Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > sci.image.processing > #4480 > unrolled thread
| Started by | dale <dale@dalekelly.org> |
|---|---|
| First post | 2020-04-07 15:05 -0400 |
| Last post | 2020-04-08 07:51 +0100 |
| Articles | 12 — 5 participants |
Back to article view | Back to sci.image.processing
spectral cameras dale <dale@dalekelly.org> - 2020-04-07 15:05 -0400
Re: spectral cameras Alfred Molon <alfred_molon@yahoo.com> - 2020-04-07 22:17 +0200
Re: spectral cameras dale <dale@dalekelly.org> - 2020-04-07 17:09 -0400
Re: spectral cameras dale <dale@dalekelly.org> - 2020-04-08 01:24 -0400
Re: spectral cameras Alfred Molon <alfred_molon@yahoo.com> - 2020-04-08 19:28 +0200
Re: spectral cameras dale <dale@dalekelly.org> - 2020-04-12 20:29 -0400
Re: spectral cameras geoff <geoff@nospamgeoffwood.org> - 2020-04-13 13:16 +1200
Re: spectral cameras Alan Browne <bitbucket@blackhole.com> - 2020-04-13 17:01 -0400
Re: spectral cameras dale <dale@dalekelly.org> - 2020-04-13 17:24 -0400
Re: spectral cameras Alan Browne <bitbucket@blackhole.com> - 2020-04-13 18:58 -0400
Re: spectral cameras Martin Brown <'''newspam'''@nezumi.demon.co.uk> - 2020-04-09 08:42 +0100
Re: spectral cameras Martin Brown <'''newspam'''@nezumi.demon.co.uk> - 2020-04-08 07:51 +0100
| From | dale <dale@dalekelly.org> |
|---|---|
| Date | 2020-04-07 15:05 -0400 |
| Subject | spectral cameras |
| Message-ID | <d2f8ea.nlu.19.1@news.alt.net> |
RGB color cameras has proved to be good on smart phones for consumers. Is sensor resolution nearing the point to measure spectral visible wavelengths instead of tricolor and is there a market for professional use? Does the research indicate that it is time to prepare the workflow for them? -- Minister Dale Kelly, Ph.D. https://www.dalekelly.org/ Board Certified Holistic Health Practitioner Board Certified Alternative Medical Practitioner
[toc] | [next] | [standalone]
| From | Alfred Molon <alfred_molon@yahoo.com> |
|---|---|
| Date | 2020-04-07 22:17 +0200 |
| Message-ID | <MPG.38f6fe4558ef051d98d0cb@news.supernews.com> |
| In reply to | #4480 |
In article <d2f8ea.nlu.19.1@news.alt.net>, dale@dalekelly.org says... > > RGB color cameras has proved to be good on smart phones for consumers. > > Is sensor resolution nearing the point to measure spectral visible > wavelengths instead of tricolor and is there a market for professional use? That doesn't depend on the sensor resolution. But you could make a sensor cell which is capable to measure light of (a limited number of) different wavelengths. Foveon is one example and the organic, multiple layer sensor design of Panasonic another one. -- Alfred Molon Olympus 4/3 and micro 4/3 cameras forum at https://groups.io/g/myolympus https://myolympus.org/ photo sharing site
[toc] | [prev] | [next] | [standalone]
| From | dale <dale@dalekelly.org> |
|---|---|
| Date | 2020-04-07 17:09 -0400 |
| Message-ID | <d2ffna.40s.19.1@news.alt.net> |
| In reply to | #4481 |
On 4/7/2020 4:17 PM, Alfred Molon wrote: > In article <d2f8ea.nlu.19.1@news.alt.net>, dale@dalekelly.org > says... >> >> RGB color cameras has proved to be good on smart phones for consumers. >> >> Is sensor resolution nearing the point to measure spectral visible >> wavelengths instead of tricolor and is there a market for professional use? > > That doesn't depend on the sensor resolution. But you could make > a sensor cell which is capable to measure light of (a limited > number of) different wavelengths. Foveon is one example and the > organic, multiple layer sensor design of Panasonic another one. > kinda figured it wasn't a unique idea -- Minister Dale Kelly, Ph.D. https://www.dalekelly.org/ Board Certified Holistic Health Practitioner Board Certified Alternative Medical Practitioner
[toc] | [prev] | [next] | [standalone]
| From | dale <dale@dalekelly.org> |
|---|---|
| Date | 2020-04-08 01:24 -0400 |
| Message-ID | <d2gcnq.6hm.17.1@news.alt.net> |
| In reply to | #4482 |
