Mathew,
Glad to see the forum is back.
I decided to build the scanner with continues film transport. I would like to know what shutter speed (on the Camera) and film travel speed you used in your scanner to arrest the movement.
Any additional information is welcome.
Thanks and Regards.
Hi! Glad you found us. The shutter speed will vary depending on the linear speed of the film and the FPS capabilities of your camera. In my experience, the best thing to do is run a few tests and you will find what works. You may also need to adjust the brightness of your light source if your shutter speed changes.
Good luck and let me know if you have any more questions!
Matthew
Thanks for the reply Matthew. But I found you years ago. I am quite familiar with your designs (I am a regular visitor). As far as I can remember you were running the machine at around 2 fps. What was the shutter speed you found best to freeze the linear movement, at 2 fps? I can start from there.
Thanks again.
I’m sorry I don’t remember and I have disassembled that machine so I could use parts to build the new one. I think it was around 1/1000th of a second. Sorry if that’s not very helpful.
Matthew
It helps. I arrived at a figure 1/2000 of a second running at 1 frame per second. I based my calculation on the Circle of Confusion theory used to define Depth of Field. (crazy)
Were your captured frames blur free and sharp?
Yes, blur free and sharp! You can see the a sample here: [Kinograph] 16mm + Micro 4/3 Camera Test 1 on Vimeo
Thanks. In this sample captures, the positioning of the frames varies. What has caused that?
Inconsistencies in the sprocket design.The bumps on the sprocket that triggered the camera where not exactly the same height. It is corrected in the software, which looked for the sprocket holes and extracted based on their location. The concept was that float in the gate was acceptable if it kept the price down, and that software could solve the issue. In the next iteration, I’ll be using an optical method of detecting sprocket holes so the capture should be much more stable.
depending on what you want to achieve, you can ignore the sprocket holes and stabilise the video based on the content using motion compensation.
it it quite cpu intensive but as long as you don’t mind non-realtime and can write code itsnit too complex.
-Steve
Sorry to say I cannot write code. But I think Matthew can. To be able to detect the frame line without the sprocket holes during capture is highly valuable to an archaival scanner. Not to stabilise. But to send a signal to trigger the camera. Some scanners do that. Very nice to have.
Agreed, @Udayarangi. From what I can tell, I believe the Kinetta has this feature. I’m unsure what kind of sensor is used to achieve this effect and how it compensates for things like fade-to-black or all-black title frames.
Does anyone know about these sensors?
Yes, that’s true. He uses the 1st camera to read the sound.
I thought it is for frame detection purpose.
