A new flashlight beam metrology

Flashlights! Ah, what a wonderful thing. A small portable object that emmits light! How enchanting!

However, capturing and quantifying the results of this light output so that it can be compared in some way with the light output of other flashlights is a very difficult task. Light is what we use to see things, or to take pictures. To turn around and try to take a snapshot of the light itself is fraught with difficulties.

I was very interested in the software mentioned on the review forum whereby a person can composite a bunch of white-wall beam shots into a single beam profile. I was also interested in coming up with some scheme whereby a lightmeter is used to measure the center of the hot spot, 50 percent out, 100 percent out (at the edge), and finally at the corona to get a spill measurement.

And I still am interested in both these things, but neither of them alone promises the sort of result I want. Because even given that we can capture what the eye really sees with a white-wall beam shot, that is only one part of what the beam is. And even if we can get a good spread of lux measurements, that also is only one piece of desribing what the beam is--much like an integrating sphere measurement of total lumens output is only one piece of information about the beam.

There is so much more to a beam. How it throws, how it renders, how "bright" it is, and so on.

And one day I had a thought. We all know that you get better pictures from your camera if you have lots and lots of light on the subject, right? So instead of standing back where the flashlight is and trying to capture what the beam looks like to the user--which we're never ever successful at--why not separate the camera and the flashlight? Why not have the camera up fairly close to the object being illuminated, be it a tree or a mail box or a shed, and have the flaslight 10, 20, 50, 100, etc, feet away? The brighter the beam and the better it throws, or the closer it is, the better the picture, right? And presumably, one could aim the light off of the object by varying degrees to capture the spill light. Why fight against what the camera wants to do? It doesn't want to capture light itself; it wants to USE light to capture an OBJECT and the more light the better you capture the object. Right?

I wanted to use this technique in my review of the A2 and L2, but I haven't gotten around to getting outside with the camera and tripods and flashlights to see what kind of results I will get. But I will do so sometime in the next couple weeks. It may be that this is a bad idea and that it won't work out very well. So be it. In any case, for the time being, I wanted to bring it up for discussion.

Because, as Ginseng said on the phone to me the very first time we talked, if we're serious about our hobby and we want to do justice to the lights we examine, we simply have to do better than a single lux measurement at 1 meter. Heck, even here it should be noted that you can move the light meter to varying distances and use the inverse square law and arrive at different lux @ 1 meter numbers. A flashlight beam is NOT a single number. And don't even get me started on how almost all flashlight companies (SureFire being the gleaming exception) exaggerate the numbers and combine them in dishonest ways--for example by quoting runtime to near absolute battery depletion with like 10 percent of starting lux, together with the lux or lumens measurement taken with super fresh hot off the charger batteries. It's lying to one degree or another. But I'll cut myself off here.

So anyway, what do people think? Does this idea have any merrit?

Hello Jim,

I am with you on this one. I use my lights to illuminate things that I am looking at. Sometimes even white walls... /ubbthreads/images/graemlins/smile.gif

The camera view would reveal how the artifacts in the beam effect the object we are looking at, the color rendition, and the amount of detail we can make out.

One difficulty may be getting a camera to record with 1 lux of illumination, but I will leave the details to you... /ubbthreads/images/graemlins/smile.gif

Tom

Interesting idea Jim, I look forward to your test results. Btw, do you have a link to the thread in the review forum?

Of course I have some ideas myself on how more useful measurements can be done.
One of the methods I have been thinking about is for estimating maximum usable throw of a light. Have a dark brown* object set against a completely black background (both should be matt), then you need a reference light that is much brighter than the test light. Move away from the test object and when you notice that it becomes difficult to see the object with the test light but still easy with the reference light, then you have found the usable throw of the test light. You can then measure your distance to the test object.

*dark brown chosen to test red spectrum which many lights are weak at.

Another method is more scientific and is what I would have liked to see from the manufacturers themselves. By measuring the beam at different points you can divide it into zones where you can state an average lux for that zone and its dimensions. I'm not sure of what the exact intensities the zones should be divided into, and different lights can have different number of zones. It is also suitable to use different number of zones when the beam is described at different distances. An example of zones would be to have one zone for 90-100% of max lux, the next zone for 10-89% of max lux and the last zone for 2-9% of max lux.

