. I have done direct capture before that,
    syncing the projector to the cam. It is definitely possible and some of it is shown in this forum with a simple setup and pretty good results. My problem was bad frames that crop up here and there. Sometimes it is hard to see it, like a smudge on the bottom of the frame.
    Here is a test setup that I ran with the Teslong microscope and a projector with modified motor. Not very good quality but just to test the sync.
    https://www.youtube.com/watch?v=CE8YAT2sSZo&t=1s

    And a setup I did several years back with slow sync i.e. 4 frame captures by the cam and lots of post processing. VERY COMPLICATED but pretty good quality but not better that what I saw with a very simple setups shown here is he posts.
    https://www.youtube.com/watch?v=sPaF3Tqauiw&t=15s

    So, my suggestion is go with a simple setup first if you have spare time. It is not much money and lots of fun. I will run a comparison between the HD setup (Olympus HD cam) and Hawkeye/Wolverine. May give us some idea on quality.

    A few more notes. You could go with the macro lens for your EOS:
    https://www.bhphotovideo.com/c/produ..._8_2_5_5x.html

    But that is expensive.

    Or go with a C mount 16mm lens.
    https://www.oemcameras.com/sv-1614h.htm

    Here is a project using the C mount 50mm lens and ImagingSource camera (Same supplier used for Hawkeye). Very sharp images.
    This lens is very expensive but that is one of teh key components.

    http://8mmforum.film-tech.com/cgi-bi...c;f=1;t=008059

    If you are going with the 16mm lens be careful because I believe your camera is full frame and 16mm may not give you optimal results.



    Last edited by Stan Jelavic; April 07, 2020, 12:33 PM.

    Comment


    • #3


      Stan - Very many thanks for your detailed response. You're right, the simple way seems preferable for the first go, and I'm very impressed with what you have achieved using fairly straightforward techniques. I'd be very interested in the outcome of side by side comparisons of real time and Wolverine transfers. Good luck!

      Since the Standard 8mm frame is 3.3mm vertical, the full frame sensor is 24mm vertical and I'll want to overscan slightly, the magnification involved is about X6, which conveniently falls within the magnification range of my 40mm Zeiss Luminar, which should do all that's required of it.

      Comment


      • #4
        Yes, that should take care of the optics. Obviously you will have to figure out how to mount the camera. The sync is another issue and you need a strong enough motor that will not change speed with the film drag. In my simple microscope setup I used a motor from a cordless circular saw with the PWM controller.
        Click image for larger version

Name:	vlcsnap-2020-04-10-08h18m08s334.png
Views:	1053
Size:	503.5 KB
ID:	7382

        Comment


        • #5

          Stan - Wow, that's some motor! But I wonder if anything like it will be necessary for the kind of set-up I have in mind, which is a close relative to the one James Miller gives at https://vimeo.com/20950590, using a Sankyo Dualux 1000. His Standard 8mm capture at https://vimeo.com/20871186, when viewed full screen, looks to be as sharp as can be captured, especially on static shots, where it's easier to judge.

          Sure, adequate cleaning beforehand, repair to film damage, "wet gate" or lubrication and post production could all improve on his result enormously, although defects in the basic cinematography will never allow it to approach the splendid results of Pol Fieldman.

          But I somewhat doubt that, with the exception perhaps of frame blending mentioned on another thread started by me, frame by frame capture would result in a very greatly improved result for James' material, or perhaps anyone else, which is the thought that kicked off this thread.

          What do you think?

          Comment


          • #6
            Hi Marcel. I had to use a hefty motor because my Chinon Whisper motor would speed up and down periodically and cause out-of-sync issues. But you can try as is anyways and go from there.
            Agree, the images are very sharp and the light distribution is pretty good but the exposure in some shots are off in my opinion. Another issue is with panning. You can see ghosting and frame by frame does not have that. Another point with panning shots.. VideoFred's does a great job on captured TIFFs. It will interpolate the frames for smooth panning, remove film grain noise and will do anti-shake.

