How To Remove Weird Colours

[flowers]

Guys, modern cameras have a UV filter built into them, part of the low pass filter stack (along with an IR Cut filter.) You don't need to filter UV. The light was probably your standard fluorescent blacklight. Cheap blacklights include a considerable amount of deep violet visible light. There isn't a UV cutoff issue here...the camera just picked up the deep violet visible light, which human eyes are naturally rather insensitive to. Thats all!

You're giving bad advice. First of all, many humans can see deeper than "visible violet", but that's a whole different point. What's relevant here is that certain cameras also seem to pick up a little bit of near-UV, between 350 and 400. Light in that region is considered UV because it's soft focusing and causes fluorescing. Some modern cameras have really bad UV filters or really bad low-pass filters that don't include a separate UV filter. Canon mk ii and iii see IR quite well up to at least 940 nm. I haven't tested higher.

 

[jrista]

Guys, modern cameras have a UV filter built into them, part of the low pass filter stack (along with an IR Cut filter.) You don't need to filter UV. The light was probably your standard fluorescent blacklight. Cheap blacklights include a considerable amount of deep violet visible light. There isn't a UV cutoff issue here...the camera just picked up the deep violet visible light, which human eyes are naturally rather insensitive to. Thats all!

You're giving bad advice. First of all, many humans can see deeper than "visible violet", but that's a whole different point. What's relevant here is that certain cameras also seem to pick up a little bit of near-UV, between 350 and 400. Light in that region is considered UV because it's soft focusing and causes fluorescing. Some modern cameras have really bad UV filters or really bad low-pass filters that don't include a separate UV filter. Canon mk ii and iii see IR quite well up to at least 940 nm. I haven't tested higher.

I never said humans couldn't see deeper than visible violet...simply that our eyes are not very sensitive to those wavelengths.

As for the visible light range, visible light (if you account for all people of all different ranges of vision) spans from 380nm through around 750nm. "Officially", the range spans from 390nm to 700nm. As an astrophotographer who has been in the market for some kind of imager for deep sky narrow band imaging for a while now, and who has done a MASSIVE amount of research on the subject recently, both UV and IR sensitivity in modern DSLR cameras is really low. The range of sensitivity fits pretty closely to the 380nm through 750nm, with a strong signal only really found from about 400nm through 690nm due to UV and IR filtration.

The UV/IR cut filters are not perfect, however they block out some 90% or more of the near bands, and nearly 100% of the far bands, and the silicon itself takes care of the rest. The IR filter in Canon cameras is actually a bit aggresive...it attenuates the deep red signal, which is part of the reason why the red channel in Canon sensors tends to be rather noisy. The natural response curve for silicon is much less sensitive to UV light either, sensitivity up to 390nm remains relatively strong (although still about half that or less than for visible visible wavelengths of light), and falls off very rapidly from there till it finally tapers out completely at 250nm. It is the near IR spectrum that silicon is more sensitive to, and that can extend well past 1100nm, with strong sensitivity up to 900nm or so without any IR cutoff filter.

In astrophotography, some of that near infrared signal is useful, however you have to modify modern DSLR cameras to enhance that sensitivity. Canon actually makes the 60Da, specially designed with a weaker IR cutoff filter that allows a much stronger hydrogen-alpha (656nm) and sulfur-II (672nm) signal to get through, (and even that still isn't as strong as green sensitivity. In all other Canon CMOS sensors with the IR cut filter, these wavelengths are less than half as strong, and falloff for wavelengths longer than that is rapid. Even if some 900nm IR does get through, the signal is so weak that it isn't going to have a significant effect...it's all probably lost to Canon's read noise anyway.

Manufacturing IR cutoff filters that allow full spectrum transparency up to 750nm, with a sharp cutoff, are much more expensive. They would technically be much better, but DSLR camera manufacturers generally don't use them. To find cameras with IR and UV filtration that is "square", with >90% signal transparency through the entirety of the desired band and <0.1% transparency outside of the band, you usually have to look to the scientific grade devices. An SBIG CCD camera for astrophotography, for example, will usually have high quality IR and UV cutoff like that, but those cameras tend to cost anywhere from $3000 bare, to over $10,000 with some additional features.

