light color / hue- Mtbr.com
Results 1 to 18 of 18
  1. #1
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902

    light color / hue

    this is something beam shots don't capture unless you aim two lights at a white wall, then you can see the difference in the hue.

    so when i compared my Fenix TK21 flashlight to my Princeton Tec EOS Bike light the EOS looked white and Fenix looked greenish. this is something i did not expect considering Fenix was 3 times the price.

    the story repeated itself again when i had Niterider 3600 and Lupine Piko side by side and Niterider looked white with Piko coming off greenish. this was again unexpected because Lupine is a better brand. and even more unexpected since they both use XML leds.

    Niterider throws rings of various hues that more or less blend into white, while the Piko presents a uniform greenish beam surrounded by purple-ish halo.

    anyway i think this gets overlooked too often, and i know some lights are far worse offenders than mine.

    how can you protect yourself when buying a light to make sure you don't end up with something weird looking ?

    at first i thought the green beam of Piko is only relative to Niterider, but then i realized it is relative to it's own Purple Halo. when i combine the two beams ( NR and Piko ) it looks like a yellow stain in the middle which i think for the price of Lupine there shouldn't be.

    i wanted to replace Piko with Wilma to get some extra lumens - but will it have the same green / yellow spot / stain problem ?
    Last edited by androgen; 10-06-2012 at 01:12 PM.

  2. #2
    mtbr member
    Reputation:
    Join Date
    May 2005
    Posts
    3,690
    Quote Originally Posted by androgen View Post
    how can you protect yourself when buying a light to make sure you don't end up with something weird looking ?
    You really can't when buying a mass produced light. The only way to control tint is to DIY or buy from a custom builder.

    I notice color shifts slightly as the drive current changes. This was most apparent when I used MCE emitters. At low current they were notably more green than at high currents. There may also be color changes due to the optics used. That's more of a guess on my part as I am most certainly not an optical engineer.

  3. #3
    mtbr member
    Reputation:
    Join Date
    Dec 2007
    Posts
    171
    Hue doesn't matter so much as getting a well balanced spectrum of light. Imagine how much easier it would be to navigate at night if objects had the same vibrant colors they had during the day. Greater lumens helps compensate. Lights are gaining a better spectrum of light, but we're still probably several years away before those lights match our dull lights of today.

  4. #4
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    Quote Originally Posted by leaftye View Post
    Hue doesn't matter so much as getting a well balanced spectrum of light. Imagine how much easier it would be to navigate at night if objects had the same vibrant colors they had during the day. Greater lumens helps compensate. Lights are gaining a better spectrum of light, but we're still probably several years away before those lights match our dull lights of today.
    i believe the sensors responsible for color vision are less sensitive than ones responsible for black and white vision, so even if your lamp had same spectrum as the sun things would still look mostly black and white unless it was bright enough.

    with street lighting the spectrum is often the bottleneck as those yellow lights are mostly just one frequency. but with white LED light i think the bottleneck of color perception is the lumens, not spectrum.

  5. #5
    mtbr member
    Reputation:
    Join Date
    Dec 2007
    Posts
    171
    Quote Originally Posted by androgen View Post
    i believe the sensors responsible for color vision are less sensitive than ones responsible for black and white vision, so even if your lamp had same spectrum as the sun things would still look mostly black and white unless it was bright enough.

    with street lighting the spectrum is often the bottleneck as those yellow lights are mostly just one frequency. but with white LED light i think the bottleneck of color perception is the lumens, not spectrum.
    You're right that the color receptors are less sensitive. So merely matching the day light color spectrum may not be sufficient. What colors do you think are most important? I would think red. It's the color of branches, rust, bricks and lots of rocks.

    Street lights are not a single frequency. White LEDs while having an abundance of frequencies, generally fall very short in the lower visible frequencies. In laymen's terms, they doesn't produce enough red lumens. Like I said, greater lumens can compensate as a brighter white LED should put more red lumens than a dimmer white LED. What I'm driving at is that if LEDs had a better light spectrum, there wouldn't be a need to over compensate.

