Incandescent/halogen ban?

[Ivo Stunga]

And I'll make a spreadsheet of what goes where and when - when it comes to it.

 

[Cinema]

CRI is BS.
Incandescent ban is BS.

CRI is based on very naive, old fashioned positivist color sensitometry. Also totally neglecting decades of new knowledge on psycho optics.

I would like to know a little more. I have worked in the film industry for a couple decades but only gleaned so much technically working at camera houses in the beginning.

I always thought CRI referred to the quality of light. Which makes sense to me because LED tvs and OLEDs generally look horrible and CRTs and plasmas look incredible vibrant and high quality (they also happen to be incandescent). What is cri referring to exactly? I have measured it before with spectroradiometers but don’t really know much more other than higher is better. My old sony pro crts for coloring were in the 90 cri range while my pioneer plasma is around 75. My LG oled is dismal i think around 40-50 cri. Yes i can read the wikipedia about CRI but i’m not smart enough to really understand why it’s BS, as the quality of light sources seems empirically obvious to me, that the higher cri a source measures, the better it looks.

 

[MattKing]

The problem with CRI is it measures something that is fairly useful when talking about viewing something with the human eye, but unreliable when trying to apply the light to things like sensors or emulsions with discrete areas of colour sensitivity, matched with areas of greatly decreased sensitivity - essentially "spiky" sensors.
It works best with "sensors" that average.

 

[Ivo Stunga]

I always thought CRI referred to the quality of light. Which makes sense to me because LED tvs and OLEDs generally look horrible and CRTs and plasmas look incredible vibrant and high quality (they also happen to be incandescent)

You're the first ever I see bashing OLED, but I kind of understand you. However, color calibrated OLED looks awesome. That said, CRTs deliver color that is just something else and forgotte, Plasma is second after that. But Plasma is Plasma, OLEDs are OLEDs : depends on the quality of the panel, all are not made equally. All CRTs aren't made equally.
I have bot LG 55" OLED and Samsung 32" CRT, have turned them on side by side just to remind myself a couple of things: CRT is useful today only as retro gaming monitor where its used as a final mastering tool instead of plain display: to hide, to blur, to mask, to make color gradient out of dithering patterns, used to get around technical limitations of retro game consoles - up to PlayStation 2.

DVD is a transition point and everything more modern looks way better on OLED, giving you more details, cleaner and sharper image - again depending on panel that is.

 

[Cinema]

The only oleds i like are the $30k sony bvms and the hold and sample motion is still really bad. I have an lg c2 that looks really plasticy and lifeless even after calibration hitting sub 1 dE just lifeless and dull and fake looking compared to blu rays on my plasmas. Cri is the only real criteria that has shown to rate light quality from crt and plasmas much higher than oleds so i guess its how i justify my preference for those panels besides my plasma looks basically off on a full black screen contrast is close to OLED and near black shadow detail is way better.

 

[MattKing]

The only oleds i like are the $30k sony bvms and the hold and sample motion is still really bad. I have an lg c2 that looks really plasticy and lifeless even after calibration hitting sub 1 dE just lifeless and dull and fake looking compared to blu rays on my plasmas. Cri is the only real criteria that has shown to rate light quality from crt and plasmas much higher than oleds so i guess its how i justify my preference for those panels besides my plasma looks basically off on a full black screen contrast is close to OLED and near black shadow detail is way better.

I would note that you are referencing CRI in relation to a presentation medium here - a medium used for viewing with the human eye.
CRI in relation to a light source which might be used in an enlarger or a scanner is a different question, because in those applications the "sensor" behaves in different ways than the human eye.

 

[Cinema]

I would note that you are referencing CRI in relation to a presentation medium here - a medium used for viewing with the human eye.
CRI in relation to a light source which might be used in an enlarger or a scanner is a different question, because in those applications the "sensor" behaves in different ways than the human eye.

Wouldn’t a higher CRI source light still provide more pleasing color and illumination? Thats why arri leds are very expensive and strive for 100 cri (i think they come in around 95-97). also, when scanning, higher CRI source light (at least when dslr and mirrorless scanning) produces a better scan based on the tests i’ve seen

 

[MattKing]

Wouldn’t a higher CRI source light still provide more pleasing color and illumination? Thats why arri leds are very expensive and strive for 100 cri (i think they come in around 95-97). also, when scanning, higher CRI source light (at least when dslr and mirrorless scanning) produces a better scan based on the tests i’ve seen

The problem with CRI is that it is based on criteria that may or may not result in high quality scanning or illumination for the purpose of "photography" - it just isn't intended to measure suitability for those purposes.
Two light sources with the same CRI rating may give significantly different results with something other than a human eye sensor.
If the "spikes" in the spectral emission curve of the source happen to align with the sensitivity "peaks" of the sensor, you will get different results than with a source whose ""spikes" align with the sensitivity "valleys" of the sensor, even though both of the sources may average out in a way that results in the same CRI.

