Jump to content

Jerry Built Spectrum Analysis


pop

Recommended Posts

Advertisement (gone after registration)

As the saying goes: Idle hands are the Devil's tools.

 

I did a brief experiment in which I wanted to see if it was possible to make a simple cheapo analysis of a camera's spectral response.

 

Part 1: Bill of materials:

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

  • One diffraction spectroscope (new, roughly USD 40)
  • One set of extension tubes (available, parts of Leica BEOON)
  • One pair of rubber eye cups (taken from a 24 year old cheap pair of Nikon field glasses)
  • One sheet of paper, white
  • One software "MEESOFT Image Analyzer" with profile line plugin

 

to be continued ....

Link to post
Share on other sites

Part 2: Assembly

 

As it happens, the eyecups fit nicely into the interior of the extension tubes while the spectroscope in turn fits into the eyecups. The following images show the major steps of the assembly.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

 

To be continued ...

Link to post
Share on other sites

As it happens (again), the contraption can be screwed (after a fashion) to the front of an elderly 50mm Elmar 1:2.8 which hasn't seen much use lately.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

To be continued ...

Link to post
Share on other sites

After some experimentation with all the associated parameters, I now can proudly present the first application of the highly sophisticated tool.

 

Case 1: taken in plain sunlight (early afternoon), manual white balance:

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

This is what the spectrum looks like, after a few simple processing steps (image rotation, cropping to visible part plus reserve, adjusting brightness, downscaling)

 

This graph is produced using the Image Analyzer software. It shows the brightness for each color component along the length of the spectrum.

 

This graph is produced using the same software. It shows the brightness of the entire image along the length of the spectrum.

 

To be continued ...

Link to post
Share on other sites

Discussion of first result

 

  • The distribution and the relative brightness of the color components can be clearly seen.
  • The spectrum is a bit under-exposed. The curves in the graphs are therefore a bit small. However, the general principle can be seen.
  • There are broad parts of the spectrum which are covered by only one of the three color filters. That would appear to imply that the camera does not reproduce all colors in the spectrum but only a smallish subset.

 

To be continued ...

Link to post
Share on other sites

Case 2: The illumination is the LED lamp hanging over my desk.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

 

 

As can be seen, I did not crop the spectrum at the same place as in the first case. The work flow can clearly stand some improvement. However, as the result can be used even so, I'll show it anyway.

 

To be continued ...

Link to post
Share on other sites

Advertisement (gone after registration)

Placing the colored graphs in close vicinity immediately makes visible the difference between the two light sources, even after manually adjusting the white balance in the camera:

 

364608d1362609446-jerry-built-spectrum-analysis-leica_sonne_man_spektrum_all.jpg

364613d1362610508-jerry-built-spectrum-analysis-leica_buro_man_spektrum_all.jpg

 

The brightness graphs reflect the same difference, but not as clearly, IMO:

 

364609d1362609446-jerry-built-spectrum-analysis-leica_sonne_man_spektrum_gray.jpg

364614d1362610508-jerry-built-spectrum-analysis-leica_buro_man_spektrum_gray.jpg

 

To be continued ...

Link to post
Share on other sites

The same procedure can be used, of course, to map the camera's rendering of the colors into a monochrome image.

 

This is case 1 again, taken in full sunlight in the early afternoon. The first graph is identical to the colored one above. The second graph shows the brightness of the BW JPG image as delivered by the camera.

 

364608d1362609446-jerry-built-spectrum-analysis-leica_sonne_man_spektrum_all.jpg

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

To be continued ...

Link to post
Share on other sites

Conclusions -

 

The procedure can indeed be used to visualize the color rendering of the camera and the light quality of different light sources.

 

Given the makeshift nature of the apparatus, care has to be taken to reproduce all steps of the procedure within some accuracy.

 

The number of parameters to be controlled is not very large:

  • Focusing on the spectrum is not trivial. The spectrum in my spectroscope seems to be at an apparent distance of some 30cm (12"). A camera with Live View would be helpful.
  • The spectrum is rather dim. Finding the proper exposure has been quite hit and miss so far; however, finding the proper value should be straight forward, given a proper light meter and some patience.
  • Turning the spectrums such that all samples have the short wavelengths on the left hand side is a bit tedious but not very demanding.
  • Since the spectroscope tends to wobble a bit in its mount, the spectrum is not always in the same place within the frame. This makes cropping the spectra to a uniform length a bit chancy.

 

The camera used is, BTW, a Leica M9.

 

Questions?

 

:D:D

Link to post
Share on other sites

Seems like you'd need to control for the "calibration profile" used in creating an image from the sensor data. Such profiles skew the primary colors (e.g. red either more yellow, or more magenta, and more or less saturated).

 

The "hump" in the red/yellow range may just be because Leica/Adobe's pre-canned profiles usually are way hot in yellow/blue saturation, and may disappear with a calibration done yourself based on a MacBeth/Gretag Colorchecker.

 

Or not - but I'd want to see what difference a calibration adjustment makes to the spectrum curves.

 

Also, what about lens color transmission? The current (since 1979) 50mm Summicron is supposed to be one of the most "neutral" of Leica M lenses, per Erwin Puts. But the 50 Summilux pre-ASPH is greener, the 50 Summilux ASPH is pinker, and the most recent 50 Elmar is a tad yellow (I don't know about the 50's version).

 

(In the accompanying chart, from Erwin Puts' e-book pdf "Leica M Lenses - Their Soul and Secrets," a neutral lens has BGR values (always normalized for Blue = zero) of 0/5/4).

 

Anyway, once you've "zeroed" the calibration and any lens tint, then you have a baseline and a good tool.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Link to post
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...