The Quest for Good Color - 4. The Diffraction Grating Shootout

Ouch. Didn’t think the difference across the spectrum would be that big.

I was on the i1Studio path last week; after all, that’s a really useful tool: 1) measuring spectral power, 2) display calibration (I already have a colorimeter, but my grandson is starting photography… ) 3) measuring target spectra, just these main uses come to mind.

Thing is, I still wonder if it would be sufficient for me to measure SPD of a few general categories of lamp, and provide those to others who just want to make a light box… indeed, that’s how I’m currently using the OpenFilmTools datasets and getting what appear to be decent results.

Here’s a plot of all the OpenFilmTools 3200K DedoLights:

is that close enough?

What counts is the shape of the curves rather than their absolute values, those are close below 700nm. However that’s if you are using a Dedolight per their instructions. What about other tungsten sources most people would have access to, built with different components. How close are they to those? As mentioned in a previous thread, one of the biggest variables to be controlled accurately and stably is lamp drive current.

Jack

PS What are the units of the vertical axis? I think energy.
PPS If you decide to go the i1Studio route, don’t be too quick to get rid of your colorimeter: spectrophotometers tend not to do too well in the deep shadows.

I have a household tungsten incandescent bulb sitting on my desk (THAT was rather hard to find…) awaiting trial. That was the main reason for whipping the RPI spectrometer into shape…

Normalized radiance. Here’s the link to their project page: https://www.hdm-stuttgart.de/open-film-tools/english/cinelight_spectra

B&H price for the i1Studio is $489US. That reaches slightly into my fiscal angst, so the colorimeter lives a while longer… :laughing: So, a question: have you measured power from a light source with it, or know that has been done? The closest application I’ve seen to that is display calibration, not evident to me that it can just be aimed into space toward light sources…

Edit: BTW, thanks for the feedback. I’m a bit out of my box in this endeavor, learning as I go…

Indeed, it has several ports so you can point it at anything, paper, screen, illuminants, including the sky:

if you use argyllCMS to run it off a Pi on a battery or a tablet/phone you can take it to the field. It has a high-res mode, which provides a reading every 10/3 nm from 400-730nm, though I think the actual resolution is closer to 10nm. Here are a few illuminants:

Halogen light:

LED:


Fluorescent:
Fluorescent SPD H vs L rez vs A

Mixed Halogen and Fluorescent:

Take a look at the spotread and illumread documentation to get an idea of its capabilities.

I think it’s a great, versatile gadget. Don’t know about absolute accuracy, but it has served me well over the last few years. Not bad though:
D50 Nov 21 1430

Jack

2 Likes

Cripes, now I want one in the worst way. I’ll provide your name to B&H so they can pay your commission… :smiley:

@ggbutcher Remember these charts?

And do not forget the CRI:

Have fun!
Claes in Lund, Sweden

2 Likes

Was waiting for you to quote yourself. :timer_clock::stuck_out_tongue:

1 Like

I do remember them; your incandescent and halogen plots contributed to my thinking that a generic profile of a lighting category would be sufficient. Still not off that, as having to include a spectrometer in the lashup will daunt a lot of folk. I am going to acquire an i1Studio, @JackH helped me cross the utility threshold. But, until it just doesn’t make sense, I’ll use that device to assess the feasibility of generic “category” profiles, at least for tungsten and halogen sources.

I so regret missing when they had that on sale (I think around $100 discount?) a year or two ago by just one day.

Edit: Refurbs on eBay are pretty well priced, around $175… I think I’m gonna grab one of those.

Edit 2: Wait, maybe not. xRite’s product naming is crap. i1Display Studio vs. i1Studio. i1Studio is $350, so not sure if worth the risk of a refurb product.

1 Like

Those sky spectrum plots have me thinking… could we use the sun/sky as light source for SSF calibration? Or is the spectrum too variable depending on latitude/date/time of day? Or not intense enough (would require mirrors or lenses)?

It would be an excellent light source for profiling, but it’s kinda hard to channel into an optical setup. You’re also at the mercy of the weather, even clouds and haze. Right now in Colorado Springs, we’ve only had maybe two days in the past month where the sky was almost completely unencumbered of clouds, or forest fire haze…

The sun might burn a hole in your sensor. Oh yes, those fires. My cousin from the west (BC, Canada) says she hasn’t been able to see the sunset. She has an autoimmune disease, so it has been tough on her.

Apparently the smoke has now gotten high enough into the stratosphere to hit some eastward winds that are blowing it all the way over to the East Coast.

Yes, even OUR skies are getting hazy and reddish now!

Edit: Which makes a perfect use case for SSF measurement since we’re losing easy access to clean D50.

1 Like

Thanks for checking out the grating corrections!

It’s a bit disappointing it didn’t improve things and also slightly surprising, given the rather large correction (factor 2 over the range of interest). At the same time I don’t know what exactly goes into the DE calculation, so it’s totally imaginable that these smooth, continuous corrections have not that high an impact.

Anyway regardless of potentially open questions, the conclusion is nice: The cheap grating is just fine for good results :slight_smile:

I still haven’t exonerated my calibration operator in ssftool; it’s a simple division, but along with 1.0 normalization I’m not sure it’s all acting correctly.

That would be ssf_powercalibrate? And calibrationfile is something like a normalized version of source powerspectrum * grid spectral efficiency * diffusor power efficiency? That looks quite straightforwardly correct to me (though not too much stock should be given to such a statement from me :slight_smile: ). Or what is there concretely that makes you doubt it’s correctness?

I don’t know yet, just looking at some normalized data out of it just didn’t look like the relationship was preserved. It’s something I’m going to pick at next week…

Now, how you express it, as a product of the three, I haven’t done; I’ve attempted to apply each individually to the SSF measurements. I don’t see how that’d yield a different result, but I’m not really a math guy and a lot of those implications escape me until I just go and try them… :laughing:

This just hit hackaday:

There’s some good discussion of calibration sources in the comments.

Most notably https://eprints.lancs.ac.uk/id/eprint/6736/1/inproc_326.pdf is linked from the comments

There’s a bit of discussion regarding line CCD vs monochrome 2D camera, TBH for our purposes a monochrome 2D camera is fine for the “reference” calibration, and obviously for the camera to be characterized itself, it has to be a non-monochrome camera

1 Like

Interesting line CCD monochrome sensor. I wonder if they compensate the CCD for angular dependence and QE in their project:

Should be relatively easy to drive directly from an ESP32 or similar via i2c? No need for a lens in our application I think.

Or from a Pi, although in this application:
Since the camera we’re measuring has a 2D sensor, most of the benefits of line sensors are lost.

The main thing there that was useful for this application is the discussion of calibration sources. Specifically that paper discussing using “consumer” tungsten bulbs for spectral reference.

In this case, we’re interested in the camera’s SSF, so we need to know:

  1. The SSF of our reference sensor (this could be hard without existing data, but perhaps for a monochrome sensor, the general SSF of silicon CMOS is close enough?)
  2. The spectral power distribution of our light source
  3. This lets us figure out the SSF of our optical system, which can be used to compensate it out when measuring
  4. The SSF of our actual camera
1 Like