Ceci n'est pas une fleur bleue.

[Andrea B.]

Editor's Note: I added the gear and exposure details.

 

Measured the WB. Made a profile in Photo Ninja.

Looks pretty good to me.

 

The CC Card and the flower were photographed between 9 and 11 am in the same location.

Nikon D610-fullSpectrum + Coastal Optics 60/4.5 + Baader UV/IR-Cut Filter + Sunlight

f/16 for 1/1600" @ISO-100

 

 

 

Applied WB and profile during conversion in Photo Ninja.

Got this.

It is not blue.

It should be blue.

Nikon D610-fullSpectrum + Coastal Optics 60/4.5 + Baader UV/IR-Cut Filter + Sunlight

f/11 for 1/2500" @ISO-100

With thanks to the pets for the cat hair contributions!

 

 

This happens a lot with blue flowers or blue-purple flowers.

[Stefano]

So, you see it blue with your naked eyes, under the same lighting (sunlight)? Both eyes (good and bad)? Stock camera? If not stock, is the UV/IR-cut filter "appropriate"?

 

The quirks of color vision. Something that seems much simpler than it actually is. The more you dive in, the messier it gets.

 

Also, I just realized I often wrote “lightning” instead of “lighting”...

[Andy Perrin]

Andrea: flower and color checker are under the same light source?

 

Also, did you know there are more elaborate color checkers? It may be that the Passport is not always enough.

 

The OTHER thing is that the passport’s colors happen to have the largest standard deviation between different editions of the passport on the two far ends of the spectrum, violet and red.

[ulf]

Maybe an individually measured target could improve te accuracy of the profile.

Here is a source:

https://www.silverfast.com/buyonline/en.html#targetfilter

 

They supply reference files for download, for not individually measured targets, like the one I use.

I got that before I bought my ColourChecker and without doing an actual comparison have assumed that more different patches would help creating a better profile.

I do not know if this type of target works with PN, but it is OK in RPP64

[nfoto]

Blue flowers turning purple have been known in colour photography for many deecades. Well back into the dark ages of film.

 

Back then the issue was caused by the colour coupling dyes picking up IR. Some films, such as Kodachrome, were horrible. Other brands less so, but still troubled.

 

Same explanation would apply today? Use an efficient UV/IR cut filter to re-establish pure blue.

[Andrea B.]

I added the shooting details above.

 

I only have the Baader UV/IR-Cut filter to use on converted cameras. But that filter together with a color profile certainly preserves the blue and purple on the color checker card. Bit of a mystery why blue is not preserved on the flower.

[Andrea B.]

So, you see it blue with your naked eyes, under the same lighting (sunlight)? Both eyes (good and bad)?

 

Yes the Blue Flax flowers are blue with the naked eye in sunlight. Blue with my good eye and with both eyes.

 

Stock camera? If not stock, is the UV/IR-cut filter "appropriate"?

 

My only UV/IR blocking filter in use currently is the Baader UV/IR-Cut Filter which has a "square" transmission. I suppose ideally that the UV/IR block should have sloped shoulders, but it is my understanding that the color profile made with the same gear would rectify any excess violet (left side) or red (right side).

 

Andrea: flower and color checker are under the same light source?

 

Yes, the usual strong sunlight here in NM. The photos were made facing the same direction within two hours of each other with identical gear. (I added this info to the first post.)

 

Also, did you know there are more elaborate color checkers? It may be that the Passport is not always enough.

 

I do know about the larger color checkers.

 

Maybe an individually measured target could improve the accuracy of the profile.

 

Interesting. I'll think about this.

 

---

 

I think the first thing I will try tomorrow is to make some photos of the Blue Flax with a non-converted camera.

I will post the results here.

[Andy Perrin]

Andrea, color and spectral distribution can change a lot in two hours. (It especially depends how high the sun is. Extreme example would be sunset vs. one hour before sunset.) I don’t know if that’s it but it would be interesting to check if the problem goes away if you have the flower and the color checker in the same photo.

[Andrea B.]

Andrea, color and spectral distribution can change a lot in two hours.

 

Of course.

I don't think there was much change.

Nevertheless, I will definitely test for that.

 

We had not much sun today, so I couldn't retest.

Hope we get some rain out of these clouds!

