Frame Averaging Function for still photography

[William Dempsey]

Hi All,

I've been playing around with the frame averaging function today for use in still photography of static objects. I was wondering what others have experienced using this feature. In my testing so far the function appears to help reduce noise in shadows and in dark objects. I am curious to know if anyone has also tested this and found any artifacts or other issues to watch out for. I have a product still photo shoot in a few days and would be interested in some feedback on this feature.


Any thoughts on this?

Thanks,

Bill

 

[David Battistella]

I would advise that you have an extra card with a couple of black shade calibrations on it. Make sure you black shade the camera at temperature in this mode, especially if you are using frame summing instead of frame averaging. If you are going to go back to a more normal setup then either have that black shading calibration ready to load or build in time to black shade the camera.

You really won't get the most out of using this mode without calibrating.

happy shooting!

Battistella

 

[William Dempsey]

Hi David,

Black shading was the first thing I did this morning. I'm in a temperature controlled studio so there is little variation once I calibrate. The only noise I'm getting is when shooting with very low light levels or on objects that are very dark in value. At the moment I'm just running tests to see what some of these newer features might be useful for.

In my tests so far , summing works effectively only for slight increases in exposure time. I'm seeing a noticeable IR issue with even a 2 second summed exposure. Since the summing process appears additive, substantial low light noise when blended makes the image mushy looking. I also am finding that summing is not necessarily eliminating accumulated noise.

Averaging was giving me better results. While it does not allow for time exposures it does seem to help with low light still applications provided the noise level is not too high to begin with. It cleans up low level noise quite nicely as far as I've seen. In one test shot of a black object it actually suppressed noise to the extent that it aided in revealing some fine detail that had been previously obscured by low level noise. It was this test that suggested that it might be a good feature to use for my upcoming project since some of the items are dark in color but have fine detail in those dark areas.

Clearly both summing and averaging have limits to their use. It is still not like the old days with film when one could do an hour long time exposure and only have to worry about reciprocity law failure.

My primary reason for this post is to see what others have experience with these newer features. Both the good , the bad and possibly the ugly.

Thanks for your feedback.

Bill

 

[paulherrin]

that's what it's for, that's what it does... you can also do summing/averaging in post (whether or not you do it in camera). it's a great technique for still subjects. second the shading of blacks being absolutely critical at temperature.

 

[Marc Dunham]

I really want to try this now hah. I sort of forgot how each function works though. I remember figuring out summing for a bit, but I haven't used it in awhile. Has anyone written up a guide for it yet that's a little more detailed then the help section?

 

[David Battistella]

I see it as more of another creative tool to get some interesting in-camera looks and to allow one to go beyond what the electronic shutter allowed previously.

I can can see how averaging would eliminate that fine low end noise as if you shaded for averaging and then switched to regular capture on the same scene you will see some funky action in the blacks. This is why the shading is very important.

Averaging is for normal shutter angles.
summing is more for extending the slow shutter capabilities.

If you are just doing still frames then you can safely shoot 3:1 compression ratios to get that noise floor as low as possible.

there are some great threads here on reduser where Trent from Red explains a lot and the latest manual explains what each mode does.

Battistella

 

[David Battistella]

Here is a frame average test I did months ago with the first beta build.

https://vimeo.com/m/56633038

[video]https://vimeo.com/m/56633038[/video]

 

[paulherrin]

averaging collects multiple exposures and literally averages (blends) them together
this gives you the exact same exposure as if you'd taken one shot, but with the motion characteristics (if any) of a long exposure and with reduced noise/artifacting
this is good for timelapse and still subjects, or for experimental stuff

summing collects multiple exposures and literally sums (adds) them together
this doubles your exposre for each frame and gives you a standard long exposure by concatenating frames
good for super low-light timelapse, still subjects, experimentation

right now, the longest exposure you can do in camera is 16 seconds (epic/scarlet, set frame processing to sum 16, 1sec exposures). there are plenty of reasons to use this long of an exposure (or longer) depending on aperture/iso/filtration for the look you are going for in a frame. maybe you want an evening landscape with everything in focus, so you close down your lens and lengthen your exposure, for example.

remember, these same things can be done post processing. a great technique, when appropriate, is to capture summed frames and average them in post, or capture averaged frames and sum them in post, or capture frames normally and sum/average everything in post to taste - remember that this will change your timing, although for some shots you can interpolate frames using optical flow/etc and then use those made up frames as part of the math.

working on going out and shooting an astro-photography tutorial that covers these concepts and techniques, so stand by

 

[Trent Lillehaugen]

remember, these same things can be done post processing.

One advantage of doing the summing or averaging in camera as opposed to in post is that the summing/averaging will take place before REDCODE compression.

 

[paulherrin]

right you are, but sometimes you need more than 16 seconds... so you can double up in post to get a 32 second exposure

the other benefit is that you get to see what you're working with sooner, of course

should we expect to see longer base exposures in firmware or are we pretty much maxed out at 1 second?