The kind of bright purple in the photos of the OP were most likely caused by that very near-UV (380nm) + deep violet light (390nm up through 400nm), a rather broad-ish band that is emitted by those cheap fluorescent tubes.

 

[flowers]

I don't know about the numbers exactly, I'm not a scientist! I did buy my 5dii slightly used (shutter count under 5000), so it's always possible the filter has been changed without my knowledge! However I highly doubt that! My test was simply to point a 940nm IR led at it, which it clearly recorded as pinkish white light with red in the middle! Many cameras can actually see IR in the 900 range. I agree with your explanation about the violet in the OP's photo, that's what it seems to be. I'm surprised OP didn't see it, that's exactly how I've seen yellow (tungsten) and blacklight violet/UV mix, I instantly recognized it! The yellow shadows in the deep purple (the object blocking the blacklight but not the tungsten) is a dead giveaway! The only difference to viewing it with your bare eyes is that the tonality is poorer

 

[jrista]

I don't know about the numbers exactly, I'm not a scientist! I did buy my 5dii slightly used (shutter count under 5000), so it's always possible the filter has been changed without my knowledge! However I highly doubt that! My test was simply to point a 940nm IR led at it, which it clearly recorded as pinkish white light with red in the middle! Many cameras can actually see IR in the 900 range. I agree with your explanation about the violet in the OP's photo, that's what it seems to be. I'm surprised OP didn't see it, that's exactly how I've seen yellow (tungsten) and blacklight violet/UV mix, I instantly recognized it! The yellow shadows in the deep purple (the object blocking the blacklight but not the tungsten) is a dead giveaway! The only difference to viewing it with your bare eyes is that the tonality is poorer

It would depend on how long you exposed for. Pointing an IR led directly at the sensor, I assume from a relatively short distance, is quite a bit different than picking up IR and UV along with whatever other light is reaching your sensor during any normal exposures. Most modern digital camera sensors use a standard silicon substrate, and silicon is naturally sensitive to a spectrum ranging from ultraviolet (maybe ~250nm) through deep infrared (up to 5000nm). The peak sensitivity range is from maybe 300nm through maybe 2000nm, beyond that the natural sensitivity levels for both deep UV and deep IR are relatively low.

It also depends on the exact characteristics of the IR cutoff filter. As I mentioned before, most DSLRs don't use high quality UV/IR filters that have high transparency to desired wavelengths and a strong and total cutoff to undesired wavelengths. They have a more gradual curve to them...filtration usually starts in the deep reds of visible light, and tapers off into near infrared. It's possible the 5D II has a rather long falloff period that allows more IR through.

 

[flowers]

I don't know about the numbers exactly, I'm not a scientist! I did buy my 5dii slightly used (shutter count under 5000), so it's always possible the filter has been changed without my knowledge! However I highly doubt that! My test was simply to point a 940nm IR led at it, which it clearly recorded as pinkish white light with red in the middle! Many cameras can actually see IR in the 900 range. I agree with your explanation about the violet in the OP's photo, that's what it seems to be. I'm surprised OP didn't see it, that's exactly how I've seen yellow (tungsten) and blacklight violet/UV mix, I instantly recognized it! The yellow shadows in the deep purple (the object blocking the blacklight but not the tungsten) is a dead giveaway! The only difference to viewing it with your bare eyes is that the tonality is poorer

It would depend on how long you exposed for. Pointing an IR led directly at the sensor, I assume from a relatively short distance, is quite a bit different than picking up IR and UV along with whatever other light is reaching your sensor during any normal exposures. Most modern digital camera sensors use a standard silicon substrate, and silicon is naturally sensitive to a spectrum ranging from ultraviolet (maybe ~250nm) through deep infrared (up to 5000nm). The peak sensitivity range is from maybe 300nm through maybe 2000nm, beyond that the natural sensitivity levels for both deep UV and deep IR are relatively low.