 

[faberryman]

Wouldn’t a higher CRI source light still provide more pleasing color and illumination? Thats why arri leds are very expensive and strive for 100 cri (i think they come in around 95-97). also, when scanning, higher CRI source light (at least when dslr and mirrorless scanning) produces a better scan based on the tests i’ve seen

The problem with CRI is that it is based on criteria that may or may not result in high quality scanning or illumination for the purpose of "photography" - it just isn't intended to measure suitability for those purposes.
Two light sources with the same CRI rating may give significantly different results with something other than a human eye sensor.
If the "spikes" in the spectral emission curve of the source happen to align with the sensitivity "peaks" of the sensor, you will get different results than with a source whose ""spikes" align with the sensitivity "valleys" of the sensor, even though both of the sources may average out in a way that results in the same CRI.

Which scanners take replaceable bulbs? Which enlargers accept bulbs with high CRI ratings?

 

[MattKing]

All tungsten bulbs (which includes halogen bulbs) are by definition 100 CRI bulbs, because they emit a continuous spectrum.
And as for scanners that use other light sources, their sensors can be matched to the non-continuous source they are designed for, so whether or not that source has a high CRI.

 

[faberryman]

All tungsten bulbs (which includes halogen bulbs) are by definition 100 CRI bulbs, because they emit a continuous spectrum.
And as for scanners that use other light sources, their sensors can be matched to the non-continuous source they are designed for, so whether or not that source has a high CRI.

Then I am not sure why we are having the conversation.

 

[Helge]

All tungsten bulbs (which includes halogen bulbs) are by definition 100 CRI bulbs, because they emit a continuous spectrum.
And as for scanners that use other light sources, their sensors can be matched to the non-continuous source they are designed for, so whether or not that source has a high CRI.

I guess that the Nikon Coolscans better colours is exactly because of RGB LEDs, and their peaky output compared to white LED scanners.

 

[Helge]

I would like to know a little more. I have worked in the film industry for a couple decades but only gleaned so much technically working at camera houses in the beginning.

I always thought CRI referred to the quality of light. Which makes sense to me because LED tvs and OLEDs generally look horrible and CRTs and plasmas look incredible vibrant and high quality (they also happen to be incandescent). What is cri referring to exactly? I have measured it before with spectroradiometers but don’t really know much more other than higher is better. My old sony pro crts for coloring were in the 90 cri range while my pioneer plasma is around 75. My LG oled is dismal i think around 40-50 cri. Yes i can read the wikipedia about CRI but i’m not smart enough to really understand why it’s BS, as the quality of light sources seems empirically obvious to me, that the higher cri a source measures, the better it looks.

Google “problems with the CRI standard”.
I don’t feel like writing a dissertation.
What Matt writes is completely correct. And what is most pertinent in this context.

My “hoppy horse” with the physiological problems is of course much more important, but perhaps outside of this forums scope.

CRI is mainly concerned with a (very) overly simplistic requirement of “does the colours looks right”.
Intended for graphical studios and production floors.
It misses all of the intricacies of how colour vision works though.

“Yellow” in the same apparent shade can be arrived at through many roads. And they “look” identical at first but how they work and “feel” on the rods, cones and photosensitive ganglion cells is crucially differently.

 

[MattKing]

I guess that the Nikon Coolscans better colours is exactly because of RGB LEDs, and their peaky output compared to white LED scanners.

The most important thing is that the light sources, any built in filtration, and the sensors are appropriately matched to each other.

 

[Cinema]

Then I am not sure why we are having the conversation.

Backlights for camera scanning vary in cri. I do have a coolscan as well—the color from that is usually pretty good

 

[MattKing]

Then I am not sure why we are having the conversation.

The thread has more than one reference to finding substitutes for incandescent/halogen bulbs. The side discussion about CRI centres on how that standard isn't particularly useful when attempting to evaluate those substitutes.

 

[Helge]

The most important thing is that the light sources, any built in filtration, and the sensors are appropriately matched to each other.

You want to match the light and sensitivity of RA4 printing. Including the peaks of the papers emulsion.

A monochrome sensor and peaky light would probably be the best overall solution.

This would work perfectly well for slide and B&W too.

 

[Cinema]

And as for scanners that use other light sources, their sensors can be matched to the non-continuous source they are designed for, so whether or not that source has a high CRI.

good point. it's much easier for me to get acceptable color inversion on a film scanner with dedicated hardware and software than a hodgepodge of software/backlight/camera when dslr scanning

 

[DREW WILEY]

CRI is still a valid system, although it's often abused to deceptively inflate the perceived quality of a substandard product using marketing tricks. It simply means the percent out of 100 representative color samples that comes out looking correct under any given light source. In my darkroom, I have true CRI 98 5000K German color matching tubes at my retouching station. Those weren't cheap, and are a specialty item. They completely apply to RA4 printing, or any other kind of color printing. The interaction of dye sensitivity spikes in the paper in relation to those in the film, and in turn, the spectral output of the enlarger or laser printer is a complex topic which must be cumulatively assessed at the endpoint of human vision itself. Bees don't typically display color prints in their hives.

