Maybe that frame vignette not being as prominent on further shots is just an illusion. So I set the camera at one setting and moved it around. One close and one far. Setting the exposure manually to its highest setting seemed to help the vignette a bit in just eyeballing it, but here is some real results.
Here is the frame closer up. Note obvious vignette.
Here is the same frame further away. the vignette doesn't seem as noticeable. I didn't change any of the settings on the webcam but you can tell that there was some brightness changes.
Thanks to the magic of image manipulation software I'm able to compare the two images side by side for comparison. Just so one doesn't appear "more correct" I rotated them sideways.
Can you tell which image is which? The one on the left is "further away shot" you can tell because the facial coloring is slightly brighter and looks a little better.
Stare at the corners real hard and you can tell that the right side does have a slightly more cropped corner. At that level does losing a little bit of corner make up for the loss of resolution? Is the slightly more cropped corner because the image is darker and a little out of focus?
It is probably in my best interest to attempt to find a halfway point that isn't too far out, but close enough to get a perfectly locked focus. It does appear like the further from the screen you are the more accurate the ability to focus and more accurate color representation.
Showing posts with label telecine. Show all posts
Showing posts with label telecine. Show all posts
Tuesday, March 26, 2013
Adventures in Telecine Vignettes
So I blogged about how I picked my camera for my telecine previously. So I've been tinkering with settings and trying to figure out just how to get the best image quality. Increase exposure, decrease gain, adjust white temperature, etc. Then I came across an interesting artifact I didn't think I'd see.
Here is what my blank canvas looks like. Nothing special. I compared the colors from the far left with the colors from the far right and the center and they match perfectly.
A typical frame capture from the box and you'll note a pretty strong vignette happening. That upper right corner is almost completely black, but it isn't on the source. I had to investigate.
I moved the camera to the right side of the frame and a lot of that vignette disappears.
Who would have thunk it. Same on the left.
Shoving the camera in real real close and letting it stay out of focus shows a slightly different vignette effect.
Pulling the camera further out shows almost no vignette of the original source. You can still tell that the blue of the sky in the middle is a lighter tone than that of the surrounding sky but it doesn't drastically drop off. Of course I'm going to lose some of my resolution, but the colors are truer which is more important to me.
*The last shot shows a slight obstruction from a piece of wood in the bottom left corner. This was just experimentation after all.
I'm going back to the camera and see if fiddling with some of the settings might increase or decrease the vignette effect. If I can't find something worthwhile at least I made some progress today.
Here is a follow up with just to make sure the vignette wasn't a complete optical illusion.
Here is what my blank canvas looks like. Nothing special. I compared the colors from the far left with the colors from the far right and the center and they match perfectly.
A typical frame capture from the box and you'll note a pretty strong vignette happening. That upper right corner is almost completely black, but it isn't on the source. I had to investigate.
I moved the camera to the right side of the frame and a lot of that vignette disappears.
Who would have thunk it. Same on the left.
Shoving the camera in real real close and letting it stay out of focus shows a slightly different vignette effect.
Pulling the camera further out shows almost no vignette of the original source. You can still tell that the blue of the sky in the middle is a lighter tone than that of the surrounding sky but it doesn't drastically drop off. Of course I'm going to lose some of my resolution, but the colors are truer which is more important to me.
*The last shot shows a slight obstruction from a piece of wood in the bottom left corner. This was just experimentation after all.
I'm going back to the camera and see if fiddling with some of the settings might increase or decrease the vignette effect. If I can't find something worthwhile at least I made some progress today.
Here is a follow up with just to make sure the vignette wasn't a complete optical illusion.
Saturday, February 16, 2013
Picking a Telecine Camera
One or the first pieces I looked to find for my telecine machine was the camera. I needed a camera that could disable auto focus, had some exposure control, and I could capture via USB.
My first option was a Canon A70 because we had one sitting around the house and it was natively supported in Canon's RemoteCapture application. Canon officially stopped supporting remote capture on their mid level cameras shortly after the A70 and decided it was a professional grade feature. It was a perfect opportunity to make use of some hardware that was sitting around gathering dust.