It also depends on the exact characteristics of the IR cutoff filter. As I mentioned before, most DSLRs don't use high quality UV/IR filters that have high transparency to desired wavelengths and a strong and total cutoff to undesired wavelengths. They have a more gradual curve to them...filtration usually starts in the deep reds of visible light, and tapers off into near infrared. It's possible the 5D II has a rather long falloff period that allows more IR through.

It was from a close distance yes, but after reading this I did another test: I asked someone to hold the IR source (a remote controller with a 940nm led, the wavelength of the led was verified) a little further away at a distance of 3.5 meters. The light seemed a little weaker but was still perfectly visible and had both the white and the pinkish component. Checked in LV at 10x magnification using sigma 35/1.4 at f/1.4, 1/50s shutter speed. Inside lighting illuminating the remote controller was very yellow CFL (around 2800K), WB was adjusted to neutralize the color temperature.

 

[jrista]

I don't know about the numbers exactly, I'm not a scientist! I did buy my 5dii slightly used (shutter count under 5000), so it's always possible the filter has been changed without my knowledge! However I highly doubt that! My test was simply to point a 940nm IR led at it, which it clearly recorded as pinkish white light with red in the middle! Many cameras can actually see IR in the 900 range. I agree with your explanation about the violet in the OP's photo, that's what it seems to be. I'm surprised OP didn't see it, that's exactly how I've seen yellow (tungsten) and blacklight violet/UV mix, I instantly recognized it! The yellow shadows in the deep purple (the object blocking the blacklight but not the tungsten) is a dead giveaway! The only difference to viewing it with your bare eyes is that the tonality is poorer

It would depend on how long you exposed for. Pointing an IR led directly at the sensor, I assume from a relatively short distance, is quite a bit different than picking up IR and UV along with whatever other light is reaching your sensor during any normal exposures. Most modern digital camera sensors use a standard silicon substrate, and silicon is naturally sensitive to a spectrum ranging from ultraviolet (maybe ~250nm) through deep infrared (up to 5000nm). The peak sensitivity range is from maybe 300nm through maybe 2000nm, beyond that the natural sensitivity levels for both deep UV and deep IR are relatively low.

It also depends on the exact characteristics of the IR cutoff filter. As I mentioned before, most DSLRs don't use high quality UV/IR filters that have high transparency to desired wavelengths and a strong and total cutoff to undesired wavelengths. They have a more gradual curve to them...filtration usually starts in the deep reds of visible light, and tapers off into near infrared. It's possible the 5D II has a rather long falloff period that allows more IR through.

It was from a close distance yes, but after reading this I did another test: I asked someone to hold the IR source (a remote controller with a 940nm led, the wavelength of the led was verified) a little further away at a distance of 3.5 meters. The light seemed a little weaker but was still perfectly visible and had both the white and the pinkish component. Checked in LV at 10x magnification using sigma 35/1.4 at f/1.4, 1/50s shutter speed. Inside lighting illuminating the remote controller was very yellow CFL (around 2800K), WB was adjusted to neutralize the color temperature.

I'd offer that 3.5 meters isn't all that far for a concentrated IR emitter like a TV remote. Those things emit a pretty powerful signal, even though we can't see it. They have to combat the ambient temperatures in peoples homes, the energy emitted by direct sunlight hitting the receivers (well, in some cases...one of my IR remotes still works when the receiver is bathed in direct sunlight, the others tend to be sketchy), etc. In terms of infrared light, those remote controls send out a pretty "bright" beam...kind of like a bright visible light of a handheld search light vs. the dimmer beam of a flashlight.

You also have to wonder if your little remote control is just emitting IR, or whether it is emitting some visible light as well. I know that I can see a faint yellow light emitted from the IR LED in one of my TV remotes if I look at it while it's emitting. I think I once had a remote that emitted bright red light as well as IR, as it doubled as the indicator light telling the user that the remote was actually indeed sending a signal.

