A ringflash or ring light gives a unique look, as all light is coming from the front, imparting a virtually shadowless look onto the front of your subject, while even shadows show around the edges of any three-dimensional surface. This look is possible in part because the camera lens sets in the very middle of the light source. Ring flash is often used in fashion photography and some more modern urban scene and portrait photography. It can be seen in a multitude of magazine and CD covers.
Since it seems to be a topic of discussion from time to time on the photo forums I frequent, and since I went ahead and wrote this up in a thread where a photo of mine sparked questions, I thought I'd share my experience so that others might be able to try this on their own.
Let it be known that others did this long before me and I was just following guidance from Daschund Woof (a moderator of the People Photography forum at FredMiranda.com - login required) and some others who were kind enough to explain some of the details in their own setups. I sort of implied the rest and made some modifications to suit my own needs.
Disclaimer and some other reference before you get started
This is by no means an absolute how-to, and I can see room for improvement over my design, but it works... Mostly.
Biggest gripe? I cheaped out on the bulbs, initially. 65W doesn't cut it. I've since upgraded to 120W incandescent floodlight bulbs with better results. Some people have also mentioned the use of daylight-balanced or full-spectrum bulbs, which should give better results (at 2-4x the cost of standard bulbs) with models and product shots alike, as they are closer in color temperature to natural sunlight and SHOULD result in better color separation due to a high Color Rendering Index (most often used in reference to fluorescent tubes, which are available with a high CRI - or at least marketed in part by their CRI). These bulbs don't quite cut it, as detailed below in "The Full Spectrum Question".
I recently stumbled across an in-depth discussion of full-spectrum light bulbs that I found to be very informative. In addition to this Photo Tip at edbergphoto.com, this information seems to validate the presumptions about color temp and CRI affecting improved performance when used in a photography setting. Bulbs can have greatly varying color temps and CRIs. CRIs can be compared from bulb to bulb but are only a good comparative indicator when the color temperature of the bulbs are close (CCT is an abbreviation referring to this comparison - Correlated Color Temperature). I'm resigned to the fact that a high CRI incandescent doesn't exist at a reasonable price. Stick with fluorescent or strobe if color accuracy is a high priority.
It is worth noting that color temperature is a less important consideration with digital photography than film, as color balance is easily corrected in post or accounted for when shooting. Color temp is a very important consideration when shooting film and your film should be balanced (or corrected for with filters) to the lighting you choose.
The Full Spectrum Question
I decided to drop the cash on some new "full spectrum" incandescents, utilizing neodymium (coated?) glass. I got 10 "Full Spectrum" bulbs (150W/R40/rated 1600 Lumens) and 10 standard bulbs (120W/R40/rated 1600 Lumens) and used a studio strobe as my baseline for color rendering. I wanted to see if these "full spectrum" bulbs were better than standard bulbs at rendering color properly.
I printed a test sheet and shot it with my Alien Bees 400 strobe and softbox to act as a baseline to compare the bulbs. The source images were shot in RAW with a custom white balance (set by a Kodak gray card), composited, adjusted the overall brightness to be equal (as equal as possible) between the three images (using highlight arrow only in Levels). I sampled the colors from each bar (eyedropper tool set to "5x5 average") to make up data points of Red Green and Blue intensities and then graphed those readings.
The source image can be found here.
What I discovered is that, essentially, with an accurately-set custom white balance, the "Full Spectrum" bulbs are not noticably better at rendering accurate color than the more modestly priced standard bulbs. At the same time, both of these incandescents fall far short of a strobe for most of the visible spectrum, which comes as no surprise.
The chart of graphed data points can be found here. I can provide the measured data if anyone's interested in re-charting this.
I'm not a lighting engineer, nor do I have a scientific background, however I do believe that this relatively crude experiment tells the story I was out to discover. If I've gone about this the wrong way and you know of a method to more accurately portray these bulbs' characteristics, please do let me know. As well, if you ARE a lighting engineer and can confirm my results, please take the time to tell me.
Other Full Spectrum Bulb Issues
- They're more expensive - usually about $10-15/each. Not worth it in this regard.
