Flowers with two false colours

[JCDowdy]

All these plots seem to show blue most sensitive in the long wavelength UV just below 400nm and then red over green. Generally blue and green are both lower than red by 350nm except for the Nikon 1 sensor I posted. I am wondering if that plot may have including some type of cover glasses or micro lens etc.

[Andrea B.]

I've been trying to eyeball the colors off the Basler chart. :D

It appears that green takes the lead "suddenly" where that red makes a sharp dip. I can't imagine how that would play out in false colour land.

 

400 = x/2, x/6, x = purple

385 = x, x/4, x = magenta

365 = x, x/3, x/3 = desaturated red = bright pink

350 = x, x/3, x/6 = orange-red

340 = x, x/2, x/6 = orange

335 = x, x/2, x/2 = pink, again

325 = x/3, x, x/3 = desaturated green (for which we have no name)

 

[Andy Perrin]

325 = x/3, x, x/3 = desaturated green (for which we have no name)

 

Ungreen? Nevergreen? Greennope? ;-)

[Hornblende]

I propose "Uvert" from UV and "vert" (french word for "green") :)

[Hornblende]

In the book Exploring ultraviolet photography by David Prutchi, there is a graph on page 7 that shows the spectral response of the Canon Tli between 300nm and 400nm.

[Cadmium]

I am not sure you are referring to the graph for the Canon Tli. Figure 11 is for the Canon Tli, but for the full spectrum, not really showing the 300nm to 400nm range very well.

However, figure 13 does show more 300nm to 400nm detail "of a full spectrum DSLR camera"., not necessarily a Canon Tli.

 

Doesn't he post on here, some of his inventions?

Maybe he will post that graph on here. Pretty Please?

[Alex H]

It would be even better if he posted the source of the information, or how it was obtained.

[Cadmium]

I sent Prutchi (on here) a message about posting his graph...

[Cadmium]

David told me he will probably be able to post that graph in about a week from now.

[Andy Perrin]

Yes, it's always best to ask the source of your information permission before putting it in a book.

[Cadmium]

Andrea, Do you have that book, or access to it on Google Books to look at that chart (figure 13)?

I think you might be able to derive more detailed Bayer filter UV-A range info from that graph than any other I have found.

It is hard to say if that graph is pertinent to all or even most sensors, but I think that graph is something you should take a look at.

[Cadmium]

Andrea, When I look at that fig 13 graph, it makes sense, because the 300-310nm color is blue green (or cerulean), and the red drops off right about there on the graph, and the green increases ever so slightly too.

Take a look, tell me what you think.

[JCDowdy]

It would be even better if he posted the source of the information, or how it was obtained.

 

I could not agree more. When one adapts a figure from another source is is necessary to say so and to cite the source, even if it is sourced from the public domain.

[Andrea B.]

Figure 11 is a chart for the Canon Tl1 which David measured himself (as per the accompanying text). But I can't make any conclusions at all about the raw colour behaviour from such a small chart. It looks like the Tl1 is basically colorless monochrome between 300 and about 365 nm? We do not see that with other cameras.

 

*****

 

The chart in Figure 13 shows a lot of raw colour. There is no text about this Fig 13 in the book to explain it?

 

At 400nm, that camera gives a slightly desaturated blue.

 

Blue-to-purple-to-magenta between about 390-to-365 as the blue decreases and the red increases.

^{For magenta I use the pattern (X,g,X) where red and blue are equal or close and there is some small amount of green (or none).}

^{For purple I use the pattern (X/2,g,X) where the red is about half the amount of blue.}

 

Magenta to blue-ish red (cerise?) back to magenta from 365-to-330 as the red increases, predominates, then decreases.

Magenta-to-blue from 330-to-315.

Greenish-blue from 315-to-300.

 

********

 

All this has left me puzzled. Going to go look up demosaicing formulas.

