Canon EOS R5
Crowdsourced Review; Manual Supplement; Known Issues; Hacks; Insights; User Forum
Sony A1 Dynamic Range < Canon R5
In a respectable showing, but perhaps disappointing to some, Bill Claff over at PhotonsToPhotos.net shows that the new flagship A1 camera only has dynamic range roughly as good as the A7r Mark IV and the A9 Mark II. For most ISO settings it sits in a very narrow range right between those two cameras.
In what is sure to inspire much online forum gnashing, the same data sets show that the A1 comes up slightly short of the Canon R5. It performs similarly versus the Canon 1DX Mark III, which itself performs slightly worse than the R5.
With some exceptions, the differences are slight. A notable exception is at ISO 400, where the R5 appears to have two stops of dynamic range advantage over the new A1. This may be the only setting where images would be noticeably affected.
[UPDATE 1/2/22: Nikon’s Z9 data can now be found among the dynamic range data set. It shows performance just shy of the Canon R3 and the Sony A1.]
PhotonsToPhotos.net is a long-standing web resource for reliable third party dynamic range testing.
Powering the EOS R5: Review of the LP-e6NH & Other Options
The Canon R5 has monster capabilities, and with it, monster power draw when used in certain high-performance modes. To account for this, rather than reduce the maximum capabilities of the camera, Canon did two things: they created different modes that the camera could fall to when it was supplied with less power; and it created a battery – the LP-e6NH – that provided a higher wattage for a longer period of time.
They also finally enabled USB charging *with* running the camera while on. Some of these learnings – like the fact that USB batteries of 30 watts or greater will power the camera in addition to charging batteries when the camera is off – are undocumented in Canon materials.
Many incorrectly believe that the LP-e6NH has slightly more usable power than its predecessor the LP-e6N, or the original LP-e6. Real-world tests using dozens of individual batteries over a half-year period with special hardware prove that the newer models don’t push more milliamp hours of power out to the camera over the course of a full charge when battery age and use is accounted for. But the newest batteries do have a special capability: they can keep their wattage up over a certain threshold through the use of about 2/3rds of the battery, maximizing the ability of battery users to employ the most power-hungry modes of the camera – particularly the highest framerate while using mechanical shutter – 12 frames per second (FPS).
The Massive Differences Between the LP-e6, the LP-e6N and the LP-e6NH
To add even more variables to the mix, sometimes the best solution for certain users is to use external batteries attached to a “dummy battery” shaped like an LP-e6 and connected via a D-Tap interface (shown below). Higher end video users will be familiar with this sort of battery arrangement, but it is less commonly seen in the stills community. This article will introduce the benefits of D-Tap and the cost-efficiency of buying the larger video-oriented batteries for some projects.
We tested all of these solutions (and several other, crazier ones that we don’t recommend) and have collected field input to chart out the capabilities, limitations and real-world notes. No EOS R5 cameras were hurt over the course of this project. The last time we conducted experimental battery tech, we blew up a 5D Mark IV and a 70D using an unregulated dummy battery cord made in China, and later one of our own (apparently poor) design. [Ed. note: Canon Professional Services replaced main boards in both cameras at no cost, despite our honesty in describing exactly how we’d abused the cameras. They’ve been a consistent ally.]
This article is the product of half a year of intense, deliberate use of two R5 bodies and one R6 body, with the cameras being left outdoors multiple times in below 0 Fahrenheit temperatures for days or weeks at a time. Earlier CFexpress card tests, burning up hundreds of thousands of mechanical shutter clicks, were exploited for double duty to get information on battery effects.
Even with all of that direct, real-world experience, there were many things we would have missed had not the online communities on FredMiranda.com, CanonRumors.com and DPReview.com hosted intelligent, data-driven conversations on battery issues. Many learnings were first observed there and confirmed with tests. This article could not have been written without those contributors in the community. We hope this article will partly return the favor.
Ways to Power an R5
Think of it this way: an R5 has two plugs into which power can flow. There is the battery compartment, and there is the USB-C connection. Both of those receptacles can attach to several different types of power sources. Both can attach to USB batteries or a wall plug. The battery compartment can also take LP-e6NH batteries and similar. It can also – via a dummy battery – connect to a seemingly endless number of other batteries in various ways.
The battery grips Canon sells just allow putting two batteries in at a time. Interestingly, whether gripped or ungripped, the USB power appears to be routed through the batteries, as in neither case will the camera power on with just a direct USB connection.
Let’s Start Simple: Real-World LP-e6NH Performance Comparisons
Camnostic tested 22 LP-e6NH batteries over the course of the six months since their release. The data over that time included up to 15 full depletion and recharges shows a gradually descending power capacity, moving from an average of 2.04 amp hours to 1.93 amp hours…
The granular data are messier than the red trend line above suggests. To the right are the individual courses over time for the batteries.