- They're not as bright - actual light output didn't come anywhere near the 120W bulbs, resulting in much longer exposures.
- They are a different, purple color - just an observation here. It may make them more comfortable for general use, but that doesn't benefit us in photography.
- Compact Fluorescent (CF) bulbs - with a potential for high CRI (as is often available with standard tube fluorescents) CF bulbs seem an attractive alternative. However, their high cost in a floodlight-style fixture ($15-$30) is prohibitive for such a budget-minded solution as this ringlight. Additionally, the output ratings of CF bulbs is relatively low (600-800 lumens, max), making for very little light output compared to our high-output 120W incandescents (1600 lumens).
- Color Temp was not considerably different. Color was. The readings from ACR:
Strobe Temp: 6550K / Tint: -13
Full Spectrum Temp: 2700K / Tint: -7
Standard Bulb Temp: 2750K / Tint: +4
Regardless of which bulbs you choose, more light is good. You'll never approach the flexibility and control you can get with a traditional ringflash, but if you don't mind wide apertures and possibly higher ISO settings (400-800), not to mention the cost of a pro ringflash and associated power pack, this home-built setup can provide some interesting and fun lighting options.
The reason you'll never get the sort of power and control with a constant-source lighting setup compared to a strobe-based setup is explained very well at this site: http://webs.lanset.com/rcochran/flash/hotorstrobe.html
Building The Ring Light
Basically, you're looking at a ring of wood (plywood in this case) with 10 (Dasch did 12) heat-resistant, surface-mount light sockets (molded plastic or porcelain) positioned evenly around the ring and wired together (all + and all - run off one wire each) to a standard light switch (also mounted to the rig). In my case, the wires are also surface-mount, so it's easy to wire up, but leaves exposed connections. A can of brush-on insulating coating (a few coats) does the trick to prevent accidental electric shock. I used a product called Plasti-Dip (typically used to dip your tool handles for grip, etc).
Although you could build it a bit larger - in the 36" outer diameter range for 12 bulbs - the outer diameter of mine is 28". Inner diameter is 22". I also hinged it at two places with spring hinges (the springs hold it open) so it folds up and takes up less space for travel or storage. Two copper pipe brackets serve as mounting points on the rear, which hook over two light stands heads. I'm going to investigate a better mount - preferably one that uses standard light stand mounts.
It's all wired to an extension cord (inside the switch box). This might better work as an outlet or short pig-tail, so your cord isn't permanently attached (a lot to lug around and wind up in my case - 25ft). It's cheaper to buy an extension cord and chop the end off - not to mention easier - than to build your own cord/plug from individual parts. You'll want to use about 14 gauge wire (i think - ask the experts) on the rig itself and something heavy (like an extension cord, as I did) from the rig to the wall. Fire safety and all that.
I was a little leary of spending too much on bulbs initially, so as I stated above, I went with inexpensive 65W bulbs... Which isn't great for exposure or distance to model. I've upgraded to 120W standard-type floodlights - not the more expensive halogen/par38 style, just incandescent bulbs. I tried 75W halogens (par38), which put out a ton of light, but even as "floodlight" bulbs, they really focus the light forward, not offering much of an even spread for full-body shots, even at moderate distances. They tend to blind your subject, as well.
At moderate distances (<1 meter), you get about 1/125 sec, f/4 at ISO100.
- All costs estimates (I don't have the receipts handy)
- (2) 2'x4' 1/2" plywood sheets: $20
- (10) 120W incandescent floodlights: $23 (special sale)
- (10) surface-mount sockets: $32
- 8' 14-2 wire: $4
- Misc electrical parts, switch, box, etc: $10
- Plastic-Dip: $5
- 25' extension cord: $10
- TOTAL: $104
I'm sure this could be done for less. I could have gotten scrap wood and gone for a less-expensive type of socket and ordered bulbs and other parts online. I'm certain you could get this done for well under $100. I'm just not particularly good at executing something like this as an experiment AND saving money. So there it is.
If you've found this useful, interesting, or the least bit inspiring, I'd love to hear from you. Drop me a line here.
Thanks for stopping by!