  6. #6
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    Many things are occurring with regard to visual acuity, low light levels, and color. Light source, intensity, contrast, and color all have an impact on how well you see at night.

    Rods and cones are our visual receptors, and the rods are mostly black-and-white wide-field vision, while cones receive color and are foveal (focal) receptors. Cones also want higher levels of light. As ambient light levels reduce, under about 5FC (if I recall correctly), your visual system slides from the photopic range to the scotopic. When the world turns blueish at dusk, your rods are taking over the majority of the work as the cones are no longer receiving enough input.

    Your visual sensitivity is basically a nice bell curve over the visual spectrum, with a photopic peak in the green range at about 555nM. At lower light levels, this bell curve shifts to blue-green and a narrower, but taller, shape and peaks at about 505nM. The blue-green wavelengths also control the dilation of the pupil. Red and yellow light will lead to a larger pupil aperture, allowing more aberrant light in, reducing visual acuity, while blue-green light will reduce the aperture. Think of a pinhole camera: a larger hole leads to a fuzzy image.

    In terms of lighting, the light source needs to produce energy in wavelengths that your eye can receive. Low Pressure Sodium streetlights are truly monochromatic and are used heavily in Europe and in cities near astronomical observatories, while High Pressure Sodium is more common in US roadway lighting. HPS has a fair amount of yellow and red in its spectral distribution, but much of its energy is at the low end of the scotopic bell curve. The image below shows the scotopic sensitivity curve overlayed on a typical HPS spectral distribution. You can see that little of the energy produced by the light source is inside the sensitivity curve. Metal Halide, Ceramic Metal Halide, and LED roadway lighting are much more blue/white, providing more energy inside the bell curve and thus, more energy for the rods to receive. In most instances, a lower level of blue/white light can have a higher visual acuity than more light in the HPS color range. So, bluer light = better sight.

    When LEDs enter the picture, the higher efficacy (lumens per watt) of the chip is typically in the blue range, especially at a pure blue. Some architectural LED fixtures use pure blue LEDs and color correct via a lens, called remote phosphor. As the LED color is warmed up to colors preferred by consumers, light output per watt drops. Color consistency between LED batches is also a major concern. Architectural manufacturers will color test and sort each batch (called "binning") to get the proper color. I do not know how careful bike and flashlight manufacturers are with this, but it's safe to assume that lower cost products have had fewer sorts.

    Light distribution is also important, as a good flat field will reduce contrast, while a field with dark spots or abrupt edges will have higher contrast. In higher contrast situations, your eye will need to adjust frequently to the high- and low-contrast areas, leading to eyestrain from adaptation (the eye flexes, muscles get tired). Aging eyes also adjust more slowly. Typically, a flat field (called a cosine field) is better than a peak field, meaning that the less difference between the light level at the center of the field compared to the edge is a smaller ratio, as opposed to all the light being within 50% of the overall diameter, will be easier to see within.

    I'm still riding a Niterider HID, which uses a Solarc lamp with pretty good 3100*K (halogen-ish) color and spectral distribution. I sanded the outside of the lamp lens to flatten out the field, and I find that the color and distribution are pretty good. However, it could croak any time and I'll have to spend a lot of time looking at LED replacements. Unfortunately, architectural LED data is easy to come by and uses fairly standard reporting procedures, but bike and flashlight data is hard to come by on paper.
    Attached Thumbnails Attached Thumbnails light color / hue-hps-scotopic-sensitivity.jpg  


  7. #7
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    good food for thought there Stockli. thanks for sharing.

  8. #8
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    No problem. this stuff can be bewildering, and frankly, there's a lot of smoke and mirrors out there. I deal with lighting every day, and what I see for bike lighting information bugs the hell out of me. I'm dreading the death of my Solarc lamp.

