Spectrum Profile lines

whisky said:

I'm not entirely sure how he did it but it looks like he took the pic with a Canon.

Will ask him next time I see him.

Cheers.

Click to expand...

Yes - I've tested with diffraction grating and what I find is the colors sort of lump together - blues overlap greens, etc.

What I seem to need is a very narrow beam of light coming through a narrow portion of the diffraction grating

I'm thinking of setting up a black box (non-reflective interior) and then cutting a small slit - maybe 1 or 2 mm wide - in one end and setting up a camera strobe set to "tele" on the other end.

That should create the narrow beam of light through a small section of diffraction grating that I would need to create a relatively overlap-free spectrum.

Then the next thing is figuring out what exposure one each camera - but I figure I can just bracket and choose the best file in post.

saltydogstudios said:

whisky said:

I'm not entirely sure how he did it but it looks like he took the pic with a Canon.

Will ask him next time I see him.

Cheers.

Click to expand...

Yes - I've tested with diffraction grating and what I find is the colors sort of lump together - blues overlap greens, etc.

What I seem to need is a very narrow beam of light coming through a narrow portion of the diffraction grating

I'm thinking of setting up a black box (non-reflective interior) and then cutting a small slit - maybe 1 or 2 mm wide - in one end and setting up a camera strobe set to "tele" on the other end.

That should create the narrow beam of light through a small section of diffraction grating that I would need to create a relatively overlap-free spectrum.

Then the next thing is figuring out what exposure one each camera - but I figure I can just bracket and choose the best file in post.

Click to expand...

Been there, done that, sd.




Blades approx 0.3mm apart




PM me if you want one ...




SD9 shot of a CFL spectrum




Nikon D50 shot of same spectrum

HTH

--
Ted

xpatUSA said:

saltydogstudios said:

whisky said:

I'm not entirely sure how he did it but it looks like he took the pic with a Canon.

Will ask him next time I see him.

Cheers.

Click to expand...

Yes - I've tested with diffraction grating and what I find is the colors sort of lump together - blues overlap greens, etc.

What I seem to need is a very narrow beam of light coming through a narrow portion of the diffraction grating

I'm thinking of setting up a black box (non-reflective interior) and then cutting a small slit - maybe 1 or 2 mm wide - in one end and setting up a camera strobe set to "tele" on the other end.

That should create the narrow beam of light through a small section of diffraction grating that I would need to create a relatively overlap-free spectrum.

Then the next thing is figuring out what exposure one each camera - but I figure I can just bracket and choose the best file in post.

Click to expand...

Been there, done that, sd.

Blades approx 0.3mm apart

PM me if you want one ...

SD9 shot of a CFL spectrum

Nikon D50 shot of same spectrum

HTH

Click to expand...

Wow - I love it when I learn that I'm actually on the right path.

No need to send me diffraction grating - I have a small bag of the same - I think maybe 2 dozen of them. Thank you for the offer, though.

This is the sort of thing I was getting and what led me to decide I'd need to build a box like the one you just posted.

This is a grey card target (spectrally neutral according to the manufacturer).



Once set up, I'd just sort of try to clamp the whole thing down and take the same photo with a series of cameras hoping things didn't move around too much.

So my goal here is to compare multiple cameras to each other - for example Canon vs Nikon - and see if there was any way I could - with this admittedly simplistic method - measure the frequency response of each CFA - how far into the blue spectrum does the red CFA go? etc.

So I'd open it in RawTherapee and export the mosaic (non-demosaiced image) and open it in photoshop & see - for just the red pixels - how far into the green/blue spectrum to they extend at a given intensity.

I know there's like a million things wrong with this technique but the other avenue that's possibly open to me is to get those modern RGB lights that have Red or Green or Blue LEDs and turn on just the reds, just the blues, just the greens and measure the ratio of Red to Green to Blue under each condition.

The LEDs may not be "pure" (narrow spectrum output) but they'll be constant at whatever color temperature they are & then I can take the histogram out of Raw Digger and compare different cameras that way.

So below is the histogram of the above image - what if I could take a "pure" blue or red or green light and photograph that and see how much the histograms overlap.



