My D300 turns purple into blue

[daniel_sandlin1]

<p>No matter what I do, I can"t get a true purple with my D300. My D200 can dtruggle with purple at times but I can go into settings and adjust the hue and get the tint I need. Not so with the D300, there's gotta be something I can do here!</p>

[elliot1]

<p>Do you shoot RAW, and if so, what program are you using to process your images with?</p>

<p>When you say your are not getting purple, is this on your computer monitor or prints? Or both?</p>

[luke_kaven]

<p>The basic guidelines involve shooting RAW and capturing the ProPhotoRGB color space, and doing all your processing in that space. In other words, use the widest gamut color space for everything up until the final step where you render it for print or the web. Profiling your output devices can help.</p>

<p>To give any more specific help, we'd have to know the particulars of your workflow, and perhaps see an example.</p>

[daniel_sandlin1]

<p>I always shoot in RAW+Jpeg Norm, and view my photos via Nikon View NX2. I do not have any program that will open a NEF other than View. I can get it to make purple sometimes, but most of the time the color purple comes out as blue in the photos right in the camera, and into the software. I learned to keep reds from clipping by shooting in manual and underexposing by a few stops, but that does not work with the purples.</p>

[luke_kaven]

<p>Purple is a funny case, as you probably know. It is a mix of colors at both the short and long end of the spectrum. It involves colors that are right in the slope where the bayer filters roll-off at each end. It's one of the hardest shades to get right. And it often lies outside of the smaller color spaces, like sRGB.</p>

<p>There is nothing wrong with the way NX2 captures. You might like to get a trial copy of Capture One for the sake of convenience and sanity. It is important to capture to a wide-gamut color space such as ProPhotoRGG in a pro workflow. Don't convert to sRGB and JPG until you're all done. </p>

<p>The best practical information for you will be over at Luminous Landscape, where this issue has been written up at length and in detail by the many painstaking landscapists there. </p>

[CvhKaar]

<p>Question : What lenss do you use and are there any filters in front of them ?<br>

If any filters ( UV ?) which ones ?<br>

Some UV fitlers / lenses /in camrea filters influence purple colored subjects becuase of the influence orf UV light which is cut of by most types of glass..</p>

<p> Therfore it is notoriously difficult to get flowers , like bluebells, in the same color blue/purple that your eyes / brains see in the same color on pictures</p>

[rodeo_joe1]

<p> "It is important to capture to a wide-gamut color space such as ProPhotoRGG" - RAW capture is colour space agnostic - as far as anything that's limited by physical transmissive filters can be. And in any case there's no point in using a space that can't be viewed with the equipment that the OP is using (or for that matter with not much else in the world either).</p>

<p>Daniel, I'm surprised that your purples are turning blue. IME it's getting true blues that don't look purplish that's difficult; like the bluebells that C.P.M. just referred to. Could you post an example? It may be that your monitor needs calibrating, or that you have a profile mismatch on your computer.</p>

<p>Try downloading these test images. <a href="http://www.photoindustrie-verband.de/_files/ftp/texte/dqtool/DQ-Tool_Monitor.jpg">This one is to check your monitor</a>, and <a href="http://www.photoindustrie-verband.de/_files/ftp/texte/dqtool/DQ-Tool_Print_13x18cm.jpg">this is to check your printer</a>. Ideally they should look identical when printed and viewed side-by-side under standard daylight, but that's probably never going to happen in real life. These are industry standard images, and (shock, horror) they're in sRGB colour space. The royal blues at bottom left are particularly difficult hues to match between screen and printer BTW.</p>

[luke_kaven]

<p>RJ, I'm speaking of doing post-processing in a color space that is as wide as possible, to avoid out-of-gamut clipping. Doing adjustments in sRGB can easily result in colors that exceed the gamut. Doing adjustments in ProPhotoRGB is much more forgiving. </p>

