Angus from Alberta

Before someone asks, Alberta is a province in western Canada.

In late November a friend came over and demonstrated his new Garrison AA flashlight. Suitably impressed, I dropped into Canadian Tire the next day and bought one for $17. A week later I dropped it somewhere between the garbage can and my back door, a distance of 10 steps. Later that night we had a serious snowfall and it's lost until spring. If it still works then, I'll write a report.

That's the time I discovered this forum and began to learn more about led flashlights and where to find a good replacement. Your enthusiastic reviewers persuaded me to buy a few items, most of which, unfortunately, still haven't arrived.

Although I've learned a great deal one thing still puzzles me. Why the enthusiasm, both in this forum and most others, for measuring light output in lumens as opposed to lux at a given distance? For someone who doesn't have a good reference point, like myslf, stating that flashlight A has an output of 100 lumens is quite meaningless. What do you compare it to? On the other hand, if a review stated that the hotspot was 100 lux at 1 metre and wikipedia tells me that "100 lux = Very dark overcast day", I'd have a very good idea of what they were talking about.

Cheers, Angus

Welcome to BLF Angus! Enjoy your time here.

Mainly because lumens indicate total output. Lux at X metres usually means the point of highest intensity in the beam. Some idea of the spatial distribution of the beam is needed with lux.

Measuring lux across the beam would be harder to do but probably more useful.

One of my lights will produce a tiny, very high intensity spot of 36500 lux at I metre but is actually quite low in output. Lumens give you a better idea of total output, but not necessarily brightness at any given point. 1 lux can be thought of as one lumen per square metre. 1 lux is also about as low as an eye can tell anything much.

Essentially lumens are measuring lux inside a standardised "room". Both measures have their drawbacks - which is why there are comparative beamshots so you get some idea of the spatial distribution of the output.

If all lights had the same distribution of output over the beam, then lux would be a more useful measure and far easier to measure. But they don't.

Not sure that I haven't just succeeded in confusing myself.

That's one advantage of being a fan of budget lights: You are delighted about them at least three times: 1. when you click the order button full in joyful anticipation, 2. when it eventually arrives long after you forgot you ordered it and 3. when it finally works!

and welcome to BLF!

Hi and welcome to BLF! : )

Lumen is to Lux what horsepower is to torque, it is a measure of sheer output and it is one point of comparison between lights. It does not tell the whole story, but it is a useful point of reference when making comparisons between lights in the same class.

A fellow Canadian! Welcome to blf:)

Those Garrison/Noma Canadian Tire flashlights are re-branded Brinkmann's - they're actually not bad lights at all. The 140 lumen one is suitably impressive, even if it takes 3xAAA.

Hello from the West Coast.

Hello and welcome.

To top the cake to Don's exuisite explanation.

Try to see it this way:

Lumens are what defines the whole light output and what you generally see. Luxes on the other hand cannot tell you anything about total brightness ,just a defined spot numbaer at some distance. Often used to compare emitters and generally used to measure thrower types of flashlights as for those luxes are more important than lumens. It's not entirely correct but for starters will suffice. To experts luxes give a rough idea what to expect (beam type) for a certain lumen flashlight. Again not entirely correct but heh... not gonna go into a territory when others have better knowledge and can tell you 1st hand the right info.

More often than not a simple zoomable flashlight rated at 300 lumens (when in reality is more like 160lm) appear brighter than a true 300+ lumen flashlight when zoomed out. At least to novices.

Anyway, you'll get a hang of it very soon when you get some flashlight to play with. If youre serious about the hobby get into li-ion territory not just AA's, it's a whole new dimension.

Here's one way of comparing. This is an animated GIF of the lights I did beamshots of last night. There are 12 of them varying from 1500-200ish lumens.

What' wrong with J4 XM-L? Was it off?

Running on NiMH as it won't work with a 14500 just now. Exposure was lower than usual as there was a near full moon (It'll be bad at work tomorrow) and was really aimed at the more XM-L lights.

The J4 will give a little less than the LT3 on max when (if) it'll work with a 14500.

Quoting Don:

"Some idea of the spatial distribution of the beam is needed with lux. Measuring lux across the beam would be harder to do but probably more useful."

Totally agree on both counts. It's actually pretty simple with ImageJ, a free opensource program that's used for image analysis. It's in java, so works on pretty much any platform. I'll try to upload some images to let you see the possibilities. Be patient while I take some pictures and figure out how to link them and so on. Cheers, Angus

First the wall shot. The light is one of the $1.50 ebay single AA units with a fresh alkaline at a distace of 25 cm from the wall. The camera was at F/4.0, 1/60 s, ISO 100, spot metering, 24 bit jpeg image.

The second image is the same except with contour lines plotted at greyscale intervals of 200, 100, 50, 25, 12, 6 and 3. If I had had a light meter I could have calibrated the contours in units of lux. Also it is easy to scale the image in cm or m or inches.

The last image is a plot of the grey scale values along a horizontal line through the hot spot. Again the plot could have had lux on the y-axis and cm or m along the x-axis. From either of these plots the beam angle corresponding to the 50% contour can be calculated.

I'll write up some detailed instructions and post a link to them here in a day or two. Also since ImageJ supports 16 bit tiff images it is possible to use a camera's RAW output and get much better than 8 bit depth. In the case of my Canon point and shoot using the chdk firmware hack, 10 bit resolution ( 1024 greyscale levels) can be utilised.

I know this all looks very sketchy right now. In the end it will result in very useful information with very little effort.

Cheers, Angus