I was wondering if I could get ye olde "red" or "green" or "blue" lens filters and use those instead of the LEDs - the kind that was used for black & white photography.

Obviously more accurate is better (something with a known frequency spectrum) - but at least these are known quantities & the tests are reproducible.

--
"Wait let me comb my hair and put on a tie."
It Gets Nerdy: https://medium.com/ice-cream-geometry
Sometimes I take photos: https://www.instagram.com/sodiumstudio/

Yes don't use colored LEDs. There are not tight.

For blue get a wratten 47 filter, for green get a wratten 58 filter and for red get a wratten 29 filter. Red 25 might be more common though.

These are really cheap these days off ebay. As people dump their old filters. You can get them in either 3x3 gel or what ever you need screw mount. The Tiffen ones are typically a wratten gel sandwich between glass plates, which is good as they rarely fluorescence.

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

whisky said:

I'm not entirely sure how he did it but it looks like he took the pic with a Canon.

Will ask him next time I see him.

Cheers.

Click to expand...

Yes - I've tested with diffraction grating and what I find is the colors sort of lump together - blues overlap greens, etc.

What I seem to need is a very narrow beam of light coming through a narrow portion of the diffraction grating

I'm thinking of setting up a black box (non-reflective interior) and then cutting a small slit - maybe 1 or 2 mm wide - in one end and setting up a camera strobe set to "tele" on the other end.

That should create the narrow beam of light through a small section of diffraction grating that I would need to create a relatively overlap-free spectrum.

Click to expand...

Been there, done that, sd.

Blades approx 0.3mm apart

PM me if you want one ...

SD9 shot of a CFL spectrum

Nikon D50 shot of same spectrum

HTH

Click to expand...

Wow - I love it when I learn that I'm actually on the right path.

No need to send me diffraction grating - I have a small bag of the same - I think maybe 2 dozen of them. Thank you for the offer, though.

This is the sort of thing I was getting and what led me to decide I'd need to build a box like the one you just posted.

Click to expand...

Here's my rig in action on a LED flood light:


Looks like an SD10.

Note the diffuser to smooth out the individual LEDs.

Here's an alternative using 1/2 a CD:



HTH.

--
Ted

xpatUSA said:

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

whisky said:

I'm not entirely sure how he did it but it looks like he took the pic with a Canon.

Will ask him next time I see him.

Cheers.

Click to expand...

Yes - I've tested with diffraction grating and what I find is the colors sort of lump together - blues overlap greens, etc.

What I seem to need is a very narrow beam of light coming through a narrow portion of the diffraction grating

I'm thinking of setting up a black box (non-reflective interior) and then cutting a small slit - maybe 1 or 2 mm wide - in one end and setting up a camera strobe set to "tele" on the other end.

That should create the narrow beam of light through a small section of diffraction grating that I would need to create a relatively overlap-free spectrum.

Click to expand...

Been there, done that, sd.

Blades approx 0.3mm apart

PM me if you want one ...

SD9 shot of a CFL spectrum

Nikon D50 shot of same spectrum

HTH

Click to expand...

Wow - I love it when I learn that I'm actually on the right path.

No need to send me diffraction grating - I have a small bag of the same - I think maybe 2 dozen of them. Thank you for the offer, though.

This is the sort of thing I was getting and what led me to decide I'd need to build a box like the one you just posted.

Click to expand...

Here's my rig in action on a LED flood light:


Looks like an SD10.

Note the diffuser to smooth out the individual LEDs.

Here's an alternative using 1/2 a CD:



HTH.

Click to expand...

Perfect.

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

--
"Wait let me comb my hair and put on a tie."
It Gets Nerdy: https://medium.com/ice-cream-geometry
Sometimes I take photos: https://www.instagram.com/sodiumstudio/

dbateman said:

Yes don't use colored LEDs. There are not tight.

For blue get a wratten 47 filter, for green get a wratten 58 filter and for red get a wratten 29 filter. Red 25 might be more common though.

These are really cheap these days off ebay. As people dump their old filters. You can get them in either 3x3 gel or what ever you need screw mount. The Tiffen ones are typically a wratten gel sandwich between glass plates, which is good as they rarely fluorescence.

Click to expand...

Thanks.

