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 :)

Dynamic Range Trick?

bryanpapierski said:
will the red one ever be able to do higher frame rates at 4k?
Click to expand...

Probably not, its been said before that this is a hardware limitation of the R1. Only enough throughput for ~36MB/s.

There is only one thing that can give us more DR, and that is MORE PIXIES.
 
OK, HERE'S AN IDEA...

Step 1: Shoot 48 fps.

Step 2: Split the file into two 24 frames per second files, one containing all of the even frames, the other containing all of the odd frames--call them File A and File B

Step 3: In File A put a curve that favors the highlights, and in File B put a curve that favors the shadows. Not too radical in either, keeping the tonal balance similar--just one favoring highlights and the other favoring shadows.

Step 4: Put them back together in their original positions, co-mingled at 48 fps.

Project at 48fps.

Is this a recipe for perceived higher dynamic range or a recipe for disaster?
 
Fuji has announced an interesting new sensor:

The FinePix F200EXR offers 3 switchable modes in one sensor:

High Resolution mode, which deploys all twelve million pixels, and is designed to offer the finest detail of intricate subjects when light is full and even

High Sensitivity and Low Noise mode, which caps two adjacent pixels together to produce 6 million large photodiodes, which are big enough to absorb light in the darkest of conditions, to produce low-light shots of extraordinary quality with minimal noise and grain; and

Wide Dynamic Range mode, which captures different exposures with two sets of six million pixels, which, when combined, gives an excellent level of detail in highlights that would otherwise be lost.
Click to expand...

http://www.dpreview.com/news/0902/09020402fujifilmfinepixf200exr.asp
 
How could you have forgotten to mention the "Film Simulation mode"? That does Black & White and Sepia!

Sounds like it's doing some sort of HDR trick. I wonder how far apart the two exposures are and how it handles fast motion...
 
Simple HDR?

Simple HDR?

I still think the simple idea of using two cameras would give good results, you can get the same shutter time and you do not need to wait for new sensors, just shoot today.

If you are shooting black and white I think you can use a blueish filter to spread the red, green, and blue images then use those as high, med, and low you put curves on them so that the "white" clip is not sharp.

You want the red filter pixels for the bright highlight images (blue filter cuts red light, underexposing the red pixels) since it will give you lighter skin tones. The blue pixels get overexposed so give you more shadow detail than without the blue filter over the lens and opening the iris.

If you sync three cameras and use three filters to offset the sensors filters by a few stops to shoot RGB with three cameras you can use the sensor spread idea also to get HDR for all three colors.

I tried this idea on film and it seemed to work well, I overexposed the blue and underexposed the red, then scanned the film and processed it for HDR black and white. The color film I was using was higher contrast than the RED ONE, so on the RED ONE you should get a longer scale.
 
Dan Hudgins said:
.. but in the RED ONE the compression limits the amount of underexposure you can use before the compression thresholds in the shadow areas become an issue, you can see the shadow issues better if you make the image negative and change the gamma.
Click to expand...

..yep
 
ronx said:
You don't need a new sensor for that. Shoot R1 4k, downsample to 2k using a GOOD downsampler. I wrote my own in Delphi. Noise floor lowers, usable dynamic range increases (a little or more, depending on downsample factor)
Click to expand...

I believe that in the wide dynamic range mode the Super CCD EXR uses two different exposure times for the two pixel arrays. The imaging elements are the same size on both grids, but the charge accumulation time is different on each grid.

EDIT: typo
 
Joe Penna said:
What about shooting in, say, 48fps with alternating shutter speeds? In post, you can HDR both of those into one 24fps stream with a very high dynamic range.

I suppose fast motion would be a problem for that...
Click to expand...

The motion doesn't need to be as fast as you might think. Even VERY small image changes produce artifacting.

I simulated a chip doing this with a 3D Rendering once for fun. It looked like garbage.

As soon as you get into any sort of non-global solutions such as multiple frame combination you're effectively into "VFX Shot" territory. Why bother with all the hastle of creating an HDR image (which is only marginally useful in the raw) when you really just want a sky replacement. Most 'hdr applications' are just small fixes.

Even with still photography I more do a sky replacement than use the HDR plate. Lots of dynamic range is extremely desaturated and flat by definition. Getting the extra 3 stops of rolloff in your highlights isn't worth ruining the rest of your picture.

