Do concerns about Sony A7 light leaks hold water?

by Dave Etchells

posted Friday, March 7, 2014 at 12:08 AM EDT


As we reported yesterday, the Internet recently has been aflutter with reports of light leaks in the Sony A7 and A7R.

If you expose the camera to bright sunlight, attach the lens cap, crank the ISO to 25,600, and take a 30 second exposure, you'll probably see streaks of light in the image. You can achieve a similar effect with a shorter exposure time and lower ISO, if you aim a high-wattage strobe at the camera as well. Referenced in our earlier article, photographer Ferrell McCollough posted a nice summary of the effect and an easy fix for the issue.

Given the level of attention the issue has received, we thought the issue deserved a critical look.

Our test of the Sony A7

Our Sony A7 is currently with Editor Mike Tomkins, who wasn't able to get out in time to catch the sun today. Instead, he pointed two high-intensity flashlights at the base of the lens flange. The results with the flashlights aren't as bright as some folks have observed, but still evident. (See Ferrell's page above for a bit more dramatic example, using a light bulb three inches from the camera.)

"Light leak" from Sony A7, 30 sec, ISO 25,600, two bright flashlights shining on lens flange.

We'll update this post with test shots from Mike first thing tomorrow morning, but this provides some frame of reference until then. 

Our test of two high-end competitors

We decided to test two high-end full-frame cameras we've reviewed exhaustively and which have never exhibited real-world problems with light leaks in our usage: the Nikon D800E and Canon 5D Mark III. The results are illuminating (excuse the pun):

Canon 5D Mark III, 30 seconds, ISO 25,600, sunlight. Sigma 70mm f/2.8 Macro.


Nikon D800E, 30 seconds, ISO 25,600, sunlight. Sigma 70mm f/2.8 Macro.

Keep in mind that these results were achieved under anything but 'normal' usage. We applied about 3 feet of gaffers tape around the lens caps of each camera to ensure we weren't getting any light leaking through the lens. We taped over and around the optical viewfinder of both cameras (the results were much worse for the 5D Mark III, which doesn't have an OVF shutter and slightly worse for the D800E which shutter isn't entirely lightproof). Importantly, as soon as we taped around the lens flange, the frames were completely black, indicating the light leaks you see above are of the exact same variety as the A7 and A7R: from the lens/flange interface. And to be clear, neither the D800E, nor the 5D Mark III were really usable with the amount of gaffers tape we applied.

The Canon LCD light leak notwithstanding, we've never had light leak problems with either the Mark III or the D800E in our own usage and we're not aware of any furor over either of these models. 

Drawing it together, is a light leak of the sort seen in our A7 in any way unusual? Absolutely not. The best Canon and Nikon models exhibit similar behavior. While a perfect camera might block every last photon, the tolerances required of lens and flange make this exceedingly difficult. 

Is this sort of light leak a problem?

So light leaks -- however small -- are not particularly unusual. The question now is: do they matter?

Let's consider what we're talking about here: a 30-second exposure at ISO 25,600, with a very bright light source shining directly on the lens flange.

Is that really something you're likely to encounter in real life? About the only time we can imagine such long exposures under bright conditions would be when you're trying for extreme motion blur in your subject, say a blurred, feathery look in moving water. But in such cases, you're going to be stacking ND filters on the front of the lens and using the lowest ISO the camera can manage.

If you love stacking neutral density filters and shoot at higher ISOs, light leaks are more likely. Photo courtesy of Lens Rentals, the best place to rent an A7 and stacks of filters.

More practically, people shooting long exposures to blur motion in clouds might use extremely long exposures, but consider that 30 seconds at ISO 25,600 is equivalent to an 128 minute exposure at ISO 100. It might take somewhat less than that for the first hint of an effect to appear in your shots, but we're still very conservatively talking about a 10-20 minute exposure at ISO 100 before you could see anything at all.

Again, on the more reasonable side, as Ferrell McCollough demonstrated in his YouTube video, a studio strobe shining directly on the base of the lens flange also can cause a problem. As he shows, though, just moving the camera back a few inches out of the direct path of the strobe instantly fixes things. (As would, we suspect, moving the camera a few inches forward, causing the strobe's light to be blocked by the camera body itself.)

As Ferrell also points out, you can completely eliminate the problem on the A7 (and the D800E and 5D Mark III, we suspect) by wrapping an elastic hair-band around the lens/flange interface.


Your mileage may vary, of course, and it's possible that some samples of the A7/A7R have poorer tolerances and light leaks that actually impact performance in real-world situations, at exposure times and ISO levels matched to ambient lighting. But to our minds, based on what we've seen with our A7 and the best competing models, this is really a non-issue.

We're definitely open to updating our findings, so welcome reader comments. Have you witnessed light leaks that have impacted your work? Have you had problems with similar light leaks in other cameras?

Stay tuned for our test of the A7 in daylight tomorrow. Check out our final conclusion on the Sony A7 and A7R light leaks.