I made the following illustration to show an example of an incan light measured this way:



To this illustration I have also added an example of runtime description that they could use. The runtime is divided into 4 zones, e.g. "75% 25 minutes" means that the light in that zone will have an average of 75% of the stated lux numbers and that the zone last for 25 minutes.

Sigbjoern

"The brighter the beam and the better it throws, or the closer it is, the better the picture, right?"

But only if the camera settings are standardized. Otherwise, I can make an L1 look much like a Beast. /ubbthreads/images/graemlins/wink.gif

Hello Jim,

You may also want to consider getting a color chart that is used in photography. The one that I have was included in my Kodak Professional Photoguide handbook.

I like Sigbjoern's beam profile idea. It can also be graphed with dimensions showing the diameter of the various beam intensities.

Tom

Hello Jim,

I was thinking along the lines of graphs like this.

I think this gives a good idea of how the beam behaves.

Tom

I agree all the way, but a beam is a beam, it depends on
what is driving the beam ie: reflector and light engine

Why not just compare beams side by side with different shudder speeds.

[ QUOTE ]
HWilliam said:
"The brighter the beam and the better it throws, or the closer it is, the better the picture, right?"

But only if the camera settings are standardized. Otherwise, I can make an L1 look much like a Beast. /ubbthreads/images/graemlins/wink.gif

[/ QUOTE ]

Well, yes and no. The problem is that when you go outside a small range of intensities, the camera has problems. I mean, haven't you noticed that no matter what the light the hotspot looks highly intense, and the rest of the majority of the field is dark, dark, dark?

I would want things adjusted so that a nice even light were falling on the object, and then I would set the camera up to take the very best picture I could. Then I would maybe (thinking out loud here) adjust the distance of the other light until I could also get a good picture. Or maybe I would also adjust the camera settings.

Because maybe the quality of the picture would tell you about the beam? Instead of trying to compare brightnesses of the beams, you compare picture quality of the objects.

I know this sounds dubious and radical, but think about it! The "beam shots" I've seen, using locked exposure, standard settings, haven't really told me much about which light is brighter, but only which hot spot is larger. It's very disconcerting, but there it is: the camera simply WILL NOT CAPTURE the "brightness" of the beam.

But I noticed that when I do outdoor beamshots with my super-incans, and let the camera just do its thing in "night mode" (it adjusts the shutter time until it gets enough light) that I get good results. Why? Well, because the beam is large enough and the spill light is massive enough, that the camera doesn't have to struggle with large intensity differences. I worried about the settings being different from light to light, and when I did my MC60 vs. Tiger11 review I had them both on side-by-side in the same shot, so it was guaranteed to be the same (cause it was the same picture!). But before that, when I took the shots for the TL upgrades offerings summary thread, I just looked at the results afterwards and said "hmm. That's pretty close to reality! I'll just go with it."

Because--and this is what I'm saying--THE CAMERA SUCKS AT CAPTURING LIGHT ITSELF. It want's to use light to capture objects, backgrounds, etc. So why not try to let it do its thing and see if we can't find a way to make comparisons possible?

[ QUOTE ]
SilverFox said:
Hello Jim,

I was thinking along the lines of graphs like this.

I think this gives a good idea of how the beam behaves.

Tom

[/ QUOTE ]

Tom, eeewwww. Yuck. Doesn't do it for me. That gave me no real sense of brightness, although, yes, beam profile is well represented. But a white-wall shot does that OK too.

One thing that always concerned me is that not all white walls are equal. The type of paint, surface treatment, etc could reflect varying amounts of light.

One idea to solve this issue would be to design a poster that could be used by everyone who is interested in reviewing and anyone who wants to compare their own lights.

With a print run of 300 or so the cost should fall lower than shipping costs. Then at least one part of light reading process could be standardized.

It seem the only true way is a completely manual setting. Pick one. And post what the settings were. If trying to meter for different lights then you are adding a variable to the results that doesn't allow for equal comparison. So say a white wall, and all pictures are taken at f5.6 with a speed of 125 on iso 100.

Bravo25 said:
It seem the only true way is a completely manual setting. Pick one. And post what the settings were. If trying to meter for different lights then you are adding a variable to the results that doesn't allow for equal comparison. So say a white wall, and all pictures are taken at f5.6 with a speed of 125 on iso 100.

Click to expand...