Also you start to encounter aliasing problems at HDR image depths. You need significant softening between a dark dark and a white white. Crush it and you get a highly aliased line.
 
I think most of the ideas presented here in this thread are not praticable. As soon as motion comes into play, two cameras are not possible to use, because of different optical axies.
Possible would be using the Fuji approach with 2 different elements on a sensor or with 2 different exposures on half of the pattern while reducing resolution.

Anyway, I do not think the dynamic range of the RED ONE is that bad. In fact it is quite good when you are able to reduce the noise floor. Have a look at some examples I made in the second and fourth posting of this thread:
http://www.reduser.net/forum/showpost.php?p=369939&postcount=2
http://www.reduser.net/forum/showpost.php?p=370111&postcount=4

Those examples are not espacially HDR, because I wanted to show what you can get out of ISO2000, but I could also improve a highlight protected shot in the darks - which in fact increases the usable dynamic range.
 
Robert Niessner said:
Anyway, I do not think the dynamic range of the RED ONE is that bad. In fact it is quite good when you are able to reduce the noise floor. Have a look at some examples I made in the second and fourth posting of this thread:
http://www.reduser.net/forum/showpost.php?p=369939&postcount=2
http://www.reduser.net/forum/showpost.php?p=370111&postcount=4
Click to expand...

Very interesting.

In the caption, it says "Denoised and unsharpmask".
Was this done in Photoshop on the still image, or on the video footage? If on the video, what software do you use?

Reading the rest of the thread, I would guess exporting to DPX images, processing one image manually in Photoshop, then using automation to apply to all images?
 
two cameras

two cameras

Robert Niessner said:
two cameras are not possible to use, because of different optical axies.
Click to expand...

You can use two cameras set to the same shutter angle with a 5%/95% beam splitter, you use four lenses and a field lens in a fixture and correct any small alignment issues with sub-pixel software, to get HDR. With the right optics used you get in effect a single optical axis for two cameras.

Please explain what you think is not possible?
 
AntonyCASAFilms said:
In the caption, it says "Denoised and unsharpmask".
Was this done in Photoshop on the still image, or on the video footage? If on the video, what software do you use?

Reading the rest of the thread, I would guess exporting to DPX images, processing one image manually in Photoshop, then using automation to apply to all images?
Click to expand...

All the processing was done in After Effects on the video footage. Photoshop wouldn't be a the right tool for this job. After Effects has got similar filters like Photoshop, hence the "Denoised and unsharpmask".


Dan Hudgins said:
You can use two cameras set to the same shutter angle with a 5%/95% beam splitter, you use four lenses and a field lens in a fixture and correct any small alignment issues with sub-pixel software, to get HDR. With the right optics used you get in effect a single optical axis for two cameras.

Please explain what you think is not possible?
Click to expand...

You are right, but I think that fixture would be expensive to build and I guess not very practical to use because you would have to avoid sub-pixel miss-alignments thru motion, lens inaccuracies aso. That will make a heavy construction with at least two cameras.
I guess it will be possible to build that device but it won't be easy. Also it will need some extra processing power when doing sub-pixel accurate alignment on 4k files.

Maybe I am wrong and this thing is much easier to do?
 
Elphel trick, split gain of two greens...

Elphel trick, split gain of two greens...

Just found out about a trick the Elphel camera can now use, they use two gain values on the greens, that gives more shadow detail on one of them, after you do some tricks to get chroma in the shadows it may give some extended range.

Maybe it would be an easy firmware change for the RED ONE. The Elphel compressed the RAW data in four images, so they can compress the two greens as high gain and low gain. If the RED ONE just has one Green image in its compression, then it would not be possable to have two gains, unless you make the Green image HDR in the camera and compress with more than 12bits resolution.

Here is the link,

http://linuxdevices.com/articles/AT4187053130.html

==

About using two cameras, there are people with money that can get things done to get a higher quality look for their projects. Easy does not matter when you can get things done, it just matters that it can be done. At any rate what people concider easy or hard depends on their abilities and experence.

The goal I would think is to imporve the dynamic range for those shots that need it so that one can shoot all Digital without feeling that he is getting a result less than what he could have shooting 35mm negative film. Film costs much more than it could cost to build a two camera fixture yourself, and the fixture can be reused any number of times for those shots the require the extra effort.
 
You must log in or register to reply here.
Back
Top