Coincidentally I was looking at Tiffen #47 #58 and #29 filters - each in 62mm which will let me use them with all of my lenses with step-up rings - or just hand hold them up to the lens to take a quick photo. They're $15 which seems reasonable. (The #25 is just $10 more and puts me over the free shipping threshold - so adding that too.)

Looking up Wratten - the frequency response is a known quantity so that's really helpful.

https://www.edmundoptics.com/p/green-58-kodak-wratten-color-filter/10792/

This gives me the confidence I need to pick these up... Now about the rest of the items in my cart. Do I really need a drone?

This method seems like it would be a "cleaner" test than the diffraction grating technique. A very well defined input & just measure the RGB values in Raw Digger.

And now my plan for world domination is complete.

Muah ha ha ha ha.

saltydogstudios said:

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

Click to expand...

I don't know the spectral emission of that model. I do see see a lot of chimping ahead with the use of the strobe. As you can see from my layout the beams come out at an angle which is quite hard to line up on.

saltydogstudios said:

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

Click to expand...

Effectively I do frame and shoot the grating at the aforesaid angle. I guess the colored rays project out similarly to those from a prism but it's not difficult to focus on them (provided that the light is continuous). For a strobe, you might need to place a small modeling light such as a narrow-beam LED at it's center-line before shooting ...

Good luck.

xpatUSA said:

saltydogstudios said:

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

Click to expand...

I don't know the spectral emission of that model. I do see see a lot of chimping ahead with the use of the strobe. As you can see from my layout the beams come out at an angle which is quite hard to line up on.

xpatUSA said:

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

Click to expand...

Effectively I do frame and shoot the grating at the aforesaid angle. I guess the colored rays project out similarly to those from a prism but it's not difficult to focus on them (provided that the light is continuous). For a strobe, you might need to place a small modeling light such as a narrow-beam LED at it's center-line before shooting ...

Good luck.

Click to expand...

Yes, my test shots were with a small LED flashlight, but wanted to switch to something more photographers had experience with and could research.

Using a strobe rather than a constant light source is more difficult - getting the exposure right, controlling the light and so on.

I just wanted to remove as many unkonwns (the spectral output of the flashlight) and things that would make it difficult to reproduce as possible.

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

Click to expand...

I don't know the spectral emission of that model. I do see see a lot of chimping ahead with the use of the strobe. As you can see from my layout the beams come out at an angle which is quite hard to line up on.

saltydogstudios said:

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

Click to expand...

Effectively I do frame and shoot the grating at the aforesaid angle. I guess the colored rays project out similarly to those from a prism but it's not difficult to focus on them (provided that the light is continuous). For a strobe, you might need to place a small modeling light such as a narrow-beam LED at it's center-line before shooting ...

Good luck.

Click to expand...

Yes, my test shots were with a small LED flashlight, but wanted to switch to something more photographers had experience with and could research.

Using a strobe rather than a constant light source is more difficult - getting the exposure right, controlling the light and so on.

I just wanted to remove as many unkonwns (the spectral output of the flashlight) and things that would make it difficult to reproduce as possible.

Click to expand...

Understood. A strobe does seem a bit peaky , though.

On the other hand, if you're feeling a bit rich:

https://www.heraeus.com/media/media...tics_1/18_209_OA_Broschuere_Gesamt_WEB_RZ.pdf

My 3500K table-top LED floods are fairy smooth with a less-obvious blue peak compared to flash-light jobs:



But do note that the Y axis is radiant power. The peak at 450nm and the nearby trough at 480nm would serve as wavelength markers for comparative shots.

Apropos of LEDs this is quite interesting but, in spite of it's title, is a bit of a red herring:

https://www.lumileds.com/uploads/801/WP36-spectral-pure-technology-white-paper-PDF

[edit] I notice you're edging toward using filters rather than a grating ... [/edit]

--
Ted

xpatUSA said:

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

Click to expand...

I don't know the spectral emission of that model. I do see see a lot of chimping ahead with the use of the strobe. As you can see from my layout the beams come out at an angle which is quite hard to line up on.

xpatUSA said:

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

Click to expand...

Effectively I do frame and shoot the grating at the aforesaid angle. I guess the colored rays project out similarly to those from a prism but it's not difficult to focus on them (provided that the light is continuous). For a strobe, you might need to place a small modeling light such as a narrow-beam LED at it's center-line before shooting ...

