Testing conducted in Seattle by our Technical Editor Richard Butler. Real-world production experiences by Jordan Drake: the director and editor of many of our ‘DPRTV’ videos.

Originally published Aug 3, updated Aug 6: conclusions and analysis revised based on additional experience with the camera

Recent Videos

If you have any interest in cameras, you may have witnessed the heated discussions lately around the new Canon EOS R5 and R6’s tendency to overheat when capturing video internally. The Internet tends to amplify the most extreme version of any story or phenomenon, which might have lead to you getting the impression that the cameras are unusable.

Jordan's EOS R5 experience

We shot for 10 hours at a variety of locations, which I thought would give the camera ample opportunity to cool down. I planned to shoot the episode in the 4K HQ mode, with occasional 4K/120P and 8K shots peppered throughout. Quickly I realized that setting up a shot and menu-diving would reduce the amount of record time I had for HQ, so I found myself spending far less time previewing the shot before rolling, adding a layer of stress.

Eventually, I realized couldn't record all the talking points in 4K HQ, and settled on using 4K HQ for wide shots and standard, line-skipped 4K for closeups. This made shooting sustainable, though I found myself avoiding trying to capture any spontaneous establishing shots or cutaways, lest I drop the dreaded overheating clock a bit lower. While our host Chris took it in his stride, I can only imagine how frustrating it would be for the talent to not know if the camera will last until the end of a take.

I also found myself heavily rationing the 4K/120P as it really chews up your remaining shooting minutes. I spent two minutes capturing the seagull footage in the episode: beforehand I the camera said it would shoot 15 minutes of 4K HQ, when I returned I had only five minutes remaining!

If the quality difference between 4K HQ and standard 4K capture were not so dramatic, this would bother me less. However, once you start viewing and editing the gorgeous 4K HQ footage, it makes it that much harder to go back to inferior line skipped 4K, and that’s a type of disappointment I don’t want to be dealing with on a shoot.

After extensive testing of both cameras, our conclusions with regards internal recording are:

  • From a cold start, the Canon EOS R5 and R6 perform in line with the company's video performance claims.
  • Non-video use cuts into available shooting time, adding significant uncertainty for video shooters

We tested a pair of R5s and an R6 in a variety of warm conditions and found they consistently performed in line with the limitations that Canon acknowledged at the point of launch. However, the practical implications are that the cameras are prone to overheating if you shoot for extended periods and if you have crew or talent waiting to re-start shooting, they may take too long to recover.

It should be noted that Canon did not design either the EOS R5 or R6 to be professional video tools, nor does it primarily market them as such. But based on our testing and real-world usage we would caution against using them as a substitute.

So why is YouTube saying the sky is falling?

Our testing suggests that the cameras perform in exactly the way that Canon said they would. However, there is an important caveat that Canon’s figures don’t address: although the cameras can repeatedly deliver the amount of video promised, they may not always do so in real-world usage.

Even set to the mode designed to limit pre-recording temperature build-up, the clock is essentially running from the moment you turn the camera on. Video recording is the most processor-intensive (and hence most heat generating) thing you can do, but any use of the camera will start to warm it up, and start chipping away at your recording times. Consequently, any time spent setting up a shot, setting white balance, setting focus or waiting for your talent to get ready (or shooting still images) will all cut into your available recording time, and you won't reliably get the full amount Canon advises.

Not only does this make R5 a poor fit for many professional video shoots, it also means that you can't depend on the cameras when shooting video alongside stills at, say, a wedding, which is a situation that the EOS R5 clearly is intended for.

Even when left in direct sunshine, the cameras continued to record for the duration Canon promised. However, this is only true when you're not using the camera for anything else.

The one piece of good news is that the camera's estimates appear to be on the conservative side: every time the camera said it would deliver X minutes of footage, it delivered what it'd promised. You can also record 4K footage for much longer if you can use an external recorder but again, this probably isn't going to suit photographers or video crews looking for a self-contained, do-everything device.

Click here if you want to see our test methods and results.

