Updated: Jun 1
What the manufacturers don't tell you
You've just purchased a 2000 lumen light. But will it stay at 2000 lumens? No! Possibly not even for 30 seconds!
LEDs create heat, lots of heat! The more current running through an LED, then more heat is produced. LEDs are more inefficient when they are pushed at higher currents. Five years ago, most 18650 battery based flashlights/torches for light painters could output approximately 1000 lumens, which is sufficient for most light painting scenes. Timer or temperature based brightness step-downs (dimming) would usually take 3-4 minutes to occur, and some lights such as the Convoy S2+ didn't even step-down in brightness and just got very hot. Thus it was possible to create light trails over and over again, and not worry too much about the light dimming half way through your light painting scene. This is still the case for most 18650/21700 flashlights up to around 1,200 lumens.
Lumens sell lights and the same way that megapixels sell cameras, and in the last few years the lumens war between flashlight manufacturers has become quite silly. Now it is possible to find many 18650 or 21700 based flashlights that can fit light painting connectors, with advertised lumen ratings of between 2000 to 4000 lumens. Just one big problem, these lights cannot handle the heat created for anything other than very short periods of time. An example is the 2,100 lumen Klarus XT2CR Pro. Based on the excellent review by one of my favourite review sites 1Lumen, the XT2CR Pro can only handle maximum lumens for 40 seconds, and reduces to just 500 lumens by 70 seconds. Once the light has warmed up, it won't return to full brightness either until it has cooled down. Thus you might create a light trail for one photo, then in the next attempt at creating a light trail, find that the light is at only a quarter of the brightness than in your previous photo.
The XT2CR Pro is a typical example, some lights are better than others due to better thermal design, more efficient LED drivers, or firmware, but no lights can defy the laws of physics. Some of the worst examples of heavy brightness step-downs are the Lumintop FW series lights, of which some models claim approximately 3000 lumens, but can noticeably step down in brightness on Turbo setting within 15-25 seconds, ending up at just 250 lumens.
The same goes for high powered lights that you might use for illumination. For example the popular Fenix LR35R is advertised at 10,000 lumens. However, it only manages this for 30 seconds, at which point it reduces to 1,500 lumens. Now, it can sustain this 1,500 lumens for the rest of its runtime, which is very impressive, but sadly sustained lumen ratings don't sell flashlights to the general public. If you set this light up for a backlit photo on full power, then by the time you come to actually take the photo, the brightness will be at 15% of the original brightness. Thankfully this light will still be bright enough enough at 15% output, but many other competing flashlights will simply not be bright enough for backlighting after a few minutes.
Are more lumens always better?
Lumens is the measurement of total luminous flux. However, it doesn't always create a useful picture of how useful that light is. Peak beam intensity is the brightness of the centre of the beam, and thus can be represented in candela, lux/distance, or throw (in metres). Peak beam intensity is affected by lumens, LED size/type, reflector/optics design. A "floody" light will spread out that light, and have low peak beam intensity. This may be great for illumination purposes, but generally not so good for use with light painting tools (maybe fiber optic brushes). If you have a "throwy" light, then for the same lumens, it will have a higher peak beam intensity. This will be much better at illuminating long narrow tools such as light sabers, swords, rods, and tubes. Due to the difference in lumens vs peak beam intensity, it is actually possible for a throwy light with less lumens to be better at illuminating longer light painting tools than floody light with more lumens. In fact one of my best lights for illuminating longer tools is just 750 lumens (Convoy S2+ CSLNM1.TG)! It seems that the candela sweet spot for use with light painting tools is around 15 to 30Kcd irrespective of the lumens. If you use strobe mode, take note of the strobe mode's brightness, it isn't always at maximum output (notably on lights with Anduril UI).
Some light painting systems such as Light Painting Brushes and Light Painting Paradise have a limited aperture for the light to pass through (25mm and 20mm respectively). Thus the larger the head size, a lower percentage of the lumens will actually enter the tool. If there are multiple emitters, instead of a single on-axis emitter, then again a lower percentage of the lumens will enter the tool. T8 Tube based light painting systems such as Light Painting Tubes and Luminosify don't have this issue as the light fits inside the tube, however the tube itself limits the maximum head size of the flashlight, which is a limiting factor in heat handling and peak beam intensity. There is no perfect solution, all options have compromises!
Also, beware of some flashlights that have a proximity sensor, yes I'm looking at you Olight. Placing a light with a proximity sensor in a light painting connector may reduce the brightness down to just 200 lumens - not good if you purchased a "4,000 lumen" light.
There are plenty of torch/flashlight features that need to be considered for light painting which in many cases are far more important than maximum lumens. These include a range of well spaced brightness levels (as you don't always want to use maximum brightness), constant frequency strobes, last mode memory for any brightness level and strobe, and accessible switches.
As previously mentioned, lumens sells flashlights. Most reputable flashlight manufacturers such as Nitecore, Fenix, Olight, Acebeam, Noctigon, and Emisar are fairly honest about their maximum output, though there is often a lack of clarity around brightness step-down time. Supposedly reputable manufacturers such as Lume Cube have been caught out making over-exaggerations - I tested the LumeCube 2.0 at 650 lumens instead of the claimed 1500 lumens. Ledlenser also have poor transparency around step-down times, with one light only lasting at its advertised brightness for 10 seconds. Then there are manufacturers such as Convoy who either state theoretical lumens (i.e. excluding lumen loss in the optics), or just state the current instead.
Sadly, there are lots of flashlights available from Amazon, eBay, and various online retailers with exaggerated lumens from mild to crazy. The most over-exaggerated (as of November 2021) being 99.9 trillion lumens!!!
The Lux Confusion
It isn't just lumens where manufacturers are less than honest. For peak beam intensity, flashlights usually use candela or throw. However, for more floody video lights and LED panels, lux is more commonly used. Unfortunately, lux is stated at a distance from the light source, and that distance isn't fixed. So one manufacturer may claim 1,500 lux at 1m, another manufacturer may claim 1,500 lux at 50cm. The former light will be 4 times brighter! There are also a lot of dubious lux claims by some LED panel lights. Another thing to consider is that lux claims are usually for white light output (most likely with warm and cool white emitters all on) and not for the colour RGB emitters. Even RGB emitter brightness can be all over the place - for example one product may have a brighter blue, another might have a brighter red. Unlike flashlights, where there are a many reputable reviewers, decent reviewers of LED panel lights are far and few between. Impartial reviews with lux measurements I have found useful are New Layer YouTube and Website, and Tech Notice.
Maximum lumens may be a selling point for flashlights, but you may also need to be aware of how long the flashlight will last at maximum brightness, and maximum sustained brightness. There are multiple other factor that may affect how many of those lumens actually end up in the light painting tool. There are plenty of torch/flashlight features that need to be considered for light painting which in many cases are far more important than maximum lumens. These include a range of well spaced brightness levels, constant frequency strobes, last mode memory for any brightness level and strobe, and accessible switches. I try to include this information in my flashlight/torch buying guide for light painting photographers. I would highly recommend reading high quality flashlight reviews such as 1Lumen, ZeroAir, and of course this website! Unless it is a new light and not reviewed yet, I would personally avoid any lights with no independent reviews.