Canon 5DIV Dual Pixel Mode - DPRSplit application

Each pixel on the sensor of the Canon 5DIV is divided into two sub pixels. If you shoot in Dual Pixel mode, two images are acquired: one composite image made by combining each of the two sub pixels; a second image from one of the sub pixels. The folks at LibRaw have made a utility that permits both images to be extracted from the resulting raw file and saved as DNGs.

Canon DPP software permits the focus and bokeh of the Dual Pixel image to be adjusted in post, based on the parallax that results from the slight offset of each sub pixel pair; however, what is even more useful is that the second image, acquired from a single sub pixel, is acquired at -1EV from the composite image - hello highlights!

Take a look here:

https://www.fastrawvie​wer.com/DPRSplit

and here is a right up about the concept:

https://www.rawdigger.​com …mode-highlights-are-there

kirk

This is definitely good news for those who are DR sensitive. I've downloaded it. Will try it out some day.

Shooting highlight tone priority is the way I have been shooting to help preserve some of the DR, based on input from John S. It seems to do quite well when coupled with auto-ISO, not blowing out highlights, but still keeping shadow noise to a minimum. I have never ran that mode on other cameras because I never saw how it could benefit me, but it has done well on the 5D4.

HTP is JPEG only and pretty much underexposes the image by a stop and then applies a tone curve to bring the image back. With the Dual Pixel mode split, you get the extra stop of highlights on top of the original exposure.

One "issue" is that the two split images will have the same EXIF data, including the shutter speed at which the shot was exposed. If you want to use the two split images and merge them into a single image in a typical HDR application, you will need to change the EXIF shutter speed or use an application that will permit you to manually enter the 1 EV difference. However, using ACR to Merge to HDR must estimate the exposure based on the data, because it merges the two exposures into a single DNG with no problems.

kirk

But dual pixel also shifts the focus between the two, correct? How do you merge and keep detail if the image is ever so slightly shifted? Canon had a chance to take a trick from Magic Lantern and could have added this functionality within their firmware and DPP, but didn't (the 2 separate exposures at different ISOs). That is a shame.

TeamSpeed wrote in post #18561683
But dual pixel also shifts the focus between the two, correct? How do you merge and keep detail if the image is ever so slightly shifted? Canon had a chance to take a trick from Magic Lantern and could have added this functionality within their firmware and DPP, but didn't (the 2 separate exposures at different ISOs). That is a shame.

It does not shift focus, it takes advantage of the parallax between the two sub pixels to slightly "refocus" the image in post - for merging the split images, if there is distinct parallax (close foreground objects and far background objects) then the shift may be noticeable, probably at edges between in-focus and out-of-focus areas. I have only tested one image (linked in the above article) and it worked well (see attached). I will test it some more with my 5DIV in a bunch of different near-far conditions and post back here with the results.

For the test image, I used the DPRSplit application to split both images. I used an EXIF editor to change the shutter speed of the -1EV image from the embedded 1/1600 to 1/3200. Then I used Aurora HDR to merge the two - the result is the attached image.

I also merged the two without changing the EXIF at all, using ACR, into a single HDR DNG.

I agree about the DualISO miss by Canon. Who knows how they make their decisions....

kirk

See this thread on Luminous Landscape too:

http://forum.luminous-landscape.com/index.ph​p?topic=123160.0

kirk

I'm interested to see how this works out. According to Canon, for best results, shoot a focal length at least 50mm, f/5.6 or lower, and ISO 1600 or lower. In my case, I shoot mostly landscapes wide angle angle and stopped down more so I wonder how well it would work for me.

In my very initial tests, the merger algorithm that is used (in the case of merging the two sub images into a single "HDR" image) makes a difference in minimizing artifact at the merger boundaries. I have found the LR/ACR "Merge to HDR" algorithm results in artifacts at the boundaries where the two images are combined; LumarRiver HDR does not. Both applications produce a merged image output to DNG.

Full scene, basic raw conversion in ACR, for context. The shadowed wall under the desk meters at about 3EV, the light face of the house outside meters at about 13.5EV. Overcast light outside. With my Sekonic meter, in spot mode, I can meter the highlights and add 3 stops and still retain them. To acquire the Dual Pixel shot, I metered the highlights (outside building) and added 4 stops, knowing that the split DNG from one sub pixel on the sensor would be about -1EV. It worked!

So basically you get an extra stop of exposure into your shadows (overexpose by 1 stop) and then add the highlights from the second split DNG back to replace the blown highlights, increasing the dynamic range of the resulting, merged DNG by 1 stop.

kirk

DPRSplit has been updated:

This is version 0.8.2.12:

http://updates.fastraw​viewer.com …it-0.8.2.12-x64-Setup.exe

http://updates.fastraw​viewer.com …lit/DPRSplit-0.8.2.12.dmg

http://updates.fastraw​viewer.com …t/DPRSplit-Manual-ENG.pdf

Changes (they are reflected in the Manual, of course, too):

- font size on the first run (and margins for Windows) are set based on the screen size. If the base font size setting is changed to smaller, the application window is also changed to smaller, accordingly. Works OK with 1200 by 800 pixels screens. The control is in Preferences.

To facilitate the use of HDR applications:
- Adjust EXIF shutter speed by -1EV for second frame (for HDR merge) – simulates reduction of exposure by 1 stop by setting the shutter speed in the EXIF of the auxiliary frame 1 stop faster. This may be useful (and even necessary) while merging first and second frame in certain HDR stacking applications that analyze frames for exposure parameters.

For the first beta, we were careful not to confuse the raw converters that do not process the extreme highlights correctly. Now we offer a finer control in Preferences:

- Data Maximum: controls the content of the DNG White Level tag:

o Normal White –«white» will be set to the Normal White value found in makernotes of CR2 file.

o Specular White (default) – «white» will be set to the SPecular White value found in makernotes of CR2 file.

o Data Max (or specular) – «white» will be set to the actual maximum found in raw data of the CR2 file or to Specular White, if Specular White is higher.

o Data format max – 16383 (minus the black level if the processing mode is set to subtracting black). This option may result in color artifacts in the highlights (such as pink clouds) if the raw converter incorrectly clips the highlights after applying white balance.

This is from the previously cited thread over on Luminous Landscape:

http://forum.luminous-landscape.com …160.msg1027081#​msg1027081

kirk

The automatic adjustment of the EXIF (shutter speed) of the -1EV image is terrific. I was adjusting it manually and must have not addressed all of the EXIF locations associated with this parameter, because with the auto adjustment of EXIF in the new version of DPRSplit, my ACR merge to HDR (DNG) is clean, with none of the artifacts in the above example.

kirk

I downloaded it and will give it a try.

So a 3 bracketed shot with 2 stop differences would yield 6 1 stop results, correct? That would yield a nice HDR result, wouldn't it?