Shooting the color purple

[SteveH]

Has anyone had probs with getting accurate color rendition of lilacs/purple?

I've noticed that there is a distinct blue shift. Can this be adjusted in camera or is it an inherent problem?

[Tom V]

I remembered a thread where the photographer could not nail down a purple.

http://www.s2pro.com/forums/showthread.php?s=&threadid=389

I suppose there could be many reasons the purple is not showing the way you expect. I know there are many colors that do not reproduce in the camera sensor, on the monitor, on inkjet prints, in film or offset printing. The S2, as well as any other camera have real problems with all the fluorescent colors, and any extreme color near (or past) the edge of the sensor's (or film's) color space.

[ballroom_boy]

That is true what Tom said. It is my understanding that the Adobe 1998 color space can handle the fluorescent colors more accurately. You can get this space by shooting RAW and when converting in EX, choosing Adobe 1998. I have read somewhere (someone correct me if I am wrong) that the EX converter simply maps sRGB to the Adobe 1998 space, thus not really solving this problem. Can someone confirm this?

Maybe the rumoured "S3" will support Adobe 1998 with JPEGS and RAW (and we choose it in camera). A boy can only hope...

[SteveH]

Thanks for the comments - at least I'm not alone.

So far my "purple" appears nice and blue on the camera lcd and on my monitor and prints with a nice blue color ... typical rgb values in ps are r35 g32 b123. A good work around in ps seems to be to use the image>adjust>replace color and then adjust the offending pixels using the hue/saturation/lightness sliders. Seems to work quite well. My offending blue is now a mauve color of r87 g57 b140.


Steve

[jhawk1000]

I try not to take any picture of anything purple since this is the color of Kansas State University---much hated by us Kansas University types. Now a nice crimson and blue is good :)

Mel

[cthornhill]

I was that photographer. Yes, in my experiance purple can be hard to capture.

This is not just an S2 thing - purple can be a tricky color for film and for the printers on the other end of the production line.

I was able to get the S2 to work, and I did use RAW and adjust my color settings on conversion to Adobe 1988.

I feel my issues had as much to do with color in my studio lights as anything, though I have seen the issue before with custom white balances too. I have also gotten good purples from the S2, so I feel it can very much depend on lighting and material.

My objects that had issues were a highly reflective plate, and a cotton shirt. I am pretty shure the shirt was very light dependant, since the dye and whiterners in many clothes also have florescent qualities.

I use dyna-lite stobes in the studio, and I know they run a lot warmer than many systems when you turn down the power (especially through a softbox).

You can get good purple from the S2, just give it a little work on color balance and on lighting, and use RAW if you need critical adjustment. Like I said, it is also an issue for film stocks, so I don't consider it a ding to digital to have issues with color now and again. It is a pain, but not very often.

[SteveH]

You're right - I can get some purples - I tried taking shots of a book cover that has a graded blue to purple. That seesm OK but a purple fabric and a purple plastic material comes out quite blue. Must be something in the chromophore of the dye I guess. It's just that I've been asked to take a few shots at a friends wedding and guess what - the in color is mauve. Hurrah - looks like some time at the computer to get the color balance right.

Steve

[cthornhill]

If the fabric is available (still on a bolt) that they are using, it might be interesting to see what results anyone else gets...Or to test it yourself under some different lights, and or with a roll of film (tranparency).

I don't know if you care that much (I mean about the why), but I would be intreaged to know if anyone else has the same problem with the same material. I suspect they would, but I have allways been to lazy to find a fabric with the issue at the store and test it. The problem is so rare (for me) that if was never top priority, but...

If there is a way to white balance it out (using a colored white card perhaps) that could save you time in post processing...Or you can make an action and batch the shots...

Cecil

[SteveH]

You've been reading my mind.
My thoughts at the moment were to try some print film. Transparency also sounds interesting - I now that with certain blue/violet flowers appear pink - blue filter needed but can give a color cast if too strong. The transparency film is sensitive to the infra-red reflected by these particular blooms - like bluebells.

This problem did occur at my niece's wedding (bridesmaids in mauve) some of the prints showed their dresses as blue - most were OK - bride's mother not pleased - luckily I didn't shoot any (guest - hurrah!)
I'm also going to try with my Olympus E20 - initial tests show that it suffers too but slightly differently! I thought I'd try a range of fabric/colors and see what happens and which one handle better E20 or S2.

