Email to Sigma America

Dear Mr. @@@@,

Thanks for taking the time to speak with me today.  As we discussed, I have made a few small discoveries since our original discussions several months ago, that may shed some light on the red channel overexposure that I have experienced in certain situations using my SD14.  Rather than a capture problem in the camera, I have discovered that the issue may be related to the color-space conversion performed by the Sigma Photo Pro software during X3F import.

I have attached two versions of an image, one that illustrates the color issue that I have experienced in scenes that push the limits of the color space (gamut) used in the current Sigma Photo Pro 2.5 software. The difference between these images is obviously intended to represent a worst case scenario, so please don’t regard the “original” over-saturated image as the work of a talentless hack — it is purely for illustration purposes!

Click to read on –>

1) The “original” image is of a red flower, shot at ISO 50 using the 1.07 firware in my SD14.  This sample image is a direct, zero adjustment 16-bit TIFF export from SPP 2.5, that I have cropped and saved to JPEG (sRGB) using Photoshop with no other adjustments.  Illumination of the original scene was by direct afternoon sunlight, and exposure was spot-metered directly on the brightest surface of the flower petal area. There is no orange whatsoever in the original scene… the flower was purely red.  As you can see, the red channel is completely overexposed across the surface of the petals, and the “orange” areas that appear in the shot seem to be due to green-channel information creeping in through some type of color conversion error.  No combination of adjustments in SPP were able to bring the red channel overexposure into control in a way that was usable as a final image.  The detail is simply “gone”, as you can see:

Crop from original image (SPP processed)

2) The “new method” image is from the same source file.  Rather than using SPP, I extracted and converted the raw data section of the X3F file to uncorrected 16-bit TIFF using a very clever and useful utility, which was coded for that purpose by an SD14 owner named Pavel Sokolov.  The uncorrected TIFF is very dark, so I brightened it using the Auto Curves function in Photoshop, then made a simple correction to color using the Hue center and Saturation controls, and finally cropped and saved it to JPEG in the sRGB color space.  Using this simple extraction utility, all of the detail in the source file was preserved, and the red-channel oversaturation experienced in Photo Pro was not a factor at all.  The results speak for themselves.  Obviously it’s not a great photo to begin with, but these results are MUCH better than what is possible with SPP given the same source X3F file.

Crop from same image (X3f Extractor processed)

It’s important to note that it is possible to capture this scene without the red overexposure, but in order to do so using the standard “Sigma” software workflow, the overall exposure must be reduced (in-camera) to the point that the blue and green channels are significantly underexposed.  That is not an ideal solution.  Obviously based on these two example images, the camera *is* capable of capturing the visual information, and in fact is doing so very well… it’s just that the range of color that the camera captures seems to be well outside the limits of the Sigma software’s color space conversion capabilities.  The camera works wonderfully, and thanks to Sigma for a fantastic design… but the software could definitely be improved.

Are there plans to update the Sigma Photo Pro software in the near future?  If so, and assuming that limitations in the available output color space options themselves are part of the problem, is it possible for a larger color space (such as the excellent ProPhoto RGB) to be offered as an output option?  That change alone could provide considerable improvements to the results of color conversion at the fully saturated edge of the CIE 1931 chromaticity space.

If such a change (or some other, unrelated improvement) could help to solve the color conversion and accuracy problems that many users have experienced, it could provide both a huge improvement in workflow and in overall image quality for Sigma owners.  Even more, it could result in a potentially huge improvement in the perception of Sigma’s camera products among professional and semi-professional photographers.

I look forward to hearing your thoughts on this, and please let me know if I can provide any additional information or assistance.

Best regards,

Jason Porter
Roanoke, VA

——————–

NOTE TO BLOG VISITORS:  Further discussion of the X3F Extractor Utility is available on the DPReview forums at http://forums.dpreview.com/forums/readflat.asp?forum=1027&thread=28576411