Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Shutterless option? 1/24 not enough

I'm aware of the astrophotography applications with CCD's and film.
But I hadn't any idea that this could be applied with any benefit to deliberately (by choice or other) underexposed moving images for noise reduction purposes. I suppose you could write a macro that would automate the process...out of my depth....
 
Frame mixing, effect on compression?

Frame mixing, effect on compression?

Spiff said:
You could do 2 frames at 50% opacity as well to get a 720 shutter. In order to sync up sound, you would want to shift it back 1/2 a frame.
You will obviously end up with some appearance of lag.
Click to expand...

In uncompressed images at high EI going bigger than 360 degrees shutter could be useful for static night shots or shots where small subjects move only 1/3 pixel per frame, about.

If you put a bell curve on the frame groups like was mentioned 25%+50%+25% the sound sync would be centered, maybe. Not go great on close lip sync, but you can use fill light on close shots, when you need less grain noise is on long shots where you cannot use enough light. This 25-50-25 gives you 200% image data so cuts the noise, you can see this type of signal processing in some of the Sony (tm) camcorders for low light dark area noise reduction, maybe, 'cause when you move around you see some lag and artifacts in the dark areas in low light.

If you are getting black clipping, you can add bias lights to shine on the camera senor to bring the dark areas above the sensor dark threshold, this gives you "infinite" EI within the signal noise and fixed pattern noise limits. Flashing used to work on film before they started adding chemicals that introduce "reciprocity failure" to control radiation fog, if the film accumulated all the exposure it would have too short a shelf life from radiation fog. I did some flashing tests a while back on EI 3 film and was able to get images about EI 400, interestingly the grain looked like EI 400 (or worse) even though the film was fine grain, I would expect that the sensor has too many fixed defects for "real time" flashing to be useful over a 1/2 too 1 stop gain. You would need to adjust the foot of the curve in post to make up for the A+B=C effect of the flash or bias lights.

I think a web cam I have does this frame mixing to reduce small sensor noise, maybe, the result does not look as bad as one might think unless you pan, or maybe it is just running at 8fps or something. Can you set the RED ONE (tm) to 8fps? And later triple print the frames up to 24fps? Not as good as mixing 24fps on a bell curve maybe since ALL (not just some) of the grain will stand still for long enough to look like a slide show.

If you cannot get 360 degrees on each frame with the rolling shutter, you could use two cameras in sync and a beem spliter. If the cameras are 180 degrees out of phase you would lose 50% of the light, but it would give you 100% time exposure for blur effects, if you use a spining mirror between the two cameras you can get close to 100% exposure time and close to 100% exposure brightness with 180 degree in each camera exposure.

I don't have any samples to show in motion, that would take an uncompressed AVI to look at since compressed media like MPEG2 would hold the grain for several frames on playback anyway, so you could not see the advantage maybe (although if the grain noise is divided the compression held grain areas should look less noisy?) Not sure if compressors blend frame to frame noise anyway? Might depend on the compression software used to encode?

BTW this is an old idea that was used in optical printing for an effect, just offset the film a frame and make another pass, then repeat, maybe adjusting the exposure to give a bell curve to the brightness of the blur. Mixing frames in film will blur the image since Movie cameras do not have good enough registration, the RED ONE (tm) should shoot steady enough on lock down shots to make frame mixing sharper than you could get with film.
 
"If you cannot get 360 degrees on each frame with the rolling shutter, you could use two cameras in sync and a beem spliter. If the cameras are 180 degrees out of phase you would lose 50% of the light, but it would give you 100% time exposure for blur effects, if you use a spining mirror between the two cameras you can get close to 100% exposure time and close to 100% exposure brightness with 180 degree in each camera exposure."

I'm interested in this as well as I need to shoot continuous footage to include an unbroken blur. I'd rather not use two cameras with a prism as it would be a logistical nightmare. Has anybody used a single RED shutterless, and what was the footage like? Was there any break in the action between frames?
 
tobyho said:
"If you cannot get 360 degrees on each frame with the rolling shutter, you could use two cameras in sync and a beem spliter. If the cameras are 180 degrees out of phase you would lose 50% of the light, but it would give you 100% time exposure for blur effects, if you use a spining mirror between the two cameras you can get close to 100% exposure time and close to 100% exposure brightness with 180 degree in each camera exposure."

I'm interested in this as well as I need to shoot continuous footage to include an unbroken blur. I'd rather not use two cameras with a prism as it would be a logistical nightmare. Has anybody used a single RED shutterless, and what was the footage like? Was there any break in the action between frames?
Click to expand...


It's very interesting that this thread was started exactly one year ago.
Too bad it wasn't 360 days ago...
 
359.75

359.75

tobyho said:
"If you cannot get 360 degrees on each frame with the rolling shutter,
Click to expand...

I think you cannot reset a line untill after it is read out, so you cannot get 360 degrees, more like 359.75, there would be a small dark band in the motion blur, maybe?

One can integrate several frames if the camera is locked down, you add motion dictation so that parts that move more than 1/2 pixel are taken only from the "center" frame, and other parts of the image are mixed from leading and lagging frames, to reduce the noise in the image. You tend to see noise less in moving parts of the image and more in still parts, so it can work to some degree... If you add motion offsets to match for pans and such, you might be able to reduce the noise even with some kinds of camera movements...
 
“Mathematics is like love; a simple idea, but it can get complicated.”

1244220183.jpg


Jim
 
since you want to keep the raw information of the senser you can still shoot 24fps at 1/24 and do the frame blending exposure post. the whole pont of the red is that you have no burnt in effects. one you have laid down on to disk the frame blended exposure you cant get it back.

now if your worried about just getting enough light in to the camera just up the asa and there are some verry clever frame belnding denoising programs that will clean the image up check this post http://www.reduser.net/forum/showthread.php?t=31433

but it sounds what your asking for is somthing that can be proessed off camera, and
attach the frame blending info in the meta data so that redcine or QT will do the work.

It's kind of like the HDR in stills, but because your dealing with a moving image, you need to work out what has changed from frame to frame. but this in combernation with more bits per channle recording we could soon be supassing what film can do. now we just need to make it look like film after we have shot it and we can shoot it almost any lighting and still get a great look.
 
You must log in or register to reply here.
Back
Top