Reading mode: Light Dark

colour purple

Go to last post
S

sylvia3

New member
Messages
1
Reaction score
0
Location
sydney, AU
Why are Sony and Canon digital cameras unable to photograph purple. I am sure they know about the problem but keep saying I am the only one they have had a complaint from. Looking around the net they had complaints back in 2002.
 
sylvia3 said:
Why are Sony and Canon digital cameras unable to photograph purple.
I am sure they know about the problem but keep saying I am the only
one they have had a complaint from. Looking around the net they had
complaints back in 2002.
Click to expand...
You deserve a decent response from someone
with an IQ far higher than mine, but here goes ...

Right or wrong, here's one of the possible explanations that have
been suggested. Consider definitions of "purple" and "violet":

Define "purple" as a mix of Red and Blue.

But then consider the spectrum of visible light; ranging from
"reds" at the low-frequency end, to "violet" at the opposite,
high-frequency end. (Call that "spectral violet", just to clearly
distinguish it versus the previous definition of "purple".)

There are NO "red-light" frequency components in such spectral
"violet" (here we are speaking only in terms of frequencies of light,
not in terms of our visual perception).

Now consider our perception of "violet".
We perceive that spectral "violet" looks something like "purple".
(Even though a pure spectral-violet light does not contain any
'red' frequencies of light, still we 'perceive it as though it did').

Speaking loosely, spectral violet stimulates our eyes' red and blue
receptors -- in other words, our eyes' response to "spectral violet" light
is somewhat similar to our eyes' response to "purple as a mix of
red and blue light".

Unfortunately a camera's sensor does not respond
to spectral violet light in the same way as a human eye:

A typical digital camera's 'Bayer type' sensor array contains pixel-sensors
individually filtered to respond to 'Red', 'Green', or 'Blue' frequency ranges.

[Note: We can ignore the additional so-called "Emerald" or cyan-like
value that Sony used in the sensor that appeared in the DSC-F828.
It doesn't matter much for this discussion. (Also, obviously
I am not including Foveon sensors in this discussion.)]

When the camera constructs the final image, those pixel sensors'
output values contribute respectively to the color-interpolated R/G/B
components of the pixel-values in the camera's final representation
of the image.

The camera's 'red receptors' (the pixel-sensor photosites which are
filtered to respond to red light) essentially don't respond to spectral
violet. Only the camera's 'blue receptors' (the pixel-sensor photosites
which are filtered for blue response) will respond strongly to spectral violet.

Result: Your photo of a "violet" object (a violet flower, for example)
may turn out too "blue", and not "purple" enough. The camera sensor
didn't 'see' much red in that mostly-violet light reflected from the flower.

The camera sensor has no straightforward way to recognize
spectral violet as anything other than 'blue' -- since only the 'blue'
pixel-sensors responded strongly to the violet light.

Those 'blue' sensors are filtered for a 'blue' range of light-frequencies
(a range including the spectral violet), AND designated to represent
a BLUE component of the final output R/G/B pixel-values.

Since only those 'blue' pixel-sensors (and none of the 'red' ones)
responded strongly to the violet light, the violet flower
unfortunately appears blue in your photo.

Also, the camera cannot (automatically) determine that some red value
should be 'artificially added' to make the result look 'more purple'.
(You can try alternative white-balance settings as noted below.)

Yet SOME "purple" flower photos do turn out OK -- perhaps because
some "purple" flowers have just-enough "redness" in their pigments,
to get an adequate response from your camera's red-responsive
photosites.

Someone who really knows this stuff can argue for or against
such an explanation. (I'm ignoring some other popular arguments,
e.g. about "ultraviolet", or about monitors or other things involved
in the process of producing the image you see, etc.).

Right or wrong, the explanation above doesn't address
the sometimes-discussed supposition that "Sony sensors"
(used in various brands of cameras, not only in Sony cameras) may be
particularly susceptible to the problem. But I don't know about that.
(I have not found information to convince me to 'blame Sony'.)

In any case ... you may be able to partly compensate for the problem:

Use "manual" white-balance (if available -- see your camera manual).
Or try another preset white-balance setting (e.g., try 'Flash' WB).