Good idea at first glance, and it is what is standard practice, but what I'm saying is that it doesn't work! The hotspot saturates the camera and all of them seem about the same, and the spill light/corona isn't enough and so they also seem about the same: dark. So the only thing you end up with is the relative sizes of the hotspots. Whereas if you have the right light to get a good picture, then even though the results may be more "sujective" or less scientific, they are, in truth, more faithful to the reality of the situation. Consider these shots of the Tiger85 vs. the stock TigerLight:





Now these tell you something! A comparison of hotspots against a white wall would also tell you something, and I am not trying to say we shouldn't do white wall shots. I'm trying to say that they are not useful when trying to gauge which light is brighter. They tell you about beam profile, and actually, only about the hotspot outline, and little more. You have to be LOOKING AT SOMETHING, and be set up to take a decent picture, in order to give people an idea of how bright a flashlight is.

I would agree with that but the same principle applies. If you make the metering subjective you add an unexplainable variable. Even if it be outside shots with distance markers or a blank wall. An open fiels with signs posted at 50 foot intervals would be a good idea as well.

Perhaps another way to look at this is the challenge we have had in the post originated by mobile 1 for the GBWR brightest light. The challenge presented by both threads is the same - metrology and defining the goals.

The measurement methods useful and important for an EDC size light which has 1 meter lux measurements between 100 - 2,000 are completely different than a high powered incan build with nominally 10 - 100x the effective lighting levels. (please forgive my poor use of the nomenclature).

I have taken Lux meter scans at 1 meter over a 2 meter X scan in front of various lights, including a SF L4, various Lux III lights, and my own 12.5mm dia. optical system. The graphs look nominally like the ones in Silverfox's link.

I can say from experience that standing in front of an L4, which is really not a high lux (but is bright) beam while conducting an X scan acrossed the diameter of the beam is about the limit of my optical pain level.

After considerable discussion in the other thread, there was a "trend" toward classifying / measuring lights based on "weight classes", which roughly converted into "output classes" if you also agreed on the one hour run time target.

The "trend direction" was that a high power light needs to be measured at something approaching 100 - 1,000 meters onto a target nominally 3 meters in diameter, while an EDC might be better measured at 10 meters. Part of this comes from the shear challenge / physical pain level of measuring a beam from one of the incan modders. /ubbthreads/images/graemlins/smile.gif

While beam uniformity is of course valuable, at 1,000 meters, it does not see quite as important as a low power light for reading in a tent - at least to me.

The color rendering question and its impact on "vision" and "recognition time" is something I have crudely experimented with using 2 and 3 color mixing. NIST and others have done much more extensive work in this area, and even they are still "discussing" how to take these measurements.

One idea is to use a spectrophotometer to scan the intensity of wavelength distribution of the light reflecting off of the target. Firms such as Ocean Optic (and others) make products which would be quite useful for our work, in the $ 3- 5 K range. That seems like a lot of money, but perhaps if we jointly funded such a purchase and rented out the meter to users, it might be a viable purchase.

I always wondered if one could simply take a telescope clock drive (one that moves the scope slowly so you can keep the same star in view as the earth rotates) and attach the light. Mount your lightmeter a meter away or whatever.
You can calculate a pseudo integration as the clock drive slowly turns the light's hotspot away from the meter. The main problem will be for lights that are not regulated and dim significantly over the course of an hour or so. Maybe a turntable that turns at so many degrees per minute (a minute on a clock hand moves 6 per minute)
repeat the operation with the light turned two or three different directions axially so you can average out any imperfections in the beam.
Example: standard 2d maglite, do first "integration" with the clicky button up. do second with clicky button at say the 4 o'clock position from vertical, and finally maybe with it at the 8 o'clock position. This would be enough info to calculate a rough total lumens and also some concept of relative brightness of hotspot vs corona output. Lights with wildly asymetric beams wouldn't be accurate, but most average beams would probably be pretty close. I suppose the biggest problem is mounting the whole mess in a flat black painted enclosure so that you dont fudge the reading by entering and exiting the room.
If the lightmeter could get several readings per second, the rotation speed could be relatively fast (one or two degrees per second) and the test could be completed in a minute or so.

Maybe, one day, we'll have practical holographic cameras with virtual displays. That would open up the flow of answers. I can imagine reaching out in the virtual image and turning the beam this way and that. Just point your finger at a particular location in the beam/picture and you get a reading display for the intensity and color. Cup your hands around part of the beam and get a lumen reading. Now that's what I want.