In practice, it worked pretty well. The A70 doesn't support manual focus but it does allow for locking focus at a particular level. The exposure, aperture, and color temperature controls are excellent so I was feeling pretty lucky. But then I hit the sticky wicket. The A70 would completely freeze up after taking exactly 3,252 pictures. I assumed that the problem was Canon's RemoteCapture App so I started looking around and found gPhoto which thankfully had a fairly active community. I installed gPhoto on my Ubuntu laptop so I could insure that it wasn't a hardware problem. Sadly, gPhoto also took 3,252 before the camera locked up.
I reached out to the gPhoto community and they replied that every point and shoot camera seems to have an arbitrary limit on number of shots taken before it locks up. The people in similar situations would either manually restart the camera or keep them on a timer to cut power to the camera when necessary. Neither of these options appealed to me because of the lack of a real manual focus meant I'd have to rig a system to try to help keep that focus consistent between restarts.
Sadly, the A70 was no longer my ideal camera.
I briefly flirted with the idea of using a DSLR instead of a point and shoot but realized that most DSLR cameras have a max shutter life in the hundreds of thousands of frames. I'd definitely kill the camera before I was done with the box of film reels I got from my Grandma.
If only there was a solid state camera that supported manual focus that allowed some kind of remote control. So I started digging into HD webcams to see what kind of support was offered on focus, exposure, etc. Specifically, I went looking to the stop motion animation communities to see if there was any kind of super camera that everybody was using. Unfortunatly, there wasn't, but the Logitech Pro HD Webcam 9xx series seemed to be popular option with people who have Windows machines. I dug deeper and found that there was even pretty good support for that series of cameras in Linux.
Now it was just a matter of deciding between the 900, 910, or 920 version of the Logitech Pro. Pricing on the cameras was pretty similar but the 920 is the most widely available as it is the newest. So what was the difference? The 910 was better than the 900 because they upgraded the camera to support 1080p HD. The 910 model was replaced with the 920 that has an upgraded video processor, but a slightly lesser camera sensor.
So I went with the Logitech Pro HD Webcam 910.
So far so good.
My first option was a Canon A70 because we had one sitting around the house and it was natively supported in Canon's RemoteCapture application. Canon officially stopped supporting remote capture on their mid level cameras shortly after the A70 and decided it was a professional grade feature. It was a perfect opportunity to make use of some hardware that was sitting around gathering dust.
In practice, it worked pretty well. The A70 doesn't support manual focus but it does allow for locking focus at a particular level. The exposure, aperture, and color temperature controls are excellent so I was feeling pretty lucky. But then I hit the sticky wicket. The A70 would completely freeze up after taking exactly 3,252 pictures. I assumed that the problem was Canon's RemoteCapture App so I started looking around and found gPhoto which thankfully had a fairly active community. I installed gPhoto on my Ubuntu laptop so I could insure that it wasn't a hardware problem. Sadly, gPhoto also took 3,252 before the camera locked up.
I reached out to the gPhoto community and they replied that every point and shoot camera seems to have an arbitrary limit on number of shots taken before it locks up. The people in similar situations would either manually restart the camera or keep them on a timer to cut power to the camera when necessary. Neither of these options appealed to me because of the lack of a real manual focus meant I'd have to rig a system to try to help keep that focus consistent between restarts.
Sadly, the A70 was no longer my ideal camera.
I briefly flirted with the idea of using a DSLR instead of a point and shoot but realized that most DSLR cameras have a max shutter life in the hundreds of thousands of frames. I'd definitely kill the camera before I was done with the box of film reels I got from my Grandma.
If only there was a solid state camera that supported manual focus that allowed some kind of remote control. So I started digging into HD webcams to see what kind of support was offered on focus, exposure, etc. Specifically, I went looking to the stop motion animation communities to see if there was any kind of super camera that everybody was using. Unfortunatly, there wasn't, but the Logitech Pro HD Webcam 9xx series seemed to be popular option with people who have Windows machines. I dug deeper and found that there was even pretty good support for that series of cameras in Linux.
Now it was just a matter of deciding between the 900, 910, or 920 version of the Logitech Pro. Pricing on the cameras was pretty similar but the 920 is the most widely available as it is the newest. So what was the difference? The 910 was better than the 900 because they upgraded the camera to support 1080p HD. The 910 model was replaced with the 920 that has an upgraded video processor, but a slightly lesser camera sensor.
So I went with the Logitech Pro HD Webcam 910.
So far so good.
Subscribe to:
Posts (Atom)