If your camera supports RAW capture, try that.

Or adjust hues in Photoshop.

Or try out a new camera, and see if results turn out any better.

Good luck. You're right, you are certainly not the first person
to complain to a camera manufacturer about this issue.
Nor are you the first to be told that nobody else ever complained,
or that the manufacturer has never heard of the problem.

-- omr
 
Seems a good possible explanation to me, omr, amatuerish as I am at this stuff. ;-)

My first experience with this type of problem was more about blues being purple, not purples being blue. I couldn't get blue denim pants in a photo to be their actual color, police car light blue. I tried manual white balance and that helped somewhat but never really lost that purplish color. By your reasoning then somehow red pixels were being introduced into the picture from what appeared in my eyes to be only pure blue (and white in the fabric threads).

I have had some success in getting purple (or violet) in flowers to appear correctly enough. Less luck in getting every blue to be right.

Of course, what also factors in is monitor and printer; which I don't ever seem to get completely right. Maybe I don't try hard enough. Heh-heh.
 
This is a Sony picture...looks like purple to me.


sylvia3 said:
Why are Sony and Canon digital cameras unable to photograph purple.
I am sure they know about the problem but keep saying I am the only
one they have had a complaint from. Looking around the net they had
complaints back in 2002.
Click to expand...
 
I found no support for your claim that red receptors are stimulated by the violet wavelength!



Going by the logic you laid out for the rest of your post, even WE shouldn't be able to perceive violet! ?!?!??
omr said:
sylvia3 said:
Why are Sony and Canon digital cameras unable to photograph purple.
I am sure they know about the problem but keep saying I am the only
one they have had a complaint from. Looking around the net they had
complaints back in 2002.
Click to expand...
You deserve a decent response from someone
with an IQ far higher than mine, but here goes ...

Right or wrong, here's one of the possible explanations that have
been suggested. Consider definitions of "purple" and "violet":

Define "purple" as a mix of Red and Blue.

But then consider the spectrum of visible light; ranging from
"reds" at the low-frequency end, to "violet" at the opposite,
high-frequency end. (Call that "spectral violet", just to clearly
distinguish it versus the previous definition of "purple".)

There are NO "red-light" frequency components in such spectral
"violet" (here we are speaking only in terms of frequencies of light,
not in terms of our visual perception).

Now consider our perception of "violet".
We perceive that spectral "violet" looks something like "purple".
(Even though a pure spectral-violet light does not contain any
'red' frequencies of light, still we 'perceive it as though it did').

Speaking loosely, spectral violet stimulates our eyes' red and blue
receptors -- in other words, our eyes' response to "spectral
violet" light
is somewhat similar to our eyes' response to "purple as a mix of
red and blue light".

Unfortunately a camera's sensor does not respond
to spectral violet light in the same way as a human eye:

A typical digital camera's 'Bayer type' sensor array contains
pixel-sensors
individually filtered to respond to 'Red', 'Green', or 'Blue'
frequency ranges.

[Note: We can ignore the additional so-called "Emerald" or cyan-like
value that Sony used in the sensor that appeared in the DSC-F828.
It doesn't matter much for this discussion. (Also, obviously
I am not including Foveon sensors in this discussion.)]

When the camera constructs the final image, those pixel sensors'
output values contribute respectively to the color-interpolated R/G/B
components of the pixel-values in the camera's final representation
of the image.

The camera's 'red receptors' (the pixel-sensor photosites which are
filtered to respond to red light) essentially don't respond to
spectral
violet. Only the camera's 'blue receptors' (the pixel-sensor
photosites
which are filtered for blue response) will respond strongly to
spectral violet.

Result: Your photo of a "violet" object (a violet flower, for
example)
may turn out too "blue", and not "purple" enough. The camera sensor
didn't 'see' much red in that mostly-violet light reflected from
the flower.

The camera sensor has no straightforward way to recognize
spectral violet as anything other than 'blue' -- since only the 'blue'
pixel-sensors responded strongly to the violet light.

