Jump to content

Freddy Van de Putte

Basic Member
  • Posts

    44
  • Joined

  • Last visited

Everything posted by Freddy Van de Putte

  1. But there is a GUI for Avisynth: AvsPmod. I'm using it all the time. Unfortunately Avisynth can not be used on a Mac. Fred.
  2. It might not be perfect, but Avisynths RemoveDirtMC() comes very close to what Guy wants: many greetings, Fred.
  3. Hi Peter, I would love to see some (color) frames, is this possible? I assume you you are capturing image sequenses , right? And how do you merge the R, G and B ? Fred.
  4. Hello everybody, I have improved my 8mm film transfer system. I'm using a custom designed RGB Led backlight source, designed and build by Frank Vine. Frank has also developed the capturing software with auto exposure and auto white balance. http://www.cine2digits.co.uk I'm also using a high quality IMI machine vision camera now. Example clip: https://vimeo.com/49963017 Please download the original AVI file. The Vimeo player is not showing the full quality. many greetings from Belgium, Fred.
  5. I agree. HD1 is more than enough for 8mm film. It's not all about resolution. A good 1280x720 transfer will look much better than a bad 1920x1080 transfer. Also, modern HD TV flatscreens are having very good internal upscaling systems. Even a good 720x576 filmtransfer will still look very good in full HD on one of those modern flatscreen TV's. It also depends on what the end target is. HD TV? Computer? Digital projector perhaps? On a HD computerscreen, you will see any artefact. But when watched the same file from a distance on a HD TV for example, those artefacts are barely visible and not disturbing at all. And then there is the aspect ratio problem. To keep 4:3, we need to add black borders. So 1280x720 is actualy 960x720 with two 160 pixels black borders. However, it is possible to crop the 4:3 picture to 16:9 as you can see here: https://vimeo.com/37394781 As you will see, I'm limited with my current 1024x768 camera. With Frank's camera it is possible to crop straight to 1280x720. PS: as for the wetgate: the Film-O-Clean used with Filmguard should do the trick. many greetings from Belgium, Fred.
  6. 1) I'm using a TIS CCD camera, not CMOS. 2) I'm using my Avisynth filmscript afterwards 3) After 5 years of testing, I have some experience by now :) But one day I will upgrade my system, and then I'm going to follow Franks advice. Fred.
  7. Hello Phil, Well.. actualy I'm using an 8 bits 400 Euro low-end 1024x768 Imaging Source machine vision camera. No, it's not monochrome but the debayering is done on the fly with the capturing software. In the future, I will upgrade to a better IMI camera. The price will be around 2000 Euro. Like you say, RGB exposures would be the max of cource. :) Frank Vine is working on this. He has developed a Led RGB lightsource. It is custom made for the use with machine vision cams with trigger. Franks site: http://www.cine2digits.co.uk/ But we are talking about getting some more detail out of the dark parts here. My camera gives already pretty good results. PS: my Avisynth script is also removing the vertical jitter pretty good. So the transfer can be done on any decent (modified) projector, no need for pin registration. Fred.
  8. I'm using machine vision cameras for 5 years now. 1974 Super-8, capturing with my machine vision camera: 1973, Rome, 16mm capturing( not by me, but also with a machine vision camera) Both pictures are post processed with my special Avisynth-film restoring script. This script removes all dirt spots etc... Fred.
  9. What about free? :) My Avisynth filmscript seems to work fine on your footage. I had to remove a few duplicate frames. The script won't work on duplicates. I have uploaded a before/after example, it is password protected. Password is 'cinematography' http://www.vimeo.com/15523600 Please download the full version to be able to examine frame by frame. many greetings, Fred.
  10. Thank you all for the kind words. :) I have uploaded some more 8mm frames. 1958 regular-8 anamorphic, Kodachrome: http://www.facebook.com/album.php?aid=2055923&id=1298449764&l=f29566379a 1974 super-8 Maroc, I assume K40: http://www.facebook.com/album.php?aid=2058406&id=1298449764&l=9830569e79 many greetings, Fred.
  11. Thank you too, Brad! Please ask them for a one film frame = one digital frame transfer. Progressive, no interlacing if possible. Like Phil already has said: Mpeg4 is a distribution format. Do not use it for a master file. If you are on a Windows machine I would suggest to ask them if they can deliver the files in uncompressed RGB24 AVI format. You will have to give them a portable HD then because those files are huge. Fred.