<p>I'm also not sure what you mean when you say "RAW capture is color-space agnostic." RAW files themselves are color-space agnostic, but capture isn't. Clearly you can capture to a too-small color space, resulting in truncated colors.</p>

<p>Lastly, it doesn't matter whether you have a device that can display the ProPhotoRGB gamut. The point is to preserve the integrity of the color values during processing up until the step where you convert to your display's color space.</p>

[Andrew Garrard]

<p>Belated quick question: What "purple" are you trying to capture? The spectral response of a DSLR is not the same as the eye's (and they're not the same as each other, sometimes within the same company). What appears purple to the eye has only a small chance of appearing purple to the camera - much as fluorescent lights often appear green when captured. There's only so much that a three-colour camera can do. This is one reason that I shoot bluebells (which are, due to the wonders of etymology, kind of purple) using Velvia, some of the time. This kind of problem <a href="http://en.wikipedia.org/wiki/Metamerism_(color)">is not unique</a> to cameras, and until we all take photos with high-resolution spectrographs, we'll be stuck with the problem.<br />

<br />

If you want to fix it, selecting the colour range in your image editing package of choice and fiddling the colour balance may be the least painful solution.</p>

[daniel_sandlin1]

<p>I have tried to post some examples but my connection sucks the big one, and falls on its face when trying to upload photos. The few I loaded here I did at Safeway's free wifi and the those on my facebook page took me days of trying , it was hit and miss, and trying to upload to photo.net from this connection is just impossible. I have been using UV filters, my 55mm micro has a 52mm Nikon UV filter, my 18-70 3.5-4.5 has a Hoya Filter on it, but my 70-210 f4,,yes the original fixed f4 has no filters. I can see what happens shooting without the filters. As for Software, I am getting ready to purchase an older version of Nikon's Capture, sadly I can't afford anything else unless someone knows of a good deal for other kinds. Sorry funds are limited at present time. So far I am not worried about printing as I can't afford a decent photo printer.</p>

[Andrew Garrard]

<p>Daniel - for software, look seriously at <a href="http://www.gimp.org/">The GIMP</a>. While the user interface is a little esoteric (and I could say the same about Photoshop), it's free and extremely powerful. I'd at least give that a go before buying <i>any</i> software, if funds are tight.<br />

<br />

And apologies if my previous post conflicted with the advice from CPM and RJ - I'm a bit jet lagged and glossed over their posts. Still, the spectrum matters, and the best way to fix "not right" is manually in editing software.</p>

[dan_south]

Sounds like a white balance issue. Try setting a custom white balance value using a gray card (inexpensive) or select a

Kelvin value using the LCD screen as a guide. You can always change the value later since your are shooting raw files.

[Palouse]

<p>Good advice and observations above. Now for something that may be off-topic a wee bit. I was reading about cataracts (of the eye) and one of the symptoms is a blue/purple confusion. Have you had someone else look at the pictures? Apologies if you have young eyes and/or perfect vision!<br>

<br />PS: In my experience, purple Petunias are impossible to capture accurately! </p>

[urban_domeij1]

<p>This is a tricky subject indeed, and the blue/purple confusion is not the only implication. <br>

We have mainly two systems of colour when reproducing reality, the additive system that the digital cameras use, and the subtractive system, in the colour film age of the last century the mostly used in photography, and which is inevitably the only one suitable for printing out photos. <br>

Both systems are represented by the "colour wheel", but those "wheels" are different of course. Our additive system has three colours, red, green and blue, while the subtractive system has the complementary colours of those, cyan, magenta and yellow. Mixing the additive colours additively will produce those same subtractive colours, and mixing the subtractive colours subtractively will produce the additive colours. <br>