Those 'blue' sensors are filtered for a 'blue' range of
light-frequencies
(a range including the spectral violet), AND designated to represent
a BLUE component of the final output R/G/B pixel-values.

Since only those 'blue' pixel-sensors (and none of the 'red' ones)
responded strongly to the violet light, the violet flower
unfortunately appears blue in your photo.

Also, the camera cannot (automatically) determine that some red value
should be 'artificially added' to make the result look 'more purple'.
(You can try alternative white-balance settings as noted below.)

Yet SOME "purple" flower photos do turn out OK -- perhaps because
some "purple" flowers have just-enough "redness" in their pigments,
to get an adequate response from your camera's red-responsive
photosites.

Someone who really knows this stuff can argue for or against
such an explanation. (I'm ignoring some other popular arguments,
e.g. about "ultraviolet", or about monitors or other things involved
in the process of producing the image you see, etc.).

Right or wrong, the explanation above doesn't address
the sometimes-discussed supposition that "Sony sensors"
(used in various brands of cameras, not only in Sony cameras) may be
particularly susceptible to the problem. But I don't know about that.
(I have not found information to convince me to 'blame Sony'.)

In any case ... you may be able to partly compensate for the problem:

Use "manual" white-balance (if available -- see your camera manual).
Or try another preset white-balance setting (e.g., try 'Flash' WB).

If your camera supports RAW capture, try that.

Or adjust hues in Photoshop.

Or try out a new camera, and see if results turn out any better.

Good luck. You're right, you are certainly not the first person
to complain to a camera manufacturer about this issue.
Nor are you the first to be told that nobody else ever complained,
or that the manufacturer has never heard of the problem.

-- omr
Click to expand...
 
It would seem to me that we see the high-frequency extreme of the spectrum as violet, violet should be a primary color instead of blue.

The three primary colors should stimulate the 3 different types of receptors in our eyes separately as much as possible. According to this chart



Violet should stimulate the blue receptor the most, with very little green or red. If you use violet as a primary color you should be able to create blue by adding a bit of green and red. OTOH if you use blue as a primary color you should never be able to create violet, since you'd have to take away more green and red from blue to create violet, which is impossible.

So did monitor manufacturers and camera sensor makes opt for blue because they couldn't make a violet light or violet sensor??

Are we not seeing the full range of possible colors with RGB monitors?
 
http://forums.dpreview.com/forums/read.asp?forum=1009&message=12631532

I've been photographing sony flowers for 4 years, and they are the truest blues and purples of any camera out there. Why?

Because it is known that sony can have a slight bluish cast on some photos. [for canon its a slight yellowish gray] Canons reds and blues suffer from that. Of course when someone gets home from a sunset shot or landscape, they aren't really going to notice the slight difference. But for garden enthusiasts, they usually are photographing their own gardens and see the color everyday. They know when its not accurate. Every camera is different, but I can assure you Sonys blues and purples hold truer than any other camera. In REAL color. I just did some comparison shots today, and my hunches were correct. Check out the first purple flower on the post above...
sylvia3 said:
Why are Sony and Canon digital cameras unable to photograph purple.
I am sure they know about the problem but keep saying I am the only
one they have had a complaint from. Looking around the net they had
complaints back in 2002.
Click to expand...
--



http://netgarden.smugmug.com/
DSC V1 Sony, Mavica FD88, Fuji s5000, Canon 20D
 
unless you print one out, you really don't know...although one printing place had too much blue in their processing, and some grass looked a little bluish compared to my camera/monitor.

Sonys have more blue available than any other camera. You can see it in sunsets a little [they won't be as pink as you wanted] and you can see it in flowers where the blues are about as true as you can get compared to other cameras/ Having my own garden and seeing true color everyday, and photographing with now up to 4 different cameras, I do know what each camera tends toward. Sony's are more accurate in flower color than canon and Nikon. I know that for sure. I just took some comparison shots with the sony and canon. Canon is off by being too pink and has a grayish yellow cast on their whites. So a palr lavender with a Canon will be a pale pinkish lavender. I was just noticing the other day, my wine colored flower was dark cherry colored on the Canon, and the sony held its violet and the color was true to my eyes.
B. H. said:
Seems a good possible explanation to me, omr, amatuerish as I am at
this stuff. ;-)

My first experience with this type of problem was more about blues
being purple, not purples being blue. I couldn't get blue denim
pants in a photo to be their actual color, police car light blue. I
tried manual white balance and that helped somewhat but never
really lost that purplish color. By your reasoning then somehow red
pixels were being introduced into the picture from what appeared in
my eyes to be only pure blue (and white in the fabric threads).