    Androgen, you and Leaftye's comments above are on the right track, I just fleshed them out while I had a cup of coffee.

  9. #9
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    Quote Originally Posted by Stockli Boy View Post
    No problem. this stuff can be bewildering, and frankly, there's a lot of smoke and mirrors out there. I deal with lighting every day, and what I see for bike lighting information bugs the hell out of me. I'm dreading the death of my Solarc lamp.

    Androgen, you and Leaftye's comments above are on the right track, I just fleshed them out while I had a cup of coffee.
    it is not particularly bewildering to me because i have an electrical engineering degree and was involved with architectural lighting at work. i was not part of our dedicated lighting team though who used software modeling for architectural lighting, and participated in competitions based on that - but i saw the kind of work they did.

    back then ( was like 5 years ago ? ) there was a huge push from Customers to switch to LEDs either for prestige or for LEED credits or both, but the fixtures weren't putting out enough lumens to do the job. not sure how it is now.

  10. #10
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    Andro, you speak the language (I'm also fluent in Engineer) so, here you go.

    Architectural LED is the norm, not the exception now. A year ago, it was 30% of my design, now it's closer to 70. All site and roadway now is LED, as well as loads of residential stuff. Municipalities and commercial clients want LED for the reduction in maintenance more than they want it for energy savings. IECC 2009 requires residences to be 50% high efficiency lamping, IECC2012 will be 75%. You can do this without breaking a sweat, but you have to know your products and have reliable documentation. The big challenge is that the products are replaced so quickly that what gets specified rarely exists by the time the contractor needs to order it. I'd say the LED industry has blown Moore's Law out of the water.

    What frosts me is that I can get a cutsheet with valid, tested photometry for most architectural LEDs, but I can't find anything like that for bike lights. To date, I've never seen "LM-&9" cited on a bike light website, let alone an ETL report. As a result, claims like "1800 lumens" have no support in my eyes. I can't trust a number without a test procedure and published report. Even with a good CREE LED, odds are very good it's not putting out anywhere near its rated lumens, so it can be pretty misleading, even to someone who deals with these topics on a daily basis.

    Some of the photos that people post are helpful for getting a sense of performance and quality, but it's really hard to shoot lighting and not have the camera flatten out the results. Fortunately for me, most of my night riding is on snow, so that can make even a lousy light work pretty well.

  11. #11
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    Quote Originally Posted by Stockli Boy View Post
    Andro, you speak the language (I'm also fluent in Engineer) so, here you go.

    Architectural LED is the norm, not the exception now. A year ago, it was 30% of my design, now it's closer to 70. All site and roadway now is LED, as well as loads of residential stuff. Municipalities and commercial clients want LED for the reduction in maintenance more than they want it for energy savings. IECC 2009 requires residences to be 50% high efficiency lamping, IECC2012 will be 75%. You can do this without breaking a sweat, but you have to know your products and have reliable documentation. The big challenge is that the products are replaced so quickly that what gets specified rarely exists by the time the contractor needs to order it. I'd say the LED industry has blown Moore's Law out of the water.

    What frosts me is that I can get a cutsheet with valid, tested photometry for most architectural LEDs, but I can't find anything like that for bike lights. To date, I've never seen "LM-&9" cited on a bike light website, let alone an ETL report. As a result, claims like "1800 lumens" have no support in my eyes. I can't trust a number without a test procedure and published report. Even with a good CREE LED, odds are very good it's not putting out anywhere near its rated lumens, so it can be pretty misleading, even to someone who deals with these topics on a daily basis.