But, and this is a strong <strong>but</strong>, as the additive colours are made up by mixing subtractive colours subtractively, they actually do not exist in the subtractive system, and likewise, the subtractive colours do not exist in the additive system - as physical true wavelengths, frequencies of light, producing those sensations of colour in our visual system, in the cortex of our brain. Because the "colours" we capture on our chip or film, are not colours. They are frequencies of optical energy, electromagnetic energy emitted from an energy source, and they do not have any colour at all, they are just energy. The colour is a figment of our imagination, it exists solely in our brain. We imagine the colour, it is our response to those energies, and as trichromates, beings sensitive to three "colours", our perception of the colour of light that reaches our vision, is always registered in our brain as a mix of those three "colours", the sensitivities of our three colour sensitive cones in the retina. <br>

It might be a bit difficult to fathom, that the strongly saturated yellow that we see in some flowers is a colour that cannot be saturated in the colour system that we use, neither in the eye, nor in digital photography. In digital photography, yellow is a colour that cannot be saturated, because it must be composed of the two colours red and green. Saturation means absense of other colours, and yellow is entirely made of other colours, there is no yellow in it, in our digital image, only red and green, while blue is absent. <br>

The colour space we use is triangular in shape, and the triangle in the diagram of the colour space for additive colour has its extreme points in other directions than the triangle for subtractive colour. In our additive system, violet, as the violet of short wave light, does not exist, but purple does, as a mixture of red and blue. If the red filtered sensels in the camera completely blocks short wave light, they cannot register violet; only the blue sensels can. So if the red channel is present in the purple representation of a violet tone, it can have three explanations: </p>

<ol>

<li>The colour actually is purple, a mix of blue and red</li>

<li>The red filters in the colour filter array has two humps in their spectral response, one for red, another for violet</li>

<li>The software introduces red where there is no red, but only a saturated blue</li>

</ol>

<p>The CRT. LCD or LED computer screen is strictly additive and cannot produce violet, to display purple, it must use both blue and red. <br>

I don't have the answer to which of the three possibilities is valid, although there's evidence of much confusion. I cannot find on the web any evidence of red filters with violet sensitivity, although possibly the red-sensitive cones in our eyes might have a raised sensitivity at the far end of violet which causes us to assess violet as similar to purple (purple being the mix of red and blue). <br>

I am fully aware that this subject is very confusing to many people, but I have tried to explain it without adding to the confusion. The "colour wheel" is cyclic, but the spectrum is straight and has two ends. The colour wheel joins the ends of the spectrum, making it cyclic, and causing violet to fall adjacent to red in the cycle. Our additive display media however will produce the colour purple to imitate violet by mixing red with blue. <br>

I'm aware that this does not answer the question why blue turns purple or why purple turns blue in images, but I think it adds to understanding the processes that display colour in images, and the problems in designing filters and software for capturing and representing colour, as well as displaying it in different media. The problem in printing is similar, but reversed. In printing, we can indeed produce a saturated yellow, but not blue, in a colour photo. The same when we capture on film, there is no blue in colour film, only cyan, magenta and yellow. When the printer needs a solid blue, as for example to a logotype, printing is done with solid blue separately, as it cannot be made to exact specification with the tricolour CMYK system. </p>

[urban_domeij1]

<p>Now I have the answer to the purple confusion. In the Colour Filter Array (called CFA, or Bayer filter), there are indeed only three "colours", transmission of either red, blue or green. But, as suspected, the red filter is a parametric filter, blocking green, but letting red and violet pass. Hence a violet colour will be registered as a mix of blue and red, the "tri-stimuli" representation of violet, which is mostly called purple.<br>

So even though violet is not present in our computer screens, a violet colour is presented as a mix of blue and red, which gives the same impression as violet would. <br>

It is not surprising that also our colour sensitive cones in the eye has such a colour perception, the "red" cones also sensitive to violet. Hence violet, also when it is a natural light frequency, will be seen as purple by the eye. <br>

There is one type of sensor that lacks this ability to produce purple from violet. The Foveon sensor has no CFA, but a different type of filtration, related to dichroic filtering. Its red sensel layer will not be reached by violet light and thus cannot register it as purple. </p>