Until then, I'm keeping an eye on this thread!

Photos of beam shots are informative in a relative way provided there are some constants in the photos and these constants are identified. A lux measure at one meter is also informative, to some extent. Measurements and photos can both be an aid but one is difficult to infer from the other. IE. you can't look at a photo and "guess" the lux of the light used nor can you be told the lux of a light and know what a photo should look like or pick beam shot photos that would match the stated lux.

I have a suggestion for a "standard" in beam quantification that would allow various members to provide data that could be compared with others. Some quick empirical study would be required to come up with a standardized lux measurement value. My proposal is that a light is positioned at a known position with its beam parallel to the ground and on an identified projection axis or identified line. This line is marked off with even spaced distances. One traverses perpindicular to the beam axis out from one of these measured distances until a hand held lux meter (same elevation as the light source) reads the "standard lux measurement", say 100 lux for example. The distance off axis is then measured. If this is repeated until the 100 lux is measured on axis, in the center of the beam, one now has a 1/2 angle map of the beam intensity. Below is a graph showing two such examples but they are just examples:



Now this map provides a grid showing the beam surface from which a constant lux measure has been read. To get an idea of a beam surface that represents a constant level of return light to the position of the light user, one would need to adjust the data based on the 1/2 square rule and target some other standardized number, perhaps 20 lux? The new map would give you a 1/2 angle curve showing you points from which you would receive the same level of return light.

For a good map, one would need to factor out ambient or use one of the meters that allows for an offset. If we are willing to take the measured map and then re-map it to the one showing calculations based on a constant of light returned to the source, the measued map would not have to use the same distances and measured lux constant. Realistically, using the sampe parameters for getting a map for an Arc AAA and a maxibeam would be a joke. However both lights ultimately have a half angle map that would show the curve of points from which 20 lux could be measured on return, at the source.

I am not sure I have explained this idea correctly but if one were to take the "adjusted" and final map and then plot that curve out, in the real world on say a parking lot and then run a roll of butcher paper, perpindicular to the ground along this curve, when the light were positioned properly and turned on, from the viewing position at the light source, the butcher paper banner would appear to be the same level of brightness and other butcher banners placed for other light sources would also appear to have the same brightnes levels. A photograph taken from the point of light source would show a homogenous white band.

I'm no a digital camera person (yet), but with years and years of photography behind me, shots like the ones needed for light comparisons need a variety of elements.

Meters and spheres for measurement are just that, measurements and data. That is useful but can be dry.

As has been mentioned, standard color charts are nice to have and need not be expensive but must be used consistantly. Just keep them at the same distance in the same location in the shots.

Bracket your exposures. Using film this can be expensive, but for digital it won't cost a thing, since you can delete the ones that don't make it. Start way underexposed and go to way overexposed. Try it before you knock it. You do need to have a lot more variation than you would think. And be sure to note your exposures and how far over or under they are.

Use a background with a grid if you'd like, but if not, use a background with a lot of contrast in both density and texture. That will give you depth perception and a comparison from light to light. Maybe see if you can find a driving range with distance markers and a few trees or shrubs. Just be sure to ask permission first and expect the security and/or cops to check up on you.

If you only photograph the item you're illuminating, and not the beam with the light next to the camera, you have to carefully measure the distance to have a quantative entry, i.e., "thus and such object illuminated by this light at X distance." Otherwise it will look OK but offer no means of comparison.

And always remember photography rule #1 - Pretty girls in pretty clothes always spruce up any picture. Always try to include pictures with pretty women to be shown here. My wife disagrees with me on this, but I say, let me be the judge of that.

Hi Don - I think you are on the right general track. The measurements I took are just a different way to obtain the same info, and my plots look more or less like yours.

I am not sure if it is a factor or not, but I took (some) care to try to avoid having reflected light come back from walls, etc back into the meter for fear of it messing up the data. I did this by keeping the potential reflecting surfaces all at least 1 meter away from the light and Lux meter.

The method you offer has the nice advantage that it takes up relatively little space and might add some consistency to the metrology side if we carefully define the paper and a few Lux target levels.

The limitation of this method is that it might be hypersensitive to flashlight angle relative to the ground. Maybe you have a work around for this ?

I also ran into a resolution problem with the Lux meter even having it pointed at the ceiling. I even tried putting the meter in a box with a slit to narrow down the acceptance angle, but the results were not substantially better - still not sure why.