    Some of the photos that people post are helpful for getting a sense of performance and quality, but it's really hard to shoot lighting and not have the camera flatten out the results. Fortunately for me, most of my night riding is on snow, so that can make even a lousy light work pretty well.
    yeah that's life. i am also frustrated with lack of information on consumer products. if you're buying a professional studio or DJ loudspeaker you almost always get a measured frequency response chart ( it is usually inaccurate, but it gives some idea ) but with home hi-fi speakers the only time i have seen a manufacturer post a frequency response is for Revel Ultima Salon 2.

    i'm not sure why they are so afraid to disclose any objective information beyond marketing babbling but that's how it is. the few products that are exceptions to this are very special.

    when it comes to bike lights though - i am pretty sure most manufacturers haven't done the testing in the first place. i'm sure Lupine did, perhaps Light & Motion - but the average light out there like MagicShine - i am sure they don't know themselves, and don't care.

  12. #12
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    I agree. I'd love a couple of Lupines on my bike, but I'd have to look into a 2nd mortgage. Maybe it's that as consumers we are driven by price before performance, and we've driven manufacturing offshore to support that. My family has a huge fleet of bikes, and all but 2 were made in China or Taiwan, even the top-of-the-line brand names.

    You're most likely correct that many of the bike light manufacturers haven't done the testing, either as a stand-alone chip or system. A lot of these tests can actually be done small scale in desktop sized devices, but you need software and a trained person to run it. I don't design or build LEDs, I use them, and I need to have real data. If the major players aren't publishing any data, there's no reason for anyone else to publish it as a sales tool. Maybe as consumers we need to demand it.

  13. #13
    mtbr member
    Reputation:
    Join Date
    Jul 2006
    Posts
    602
    Sounds like you guys might be the right people to ask this ?.

    I have been wondering what the color of sunlight is, 4000K?? Also, at what point does the light start to look blue? 6000K??

    Thx

    MB

  14. #14
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    Sunlight is more like 20,000K, and cloudy days can be higher in some conditions. Lightbulbs to emulate sunlight are usually 4100K or sometimes 5000K. The perception of when light starts to look blue will vary based on ambient light, colors of lit objects, time of day, light levels, etc. You can usually perceive a blue-ness somewhere above 3500K if you also see a lower color temperature source as a comparison. I know that's not a very firm answer, but at night, a 3500K source can be very blue, but in daylight it can look pretty pink. It's really hard to make an electric light source look like sunlight, partly due to the mount of light an electric source can produce, partly due to the variation in sunlight due to solar angle, atmospheric conditions, etc.

  15. #15
    just a Newbie
    Reputation: colleen c's Avatar
    Join Date
    May 2011
    Posts
    252
    I find this topic interesting. Vancbiker mentioned about optics changing the color. I have seen this on several different light. I can see greenish tint with the lights but yet, I get pure white from the Led simply by removing the reflector or the optics and shine the bare Led at a wall. I have gotten similar results from some lights that have interchangeable optics and reflector. Using an optics I get white tint but yet using a throw reflector it goes toward greenish from the same light.

    I recall there was this thing called "tint shifting" where you get a differnt color from the spill vs the hot spot. From memory, it was explained to having to do with something about the angle of incidence from the light emitted from the LED and the way it strike the phosphor. Those that comes straight out vertically will differ from those that travels horizontally thus causing a shift in tint. From what I heard, reflector design is very important to help correct this by using somenthing like an OP. This almost sound like what Androgen has experience with the Piko.

    Is this true? Can anyone explain more in detail. Thanks in advance.

  16. #16
    mtbr member
    Reputation:
    Join Date
    Feb 2012
    Posts
    263
    Hoooboy, now you're getting into the fun stuff. Odds are that a color shift due to a lens is caused by refraction through the lens. All lenses, regardless of material, have some losses (photons go in, but they don't come out, just like a Jogger Motel).

    What you are describing sounds more like chromatic aberration caused by refraction. Basically, as the light travels through a filter, some light will go straight through and not bend, while other photons will hit the material at an angle, bend, and exit at a different angle (think straw-in-glass-of-water illusion). As the light travels farther due to the bends at each surface of the material, it may shift in wavelength, changing color slightly, based on the material's index of refraction. The thickness and type of material will have an impact on the amount of shift and color created, so polycarb may act differently than glass or acrylic.