 

Like at 400nm we have Violet light. But it would appear that the T12i records spectral Violet as blue. Our eyes would record spectral Violet as a kind of "dark reddish-blue" because our red receptor is stimulated along with the blue. If you go in a dark closet and shine an unfiltered Nichia 365nm UV-LED onto the back of a spoon, you can see spectral violet. It is very vibrant.

 

But wait.......if I'm in the closet in the dark, aren't I using my scotopic vision?

[Andrea B.]

Here is what I do not know about all these charts and curves we see: how is that "raw" colour response extracted from the experimentimental data? Is the actual pixel record being read? Is a demosaicing algorithm used? If so, which one?

 

These few spectral charts we've seen across the internet universe do not always match up with the raw colours we actually see in our files in UV region between 350-400nm.

 

There simply MUST be a clearly described methodology accompanying each chart about how the response was generated and about how the response was interpreted to be come a colour on a chart.

 

What am I missing here?

[Alex H]

Such graphs should in theory show the information recorded by individual R, G and B pixels, so no demosaicing and conversion. How these particulars graphs were created - only the author can answer.

[Cadmium]

Yeah, that is what I said, figure 13. Sorry, it looked like a good graph of 300nm to 400nm to me. He may post it and tell us more info about it too, I mean if we haven't all scared him away...

I think his book is a great book for people who want to learn how to do UV photography.

 

Here is something that I would like to play with, but it is not bandpass, it is longpass, so colors would be different.

http://www.edmundopt...-filters/88363/

[Andrea B.]

That is rather a fascinating filter but I don't get how one would use it?

[Cadmium]

They use those in spectrometers, and I don't know how they could be used other than that, but the colors you see in the photo are visual colors, so part of that would transmit UV-A, which would be colored if you took that photo with your UV filter on the lens.

https://www.edmundop...log/1006234.jpg

 

So it might work similar to the Sparticle by showing UV-A Bayer colors per wavelength, in a way, but more like a UV-A Rainbow. Just guessing.

But kind of expensive.

[Cadmium]

Pretty huh? This is using a linear variable bandpass filter.

 

That one they are using in that video, works basically the same way the Sparticle works, except just in visual, as they are using a more visual range oriented linear variable bandpass filter, and they are using a visually filtered camera. So if you had one of these with a 300nm and above range, and you shot it with your UV-only filter on your lens, then you would see the Sparticle colors of the UV-A / Bayer filter range.

 

The one they are using in the video is this one:

http://www.deltaopti...w%20LVVISBP.pdf

 

They have some that are 300nm and above.

http://www.deltaopti...riable+bandpass

[prutchi]

Thank you Steve for contacting me regarding Figure 13, and thank you for your patience with my answer. I’ve been abroad on vacation, travelling only with my iPad.

Please note that I don’t state that Figure 13 represents measurements for my converted T1i. This graph shows a very stylized response for some Bayer filter dyes in the near-UV. The point of the figure is to convey the idea that the Bayer filter array is another heavy attenuator of near-UV in the path to the sensor.

I’m currently trying to reconstruct my sources from when I wrote this some 3 years ago. Unfortunately, I had to cut my original manuscript by more than half to fit the book in Amherst Media’s fixed 127-page format. One of the casualties was my extensive source of references, which I’m thinking about re-compiling and posting on my blog when time permits.

Cheers,

David

[Cadmium]

David, Would it be OK with you and your publisher to post a copy of that graph here so we can all see it?

[prutchi]

[Cadmium]

Thanks David! That is the most detailed graph I have ever seen of the Bayer filter transmission in UV. I have not seen one before that goes lower than about 350nm.

I would really like to know more info about that graph, who made it, how it was made...

Perhaps it is just a hypothetical graph, but even so it would be good to know if that were the case, and on what evidence that plot may have bee proposed.

It seems to have a different plot from the one you have for the Canon camera, albeit the Canon graph is much more limited in detail and range.

If you could find some documentation of this graph it would be really interesting.

Thanks!