spectral response from D800 to D500 Locked

this is a continuation of Marianne Oelund's "how our cameras see color"  thread. I have used an old Leitz Pradovit 2500 slide projector (250 W halogen lamp) to illuminate the slit of my ad-hoc spectrometer. The grating with 500 lines/mm was mounted directly on the front of the lens. The same lens was used with each camera and all raw data were taken in the DX format.

The D800 spectral response strongly resembles the D3 famly responses taken by Marianne.  The newer models covered her differ slightly but noticeably.

spectral response for some Nikon cameras

Wow! Quite a lag from 2009...

But I love data like this, so "thank you" for providing it.

I can see the differences, but it's difficult to accurately compare them. What I would ideally want is each of the 3 colors over-laid.

The D800 red channel has a noticeable bump at 525 nm. Marianne’s data on the D3 has this bump. I have always guessed it was intentional? Not sure that makes sense though. But the other sensors don't seem to have this bump.

Don’t the filters in digital cameras change w/ age and use? Is your D800 degraded slightly?

The D500 appears to have much broader curves for green and blue?

chuxter wrote:

Wow! Quite a lag from 2009...

But I love data like this, so "thank you" for providing it.

I can see the differences, but it's difficult to accurately compare them. What I would ideally want is each of the 3 colors over-laid.

superpositioln of response curves for D800 D7100 D7200 D500 : green and blue with increase slightly, red D800 bump gets reduced and D800 red sensitivity around 450 disappears

The D800 red channel has a noticeable bump at 525 nm. Marianne’s data on the D3 has this bump. I have always guessed it was intentional? Not sure that makes sense though. But the other sensors don't seem to have this bump.

The tweaks are such that green and blue integral quantum efficiency is helped.

Don’t the filters in digital cameras change w/ age and use? Is your D800 degraded slightly?

I think these filter colors in a DSLR show no relevant degradation during the little exposure the may get in the most  heavy professional use. Even for EVF cameras, where the sensor is exposed to light all the time while the camera is in use, I have not seen any mention of CFA bleaching.

The D500 appears to have much broader curves for green and blue?

a little for sure, a tweak.

Bernard Delley wrote:

chuxter wrote:

Wow! Quite a lag from 2009...

But I love data like this, so "thank you" for providing it.

I can see the differences, but it's difficult to accurately compare them. What I would ideally want is each of the 3 colors over-laid.

superpositioln of response curves for D800 D7100 D7200 D500 : green and blue with increase slightly, red D800 bump gets reduced and D800 red sensitivity around 450 disappears

The D800 red channel has a noticeable bump at 525 nm. Marianne’s data on the D3 has this bump. I have always guessed it was intentional? Not sure that makes sense though. But the other sensors don't seem to have this bump.

The tweaks are such that green and blue integral quantum efficiency is helped.

Don’t the filters in digital cameras change w/ age and use? Is your D800 degraded slightly?

I think these filter colors in a DSLR show no relevant degradation during the little exposure the may get in the most heavy professional use. Even for EVF cameras, where the sensor is exposed to light all the time while the camera is in use, I have not seen any mention of CFA bleaching.

The D500 appears to have much broader curves for green and blue?

a little for sure, a tweak.

Mostly, I see a bigger overlap of the blue and green.

Back in 2009, when Marianne posted her initial data, she was shocked at the lack of differential in the broad green area. Perhaps pulling the blue response down [in wavelength] a bit has helped?

David Lal wrote:

Bernard Delley wrote:

The D800 spectral response strongly resembles the D3 famly responses taken by Marianne. The newer models covered her differ slightly but noticeably.

Great composition. My favourite is #1.

Thanks for your positively phrased feedback! My first thought was, you ironically say 'who cares' !

Funny enough, I do apply similar criteria to my 'science' graphs as I do to images. First of all is content, of interest to some audience, hopefully wider than an autistic self. Here the interest is with the paucity of such published camera sensitivity spectra and its relation to failed color rendition for some objects. The compositional part for the graphs involves framing, bringing a message visually across, choices of colors, and leaving away clutter.

The making of this might be compatible with your posting history. Cheap extra hardware: a $2 grating, and already available old slide projector. Materials like cardboard, adhesive tape, aluminum foil and use of a scalpel. And a bit of computer programming for the data analysis and graphics. Lacks: humorous presentation.

Bernard Delley wrote:

this is a continuation of Marianne Oelund's "how our cameras see color" thread. I have used an old Leitz Pradovit 2500 slide projector (250 W halogen lamp) to illuminate the slit of my ad-hoc spectrometer. The grating with 500 lines/mm was mounted directly on the front of the lens. The same lens was used with each camera and all raw data were taken in the DX format.

The D800 spectral response strongly resembles the D3 famly responses taken by Marianne. The newer models covered her differ slightly but noticeably.

spectral response for some Nikon cameras

I'm inclined to think that the profiles from a Color Checker would even all that out in the finished/raw converted photograph. Can someone more technically informed please verify?

Ken Seals wrote:

Bernard Delley wrote:

this is a continuation of Marianne Oelund's "how our cameras see color" thread. I have used an old Leitz Pradovit 2500 slide projector (250 W halogen lamp) to illuminate the slit of my ad-hoc spectrometer. The grating with 500 lines/mm was mounted directly on the front of the lens. The same lens was used with each camera and all raw data were taken in the DX format.

The D800 spectral response strongly resembles the D3 famly responses taken by Marianne. The newer models covered her differ slightly but noticeably.

spectral response for some Nikon cameras

I'm inclined to think that the profiles from a Color Checker would even all that out in the finished/raw converted photograph. Can someone more technically informed please verify?

These changes represent tweaks to the photo response. Through the raw processing the accuracy for the color checker test patches comes out quite high. The tweaks produced a slight improvement for to metamerism index as shown by DXOmark color response like from 78 to 82 for the index. From great color rendition to even slightly better. Yet it is not perfect, highly accurate color matching remains a challenge (anecdote:  color matching for car paints) . Yet for some difficult colors like blue-green all these cameras have a more or less  pronounced blue hue shift.