    So, in essence, the angle of incidence matters, but not necessarily in regard to phosphor, since this can happen without any phosphor. It's pretty common to see in theatrical spotlights with glass lenses. If you have eyeglasses with polished edges, you can often see it in your periphery as well.

    Phosphors are used more to intentionally shift photons to particular wavelengths. That white coating on the inside of a fluorescent tube is a phosphor, it shifts the photons from a UV wavelength to a visible one. LEDs work differently, the color is determined in the chemistry of the gate on the diode and shift the photon upon emission (LED engineers feel free to correct this).

    LED optics are also different than reflector optics, using columnators and other more exotic lensing, but the transmission of the light through any medium can create color shifts. All media will do this, the relationship of the angle of incidence, thickness, material, and such will vary. Dichroic glass is pretty remarkable with how it reacts to light at angles greater than 45 degrees, I love playing with it in architectural lighting to make wacky color splashes and such. It takes what you describe and steps it up across the entire spectrum.

  17. #17
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    Quote Originally Posted by mb323323 View Post
    Sounds like you guys might be the right people to ask this ?.

    I have been wondering what the color of sunlight is, 4000K?? Also, at what point does the light start to look blue? 6000K??

    Thx

    MB
    daylight temperature varies with weather and time of day. it is usually in the 5,000K to 10,000K range. for photography and video lighting 5,500K is the standard daylight bulb ( fluorescent or LED ). for LCD monitor calibration 6,500K is the standard.

    so basically sunlight doesn't have any temperature because the atmosphere changes it dramatially. generally speaking 5,500K will be a good value to go with for a light, unless you're looking at human skin or food - then warmer light will be more pleasant. that's why in addition to 5,500K "daylight balanced" studios also use "tungsten" bulbs at 3,200K. these "tungsten" bulbs refer to the 3,200K and could be Tungsten, Fluorescent or LED.

    i would put 3,200K ( preferably halogen ) in a residential setting and 5,500K in a warehouse or assembly line etc.

    Color temperature - Wikipedia, the free encyclopedia

  18. #18
    mtbr member
    Reputation: androgen's Avatar
    Join Date
    Apr 2005
    Posts
    902
    Quote Originally Posted by colleen c View Post
    I find this topic interesting. Vancbiker mentioned about optics changing the color. I have seen this on several different light. I can see greenish tint with the lights but yet, I get pure white from the Led simply by removing the reflector or the optics and shine the bare Led at a wall. I have gotten similar results from some lights that have interchangeable optics and reflector. Using an optics I get white tint but yet using a throw reflector it goes toward greenish from the same light.

    I recall there was this thing called "tint shifting" where you get a differnt color from the spill vs the hot spot. From memory, it was explained to having to do with something about the angle of incidence from the light emitted from the LED and the way it strike the phosphor. Those that comes straight out vertically will differ from those that travels horizontally thus causing a shift in tint. From what I heard, reflector design is very important to help correct this by using somenthing like an OP. This almost sound like what Androgen has experience with the Piko.

    Is this true? Can anyone explain more in detail. Thanks in advance.
    from the beamshots i see serfas true 1500 is probably the worst in this yellow stain department.

    i knew that a prism will separate color into a rainbow, and in fact rainbow is produced by round water droplets doing the same but for some reason i did not expect to see this on the lupine ... even though the lens is not much different in shape there from the rain drop - i just didn't expect it for some reason.

    i wonder if mirror type reflectors ( like a Fenix flashlight ) are better in this regard than lens type optics ( like Lupine ).

    unfortunately i don't have any answers on this subject ... only questions.

Members who have read this thread: 0

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

THE SITE

ABOUT MTBR

VISIT US AT

© Copyright 2019 VerticalScope Inc. All rights reserved.