 

[flowers]

I'd offer that 3.5 meters isn't all that far for a concentrated IR emitter like a TV remote. Those things emit a pretty powerful signal, even though we can't see it. They have to combat the ambient temperatures in peoples homes, the energy emitted by direct sunlight hitting the receivers (well, in some cases...one of my IR remotes still works when the receiver is bathed in direct sunlight, the others tend to be sketchy), etc. In terms of infrared light, those remote controls send out a pretty "bright" beam...kind of like a bright visible light of a handheld search light vs. the dimmer beam of a flashlight.

You also have to wonder if your little remote control is just emitting IR, or whether it is emitting some visible light as well. I know that I can see a faint yellow light emitted from the IR LED in one of my TV remotes if I look at it while it's emitting. I think I once had a remote that emitted bright red light as well as IR, as it doubled as the indicator light telling the user that the remote was actually indeed sending a signal.

It doesn't have to be that strong because window glass is mostly opaque to near-IR and quite a wide band of IR and UV, all sunlight in homes is diffuse, and IR receivers in TVs and other devices only accept very specific signals (repeating patterns of very specific duration and interval) of an extremely narrow band of IR, namely the wavelength of the remote controller.
There was no visible light, I've tested the remote controller I used on another person with normal vision, that person could see nothing with the lights on and nothing in complete darkness. I could, but that's beside the point, my eyes aren't normal.

You should also note that the different types of cones in people can have different spectral sensitivity curves, the cones in some people's eyes can simply have a much wider spectrum they detect than some others. There can be significant variance to spectral sensitivity in the cones, depending on genetic and environmental factors (smoking, retinol intake, betacarotein intake, exposure to UV, genetic sensitivity to UV, physical properties of the lens, refractive index of the lens, cataracts, how easily the lens "clouds" in response to stimuli and so on). Like I said, I'm no scientist, but I know that much from reading and participating in conversations. The 380-750 is by no means a hard limit. In the case of extremely bright IR sources like lasers, there are even military reports of people being able to see the beams of 1064 nm lasers on a moonless night in some cases. Like you already pointed out yourself, sensitivity is different from absolute ability.

 

[jrista]

I'd offer that 3.5 meters isn't all that far for a concentrated IR emitter like a TV remote. Those things emit a pretty powerful signal, even though we can't see it. They have to combat the ambient temperatures in peoples homes, the energy emitted by direct sunlight hitting the receivers (well, in some cases...one of my IR remotes still works when the receiver is bathed in direct sunlight, the others tend to be sketchy), etc. In terms of infrared light, those remote controls send out a pretty "bright" beam...kind of like a bright visible light of a handheld search light vs. the dimmer beam of a flashlight.

You also have to wonder if your little remote control is just emitting IR, or whether it is emitting some visible light as well. I know that I can see a faint yellow light emitted from the IR LED in one of my TV remotes if I look at it while it's emitting. I think I once had a remote that emitted bright red light as well as IR, as it doubled as the indicator light telling the user that the remote was actually indeed sending a signal.

It doesn't have to be that strong because window glass is mostly opaque to near-IR and quite a wide band of IR and UV, all sunlight in homes is diffuse, and IR receivers in TVs and other devices only accept very specific signals (repeating patterns of very specific duration and interval) of an extremely narrow band of IR, namely the wavelength of the remote controller.

Well, first off, you have a few things wrong here. Window glass is mostly opaque to UV, and tends to block around 90% of it. However given the intensity of the sun, a LOT of IR gets through. IR is one of the primary means by which heat is transferred from the sun, and the sun is massively intense. Even if window glass blocks 70% of it, the 30% that gets through is still considerable. Next, NOT all sunlight in homes is diffuse. Sunlight shining directly through a window is not diffuse at all, it's direct. You put your hand under it, and you can feel the heat, which is primarily transferred via IR radiation.

There was no visible light, I've tested the remote controller I used on another person with normal vision, that person could see nothing with the lights on and nothing in complete darkness. I could, but that's beside the point, my eyes aren't normal.

I figured. If you allowed normal light to expose along with the IR beam, you probably wouldn't see much of the IR beam at all. Because of the IR cut filter, the camera is going to receive much more visible light than IR. It isn't a hard cutoff, it gradually falls off from around 630nm or so until maybe 780nm, then it falls off more quickly until around 1100nm. The filter is still blocking the vast majority of IR from around 750nm through 1100nm though.