On 4/7/2020 5:09 PM, dale wrote: > On 4/7/2020 4:17 PM, Alfred Molon wrote: >> In article <d2f8ea.nlu.19.1@news.alt.net>, dale@dalekelly.org >> says... >>> >>> RGB color cameras has proved to be good on smart phones for consumers. >>> >>> Is sensor resolution nearing the point to measure spectral visible >>> wavelengths instead of tricolor and is there a market for >>> professional use? >> >> That doesn't depend on the sensor resolution. But you could make >> a sensor cell which is capable to measure light of (a limited >> number of) different wavelengths. Foveon is one example and the >> organic, multiple layer sensor design of Panasonic another one. >> > > kinda figured it wasn't a unique idea > > back to resolution ... the visible spectrum https://en.wikipedia.org/wiki/Visible_spectrum is from 380nm-720nm wavelengths if you sample that in 10nm increments, that a low end spectrometer 20 some years ago for approx $10K would do ... wouldn't you need 34 dots per pixel? In printing and publishing a typical rosette pattern CMYK pixel requires 8 dots ... -- Minister Dale Kelly, Ph.D. https://www.dalekelly.org/ Board Certified Holistic Health Practitioner Board Certified Alternative Medical Practitioner
[toc] | [prev] | [next] | [standalone]
| From | Alfred Molon <alfred_molon@yahoo.com> |
|---|---|
| Date | 2020-04-08 19:28 +0200 |
| Message-ID | <MPG.38f82819692d38c098d0ce@news.supernews.com> |
| In reply to | #4483 |
In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org says... > back to resolution ... > > the visible spectrum > > https://en.wikipedia.org/wiki/Visible_spectrum > > is from 380nm-720nm wavelengths > > if you sample that in 10nm increments, that a low end spectrometer 20 > some years ago for approx $10K would do ... > > wouldn't you need 34 dots per pixel? In printing and publishing a > typical rosette pattern CMYK pixel requires 8 dots ... It wouldn't be a good idea to split up each pixel into 34 subpixels, each only capturing a very narrow wavelength range. Most incoming photons would not be used, i.e. the sensor would have a low spectral efficiency. -- Alfred Molon Olympus 4/3 and micro 4/3 cameras forum at https://groups.io/g/myolympus https://myolympus.org/ photo sharing site
[toc] | [prev] | [next] | [standalone]
| From | dale <dale@dalekelly.org> |
|---|---|
| Date | 2020-04-12 20:29 -0400 |
| Message-ID | <d2t1b6.a12.17.1@news.alt.net> |
| In reply to | #4485 |
On 4/8/2020 1:28 PM, Alfred Molon wrote: > In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org > says... >> back to resolution ... >> >> the visible spectrum >> >> https://en.wikipedia.org/wiki/Visible_spectrum >> >> is from 380nm-720nm wavelengths >> >> if you sample that in 10nm increments, that a low end spectrometer 20 >> some years ago for approx $10K would do ... >> >> wouldn't you need 34 dots per pixel? In printing and publishing a >> typical rosette pattern CMYK pixel requires 8 dots ... > > It wouldn't be a good idea to split up each pixel into 34 > subpixels, each only capturing a very narrow wavelength range. > > Most incoming photons would not be used, i.e. the sensor would > have a low spectral efficiency. > what about analog capture through a prism then digital rendering in the camera's software driver to the seven optical wavelength colors? https://en.wikipedia.org/wiki/Prism#/media/File:Light_dispersion_conceptual_waves.gif red orange yellow green blue indigo violet -- Minister Dale Kelly, Ph.D. https://www.dalekelly.org/ Board Certified Holistic Health Practitioner Board Certified Alternative Medical Practitioner
[toc] | [prev] | [next] | [standalone]
| From | geoff <geoff@nospamgeoffwood.org> |
|---|---|
| Date | 2020-04-13 13:16 +1200 |
| Message-ID | <E6idnSyQOYMUIA7DnZ2dnUU7-ffNnZ2d@giganews.com> |
| In reply to | #4487 |
On 13/04/2020 12:29 pm, dale wrote: > On 4/8/2020 1:28 PM, Alfred Molon wrote: >> In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org >> says... >>> back to resolution ... >>> >>> the visible spectrum >>> >>> https://en.wikipedia.org/wiki/Visible_spectrum >>> >>> is from 380nm-720nm wavelengths >>> >>> if you sample that in 10nm increments, that a low end spectrometer 20 >>> some years ago for approx $10K would do ... >>> >>> wouldn't you need 34 dots per pixel? In printing and publishing a >>> typical rosette pattern CMYK pixel requires 8 dots ... >> >> It wouldn't be a good idea to split up each pixel into 34 >> subpixels, each only capturing a very narrow wavelength range. >> >> Most incoming photons would not be used, i.e. the sensor would >> have a low spectral efficiency. >> > > what about analog capture through a prism then digital rendering in the > camera's software driver to the seven optical wavelength colors? > > https://en.wikipedia.org/wiki/Prism#/media/File:Light_dispersion_conceptual_waves.gif > > > red > orange > yellow > green > blue > indigo > violet > > That's still quantisation into discrete bands, rather than a truly linear . geoff
