Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: piglet <erichpwagner@hotmail.com>
Newsgroups: sci.electronics.design
Subject: Re: "Colorimeter"
Date: Sun, 18 May 2025 12:55:37 -0000 (UTC)
Organization: A noiseless patient Spider
Lines: 104
Message-ID: <100clc9$10a53$1@dont-email.me>
References: <100ao53$hkhu$1@dont-email.me>
 <100atil$ik7q$1@dont-email.me>
 <100bgn8$ls2c$1@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Injection-Date: Sun, 18 May 2025 14:55:37 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="87ae7d5c376dd9cf12c0140931d38dca";
	logging-data="1058979"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+uONMgikq4e4qH7luGRYm/"
User-Agent: NewsTap/5.5 (iPhone/iPod Touch)
Cancel-Lock: sha1:+H9wsON30nnjc4D+54+8f3XMo00=
	sha1:oRMQN4YH7JaponPkUf4m5Y2XGFM=

Don Y <blockedofcourse@foo.invalid> wrote:
> On 5/17/2025 2:03 PM, Martin Brown wrote:
>> 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'm not looking for a laboratory grade instrument.  (hence the
> "corner cutting" caveat).
> 
> Rather, "how does the light falling on THIS body compare to the
> light on this OTHER body" (using the same measuring instrument)
> 
>>> 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.
> 
> Ours controlled the color temperature of an incandescent lamp
> "seen" through a pair of filters.  Then, compared the detected
> signal from the sample under test (inserted between the emitter
> and detector) in the same short time interval, looking for a
> particular color shift (analyzing blood assays)
> 
> Again, you don't care WHAT "color" it is, just how the chemistry
> altered the color within a band of expected results.
> 
> But, that system KNEW what to expect (expectations were dependent
> on the actual assay being run)
> 
>> 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.
> 
> How durable are the CCDs used in phones?  Especially to high intensity light
> sources?
> 
>>> 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).
> 
> Again, not looking to make an "instrument".  The phone idea may work
> if the CCDs don't freak out with high intensity sources.
> 
>> 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.
> 
> Also folks who are truly colorblind.  Camouflage looks different than
> natural foliage when you are just looking at the values without the
> hues to distract.
> 
>> 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.
> 
> Or, leverage the fact that the spectrum won't be changing in
> the short term (for some value of "short") and cycle a set
> of filters (rotating disc?) between the detector and source.
> 
> Again, if you aren't looking for repeatability instrument to
> instrument, this may be good enough to answer the question above.
> 
>> PE OES instrument in the early 1990's was the first with this.
>> (I forget the model number) I was seriously impressed with it.
>> 
> 
> 

If you just want to color match then your phone camera is dandy. There are
apps used by printers and film lighting cameramen to do just that. ISTR
chromlink ?


-- 
piglet