EOS R5 suggestions:

  • Expect to shoot line-skipped 30p for the bulk of your footage
  • Only use 8K or oversampled HQ 4K for occasional B-Roll
  • 4K/120 and 8K will cut into your shooting time quickest of all
  • Be aware of your setup time and cumulative usage (including stills shooting)

EOS R6 suggestions:

  • Don’t expect to be able to shoot video for extended periods
  • Be aware of the need for extensive cooling periods between bursts of shooting

Analysis: Why hadn’t Canon thought about this?

It’s easy to fall into the trap of thinking this means Canon didn’t put enough thought into thermal management for these cameras. Our testing suggests this isn’t the case, but that the cameras’ specs are rather over-ambitious.

Jordan's EOS R6 experience

I had done some testing prior to my shoot, and was comfortable that overheating wouldn't be a problem if I stuck to 4K/24p. Unfortunately, my experience on a warm day was quite different to that room-temperature test. There's no line-skipped 4K mode on the R6, so if the camera overheats, you're back to 1080P, which will be a jarring transition for viewers watching on larger screens.

While I was able to record much longer with the R6 before encountering the overheat warning, once it appears the camera takes far longer to cool down again than the R5. Our regular drives in an air conditioned car allowed Chris and Levi’s R5 to function throughout the day, but at one point I was left sitting in the car, babysitting a hot R6 while they went out to shoot. During a one hour lunch, the R5 had returned back to normal, but the R6 had a twenty minute warning still on.

This was hugely disappointing as, rolling shutter aside, the R6 video quality is excellent, and I’d be perfectly happy using it over the R5. However, the longer cool down times would probably lead me to use the R5, dropping to line-skipped 4K from time to time.

While I enjoyed most aspects of using these two cameras, I have no intention of using either of them as a primary video camera. They would be great for grabbing occasional, very high quality video clips, but I’d never want to rely on them for paid work.

With the exception of specialist video models, most cameras that shoot 4K are prone to overheating, regardless of the brand. Some companies let you extend the recording time by ignoring overheat warnings (and risk ‘low-temperature burns’ if you handhold the camera), while others simply stop when they get too hot. This should make it clear that shooting 4K for an extended period is difficult. For instance, Sony says the a7 III will shoot around 29 minutes of 4K video with the temperature warnings set to ‘Std,’ while the Fujifilm X-T4 promises 30 minutes of 4K/30 and 20 minutes of 4K/60.

The cumulative heat is constantly counting against you

8K is four times as much data as natively-sampled 4K and seventeen times more than the 1080 footage that older cameras used to capture so effortlessly. Perfect 2:1 oversampled 4K (downsampled 8K) requires this same amount of data, which is still 1.7x more data than is used to create 4K oversampled video from a 24MP sensor. Data means processing, which means heat.

What's interesting is that the exterior of the cameras don't get especially hot when shooting for extended periods. We're only speculating, but this could indicate that Canon has tried to isolate the camera's internals from external temperature fluctuations, with the down-side that they can't then dissipate internally produced heat.

This would be consistent with us getting the full recording period out of the camera, even when tested well above the 23°C (73°F) conditions specified by Canon. And with the fact that leaving the camera's doors closed and battery in place didn't change the recovery time. However, while this appears to be workable for the line-skipping 4K mode, the added workload of the higher quality settings seems to present a problem. Dealing with 1.7x more data than the a7 III and X-T4 is a step too far: the R5 will match them for promised recording duration but only from cold. This leaves it much more sensitive to any other use beyond video recording.

The EOS R6 is a slightly different matter. It can shoot 40 minutes of 4K taken from 5.1K capture, which is a pretty good performance and may be enough that you won't often hit its temperature limits. However even after a 30 minute cooling period, it has only recovered enough to deliver around half of its maximum record time, whereas the EOS R5 recovered nearly its full capability. The more extensive use of metal in the construction of the R5 seems to help it manage heat better than the R6 can.

And, as both Jordan's and Richard's experiences show: if you don't have time to let the cameras cool, that cumulative heat is constantly counting against you.