[toc] | [prev] | [next] | [standalone]
| From | Alan Browne <bitbucket@blackhole.com> |
|---|---|
| Date | 2020-04-13 17:01 -0400 |
| Message-ID | <dsCdnQ8HaYi6TgnDnZ2dnUU7-c_NnZ2d@giganews.com> |
| In reply to | #4487 |
On 2020-04-12 20:29, dale wrote: > On 4/8/2020 1:28 PM, Alfred Molon wrote: >> In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org >> says... >>> back to resolution ... >>> >>> the visible spectrum >>> >>> https://en.wikipedia.org/wiki/Visible_spectrum >>> >>> is from 380nm-720nm wavelengths >>> >>> if you sample that in 10nm increments, that a low end spectrometer 20 >>> some years ago for approx $10K would do ... >>> >>> wouldn't you need 34 dots per pixel? In printing and publishing a >>> typical rosette pattern CMYK pixel requires 8 dots ... >> >> It wouldn't be a good idea to split up each pixel into 34 >> subpixels, each only capturing a very narrow wavelength range. >> >> Most incoming photons would not be used, i.e. the sensor would >> have a low spectral efficiency. >> > > what about analog capture through a prism then digital rendering in the > camera's software driver to the seven optical wavelength colors? > > https://en.wikipedia.org/wiki/Prism#/media/File:Light_dispersion_conceptual_waves.gif > > > red <s> Everything is trade space. If you sample wider spectrally then you'll likely be under sampling spatially. Further contribution to noise as well from each discrete sampling location. IAC, whether today's cameras are RGB, RGBh or CYGM they seem to capture the colour range more than adequately enough.
[toc] | [prev] | [next] | [standalone]
| From | dale <dale@dalekelly.org> |
|---|---|
| Date | 2020-04-13 17:24 -0400 |
| Message-ID | <d2vaqk.q8m.19.1@news.alt.net> |
| In reply to | #4489 |
On 4/13/2020 5:01 PM, Alan Browne wrote: > On 2020-04-12 20:29, dale wrote: >> On 4/8/2020 1:28 PM, Alfred Molon wrote: >>> In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org >>> says... >>>> back to resolution ... >>>> >>>> the visible spectrum >>>> >>>> https://en.wikipedia.org/wiki/Visible_spectrum >>>> >>>> is from 380nm-720nm wavelengths >>>> >>>> if you sample that in 10nm increments, that a low end spectrometer 20 >>>> some years ago for approx $10K would do ... >>>> >>>> wouldn't you need 34 dots per pixel? In printing and publishing a >>>> typical rosette pattern CMYK pixel requires 8 dots ... >>> >>> It wouldn't be a good idea to split up each pixel into 34 >>> subpixels, each only capturing a very narrow wavelength range. >>> >>> Most incoming photons would not be used, i.e. the sensor would >>> have a low spectral efficiency. >>> >> >> what about analog capture through a prism then digital rendering in >> the camera's software driver to the seven optical wavelength colors? >> >> https://en.wikipedia.org/wiki/Prism#/media/File:Light_dispersion_conceptual_waves.gif >> >> >> red > <s> > > Everything is trade space. If you sample wider spectrally then you'll > likely be under sampling spatially. Further contribution to noise as > well from each discrete sampling location. > > IAC, whether today's cameras are RGB, RGBh or CYGM they seem to capture > the colour range more than adequately enough. > RGBW might be something to consider since for example pink hues require R and W, that would be R and (RGB) -- Minister Dale Kelly, Ph.D. https://www.dalekelly.org/ Board Certified Holistic Health Practitioner Board Certified Alternative Medical Practitioner
[toc] | [prev] | [next] | [standalone]
| From | Alan Browne <bitbucket@blackhole.com> |
|---|---|
| Date | 2020-04-13 18:58 -0400 |
| Message-ID | <x7qdnfrU1PUTcwnDnZ2dnUU7-L-dnZ2d@giganews.com> |
| In reply to | #4490 |