You should also note that the different types of cones in people can have different spectral sensitivity curves, the cones in some people's eyes can simply have a much wider spectrum they detect than some others. There can be significant variance to spectral sensitivity in the cones, depending on genetic and environmental factors (smoking, retinol intake, betacarotein intake, exposure to UV, genetic sensitivity to UV, physical properties of the lens, refractive index of the lens, cataracts, how easily the lens "clouds" in response to stimuli and so on). Like I said, I'm no scientist, but I know that much from reading and participating in conversations. The 380-750 is by no means a hard limit. In the case of extremely bright IR sources like lasers, there are even military reports of people being able to see the beams of 1064 nm lasers on a moonless night in some cases. Like you already pointed out yourself, sensitivity is different from absolute ability.

There are exceptionally few human individuals on the planet who can see beyond the range of 380nm to 750nm. That would be more of a maximum range, factoring in the weakest sensitivities under stringent testing conditions, which is quite different than normal vision on a normal day to day basis. Realistically, I'd say human vision on the average case is from ~400nm to ~700nm or so.

I used to do extensive research in color theory and how it pertained to human vision. There are some rare genetic aberrations that give some people much more limited sensitivity to light, usually the loss of function in one type of cone (this usually leads to one of the various forms of color blindness). A very small percentage of women seem to have an extra cone color, called the orange cone with an 'orange yellow-orange' color, within the 2° foveal spot. It isn't well understood how this extra cone works with the normal process of vision, although it seems to be sensitive to a narrower band than standard red cones. Some scientists suspect it is simply a differently or malformed red cone, and is therefor treated as a red cone by the brain. Being sensitive to light wavelengths closer to green than standard red cones and not sensitive to blue light at all, it may not significantly change the range of wavelengths these women are sensitive to, if it changes the range at all. I don't know of any human who has even more than the most minimal sensitivity to frequencies around 380nm, although I've read some obscure things about other rare diseases that might enhance sensitivity to blue and purple hues. Some humans are rarely sensitive down to 780nm, however sensitivity levels are extremely low (notice the right-hand shoulder of the green sensitivity curve below, how it tapers off very slowly to the end of the graph...something like that, only with the red sensitivity curve.)

Human vision is a tristimulus, we see three primary bands of light...the reds, the greens, and the blues, however our red cones are also visible to some blue as well (which is what gives us the ability to see magenta.) The spectral response of the cones in the vast majority of human eyes is as follows:

The blue curve produces a narrower peak, with little overlap with green and red sensitivity (although do note that red sensitivity reaches almost to 400nm). You can see how rapidly blue sensitivity falls off after ~420nm, down to nothing by 380nm. Even if that faloff curve was extended or even shifted to 370nm or 360nm, it is not going to greatly enhance our sensitivity to near-uv. Official range measurements for cone sensitivity in humans is as follows:

Cone	Range	Peak
S (Blue)	400-500nm	420-440nm
M (Green)	450-630nm	534-555nm
L (Red)	500-700nm	564-580nm

The widest variations in human vision are primarily caused by defects in the way rods and cones form. This is primarily dominated by color blindness, which basically punches holes in the spectrum of visible light which color blind individuals can see. The various forms of color blindness each affect anywhere from 1% to 5% of the population, depending on the kind. An ultra-rare group of people on earth have monochromatic vision (lack of functioning cones entirely), therefor they see the world in black and white with only their rods. Another exceptionally rare forms of color blindness result in random interpretation of various wavelengths of light (knew someone like this when I was a kid...he would organize crayons into a "rainbow" according to how his brain interpreted each color...it looked completely random and without pattern or sequence to everyone else...weirdest vision impairment I've ever encountered in my life, and his impairment was around one in a billion, so exceptionally rare.)