Good luck.

Click to expand...

Yes, my test shots were with a small LED flashlight, but wanted to switch to something more photographers had experience with and could research.

Using a strobe rather than a constant light source is more difficult - getting the exposure right, controlling the light and so on.

I just wanted to remove as many unkonwns (the spectral output of the flashlight) and things that would make it difficult to reproduce as possible.

Click to expand...

Understood. A strobe does seem a bit peaky , though.

On the other hand, if you're feeling a bit rich:

https://www.heraeus.com/media/media...tics_1/18_209_OA_Broschuere_Gesamt_WEB_RZ.pdf

My 3500K table-top LED floods are fairy smooth with a less-obvious blue peak compared to flash-light jobs:

But do note that the Y axis is radiant power. The peak at 450nm and the nearby trough at 480nm would serve as wavelength markers for comparative shots.

Apropos of LEDs this is quite interesting but, in spite of it's title, is a bit of a red herring:

https://www.lumileds.com/uploads/801/WP36-spectral-pure-technology-white-paper-PDF

[edit] I notice you're edging toward using filters rather than a grating ... [/edit]

Click to expand...

Thank you for the pointers, and yes I'm probably going to do the filters.

My goal is to test the CFAs of various cameras (and the layers on Foveon cameras) to see how/if different camera companies tune their CFAs.

Using filters seems a cleaner way to do it than with diffraction grating.

So my current plan is to

• Put each filter in front of each camera along with a bit of diffusion glass (meant for setting white balance).
• Take a series of bracketed photos at various apertures - basically to ensure ETTR.
• Select the most "well exposed" photo in each series.
• Open the files in RAW Digger to see the R to G to B ratio among different cameras.
• See if any of it makes sense.
• Steal Underpants
• ....
• Fortune.

--
"Wait let me comb my hair and put on a tie."
It Gets Nerdy: https://medium.com/ice-cream-geometry
Sometimes I take photos: https://www.instagram.com/sodiumstudio/

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

xpatUSA said:

saltydogstudios said:

I would most likely be using a Vivitar 285 HV strobe - it's a well known light source and should be fairly even.

Click to expand...

I don't know the spectral emission of that model. I do see see a lot of chimping ahead with the use of the strobe. As you can see from my layout the beams come out at an angle which is quite hard to line up on.

saltydogstudios said:

So - what are you photographing? Not the diffraction grating itself - some sort of projection from it?

Click to expand...

Effectively I do frame and shoot the grating at the aforesaid angle. I guess the colored rays project out similarly to those from a prism but it's not difficult to focus on them (provided that the light is continuous). For a strobe, you might need to place a small modeling light such as a narrow-beam LED at it's center-line before shooting ...

Good luck.

Click to expand...

Yes, my test shots were with a small LED flashlight, but wanted to switch to something more photographers had experience with and could research.

Using a strobe rather than a constant light source is more difficult - getting the exposure right, controlling the light and so on.

I just wanted to remove as many unkonwns (the spectral output of the flashlight) and things that would make it difficult to reproduce as possible.

Click to expand...

Understood. A strobe does seem a bit peaky , though.

On the other hand, if you're feeling a bit rich:

https://www.heraeus.com/media/media...tics_1/18_209_OA_Broschuere_Gesamt_WEB_RZ.pdf

My 3500K table-top LED floods are fairy smooth with a less-obvious blue peak compared to flash-light jobs:

But do note that the Y axis is radiant power. The peak at 450nm and the nearby trough at 480nm would serve as wavelength markers for comparative shots.

Apropos of LEDs this is quite interesting but, in spite of it's title, is a bit of a red herring:

https://www.lumileds.com/uploads/801/WP36-spectral-pure-technology-white-paper-PDF

[edit] I notice you're edging toward using filters rather than a grating ... [/edit]

Click to expand...

Thank you for the pointers, and yes I'm probably going to do the filters.

My goal is to test the CFAs of various cameras (and the layers on Foveon cameras) to see how/if different camera companies tune their CFAs.

Using filters seems a cleaner way to do it than with diffraction grating.

