Quote:
Originally Posted by Don Miller
Richard and everyone, what spectrum differences to you see between different light sources? What do we give up with high efficiency lights? Will there be variable temperature LED in the future?
|
Good question. LEDs right now are the LEAST closest to full spectrum. We have a ways to go to get daylight color small 5mm LEDs to be great without paying super big money. If you want to stay within manageable costs, 5mm LEDs in the range of 3000K to 3500K or 7000K to 8500K and above are the best. 5600K is doable with good color rendering but is best with a really wide beam angle and thus will be weaker than you would hope. Getting 5mm LEDs to play with other light sources well is a real trick though. 5mm LED panels have special uses and shouldn't be mistaken as necessarily better for all uses because they're the new thing.
It's hard to beat LEDs though when you're talking about volume of light output considered by itself. LEDs have an unfair advantage over other types of lighting that goes well beyond what lumen specification comparisons can tell you. Even lumens per watt doesn't tell the full story. On paper, comparing lumen values of various bulbs and LEDs makes it seem an LED can't do much at all. You have to get away from measuring lumens and look at foot candles or lux to get the real story with LEDs due to the fact that each LED is really a miniature fixture (with its own lens), not a bulb. It simply wouldn't be practical to build lenses into most bulbs and when they've tried, it never works out too economically (par lamps for instance--the lens is simply thrown away with the bulb when its exhausted and that's wasteful). Lumen values tell you about how bulbs are and allow for rational comparisons between bulbs, but not really how fixtures are, due to lenses and mirror configurations which magnify the performance of a bulb. And given two different fixture configurations, both using the same bulb, we know one fixture may beat the other one thanks to its mirror and lens. See how lumens were irrelevant in that comparison?
Once you put a bulb in a fixture, that's when a light meter reading is far more appropriate for comparisons. So comparing an LED to a tungsten or metal halide bulb using lumen specs is unfair to the LED since it already is a micro fixture in and of itself. Because each LED is too small to be of much value, we build them into matrices to make a really usable "macro" fixture just as we build displays out of pixels. Once you build such a fixture and compare lux values of real world manifestations of fixtures, you start to see all this more clearly. So, LEDs have the unfair advantage in light output but in color rendering and considering cost to deliver whatever light output each technology can give, fluorescent and metal halide are still way ahead for conventional use.
1w LEDs are where it starts to get interesting to try and do daylight really well, mix with other sources and keep costs down. 3w and above are fine but not as efficient in most cases as a 1w LED. We'll do some 1w projects next year and you'll like those too.
As for variable color temperature, yes I think that will be doable in the future. There's a company now that makes an LED that can switch between 3200K and 5600K but it's super expensive so unusable for practical purposes. Since we can put multiple chips on one LED, that's the key there. But, for practical purposes, I think the RGB mixture method is where we'll be stuck for a while, if you want variable color temperature choices in one fixture. Kind of a simulation of full spectrum. My idea of putting both 3200K and 5600K LEDs in alternating columns is another poor man's way to do it too, each set with its own independent dimmer so you can custom mix color temperatures. It just takes a lot more room to do that and you have a lot of wasted LEDs if you end up not using the mixture feature or one of the two sets of LEDs on the panel.
Other types such as fluorescent and metal halide (CDM and HMI) are the closest to full spectrum and are being used every day of course by many of you with great results. They all have their places and use, but remember that they are measured in "correlated color temperature" not color temperature (which means they produce a simulation of a given color temp). If you have a color temperature meter that only responds well to full spectrum sources (like real daylight and real tungsten), you can use that to tell you the story of how your fixtures are doing. If it's giving you readings that are what you expect, you know you have something that approximates a full spectrum source really well. If it is giving you other readings, differing from what was advertised as the correlated color temperature, you don't necessarily have a source with the wrong color temperature, but rather a source that's not full spectrum. It could be the right correlated color temperature, but just can't be read right by your meter.
Regardless of all that, it will still be a very usable light, given the ability to white balance using digital mediums (which most of you are using today). And that's why many of you are having such good results with high efficiency, ballast-oriented lighting because it is so easy to get rid of any biases there may be in the spectrum thanks to our very advanced cameras.