[Andy Perrin]

I don't think there was much change.

I mean, how would you know, if it's IR (per Birna's suggestion) or even just very deep red (which our eyes wouldn't necessarily notice, even for the non-colorblind)? The Baader over-emphasizes the red end. Like, if the deep red tripled in intensity, it might look the same to you but not to the camera.

[Stefano]

From here:

 

"Green flowers are very rare (i.e., red-absorbtion is very rare in flowers)."

"Most blue flowers have a red component."

 

This means that one has to be careful about red. I don't like the "square shoulders" of the Baader UV/IR cut filter, this is not how our eyes see. BG glass is much better, a gradual attenuation.

[Cadmium]

Stefano, Indeed, I personally don't tend to use Baader UV/IR Cut filter, because it doesn't have a natural curve, and thus has more red content than natural to my eyes with my cameras.

BG38 and BG40 looks much more natural to me.

[Andrea B.]

The Baader over-emphasizes the red end. Like, if the deep red tripled in intensity, it might look the same to you but not to the camera.

 

Even if there is too much red, the camera profile would correct it -- maybe?? That's why we make profiles for each combination of converted camera + lens + filter -- to try to compensate for the differences in filters, sensors, lenses and so forth.

 

---

 

BG38 and BG40 looks much more natural to me.

 

But they are passing too much violet, blue and UV as compared to the camera's original visible curve and so are also "incorrect" in their own way. So we would need a profile for these filters also.

 

---

 

Here is the transmittance and retinal absorption of the human eye. Supposedly our cameras attempt to match this. (They don't completely match this.) Anyway the Baader UV/IR cut is a bit closer to the eye's absorption in the visible red range than is either BG 38 or BG 39. But none of these filters completely match the human eye.

 

From Corneal Structure and Transmittance by Keith Nichael Andrew Meek

Article in Progress in Retinal and Eye Research July 2015

 

You can see that there are no shoulders in the visible portion of human eye cornea transmittance.

But see next chart.

 

However, when the transmitted light hits the retina, there is some falloff, more on the left/blue side than the right/red side.

 

---

 

The camera does not completely match the eye.

The filters don't completely match either the camera or the eye.

 

The only thing I know to do about this is make a profile which includes the camera + lens + filter.

But the profile is made against some standard which of course

doesn't really match the eye or the filter or the camera --- or reality for that matter.

 

I think you see where I'm going with this.

We try to come as close as we can, but the effort is not always perfect.

 

I will still be trying to track down where the "error" is induced for my Blue Flax.

[Andy Perrin]

The Baader over-emphasizes the red end. Like, if the deep red tripled in intensity, it might look the same to you but not to the camera.

 

Even if there is too much red, the camera profile would correct it. That's why we make profiles for each combination of converted camera + lens + filter.

No, not if the profile is made with a color checker under DIFFERENT LIGHT (I said if the red light tripled in intensity.). The profile is not psychic and can't account for changed scene lighting. They absolutely must be done under the same illumination if you want it to work properly. Not two hours later.

 

You can see that there are no shoulders in the visible portion of human eye transmittance.

That's just the transmittance of the eye, not the overall response of the eye to each wavelength. We can't see the shoulder because it's not in the transmission, it's in the L cones. My statement was that YOU (Andrea) wouldn't notice the change in red light over two hours but your camera probably does.

[Andrea B.]

I wasn't quite finished. I stopped for a minute to enjoy some ice cream.

I should lock these posts up until I finish!!!

**********

 

Please refer to the second chart above which shows how the transmitted light is absorbed by the retina. There *IS* a red shoulder for retinal absorption, but it is not at all steep. It is quite high because absorption continues on into the infrared waveband.

 

 

 

They absolutely must be done under the same illumination if you want it to work properly. Not two hours later.

 

Yes, Andy, I hear you. I can only say that I've done enough of these profiles (like literally hundreds) to know that they don't really change much during the main portion of the day. But, again, as I mentioned above, I will definitely test for this specifically.

[Andrea B.]

Funny thing is, in the tutorial I wrote up about how to make a color profile in Photo Ninja, the photograph I used to illustrate the before & after affect of applying a profile happens to be a photograph of a blue flower !! :lol:

Photo Ninja: How to Make a Custom Light Profile

 

In that example the blue flower looked slightly more purple than it should have, but the application of the profile corrected that and the flower became blue again.