On 2020-04-13 17:24, dale wrote: > On 4/13/2020 5:01 PM, Alan Browne wrote: >> On 2020-04-12 20:29, dale wrote: >>> On 4/8/2020 1:28 PM, Alfred Molon wrote: >>>> In article <d2gcnq.6hm.17.1@news.alt.net>, dale@dalekelly.org >>>> says... >>>>> back to resolution ... >>>>> >>>>> the visible spectrum >>>>> >>>>> https://en.wikipedia.org/wiki/Visible_spectrum >>>>> >>>>> is from 380nm-720nm wavelengths >>>>> >>>>> if you sample that in 10nm increments, that a low end spectrometer 20 >>>>> some years ago for approx $10K would do ... >>>>> >>>>> wouldn't you need 34 dots per pixel? In printing and publishing a >>>>> typical rosette pattern CMYK pixel requires 8 dots ... >>>> >>>> It wouldn't be a good idea to split up each pixel into 34 >>>> subpixels, each only capturing a very narrow wavelength range. >>>> >>>> Most incoming photons would not be used, i.e. the sensor would >>>> have a low spectral efficiency. >>>> >>> >>> what about analog capture through a prism then digital rendering in >>> the camera's software driver to the seven optical wavelength colors? >>> >>> https://en.wikipedia.org/wiki/Prism#/media/File:Light_dispersion_conceptual_waves.gif >>> >>> >>> red >> <s> >> >> Everything is trade space. If you sample wider spectrally then you'll >> likely be under sampling spatially. Further contribution to noise as >> well from each discrete sampling location. >> >> IAC, whether today's cameras are RGB, RGBh or CYGM they seem to <--[1] >> capture the colour range more than adequately enough. >> > > RGBW might be something to consider since for example pink hues require > R and W, that would be R and (RGB) See the 'h' above [1]. Really a highlight sensor so it's colour blind. Anyway, nothing to see here. Running along.
[toc] | [prev] | [next] | [standalone]
| From | Martin Brown <'''newspam'''@nezumi.demon.co.uk> |
|---|---|
| Date | 2020-04-09 08:42 +0100 |
| Message-ID | <r6mjl1$19jg$1@gioia.aioe.org> |
| In reply to | #4483 |
On 08/04/2020 06:24, dale wrote: > On 4/7/2020 5:09 PM, dale wrote: >> On 4/7/2020 4:17 PM, Alfred Molon wrote: >>> In article <d2f8ea.nlu.19.1@news.alt.net>, dale@dalekelly.org >>> says... >>>> >>>> RGB color cameras has proved to be good on smart phones for consumers. >>>> >>>> Is sensor resolution nearing the point to measure spectral visible >>>> wavelengths instead of tricolor and is there a market for >>>> professional use? >>> >>> That doesn't depend on the sensor resolution. But you could make >>> a sensor cell which is capable to measure light of (a limited >>> number of) different wavelengths. Foveon is one example and the >>> organic, multiple layer sensor design of Panasonic another one. >>> >> >> kinda figured it wasn't a unique idea > > back to resolution ... > > the visible spectrum > > https://en.wikipedia.org/wiki/Visible_spectrum > > is from 380nm-720nm wavelengths > > if you sample that in 10nm increments, that a low end spectrometer 20 > some years ago for approx $10K would do ... Here is how they actually do serious spectroscopy with plenty of light. https://www.noao.edu/image_gallery/html/im0609.html Its a low dispersion prism and a high resolution grating combo to illuminate a conventional 2D CCD with the entire stellar spectrum. The other technology uses fibre optics to route positions on the sky onto a line at the input to a classic high dispersion spectrometer so that individual targets of interest can be done simultaneously. You would gain very little useful improvement by sampling at 10nm since most LED emitters (apart from spot frequency lasers) have FWHM of 50nm and dyes and pigments are considerably worse. -- Regards, Martin Brown
[toc] | [prev] | [next] | [standalone]
| From | Martin Brown <'''newspam'''@nezumi.demon.co.uk> |
|---|---|
| Date | 2020-04-08 07:51 +0100 |
| Message-ID | <r6jsa0$1j8b$1@gioia.aioe.org> |
| In reply to | #4480 |
On 07/04/2020 20:05, dale wrote: > RGB color cameras has proved to be good on smart phones for consumers. > > Is sensor resolution nearing the point to measure spectral visible > wavelengths instead of tricolor and is there a market for professional use? They are called spectrometers and already in use where there is a scientific requirement to do so. > > Does the research indicate that it is time to prepare the workflow for > them? Utterly pointless. So many combinations of mixtures of different wavelengths map to the same perceived colour to be worthwhile. -- Regards, Martin Brown
[toc] | [prev] | [standalone]
Back to top | Article view | sci.image.processing
csiph-web