I hadn't thought of trying your suggestion of using a custom white balance - worth trying.

Do companies like Fuji produce tech data to show the spectral sensitivity of the sensors in these cameras? Or how they react to different colors? They must have the data but do they tell...?


Steve

[cthornhill]

Steve,

Good question for Fuji, but as far as I know neither Fuji nor any other camera manufacturer is willing to publish this data openly.

I don't think it should be such a secret, but these guys are in a fierce race with big money at stake, so I guess none of them want another area for people to compare numbers - especially one that means little to most consumer types.

I have threatened to build a gamut map for the camera, and I have a friend with the gear, but so far we have never needed it that badly, so we just never did it. Our experiance was that out of the box images from the S2 were close enough to his print target that a device profile would give us very little. If we do any fine artwork shots, we will most likely do one, but not so far...

BTW - I just happened to do a test shot with a new strobe I know is cooler than my Dynalight, and my wife happened to be wearing the purple shirt in question, and it stayed pruple, so I think light color is it for me on that fabric. The same day I noticed someone on Food Network using the same set of plates I had trouble with, and in their shot it was blue not purple...Ha! It I am not alone!

One other thing - I recently put my monitors under profile control, and that has been a big help to me. If you are not using a monitor profile, I would suggest it. It is not cheap, but does return much more accurate color previews for the printing process.

I would also note that the GretagMacbeth chart has a purple square, and is a good test tool. They make a little tiny one (still $75), that people I know use in product shots a lot. I still just have the big one, but both are nice.

Also - I am not saying anything bad about DynaLites! Mine are great, and I love them. I have older packs, and I often dial them down and most of the time have a softbox on the heads, so of course they are a lot warmer that way. Still daylight, but not as cool as other brands (or the newer packs when dialed down). For people and most food this is mostly good, but sometimes I ought to do a custom balance don't since I am often working in RAW mode and get anther chance to set color anyway. That is my problem, not the light's.

Cecil

[KeithM]

Kodak seem happy to publish data on their sensors...

http://www.kodak.com/global/en/digital/ccd/

so not sure why Fuji don't ( well put it this way, I haven't seen any ).

Makes interesting reading if only to get an impression of the technology !

Keith.

[cthornhill]

Keith,

Good point - I should have remembered Kodak is very open about their line of sensors. I worked with this chip a lot as a developer, and really love this line of equipment. Kodak also has lots of other great information on digital image chips - theory and practice. A great paper they did can be found at:

http://www.kodak.com/global/plugins/acrobat/en/digital/ccd/papersArticles/ultimateSensor.pdf

I notice a few new articles in the papers sections including one on 4/3 technical development.

Kodak has done some very wonderful work, and they are very open and helpful in teaching people about the technology. I should have given them credit for this.

Again - I would point out that this issue is very similar to film stock color rendering issues. One of the most interesting exhibits I saw this year was an exhibit of 30" x 40" prints from most major camera systems (it took place in conjunction with the 14n release). In each print we had colorful subjects, people and flowers, high levels of detailed material at a range of scales, and a color chip chart. They shot with Canon's gear (several camers including the 1Ds), Nikon's gear, Fuji's gear, the Kodak 14n, DCS 35mm's, and DCS Pro Back. They also shot with film in MF and 35mm. You could walk the room and compare, and for me the shocking thing was how much better the digital did than film in any format as far as color. I had just forgotten how much work it is to get the color you expect from many film stocks.

I really don't miss that big box of CC filters now that I shoot digital.

Cecil

[lightwrangler]

As well as purple going blue, I have had the reverse happen with a blue african violet ending up purple. No matter what I tried it was not possible to get the real colour. Very frustrating. I might redo this sometime and try CC filters in front of the lens. Perhaps as a combination with Photoshop I can get the colour that actually exists. However I do agree that digital is far more easy to control than film ever was. I have had times in the days of film when entire days were spent on the pursuit of the correct colour, only to end up with something that couldn't be printed anyway. Digital is easier since the CCD seems to be more tolerant and you don't have to keep sending film to a lab and waiting for results.

Also just because we can capture it, it still doesn't mean that it can stand the conversion to CMYK and then offset printing and still look right. Fabrics and wall papers can be especially difficult in this area.

[SteveH]

So what happens if you're red/green color blind?

Just a thought...

steve

[SteveH]

Just had chance to look over the Kodak links you guys have posted - very informative. At least now I'm getting a basic understanding of how the thing works.

Thanks :)