So my current plan is to

• Put each filter in front of each camera along with a bit of diffusion glass (meant for setting white balance).
• Take a series of bracketed photos at various apertures - basically to ensure ETTR.
• Select the most "well exposed" photo in each series.
• Open the files in RAW Digger to see the R to G to B ratio among different cameras.
• See if any of it makes sense.
• Steal Underpants
• ....
• Fortune.

Click to expand...

As far as I understand your posts, you want to see the filter response of CFA or Foveon layers over frequency. Three fixed filters will not help. You either need to be able to scan through the spectrum with a narrow bandwidth or to spread out the frequencies over the sensor like you do with the diffraction gratings or refraction prisma.

FDecker said:

FDecker said:

<snip>

Click to expand...

As far as I understand your posts, you want to see the filter response of CFA or Foveon layers over frequency. Three fixed filters will not help. You either need to be able to scan through the spectrum with a narrow bandwidth or to spread out the frequencies over the sensor like you do with the diffraction gratings or refraction prisma.

Click to expand...

My premise is that cameras have stronger or weaker CFAs - they "leak" in various frequency ranges - that is, the blue filter allows in some red light (for example).

Camera makers have made their CFAs weaker to improve high ISO performance.

Some posit that this was at the heart of "CCD Colors" - stronger CFAs yielded more vibrant colors.

This video is hugely helpful in explaining these things.

https://luminous-landscape.com/phase-one-trichromatic-sensor-explained/

What I'm expecting to find is that when I put on the strong red filter - green will respond (green naturally overlaps with red) - but do the blue pixels respond? How much?

And the reverse - will blue trigger red. Some cameras have a red CFA with a "second hump" in the blue region (much like the human eye).

If different cameras "leak" more or less - red light triggering blue pixels - then I'll take it as evidence of a weaker or stronger CFA.

This is an oft touted graph - the difference between the original Canon 5D and the 5Dmk2.

If I find what I think I'm going to find - it's a first step in "proving" that different cameras have different responses to light, fundamentally.

The second step may be creating a shoebox diffraction grating viewing tool - I've been down that road and it's more annoying - I'm hoping this will be a simpler, cleaner experiment.

So I'm not looking for the overlap regions per-se, rather I'm just looking for evidence that different sensors respond differently.

The third step (which I've already done) is to photograph normal objects (in my case people) and use various calibration tools to see if the output is identical (it's not) to disprove all the people who say "RAW is RAW" and "If you don't like a camera's colors, you just don't know how to work with colors."



Various other images found around the web of dubious utility - many of these are in astrophotography forums so likely with the IR/UV hot mirror removed.







--
"Wait let me comb my hair and put on a tie."
It Gets Nerdy: https://medium.com/ice-cream-geometry
Sometimes I take photos: https://www.instagram.com/sodiumstudio/

saltydogstudios said:

FDecker said:

saltydogstudios said:

<snip>

Click to expand...

As far as I understand your posts, you want to see the filter response of CFA or Foveon layers over frequency. Three fixed filters will not help. You either need to be able to scan through the spectrum with a narrow bandwidth or to spread out the frequencies over the sensor like you do with the diffraction gratings or refraction prisma.

Click to expand...

My premise is that cameras have stronger or weaker CFAs - they "leak" in various frequency ranges - that is, the blue filter allows in some red light (for example).

Camera makers have made their CFAs weaker to improve high ISO performance.

Some posit that this was at the heart of "CCD Colors" - stronger CFAs yielded more vibrant colors.

This video is hugely helpful in explaining these things.

https://luminous-landscape.com/phase-one-trichromatic-sensor-explained/

What I'm expecting to find is that when I put on the strong red filter - green will respond (green naturally overlaps with red) - but do the blue pixels respond? How much?

And the reverse - will blue trigger red. Some cameras have a red CFA with a "second hump" in the blue region (much like the human eye).

If different cameras "leak" more or less - red light triggering blue pixels - then I'll take it as evidence of a weaker or stronger CFA.

This is an oft touted graph - the difference between the original Canon 5D and the 5Dmk2.

If I find what I think I'm going to find - it's a first step in "proving" that different cameras have different responses to light, fundamentally.

