Revolutionising Tripods: The Benro Theta Is An Auto-Levelling Travel Tripod

Poor tint rendition of strong reds by CCD

27 Apr 2007 5:02AM
I do mostly floral pictures, so I run into a lot of strong reds, meaning (230-254)/255 bits and fully saturated. I have major problems on my Pentax *istDL capturing delicate tints away from pure red, especially towards the blue side. It's as if the firmware is overcorrecting for unwanted red spillover into the blue detector of each pixel - and ignoring negative values - so that a gentle gradation from pure red to very slightly magenta creates a "lake" of pure red then a boundary into bluish tints. It's very hard to correct for this by post-processing. I have noticed what seem to be similar effects with some other people's ephotozine pictures. Anybody want to talk about it?
mattmatic 19 598
27 Apr 2007 11:57AM
Personally, I would say shoot RAW and process in something like Silkypix (the latest Pentax Browser software uses the Silkypix engine, but the full package has many more options).

It's not just exposure, but also down to colour space gamut.

If you have to shoot JPG, then at least switch to the AdobeRGB colour space. That will give you a lot more "headroom" in terms of colour.

But, again, RAW will give you the best chance of an accurate capture, with the ability to adjust during post-processing.

The sister site, is probably a better place if it's Pentax specific - I'm on there more often than epz now Wink

Hope that helps!

(same username on Pentax User)
29 Apr 2007 3:48AM
Thanks Matt. I do seem to be doing the important things youmention already, and I am still left with the problem. I always expose for highlights in individual R,G,B channels (not the grayscale meter which is pretty useless for this sort of photography), unless I am delibreately planning HDR. I always save in Pentax so-called 3x12-bit colour RAW, and always check each image in the Pentax RAW browser before saving it as a 16-bit TIFF. These artefacts [lose your little bit of blue when there is lots of red and no green in that pixel] are already visible in the RAW file, so I am firmly convinced that this is a capture problem, it is something lurking in the firmware, and NOT a post-processing problem.
tepot 18 4.4k United Kingdom
29 Apr 2007 3:58AM

Quote:But, again, RAW will give you the best chance of an accurate capture, with the ability to adjust during post-processing.

lol...i have got into a bad habit of relying on RAW to bale me out! much better to think and get it right in camera though.
29 Apr 2007 4:19AM
Thanks for your interest. Perhaps I am not explaining myself well enough. My problem is that (on a scale of 0-255 per colour channel) an image which I can see with my eyes should be approximately Red:250, Blue:8 is captured in the RAW file as Red:250, Blue:0. In a different part of the subject, where the colours look to be approximately Red:220, Blue:7, the captured pixel is Red:220, Blue:4or5. This does not come about because I have "pumped" the colours in the computer. The ugliness arises from the manifest boundary between no-blue and a-little-blue in the captured image, when simple observation indicates a continuum in the subject. I plan to try using a blue A filter to deliberately load up the image with extra blue, then back correct it digitally.
mattmatic 19 598
29 Apr 2007 7:58AM
There are a number of technical items that you should be aware of (bayer matrix, blue channel noise, colour space etc) that I can explain...

But can I suggest you email me:
a) a RAW file
b) a JPG highlighting the problem (annotate too)
and I'll dig in and see what's happening.

(email to matt at mattmatic dot co dot uk)

Sign In

You must be a member to leave a comment.

ePHOTOzine, the web's friendliest photography community.

Join For Free

Upload photos, chat with photographers, win prizes and much more.