 

So we have at least one example of this working. La !!!

 

I have another link to the purple/blue flower problem to dig out. LINKIE HERE if I can find it.

[Andy Perrin]

Andrea, absorption in the retina still isn’t signal that goes to the brain. Some of the absorbed light turns into heat. That’s why the chart I linked shows signal going to zero at 700nm not 1000nm as you seem to be implying. Beyond 700 isn’t being made into any kind of color signal. Also I think the rods have sensitivity further out but can’t detect color.

[Andrea B.]

The thing is --- none of the following match each other completely: camera, camera profile in Photo Ninja, the human eye, the UV/IR-Cut filters (whether Baader UV/IR-Cut or BG glass), the converter software, the monitor. ( I probably left something out.)

 

That was supposed to be my point somewhere up above which is why I went looking for eye charts.

 

Anyway, we can only do so much to correct all this mismatch-ery. And sometimes our best efforts fail, and we get a purplish flower when we should have gotten a blue one.

 

At this point the only working theory for this particular color anomaly is Andy's suggestion that the light changed in that two hour interval. I don't think it changed all that much, but I have said that I will look again.

 

woo. I'm falling asleep now. See everyone tomorrow.

[dabateman]

Slightly off topic, but I like how Vivek Iyer has determined the response for the Nikon Z6 camera with a Baader venus u filter:

 

Something to remember that camera response plays a factor and its just not the filter transmission curve that you need to think about.

[JMC]

Slightly off topic, but I like how Vivek Iyer has determined the response for the Nikon Z6 camera with a Baader venus u filter:

https://www.flickr.c.../in/photostream

 

Something to remember that camera response plays a factor and its just not the filter transmission curve that you need to think about.

David, that graph of Vivek's is sensor coverglass +Baader U - it's not camera response. That suggests the coverglass is absorbing a lot of the short wavelength UV and starts absorbing just below 400nm. Similar to what I saw with the Sony A7III sensor coverglass.

[ulf]

David, that graph of Vivek's is sensor coverglass +Baader U - it's not camera response. That suggests the coverglass is absorbing a lot of the short wavelength UV and starts absorbing just below 400nm. Similar to what I saw with the Sony A7III sensor coverglass.

On top of that comes the sensor's photodiode's sensitivity-curve and if still present, the losses in CFA and micro lenses.

There is a reason for the big sensitivity gain seen after a monochrome conversion.

[Stefano]

Anyway, I still think that the UV/IR cut filter being used is important. UV can be cut sharply with say a GG400, since our eyes quickly lose sensitivity going below 400 nm, also because UV is absorbed by our lenses, but IR is different, the eye doesn't absorb it up to SWIR. A 365 nm illumination is almost invisible, but a far red LED (730-740 nm) of the same power easily lights up a room, although dimly. I don't think a color profile can correct colors when you have captured too much deep red. Imagine an extreme case, where your filter cuts IR at 800 nm instead of 700 nm. That would have a huge impact on the colors, and you can't fix that.

 

Try a stack of GG400-GG420 and S8612/similar. It should improve it, I hope.

[dabateman]

Stefano I am passively arguing the other way. A LP 420 isn't needed, but rather a Short pass 600nm. You have to remember camera sensitivity. In the IR all the dyes light up. That will lead to purple or odd magenta as the blue and red are taking over the blue signal.

This is also what Birna was indicating.

 

Just scroll through the Lee filter pack spectra to get an understanding of how the spectrum leads to color. Magenta and purple are similar with 400nm and 600nm bumps.

[Stefano]

A camera sensor is much more sensitive to IR than UV, but I remember I did notice a difference in the sky with my Panasonic DMC-F3 when I used my Chinese BG39 alone or with a pair of polycarbonate goggles on the lens. If I didn't use the goggles, the sky was a bit violet. With the goggles, it was blue. If a flower reflects UV and you use a full-spectrum camera, especially with a UV-capable lens, then you might see some difference.

[colinbm]

Magenta & Purple are not in the light spectrum.

I fail to understand the Colour Wheel we use, when colour is linear !

A short pass filter....what a dream, I am waiting for one.