which camera has most accurate colours?

I am looking to buy the DSLR that has the most ACCURATE colours not necessarily the most vibrant, etc....I am

most concerned with colour accuracy whe i take pictures....thanks.

I'd say that's an expert question, not a beginner question!

 

If you shoot RAW rather than JPEG, you can make the output from any given camera look like almost anything, from

natural to psychedelic. And even with JPEG, most cameras have lots of different ways that you can tweak the

output to make it more pleasing to you.

 

Also, you have to realize that no (production) camera can match what your eye sees for all subjects and all

lighting conditions. The spectral responses of the cone cells in your eye are different than the spectral

response of the dyes used in a camera's color filters. And once a particular mix of wavelengths of light has

been turned into RGB values by the camera, there's no way to recover the original mix in order to simulate the

RGB response of your eye. It's almost a miracle that cameras work as well as they do.

 

This is all a long way of saying that I have no particular advice, other than don't worry about it too much. You

can find tests of color accuracy at places like Imaging Resource, but you have to realize that those tests are

very limited.

The concept of accurate colors is pretty nebulous. Do you want this for a technical purpose, like a spectrophotometer?

 

If you mean accurate to your eye, then it's all subjective, and depends more on things like your monitor, color calibration methods, and final output.

 

a RAW file from a Canon or Nikon DSLR is unprocessed, and therefore represents a set of measurements of linear intensity of light falling on the photosites. Each photosite has a filter in front of it... so there are issues with the spectrum of the filters... But unless you're trying to take pictures of say two different microorganism bioluminescences, and then compare the quantity of bioluminescence in one wavelength to the quantity in another.... you're going to do fine with RAW from Canon or Nikon... or Olympus for that matter...

 

Maybe you can elaborate on what aspect of accuracy you're interested in. Accuracy of display on a monitor (depends more on the monitor than the camera) accuracy of display on prints (Depends on color calibration of the printer) accuracy for measurement purposes (like I mentioned above) then it will depend on doing calibration of pictures from a standard color target...

 

Basically, DSLR cameras in RAW mode all do a good enough job that it is possible to correct any practical inaccuracies with the proper calibration software when used properly.

Oh, and let's not forget the light... shining a tungsten hot light on a color target will produce a very different condition than hitting it with a strobe, or a fluorescent, or daylight through a window, or direct daylight, or evening light....

John, it just depend how you set your camera's white balance and how correct you exposed the scene, I don't think any decent model of DSLR with good sensor of any well reputed brand has inaccurate color renderings. Off course the film SLRs are totally different and still the DSLRs are waiting to achieve the color quality and dynamic range of films. But I think you are asking this question because you just have used the point and shoot or mobile cameras so far and not used any DSLR. Color inaccuracy is the characteristic of those P&S and mobile phone cameras, not of the DSLRs so don't worry and just buy a good DSLR like Nikon or Canon without any hesitation. Later on after the usage you would be asking queries about lenses not for the camera. Lenses are the most important part while concerning the image quality, if you sensor is good enough.

This is a really interesting question!

I am always amazed how inaccurate most DSLR models are

on the "Imatest" color accuracy graphs and Macbeth chart.

As David said, you can check ImagingResource.com for details.

They test only in-camera sRGB and AdobeRGB however.

If you go with a RAW mode converter, other tests are needed.

White balance does not affect RAW mode, by the way.

 

Most DSLR models have enough adjustments so you can produce

relatively accurate color, usually by reducing color saturation.

I am not familiar with all the RAW converters on the market,

but the same truism probably holds. So provided your

P&S digicam produces RAW mode, you probably do not even

need a DSLR to produce accurate color.

First questions first:

 

What are you photographing and how is it being illuminated?

 

if you are always photographing with the same lighting set up you can begin your work by calibrating the set up (lighting + camera+ lens) . By same lighting set up I mean the light source and modifier (if any) combination.

 

ColorEyes Camera seems to be the state of the art here http://www.integrated-color.com/cecamera/index.html -- along with X-Rite' s Profile Maker 5 and an X-Rite EyeOne Pro photospectrometer and X-Rite SG target.

 

And of course you'll need a highend NEC or EIZO monitor.

John, here's the answer that nobody wants to hear: As a sensory physiologist, I can assure you that color is a perceptual phenomenon, and the concept of color "accuracy" is little more than a question of what most correctly mimics the response of the human visual system. Because we currently do not have sensors with absorption curves similar to R, G, and B cones, there are always going to be some inconsistancies in how colors are recorded and depicted, at least with respect to human vision. There will always be some colors that don't match or look right. It doesn't matter how good our color management is, 100% color accuracy can't be achieved.

 

If we did have photodiodes with human-like absorptance spectra, then we could create a trichromatic output that would work. But we're not there.

Sarah describes the principles underlying metamerism wherein colors of objects, including prints, change appearance depending on the source of illumination. In the case of prints, the reflectance spectrum of the component pigments or dyes don't coincide with that of the eye. A similar phenomena occurs with a digital (or film) camera.

 

One way to overcome the "accuracy" issue with a camera is to shoot a color chart with a low level of metamerism, such as a Gretag-MacBeth "Color Checker" chart. They are expensive ($80 and up) and come in several flavors. You can then analyze the results with a variety of programs, including Eye-One Match (X-Rite) and Incamera (Pictocolor). This only works for a particular lighting situation, for the reasons cited above. The results are corrected for daylight balance and the process works with scans of reversal and color negative film as well as digital cameras.

I should add that "accurate" is not always subjectively desirable. We become accustomed to the over-sharpened, over-corrected prints from a minilab so that home scans and digital cameras seem flat by comparison. We like our music loud and compressed (which increases the average volume), fluorescent dyes in our laundry soap so whites appear even whiter, and more fat and salt in our food than desirable for healthy living, let alone flavor.

 

An image color-corrected using a chart tends to look flat and lifeless, until you hold the test chart next to the (calibrated) screen.

Sunlight and shade? Simplifying the lighting question.

 

One thing that bothers me about color accuracy

is that although Adobe Lightroom and Adobe Camera Raw

are extremely popular here on photo.net, having developed

almost cult status, I note that in the DPreview.com tests,

the Macbeth chart usually looks less (or much less) color accurate

with ACR than with in-camera JPEG. Sometimes

the ACR conversion borders on unacceptable, e.g. Pentax K200D.

There is no such thing as the most accurate color. A digital camera does not see "color" the sensor senses wave length

and convert that into zeros and ones. That is the "RAW" information that the sensor captures. The collection of RAW

data is not quite a picture with colors yet until you run the data through some software. The software is partially in your

camera processing chip and partially in your computer.

 

Then data has to be translated into something visible to the human eye. That can be on the LCD at the back of the

camera, your computer monitor and/or your printer.

 

Depends on what you need to be the most accurate you will have to calibrate the output device to faithfully duplicate

what is the original color of your subject and the representation from your output device.

Reality is outside of commercial product photographers that need to get their client product colors correct, and often

use very pricy Pantone or X-Rite SG/ Eye-one etc to do so, the vast majority of DSLR users only regularly white

balance. Doubtless not long out of the box, they will have tweaked all the myriad settings or loaded complementary

software that changes whatever semblance of manufacturer's color and luminance settings even if they were accurate

as RAW. As long as users who buy DSLRs haven't an interest in color accuracy and don't demand it, manufacturers

will continue to ignore it and leave it to piece meal solutions by third parties. I'll be glad to see the day when they

integrate that into higher end DSLRs so that one can easily return to defaut color settings. DSLR's certainly have

the theoretical capability of being much more accurate than the most accurate films like EPN-100 or Provia, but until

manufacturers bother results will be "in one's mind eye". I don't keep up on what is currently available and will be

glad to be corrected someday with the above no longer being the case.