Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: Martin Brown <'''newspam'''@nonad.co.uk>
Newsgroups: sci.electronics.design
Subject: Re: "Colorimeter"
Date: Sat, 17 May 2025 22:03:07 +0100
Organization: A noiseless patient Spider
Lines: 61
Message-ID: <100atil$ik7q$1@dont-email.me>
References: <100ao53$hkhu$1@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Sat, 17 May 2025 23:03:18 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="00ff4f060a7ae7fdb741d61b2b22da30";
	logging-data="610554"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/66vifkU+0rQPmRjF4IsMUWCf7nHuDyjsHvl+VxVomdw=="
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:j6JQmAWXeX1n7cm+8TQU63HiPBs=
In-Reply-To: <100ao53$hkhu$1@dont-email.me>
Content-Language: en-GB

On 17/05/2025 20:30, Don Y wrote:
> Not quite, but, close enough...
> 
> How can I determine the spectrum of incident light on a sensor,
> in general?  Then, how many corners can I cut to sacrifice resolution
> and accuracy?

Short answer is you can't - at least without making some *very* 
questionable assumptions. It is even worse now with narrowband LEDs.

If you are allowed to make the assumption of a radiant perfect black 
body (something that doesn't exist) then it is much easier.

> I've worked with true colorimeters (dual wavelength) in the past.
> But, they were optimized to look for specific wavelengths.

True colorimeters were designed to match visible colours pretty much 
exactly under *any* lighting conditions (extremely tough problem). The 
first that actually worked well enough was the Imperial Match Predictor 
which ISTR was an analogue computer made in the UK by ICI strictly for 
internal use only. I don't think any documentation survives.

There was a US made spectrometer which formed a part of it whose 
manufacturers name escapes me for the moment. Got it Hardy 
Spectrophotometer:

https://collection.sciencemuseumgroup.org.uk/objects/co11842/ge-hardy-spectrophotometer-c-1940

That model isn't quite the right one but it is close.

Now any suitable paint test chart and a mobile phone will do the job.

> I calibrate the light emitted by my monitors with a device,
> but it controls the light source to do so.

If you are serious about doing this right then a 2D CCD sensor and a 
prism hires grating combo at right angles will allow you to quantify the 
entire visible spectrum at ultra high resolution. Be careful though 
Perkin-Elmer (and others) have some very good lock out patents on this 
trick (may be about to expire).

A few people can see longer wavelengths than most with an extra type of 
cone cell. They were sought after in WWII (pre thermal IR band imaging) 
because they could see the difference between live foliage still growing 
and cut down dying foliage used as gun emplacement camouflage.

Denatured chlorophyll looks much darker to them.

> With no knowledge of the actual (visible) spectrum impinging on
> a sensor (and a bit of time to integrate results), how can I
> do this short of swapping individual filters in front of the
> sensor(s)?

Measure the intensity at all wavelengths in a single shot.

PE OES instrument in the early 1990's was the first with this.
(I forget the model number) I was seriously impressed with it.

-- 
Martin Brown