I had the same kinds of tubes installed at the paint matching counter where I worked; but matching is just as much a matter of psychology and training in what to specifically look for than it is physiological. No machine, no matter how sophisticated, is a substitute for the human eye in that respect. And ideally, one matches something to those same lighting conditions it will be seen under; hence light booths containing multiple compartments, each with its own kind of representative illumination. And when daylight is involved - what season, what time of day, what engle of the sun, what weather conditions, are given the priority? All that has to be taken into account if you're dealing with a serious architect and his client.

But parallel systems of statistical rating have developed which are more accurate in other contexts; and some high-end bulb makers specify things that way. And different modes are essential when machine vision comes into play. An example would be those art forensics digital cameras which scan paintings at multiple discrete wavelengths, including shorter and longer wavelengths than human vision can detect. It's also necessary when plotting respective pigments within an analytic geometry color mapping program, like when getting color inkjet printers doing what they do.

 

[Helge]

CRI is still a valid system, although it's often abused to deceptively inflate the perceived quality of a substandard product using deceptive marketing. It simply means the percent out of 100 representative color samples that comes out looking correct under any given light source. In my darkroom, I have true CRI 98 5000K German color matching tubes at my retouching station. Those weren't cheap, and are a specialty item. They completely apply to RA4 printing, or any other kind of color printing. The interaction of dye sensitivity spikes in the paper in relation to those in the film, and in turn, the spectral output of the enlarger or laser printer is a complex topic which must be cumulatively assessed at the endpoint of human vision itself. Bees don't typically display color prints in their hives.

I had the same kinds of bulbs installed at the paint matching counter where I worked; but matching is just as much a matter of psychology and training in what to specifically look for than it is physiological. No machine, no matter how sophisticated, is a substitute for the human eye in that respect.

But parallel systems of statistical rating have developed which are more accurate in other contexts; and some high-end bulb makers specify things that way. And different modes are essential when machine vision comes into play. An example would be those art forensics digital cameras which scan paintings at multiple discrete wavelengths, including shorter and longer wavelengths than human vision can detect. It's also necessary when plotting respective pigments within an analytic geometry color mapping program, like when getting color inkjet printers doing what they do.

CRI was defined for assessing colours of art, products and produce under artifical light. And that’s what it’s still good for. Nothing less and nothing more.

 

[DREW WILEY]

Doesn't that "nothing less and nothing more" set of parameters apply to just about everything concerning the visual output of photography, either front-lit or backlit?

 

[Helge]

Doesn't that "nothing less and nothing more" set of parameters apply to just about everything concerning the visual output of photography, either front-lit or backlit?

It doesn’t concern itself with the intricacies of how the sensors involved actually work, including those of the human eye and the exact nature of the light.
That is crucial in copying work and WRT long term health and cognition.

 

[DREW WILEY]

Of course it does. CRI is a useful shorthand method for pointing out the effects of discontinuous light sources from blackbody ones. No, it's not like CIE color mapping and so forth, but that kind of thing is more for specialists. Consumers need something simple to understand. And human vision isn't a fixed sensor system. The cones fatigue and interact in the brain- hence those perceptual characteristics like successive color contrast and simultaneous contrast, which every good painter knows, but alas, few color photographers apparently. Machine sensors and light sources don't respond like that.

 

[Helge]

Of course it does. CRI is a useful shorthand method for pointing out the effects of discontinuous light sources from blackbody ones. No, it's not like CIE color mapping and so forth, but that kind of thing is more for specialists. Consumers need something simple to understand. And human vision isn't a fixed sensor system. The cones fatigue and interact in the brain- hence those perceptual characteristics like successive color contrast and simultaneous contrast, which every good painter knows, but alas, few color photographers apparently. Machine sensors and light sources don't respond like that.

Human vision and especially colour vision is not very comprehensively understood.
And it certainly wasn’t when CRI was worked out.
There has be attempts at replacing CRI with something radically better, but so far the result has only been iterative tweaks.

 

[Alan Edward Klein]

Human vision and especially colour vision is not very comprehensively understood.
And it certainly wasn’t when CRI was worked out.
There has be attempts at replacing CRI with something radically better, but so far the result has only been iterative tweaks.

The eye's cones on the retina only perceive red, blue and green. Other colors transmit frequency signals to more than one cone. The brain then processes them to come up with another stimulation in the brain to reporesent the other unseen color. But the brain really doesn't :see" those colors as they are in the natural world. So how one perceives color can be different from one person to another.

Maybe the whole universe is all perception in our brains. The universe really doesn't exist like we imagine it which is why we really can't understand it. It;s like gravity which is not a force as was originally believed. Its now said that it's curvature of the universe. Well explain that? Explain color. Explain the color purple to a blind person.