In all honesty, outside of the orange cones in a very small percentage of women, I don't really know of any regular wild variations in cone type and spectral sensitivity, nor anyone who can really see UV light...even if they had a sensitivity to it, it would be so minimal that it wouldn't affect their vision much in a regular basis. They would probably only learn they had such a sensitivity in a carefully controlled scientific test space and shown deep violet and UV lights of varying frequencies. It certainly isn't common enough to be of concern in the context of photography. Extreme exposures to UV light can and will certainly damage your eyes, and indeed it can cause terrible ailments like cataracts...but that doesn't actually have anything to do with how sensitive our cones are to UV frequencies.

As for people seeing 1064nm lasers...you gotta provide some references to that one! I might be able to understand the use of such a laser to stun someone or temporarily blind them with a short pulse, however that is well beyond the range of human eyesight. Well beyond. I highly doubt more than a handful of people on earth could see that, if even one person at all could see infrared energy if that frequency.

 

[flowers]

There are exceptionally few human individuals on the planet who can see beyond the range of 380nm to 750nm. That would be more of a maximum range, factoring in the weakest sensitivities under stringent testing conditions, which is quite different than normal vision on a normal day to day basis. Realistically, I'd say human vision on the average case is from ~400nm to ~700nm or so.

I used to do extensive research in color theory and how it pertained to human vision. There are some rare genetic aberrations that give some people much more limited sensitivity to light, usually the loss of function in one type of cone (this usually leads to one of the various forms of color blindness). A very small percentage of women seem to have an extra cone color, called the orange cone with an 'orange yellow-orange' color, within the 2° foveal spot. It isn't well understood how this extra cone works with the normal process of vision, although it seems to be sensitive to a narrower band than standard red cones. Some scientists suspect it is simply a differently or malformed red cone, and is therefor treated as a red cone by the brain. Being sensitive to light wavelengths closer to green than standard red cones and not sensitive to blue light at all, it may not significantly change the range of wavelengths these women are sensitive to, if it changes the range at all. I don't know of any human who has even more than the most minimal sensitivity to frequencies around 380nm, although I've read some obscure things about other rare diseases that might enhance sensitivity to blue and purple hues. Some humans are rarely sensitive down to 780nm, however sensitivity levels are extremely low (notice the right-hand shoulder of the green sensitivity curve below, how it tapers off very slowly to the end of the graph...something like that, only with the red sensitivity curve.)

Human vision is a tristimulus, we see three primary bands of light...the reds, the greens, and the blues, however our red cones are also visible to some blue as well (which is what gives us the ability to see magenta.) The spectral response of the cones in the vast majority of human eyes is as follows:

The blue curve produces a narrower peak, with little overlap with green and red sensitivity (although do note that red sensitivity reaches almost to 400nm). You can see how rapidly blue sensitivity falls off after ~420nm, down to nothing by 380nm. Even if that faloff curve was extended or even shifted to 370nm or 360nm, it is not going to greatly enhance our sensitivity to near-uv. Official range measurements for cone sensitivity in humans is as follows:

Cone	Range	Peak
S (Blue)	400-500nm	420-440nm
M (Green)	450-630nm	534-555nm
L (Red)	500-700nm	564-580nm

The widest variations in human vision are primarily caused by defects in the way rods and cones form. This is primarily dominated by color blindness, which basically punches holes in the spectrum of visible light which color blind individuals can see. The various forms of color blindness each affect anywhere from 1% to 5% of the population, depending on the kind. An ultra-rare group of people on earth have monochromatic vision (lack of functioning cones entirely), therefor they see the world in black and white with only their rods. Another exceptionally rare forms of color blindness result in random interpretation of various wavelengths of light (knew someone like this when I was a kid...he would organize crayons into a "rainbow" according to how his brain interpreted each color...it looked completely random and without pattern or sequence to everyone else...weirdest vision impairment I've ever encountered in my life, and his impairment was around one in a billion, so exceptionally rare.)