I have had some success in getting purple (or violet) in flowers to
appear correctly enough. Less luck in getting every blue to be
right.

Of course, what also factors in is monitor and printer; which I
don't ever seem to get completely right. Maybe I don't try hard
enough. Heh-heh.
Click to expand...
--



http://netgarden.smugmug.com/
DSC V1 Sony, Mavica FD88, Fuji s5000, Canon 20D
 
I work as a fine-art repro photographer, and I agonzied about this issue for years. I work with museums around the country(US) and use every camera, from my Sony point and Shoot to the BetterLight scan back.

I use all the color programs available, and calibrate everything according to spec. The problem as I see it is not a failure of the sensor or the camera, it is a logic error.

The RGB color model is not a model designed to simulate human color perception, it is a model used to create illumination using a trichromatic emitter. It is a fundementally linear model. Colors are not arranged in this model the way we see, but rather as linear fractions of RGB output. The chromaticity of the illumination stays constant as the intensity changes, which is not the way we see, at all. I am aware that RGB values are deisgned to match CIE chromaticity coordinates, but the CIE system is not a model of human color perception, it was a test of the laws of trichromaticity. CIE coordinates provide a device independent, mathmatical abstract useful for matching colors, but not a model of how to create colors.

I have a Photoshop plug-in that fixes this. It was developed for my work with museums to produce color accurate images. At risk of sounding like advertising, which I legitimately want to do on this site, I would refer you to http://www.tribecalabs.com for more information regarding this problem. Please visit the technology page, http://www.tribecalabs.com/technology.htm and click on the link to see an actual photographed spectrum. You can clearly see how the digital reproduction of the component color of white light is lacking violet and deep blue.
 
If you are having a problem with purple then it is most likely that your camera white balance needs to be adjusted. This can be done after the fact in post processing. If you provide an example of a problem shot, I can make the adjustment.
 
I noticed my nikon had problem with reproduciong black. a black polar jacket always had sort of violet cast. i don't know whether it was a white balance problem.
--
mc
 
"I can assure you Sonys blues and purples hold truer than any other camera."

Why make the above statement when everyone knows you haven't shot blues and purples with all other cameras in existence??
 
There are three types of cones (the color receptors in the eye) and these are sensitive to curves that correspond to red, green and blue (tristimulus values).

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/rodcone.html

http://hyperphysics.phy-astr.gsu.edu/hbase/vision/cie.html#c2

The eye, the camera's sensor, the colorspace used when processing the image, various monitors and printing devices all have different color gamuts. The eye/brain system has a much wider gamut than any device. There is no way that all the hardware gamut limitations can equal the color gamut of human perception.

What color reproduction professionals do to get better results is selectively remap colors that can't be reproduced via display or print technology into color ranges that are possible.

--
BJN
 
It happened to me with two of Canon cameras! The photos of a violet flower becomes blue. I repair the color in PP but the surroundings becomes strange. I add an Olympus E500 and the same kind of flower have the right color. Is not the WB but is the camera processing.
 
You must log in or register to reply here.

Similar threads

tarapada focus
Lake District Round 2 - Autumn in the North Lakes
Replies
14
Views
18K
tarapada focus
tarapada focus
D
Portraits of Flying Purple Martins. OM1.2+MC20+300mm.
Replies
4
Views
576
drj3
D
Yannis1976
Which of those 35mm lenses would you recommend?
2 3
Replies
44
Views
25K
Paulmorgan
P
A
How do I make my film photos less "digital"?
Replies
16
Views
37K
alex_virt
A
H
ON1 Photo RAW: small and medium lossless compressed RAW unsupported?
2
Replies
24
Views
6K
Htbaa
H

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top