            Comment


            • #7


              Stan, thank you so much for reminding me of VideoFred's website. I had looked it over before and concluded it was way above my head. But, this time, I looked more closely at his videos https://vimeo.com/user678523/videos/all/sortlays. The ones showing before and after AviSynth processing were a real revelation! I had no idea that it was capable of smoothing out shaky hand-held material etc. This convinces me more than ever that frame by frame capture of my family films is unnecessary. The huge improvements demonstrated by Freddy Van de Putte in using AviSynth are very considerably greater than any incremental improvements that might be obtained, at either great labour or cost, from frame by frame.

              Since I now have a rational way forward, I shall retire from the forum for the present and thank you for your valuable advice. But I'll check back from time to time just to see if you have any luck in doing comparative tests of frame by frame versus real time video capture. Many thanks Stan.

              Comment


              • #8
                OK Marcel. Good luck. I will post the comparison here.

                Comment


                • #9
                  Marcel Friedman Freddy's Avisynth scripts were designed for frame-by-frame captures not realtime. If you use them with realtime captures you will probably notice some ghosting and frame blending depending on which script and variables you use. There is a VirtualDub filter called De-shaker that does a nice job of stabilizing shaky footage. It's very easy to use.

                  Comment


                  • #10
                    Good point Janice. I actually tried the VideoFred script with frame-by-frame and noticed a big improvement when I switched from jpg to tiff. Have not delved into it too much and just stayed with tiff captures. Additionally I never tried it with a video obtained with the video capture(projector to video cam). I actually have some and will try it but would not put my money on it.
                    De-shaker works very nice.

                    Comment


                    • #11
                      Janice, thank you so very much for that information. Many bear traps await the unwary newbie and you have undoubtedly saved me a great deal of wasted time. VirtualDub has the big advantage of having several YouTube tutorials devoted to it and I've just invested in a rather elderly book on it, too. It probably contains more than enough for my probably modest needs and, if not, there are plenty AviSynth scripts other than by VideoFred that seem suitable. Many thanks to all of you!

                      Comment


                      • #12
                        Marcel Friedman I suggest you use VirtualDub2. This is a newer more updated version that can read and write multiple video formats in addition to AVI. It comes with some codecs and you can add others. If you subscribe to the SoundForge update service you will be alerted when a new version of the sofware is released.

                        http://virtualdub2.com/

                        Comment


                        • #13
                          Again, many thanks for that Janice. It may be that your kindness was prompted by my remark "..elderly book..", but I had got to the VirtualDub 2 destination before your post and was glad to realise that my Canon's MP4 output would not need conversion to an AVI file, it being acceptable to the latest version. Although using a computer for the last fifteen years, it's only been for relatively simple things like email, surfing the internet and shopping, so using a sowtware suite like VirtualDub 2 will entail a rather steep learning curve, and it's reassuring for oldies like me to work from a printed text, even elderly ones, than to be solely reliant on the look-before-it-vanishes speed of many of the YouTube tutorial cursors.

                          Comment


                          • #14
                            Here is the side by side comparison of the frame-by-frame Hawkeye scanner and real time transfer, modified Chinon into Olympus Pen with the 100mm macro + extension rings.
                            https://photos.app.goo.gl/M6jhD5Wmx7PkcLAX7
                            It is pretty obvious that the Hawkeye optics are superior. The panning and zooming is pretty close between the two. The Pen camera exposure is set too high. I remember I had it in manual mode but probably did not tweak it for this clip (7 years back).
                            So, my two cents here:
                            You can probably get a very good scan with direct projection into the video camera but make sure your optics are designed right and good quality. Adding extension rings strains the macro lens and pushes it beyond its specified range.
                            Your setup Marcel should be good with your macro lens.
                            DSLR exposure has to be set properly. If manual you may have to tweak it from time to time. If auto you will get the transition effects.
                            Sync issues. Get a decent motor with stable speed plus controller.
                            BTW - the Pen Camera video was processed by Fred's script and it seems to help. Have not spent much time on it.

                            Comment


                            • #15
                              Stan, thanks for that. An interesting comparison. I agree that the panning and zooming (my biggest concerns) are about equal between the two. Given that it's a Luminar on to a 21 Megapixel full frame sensor, I'm fairly relaxed about the optical train. Ditto the variable speed control of the Sankyo. But exposure, colour rendition and camera unsteadiness software corrections are all issues I'm going to have to learn about the hard way. This may take some time! See you around.

                              Comment

                              Working...
                              X