  12. Thank you Phil! Sure: I am using a relative cheap machine vision camera, 1024 x 768 pixels. In trigger modus, a machine vision camera takes one picture when it gets a pulse on the trigger input. Then it waits for the next trigger pulse and so on. The trigger pulse is generated by the projector, with an optical IR switch. So every time a new frame arrives, the camera takes a picture of it. Then the software is saving this picture as an AVI frame to a 'growing' AVI file. I always use uncompressed RGB for my masters. Play speed of the projector does not matter. The result is always one film frame = one digital frame. I can even stop the projector while in capturing modus. Then the camera and software will just wait for a new frame to capture. This stop will be not visible in the end result. Variable speed has also no effect whatsoever on the end result. The maximum capture speed depends on the computer system. With a fast computer and Raid-O harddisks, it is possible to capture at 30fps in full resolution. It is also possible to save an image sequence (TIF or BMP). The software is auto numbering the files. The machine cam is sending the frames over the fire wire to the computer in an uncompressed, not debayered format: Y800. The capture software is doing the debayering on the fly. So the preview is in full color, and the saved result too. For more information please visit my website Some pictures are outdated, I must urgent do a website update. Frank Vine has developed special capture software, based on the ActiveDcam engine. This software has an exellent HQ debayering option, live RGB histogram while capturing etc... http://www.cine2digits.co.uk/ Frank is also developing a special RGB backlight source. I have one here and I will test it as soon as possible. I'm using a Computar macro zoom lens, so my camera is 'looking' straight to the film frame. For the backlight I'm using 5W White Luxeon Star Leds. Later I will use Franks RGB light source of cource. To diffuse the backlight I'm using a piece of 'scanhancer' plate. http://www.scanhancer.com/ Many greetings from Ghent, Flanders, Belgium Fred.
  13. For computer use and for the digital master files of your transfered film, the best way to go is one film frame = one digital frame. Progressive, no interlacing. This way, your master file is a digital copy of your original film. Frame 1022 on your film is frame 1022 on the digital file. Use a losless codec for the master files. Huffyuv is a very good one and it's free. Later, if you need 25fps for DVD for expample, you can do this in post. It is important to have the original film frames on the digital master file. The problem with NLE's is the fact that it all happens on the background. It is better to have full control over the frame rate conversion. Play speed is one thing, frame rate is another thing. :o Assuming a one film frame = one digital frame file, then play speed is just a number, telling the computer how many frames/sec it must display. Easy enough to change this with VirtualDub for example. Changing the frame rate is only needed if one wants to make a copy of the file in a standard format that needs another frame rate like standard PAL (25fps) for example. The problem is always the same: we must change the frame rate without changing the original play speed. There are three ways to do this: --------------------------------------------------------------- 1) adding duplicate files, this means some frames will be showed twice. 2) blending: the computer will create intermediate files with info from the previous and the next file, but only on parts with high motion. The more motion, the more blending. 3) interpolation: the computer will draw complete new frames, based on motion analysis. Only a few original frames will be left over. It is possible to convert any play speed to any frame rate, and it can be all done with Avisynth. A correct prepared 18-25fps clip will not be changed by your NLE any more. The NLE will accept it as a standard 25fps file. If you render it to Mpeg2 (DVD) it might add interlacing, but you can force it to render progressive. On high motion scenes, I use frame blending. On low motion scenes, I use interpolation. See my examples on Vimeo. Because I always keep my frame accurate progressive masters on portable HD's and on several backup computers, I can always change them later to another play speed/frame rate. (HD for example) Hope this helps, ;) Fred.
  14. I have just uploaded a new example clip on Vimeo: http://www.vimeo.com/11133342 Imho it shows very well the power of Avisynth. Fred.