The second step may be creating a shoebox diffraction grating viewing tool - I've been down that road and it's more annoying - I'm hoping this will be a simpler, cleaner experiment.

So I'm not looking for the overlap regions per-se, rather I'm just looking for evidence that different sensors respond differently.

The third step (which I've already done) is to photograph normal objects (in my case people) and use various calibration tools to see if the output is identical (it's not) to disprove all the people who say "RAW is RAW" and "If you don't like a camera's colors, you just don't know how to work with colors."



Various other images found around the web of dubious utility - many of these are in astrophotography forums so likely with the IR/UV hot mirror removed.

Click to expand...

Thank you for explaining your motivation in detail. For what you describe as a first step, your approach is fine in my point of view. Have fun with your experiments.

FDecker said:

FDecker said:

<snip>

Click to expand...

Thank you for explaining your motivation in detail. For what you describe as a first step, your approach is fine in my point of view. Have fun with your experiments.

Click to expand...

What's really interesting is that - whatever this chart shows (I think it's strength of the CFAs between models - they weakened it for the Mark 2, which is why it shows more quantum efficiency? - someone can help me with the actual science here).... Whatever the change, it significantly changed where the frequencies overlap, shifting things to the left.

This means that "small" changes to CFA strength can lead to big changes in how cameras "see" color.
--
"Wait let me comb my hair and put on a tie."
It Gets Nerdy: https://medium.com/ice-cream-geometry
Sometimes I take photos: https://www.instagram.com/sodiumstudio/

I didn't know you wanted to go down the filter only route. I thought you were going to use them to validate the spectral ranges of a grating.

If this is the goal a simple set up would be to order 2 packs of Lee swatch books. From the Lee website it would be about $10 USA including shipping in the USA. (I don't know where you are).

https://shop.leefiltersusa.com/LEE-Swatch-Books_c4.htm

These are roughly 38mm x 80mm usable filter. Perfect for fitting over a flash like a cheap Canon 199A. The head on the 199A is 35mmx60mm. It also has really good spectral out put.

https://www.ultravioletphotography....and-godox-ad200/page__hl__199a__fromsearch__1

Point the filtered flash at a PTFE white target. Then take your pictures and seperate the Raw channels using 4channels command line program from libraw package. PTFE will reflect all wavelengths from UV to IR.


The swatch book has many filters which will cover your ranges. You can download the Android/Apple Lee app to preview the spectra to get an idea. Or just look on the Lee site. But the swatch book includes a better spectrum for each filter from 300nm to 800nm.

dbateman said:

I didn't know you wanted to go down the filter only route. I thought you were going to use them to validate the spectral ranges of a grating.

Click to expand...

It just seems like a cleaner experiment to me than the spectral ranges from diffraction rating.

I set up the diffraction grating and was using an LED flashlight as alight source - strobes were too messy and light went everywhere.

So I was taking long exposures in total darkeness, and any slight bump to the system led to a major shift in where the "rainbow" showed up - ruining my ability to compare cameras against each other.

I was opening up files in Photoshop, trying to align photos of different resolutions with different framings - it was a long and messy process.

Having 3/4 filters with known frequency responses allows me to take quick snaps with each camera & then compare them against each other looking at just the histograms.

Does the red filter trigger more or less of the green histogram? the blue histogram?

Besides as simplistic as this sounds - histograms look like science! My goal is to convince people there are quantifiable differences between CFAs and this is a clean way to do that. I think.

dbateman said:

If this is the goal a simple set up would be to order 2 packs of Lee swatch books. From the Lee website it would be about $10 USA including shipping in the USA. (I don't know where you are).

https://shop.leefiltersusa.com/LEE-Swatch-Books_c4.htm

These are roughly 38mm x 80mm usable filter. Perfect for fitting over a flash like a cheap Canon 199A. The head on the 199A is 35mmx60mm. It also has really good spectral out put.

https://www.ultravioletphotography....and-godox-ad200/page__hl__199a__fromsearch__1

Point the filtered flash at a PTFE white target. Then take your pictures and seperate the Raw channels using 4channels command line program from libraw package. PTFE will reflect all wavelengths from UV to IR.

https://www.ultravioletphotography....7520__hl__4channels__fromsearch__1#entry27520

The swatch book has many filters which will cover your ranges. You can download the Android/Apple Lee app to preview the spectra to get an idea. Or just look on the Lee site. But the swatch book includes a better spectrum for each filter from 300nm to 800nm.