In all honesty, outside of the orange cones in a very small percentage of women, I don't really know of any regular wild variations in cone type and spectral sensitivity, nor anyone who can really see UV light...even if they had a sensitivity to it, it would be so minimal that it wouldn't affect their vision much in a regular basis. They would probably only learn they had such a sensitivity in a carefully controlled scientific test space and shown deep violet and UV lights of varying frequencies. It certainly isn't common enough to be of concern in the context of photography. Extreme exposures to UV light can and will certainly damage your eyes, and indeed it can cause terrible ailments like cataracts...but that doesn't actually have anything to do with how sensitive our cones are to UV frequencies.

As for people seeing 1064nm lasers...you gotta provide some references to that one! I might be able to understand the use of such a laser to stun someone or temporarily blind them with a short pulse, however that is well beyond the range of human eyesight. Well beyond. I highly doubt more than a handful of people on earth could see that, if even one person at all could see infrared energy if that frequency.

Wow you write a lot! I really don't think it's as rare as you say it is, if it was I couldn't have found so many people online who can also see the 940nm LEDs in remote controllers that their friends/relative can't and in broad daylight! You got some things wrong, I asked that other person if he could see it 1) first with the lights on, 2) then in darkness after some adaptation. He couldn't. I can see it as faintly red with the lights on, and to some extent even in sunlight flooding the apartment (but it's very dim). I've proven this to many people IRL and I'm always willing to prove it! Though I won't give random people my home address online just to prove a point! I used to see UV better when I was a child, now I see it poorly, but I know it's UV because of the way I see flowers compared to how others see them. I don't have aphakia or cataracts removed so I can't explain that one. I've actually contemplated having the lens removed from one eye because I'm slightly bothered by my impaired UV vosopn, it certainly affects my color vision! It's a little wonky anyway, green is supposed to be the color normal vision is most sensitive to, my vision is the least sensitive to green.
On seeing UV: http://www.komar.org/faq/colorado-cataract-surgery-crystalens/ultra-violet-color-glow/
(check out the two fluorescent tube images, what I see is between the two, closer to the non-UV one but I also see some glow around the tube, in direct sunlight)
http://www.downloadtheuniverse.com/dtu/2012/04/monets-ultraviolet-eye.html

Theory is theory, practice is practice. I've talked to enough people (in real life as well) to know the difference between the two.

https://hplusbiopolitics.wordpress.com/2008/09/15/therma-vision/

Actually, people can see infrared, at least up to 1064 nm, and likely up to 1152 nm. In the mid-20th century vision researchers documented human ability to see well into the 900s. The definitive study on the subject was done by Sliney et al (1976) using heavily filtered infrared lasers, all with 3 mm beams. At 1064 nm the average power needed to see the light was 0.069 mW (entire beam, measured at the pupil). Above 632,8 nm the cones are more sensitive than the rods so that color is at maximum saturation at threshold.

As for what it looks like, basically red, but trending towards orange at higher wavelengths. Above 704 nm the sensitivity of the green cones tapers off more slowly that that of the red cones, while the blue cones play no significant role. This is known as infrared color reversal.

With the last laser they tested was at 1152 nm, none of the test subjects saw the light. Researchers calculated that it should have been visible at 3 – 4 mW, but thought it would unsafe to go any higher than 2.5 mW, mainly for thermal reasons. They didn’t want to go above a rise of 1.0 degree C in the hottest part of the retina over the course of 10-second exposure during which time power would be increased by a factor of ten in ten steps.

In the mid range of the near infrared, sensitivity drops off by about one order of magnitude every 50 nm, with minor variations in the opacity of the eye in the high 900s and low 1000s, and a major increase starting around 1200 nm to 1300nm. To see 1200 nm you would need to shine a 40-mW laser in your eye (not safe). At 1250nm it would have to be about 400 mW (guaranteed eye injury).

By the way, the fourth cone leads to increased color discrimination and it's really really rare and only encountered in females or chimeras/mosaics with XX chromosomes. It seems that increased sensitivity to IR/UV and changes in the lens that allow the UV to pass through are a lot more common. For some reason it doesn't seem like it's widely known or talked about. Maybe because it would endanger "national security" or something like that? After all the military uses IR lights and lasers to detect intruders! But they also use thermal imaging so not much use!

More links:
http://www.fotopedia.com/wiki/Light