tips for photographing conference speakers
This page is a copy of Igal's tips for photographing conference speakers from the Open Source Bridge wiki.
Igal Koshevoy put together the following tips for conference photography, specifically focusing on capturing speakers.
I'd like to write a bit about what goes into taking and processing these photos in hopes that it will help others. I'm not an expert, but do take this fairly seriously and have learned a few things over the years. If you have other ideas, I'd like to hear them.
Thou shalt not post crappy pictures of people onto the Internet
People hate having bad photos of them posted to the interwebs. Instead, you should take many photos and only post the best of those that look good. I posted maybe 1/10th of the photos I took because I didn't want to post the countless pictures of the speakers looking ridiculous. Getting a single good photo of an well-caffeinated speaker can take dozens of tries, but can be very rewarding when you succeed.
Follow the speaker's pace
You can't just shoot a bunch of photos in continuous mode and hope to get a nice photo. Instead, you must figure out the pattern of the presenter's speaking style so you can predict when they'll do something like smile and pause for a moment, which gives you maybe a quarter of a second to press the shutter and get a good photo. When that moment comes and your camera isn't already fully dialed in, photo isn't focused and composed, or you're not ready, you'll miss that fraction of a second needed to get the shot.
Calibrate for each room
You are unlikely to get good photos in a room you've just stepped into because your camera will typically not be calibrated for the conditions. Instead, you should start off by shooting a photo you don't intend to keep just to calibrate your camera and only then begin trying to take photos you intend to keep.
Photograph the room sign before photographing the presenter
You won't post this reference photo of the sign, but it'll help you title your photos with the name of the speaker and talk, and later aid you in identifying this person outside of their talks.
Spot Exposure and Exposure Lock
You can't expect your camera's automatic settings to figure out how to properly expose a difficult scene with a blinding window, black curtains, bright projector, dim presenter, pale skin, dark shirt, etc. Instead, you need to turn on Spot Exposure to get pinpoint exposure measurement and use Exposure Lock to fixate it on the one point in the photo that's the midpoint for your exposure, which is typically the presenter's face, but it can also be the pants or shirt if these are a neutral value since these are a bigger target. You'll then use this same locked exposure between photos, rather than setting it anew each time. Exposure Compensation or Manual mode can be used as well, but typically take much longer to adjust.
AF-C, Spot Focus and prefocus
You can't expect your camera to figure out which part of a cluttered scene to focus on when there's a speaker bouncing around and waving their hands, their desk is covered with stuff and so on. You also won't get a good photo by smashing down down on the shutter in one motion because that shakes the camera, takes a fraction of a second too long, and is unlikely to properly focus on the speaker's face. Instead, you should turn on AF-C (Auto Focus Continuous) so that the focus will follow the subject that it's locked on, e.g. a speaker's face. Next set Spot Focus to get pinpoint focusing at the very center of your viewfinder. Next point the center focusing mark at the speaker's face and half-depress the shutter to lock the focus, recompose the shot, hold it there until the speaker's presentation pattern suggests that they'll do something worth photographing, and only then fully-depress the shutter. You'll have to repeat this every single time, which gets tedious. If you have great lighting, which I never have indoors, then you can use the "zone focusing" trick from the manual focus street photography hipsters where you set a small aperture (high f-number) to increase depth of field and prefocus on a wide range that covers all of the speaker and then some so you don't need to refocus, rather than trying to keep focusing on the narrow sliver just featuring their face.
Shoot RAW, not JPEG
This increases the chance of you getting a usable photo in tricky lighting by 3-10x because the camera saves more data for use during processing.
Fast, stabilized, limited-zoom lens
I used a Sigma 17-50mm f/2.8 EX DC OS HSM. This big lens offers a f/2.8 constant aperture over its entire zoom range, so although it has much less zoom range than a cheaper super zoom and weighs a lot more, it has superior light gathering powers that let me take photos in poor lighting. The image stabilization lets me shoot at lower shutter speeds, which also helps with poor lighting. I took most of my photos at about 1/50th of a second to limit the ISO noise on my older DSLR's sensor, which has the charming side-effect of capturing nice, dynamic movement blurs from the presenter's hands. I use higher shutter speeds for hyperactive speakers that flail as they move, at the expense of significantly more ISO noise. If you have a newer, fancier camera, then you can use a higher shutter speed to get crisper photos without having them drown in ISO noise.
Crop later
Don't zoom in all the way because otherwise the slightest error in tilt or framing will ruin the image, e.g. a perfect photo of a speaker can be ruined if their hand is cut off the edge of the frame. Instead, zoom out a bit and crop later using your photo processing software.
White balance from name badges
Sometimes the auto white balance of the camera and photo processing software fails, but the neat thing is you can set white balance based on the white name badges which are included in most photos and adjust from there.
Adobe Lightroom
I love open source, but this photo processing application is magnitudes more awesome than any open source app available. The automatic mode does about 90% of the work on a photo for you. Exposure/Highlights/Shadows features rescue badly exposed photos, which are very common in tricky lighting. Clarity/Sharpness features pop details out of an image, which is helpful in making blurry photos taken at slow shutter speeds in dim rooms clearer. Noise Reduction reduces blotches from high ISO noise without destroying detail. Pick/Reject and Compare/Survey help compare a bunch of similar photos and find the best version. Spot Editing lets you tweak things like the Exposure and Clarity on just a face, or quickly obliterate a distracting clump of lint on the speaker's shirt. Etc.
Adobe Photoshop
I love open source, but this photo editing and digital painting application is decades beyond GIMP. I had to do major surgery on about a dozen photos, like editing out a person, pasting heads from various photos together so that an entire group of people looks good, remove vast amounts of cat hair on a person's shirt because it made their close-up look silly, etc. Yes, I realize that this makes the purist "Camera Verit=E9" crowd of "photojournalists" upset, but that's not what I'm striving for here. Some Photoshop features that make fixing photos easier include... Smart Objects provide non-destructive scaling and rotation of layered elements. Vector Mask allow crisp cutouts that you stack on top of each other, while classic Masks let you feather cutouts delicately. Filter Layers provide non-destructive color and exposure matching of elements from different photos that you can keep fussing with. Distort and Liquify fix perspective errors due to differences in the camera position between photos. Groups combine complex layer together to ease editing. Layer Comps let you compare various combinations of tweaks quickly. Patch Tool and Clone Stamp eliminate unwanted elements by replacing them with other bits from the image. Comprehensive Wacom tablet support provides precise control of opacity and position while fixing. Etc.
This content is reproduced here under the Creative Commons Attribution Non-Commercial Share Alike license, and is not intended to be re-licensed under the IndieWeb CC0 license.