Click to expand...

Interesting, I've played with libraw - I'll try that too.

I have the "strobist" gel collection - but I like that the filters I have are well known & their frequency response is published on the web.

I have a frosted glass filter that's designed to set white balance - very spectrally neutral & will produce a nice even light across the frame.

I'll use a studio strobe for consistency of color temperature. The combination of color filter + frosted glass should produce a frame that's pretty much all just one color.

Bracket the exposures to find the most "ETTR" one and open the files in Raw Digger and look at the histograms.

Did the Red filter trigger different amounts of green/blue in different cameras? Hopefully 1/3 stop exposure bracketing will allow me to produce files from each camera with comparable histograms - but even if they're close I should be able to look at the ratios.

saltydogstudios said:

dbateman said:

I didn't know you wanted to go down the filter only route. I thought you were going to use them to validate the spectral ranges of a grating.

Click to expand...

It just seems like a cleaner experiment to me than the spectral ranges from diffraction rating.

I set up the diffraction grating and was using an LED flashlight as alight source - strobes were too messy and light went everywhere.

Click to expand...

Off topic now with your selected method, but you've just reminded me that I sometimes had to put a thick dark cloth over the camera and the diffraction end of my spectroscope in order to block out daylight from a nearby window or if, for some reason, I decided to leave an overhead lamp on.

My sure-to-be controversial blog post is live! And not on that wretched site that Ted hates.

(no Sigma content - but will test Sigma cameras in a later post)

http://sodium.nyc/blog/2019/12/camera-color-science-does-it-exist-and-if-so-what-to-do-aboutit

saltydogstudios said:

My sure-to-be controversial blog post is live! And not on that wretched site that Ted hates.

(no Sigma content - but will test Sigma cameras in a later post)

http://sodium.nyc/blog/2019/12/camera-color-science-does-it-exist-and-if-so-what-to-do-aboutit

Click to expand...

Yes you do say somethings that will upset some people. Also you don't isolate all the variables.

What got to me in first read.

1. All humans actually perseve color differently. So your graph of human perception is only for one human. The range people can see is 380nm to 750nm and you can train yourself to see different colors over that range. Also as we age, no doubt our color perception changes.

2. We have know idea how we evolved to see different colors. You present only your personal hypothesis, which may be valid or completely wrong. Did we once see many colors and selected down to just these three cones? Maybe we had UV cones, but adapted them to blue or natural section favoured there loss. No idea.

3. You are actually testing different things than I think you want to know the results of. If you are interested in moving to Z7 or RP or FP cameras. You would need to specifically test those cameras. Manufacturers don't hold constant their CFA? Or do they? You actually have that bias within your article. As in, look Canon changed the color between two models. But I am going to look at old models to infer into what the new models color will look like? See the problem.

So how have color filter arrays changed over time? Are all companies moving to the same Fuji colors as we are all moving to Sony sensors? Are all newest models the same between manufacturers? Or still different and different within a manufacturer?

Many more variables. Are the color between a sony sensor Nikon or Olympus the same but different than when these manufacturers use different company sensors? See it gets complex quickly.

Best would be to contact Lensrentals, whom have many different cameras. See if you could set up the idea to do this quick test with them. Maybe do it with them. Use at least 3 "same" cameras each, as in 3 Z6s, and shoot your charts with as many cameras as you can. Then see if the data evolved, is the same or becomes interesting.

Shoot the color checker passport with no filter, 47, 58, 29, could save time to analyze the data.

This would also help you see, literally, which specific camera has your favorite color and is needing the smallest correct.

But as you pointed out your test chart may need also standard pictures of people with different skin colors. The ones you may most want to image to also get that color "correct".

Or just skip the whole exercise and boost the saturation in Instagram, as most people don't care if their photos look ugly. They just want them punchy.

Its more for production photography, where a company has spent millions on the correct red. And you bet your life you have to correctly reproduce that specific red to their eyes.

Also here is a link to a tiffen catalog with their filter spectrum: