What objects has the most saturated colors?

shutterhappens wrote:

I want to put together some color test samples, to test the color capability of my cameras.

What objects have the most saturated colors (red, green, blue, yellow, purple, orange...)

Or to rephrase, what objects have colors at or near the edge of the visible color gamut?

Oil paint? Crayons? Children toys? Laser? Embroidery threads? Minerals?

I also want the color to be real color, not dithered or simulated color using RGB or CMYK.

I don't think super saturation will help.

Pigments & dyes will probably be the most saturated, but one I work with shows the issue. - 'Solvent Yellow 124' (a European Customs marker for kerosenes) gives a yellow colour at normal dilutions but the concentrate looks more like red presumably due to it's intense saturation.

Most of the options you list will be coloured using pigments or dyes, but can be saturated or pastel shades. Minerals will have natural colours but some are white & some are strongly coloured (even the source of pigments).

Lasers will usually be a single very specific wavelength, but IIRC green laser diodes also have UV output. I'm sure they won't be the only ones with secondary wavelengths (perhaps at half the main wavelength)

To get pure single wavelength light you need to use a element specific light and a filter or other monochromator. Each element when excited will emit a number of very specific wavelengths, & measures can be taken to block out any that are not desired. This approach is used in atomic spectroscopy where wavelengths need to be accurate to 0.01nm or better.

For your purposes it will probably be sufficient to use an incandescent light source with a range of monochromator filters. The standard Tri colour set of wratten #29, #47 & #61 might be adequate but there is also a strong Tri colour set #25, #47B, & #58 or even the technical set of 7 or 8 filters giving narrower transmission windows (& lower transmission).

Perhaps the easiest approach might just to project a spectrum of your light source with a prism (or diffraction grating) & then photograph that. With the right set-up this could cover all the way from UV to IR - Even modified full spectrum cameras can't see outside of the 190-1100nm range (below 190nm air block UV very rapidly, above 1100nm Silicon is transparent to IR). Getting below 300nm will require rather special lenses

shutterhappens wrote:

Wow. Thanks for the detail response. Will try the prism method.

Using multiple converted cameras & working in spectroscopy, has had me think of options for this sort of thing in the past. If the set up is right it should be able to give you the results you want. However the 'quick & dirty approach' I tried just using a diffracting filter on the lens to shoot a lightbulb didn't work very well...

A bright light masked to a narrow slit before the prism then a fair space to a screen, should help spread out the spectrum. You'll probably need long exposures & restricted ambient light.

I've never got round to setting that sort of thing up, though it could be useful for looking at how my converted bodies work outside the visual range. Filters & to some extent lenses I can measure properly on the spectrometer at work

If you can find one of the old sodium street lamps, it is a good monochromatic source.

Ed

In my experience, flowers.

However, the Color Checker is always useful, and it has free downloadable software for making profiles for use in Adobe Raw Converter and other programs.

The Winsor and Newton Artists' Water Colours chart, made with the real paints, is good but hard to come by. You might find a specialist art supplies dealer that has one to sell.

The raw conversion, and especially the profile used, makes more difference than the camera hardware. However, a camera with a wide dynamic range is a Good Thing for recording saturated colours.

This is a huge subject and if one gets too involved in it there will be no time left for taking photos.

Don Cox

shutterhappens wrote:

Wow. Thanks for the detail response. Will try the prism method.

spectral colors form the curved edge of the visible color gamut . You can produce them using an optical  grating or a prism. This kind of investigation can be pushed quite far with modest means, see this thread .

LEDs without phosphor coatings (e.g. most color LEDs, but not white) are strongly monochromatic.  If you're REALLY picky you will want to additionally filter the wavelength because they do have a few nanometers of bandwidth, but for the most part, color LEDs are going to be the cheapest and simplest way to get strongly monochromatic/saturated light sources.

Entropy512 wrote:

LEDs without phosphor coatings (e.g. most color LEDs, but not white) are strongly monochromatic. If you're REALLY picky you will want to additionally filter the wavelength because they do have a few nanometers of bandwidth, but for the most part, color LEDs are going to be the cheapest and simplest way to get strongly monochromatic/saturated light sources.

Colored LEDS can indeed be fairly good, but the spectral power distribution for standard color LEDS is typically about 50nm bandwidth see wikipedia. They'll very probably be good enough for the OP.

Atomic lamps will have emission bands specific to fractions of a nm when I used to determine trace lead in our products I used the emission at 283.3053nm (though the spectrometer probably detected ~1nm either side of 283nm). More familiar to most people would be the Sodium D lines at 588.9950nm & 589.5924nm as produced by old fashioned street lights.

LEDS are certainly going to be much cheaper & simpler than atomic lighting, which would often need lamps costing hundreds of dollars, along with dedicated power supplies...

petrochemist wrote:

Entropy512 wrote:

LEDs without phosphor coatings (e.g. most color LEDs, but not white) are strongly monochromatic. If you're REALLY picky you will want to additionally filter the wavelength because they do have a few nanometers of bandwidth, but for the most part, color LEDs are going to be the cheapest and simplest way to get strongly monochromatic/saturated light sources.

Colored LEDS can indeed be fairly good, but the spectral power distribution for standard color LEDS is typically about 50nm bandwidth see wikipedia. They'll very probably be good enough for the OP.

Laser pointers are quite cheap and have negligible bandwidth for the present purpose.

violet laser 405 nm  (GaN)

green laser 532 nm (frequency doubled from Nd:YAG diode pumped)

red laser ~ 650 nm ?

and further types are easily available.

Uneven illumination, speckle, may be an issue. Shining a blue laser on a white surface may easily add to the spectrum due to fluorescence by the surface material.