  15. Hi Bruce, As far as I know, no one is transfering 8mm films like I do it (with machine vision cameras)commercially. But the HD Sniper from Moviestuff comes close. You can ask them for a frame accurate transfer: one film frame= one AVI frame (progressive). Uncompressed to harddisk if possible. Those uncompressed frame accurate AVI files are ideal to keep as master files and for later Avisynth use. The Huffyuv lossless codec can reduce the size of these files without any loss. The play speed of a frame accurate AVI file is not important, it's just a number in the header of the file, telling the computer how many frames/sec to show. Easy enough to change this in post. Fred.
  16. Avisynth is basically a frame server. It opens a video file, applies all kinds of filters and modifications and serves this virtual clip frame by frame to another application. It works with plain text files. Avisynth has his own scripting langue. It's open source.. free of charge. It works very fine together with VirtualDub. Avisynth has his own, build in plugins. But lots of other plugins are available as *.dll files. They are all written by amateurs. Some of those amateurs are very talented. With the MVTools package for example it is possible to do very sophisticated grain removing and frame interpolation. See my example files on Vimeo: http://www.vimeo.com/user678523/videos/sort:plays Fred.
  17. Wonderful !! 'Made on Kodachrome'.... K40 perhaps? What camera? Fred.
  18. Of cource preserving moving pictures from the past is a relative new issue when we look at the entire history of this earth and human civilisation. 100 years is nothing compared with the age of this planet. So preserving pictures for the future on the long run is not done yet. We will have to wait another 500 years to know if it works or not. A question is: how long do we want to preserve those pictures? 100 years? 500? Forever? Then digital is the way to go because at some point film will fade out and finaly desintegrate. With digital there is at least a chance that the picture still can be seen in the far future. Another question: what do we want to preserve? Dancing uncles and aunts on birthday partys? :D On the other hand.. I would find it very interesting to see common life from -say- the year 1500. Can you all imagine? If the digital migrating system works, then - in 2500- it will be able to see how life was in our time. In the same quality as the original. Let's imagine the year 10.000. Someone is watching a film from 2009, in full HD quality, migrated to whatever the system then will be. (downloading straight to the brain perhaps :P ? ) Now that would be the same as if we could watch -in 2009- a full HD film from the very beginning of human civilisation. Fred.
  19. Thank you! I have a theory about this: the information is hidden in the grain. And because the grain is constantly moving and changing, frame 1 contains other information then frame 2 and so on. Thus averaging multiple frames makes more information visible on all frames.. On the 2009 E64T clips I have averaged 8 frames. PS: The source was very grainy and there was something wrong with the exposure of my Canon 1014. There was random flickering on the original film. (no pattern). The Avisynth script also takes care of this. I will upload a before/after clip to show you all the effect of the script Fred.
  20. Visit My Website ;) And for the post processing I have developed an Avisynth script: film restoring script Fred.
  21. Sorry... It was not possible to edit my posting any more.. That clip is pasword protected. Here is what I wanted to post: Here's a HD 720p25 test, aspect ratio 16:9. Cropped and upscaled from a DIY 1024 x 768 transfer. http://www.vimeo.com/6751675 pasword= videofred From the same film, but this time in 720p50: http://www.vimeo.com/6377661 Format is 980 x 720, so with two 150 pixels black borders we have 1280 x 720. many greetings from Belgium, Fred.
  22. Here's a HD 720p25 test, aspect ratio 16:9. Cropped and upscaled from a DIY 1024 x 768 transfer. http://www.vimeo.com/6751675 many greetings from Belgium, Fred.
  23. Hello Shane, This is typical for the old regular-8 format. You had to open the camera to turn the film to expose the other half. It is caused by light coming on the film. Very often this effect is only visible on one side of the film, at the end of the reel. What's the film-chain method? There should be no flicker.. I bet they have added it to have a more vintage look. PS: on a second look... It might be a mix of different films too. The red/brown scenes looks like faded out colors to me. Fred.
×
×
  • Create New...