If extrapolated for a decade or so, one might expect the NH version of the batteries to meet the current performance of the aged LP-e6 and LP-e6N versions. We had the opportunity to test unused LP-e6 and ‘N variants, and they behaved similarly to the new ‘NH versions.
The blue and orange bars to the left show the average recharge performance of about a dozen used and abused LP-e6 and LP-e6N batteries respectively. The LP-e6N batteries do outperform the LP-e6 batteries, but this is likely due to the “N” batteries being about five years younger on average. When new versions of those batteries were tested, they were all about equivalent to the ‘NH batteries.
The main downside to using one of the older batteries is that the R5 and R6 cameras’ firmware treat those older batteries differently, disallowing the fastest mechanical shutter frame rate modes. This means that your 12 frames per second (FPS) shooting – considered by many to be a perfectly acceptable or preferable to the electronic shutter’s 20 FPS – suddenly drops to 9 FPS, or heaven forbid, 7 FPS.
The same disadvantages seen with obsolete Canon batteries happens with third party batteries imitating the LP-e6 and LP-e6N types. But, worse, more modern camera firmware will put up a dialog box upon first turning on the camera with one of these third party batteries, asking if you recognize that this isn’t a genuine Canon battery. This “feature,” seemed at first a nice effort to try to winnow out of the market manufactures who were selling fake Canon batteries. But the main result today is that if you have a remote camera out in the field, and it needs to restart for any reason, your remote rig is now useless unless a solution is figured out to keep remote cameras on indefinitely.
Unfortunately, all current dummy batteries are treated as third party LP-e6 batteries, not the native NH variant, no matter the power source to which it connects. This means that no dummy battery solution will access the full framerate for mechanical shutter – although silent shutter will of course still provide the even higher 20 FPS. We have found that most use cases for the remote cameras – where this problem would toll the greatest – are exactly the ones where electronic shutter is appropriate, such as remote cameras outdoors on wildlife sites, or remote setups in sports arenas, so this does mitigate the number of times this problem proves a frustration.
Sweet, Sweet GH+
Mwah, Mwah, Mwaaah
Below you can see a bar chart of the frame rates achieved by the various modes. Canon uses an icon (seen above) on the left-hand side of the viewfinder or back screen to indicate what frame rate mode the camera is in at the time. Electronic shutter shows 20 frames per second across the board. It’s lower power draw allows it to be unhindered by the batteries. We’ll concentrate on trying to achieve the fastest mechanical shutter frame rates, particularly the Green H+ (above left and referred to in this article as GH+), which indicates shooting at 12 FPS. If your H+ turns white (above right), you’re shooting at about 9 FPS, and if it starts to blink on you, you’re already down to 7 FPS.
Even though the older batteries can provide the needed wattage – if for a shorter period of time – they do not identify themselves as an LP-e6NH battery, so the R5 does not allow them to access the larger power draw modes. Interestingly, four companies are now producing third party “NH” batteries. Those brands are Big Mike (BM), Watson, Powerextra and Neewer. All of them will identify to the camera as an LP-e6NH, and all of them will access all framerate modes. However, they will do so for only about half of the duration of an actual Canon NH battery before the camera automatically reverts to lesser modes with lesser power draw. Below we show a comparison of all five available “NH-reporting” batteries.
All Currently Known LP-e6 Form Factor Batteries that Successfully Report as LP-e6NH
Real-World LP-e6NH Performance Comparisons
The chart shows the blue columns representing the gross usable power in the battery, with Canon winning by about 10-15 percent (left scale). The orange columns (right scale) represent the percentage of the battery’s power that is applicable in the higher wattage mode allowing for 12 FPS before the waning wattage fails over to a lesser framerate mode.
It is here that the Canon batteries are plainly superior. Users who shoot only in electronic shutter may never notice a functional difference between battery brands, but those who use mechanical shutter and require 12 FPS certainly will. A third party battery causing the GH+ 12 FPS mode to go away isn’t as terrible as it might seem when one can always just switch to electronic shutter and get a consistent 20 FPS. Few applications (rapid panning, for one) do call for mechanical shutter necessarily.
The Battery Grips
The Canon WFT-R10A grip – the $1,000 one – is significantly less capable than the $350 BG-R10 version in some key aspects. This is primarily because even if the wifi feature is turned off, it forces the R5 to consider any limitations associated with wifi to be in force. This means that the maximum framerate of the R5 employing this grip is 9 FPS, down from 12. There is no way to make the camera fully perform in mechanical shutter using that grip. Several calls to Canon Professional Services eventually confirmed this with some surprised CPS reps having to repeatedly double-check with higher-ups. This limitation does not seem rational, especially considering the specific market for which the WFT-R10A was designed: pro shooters with live hookups to photo desks. In the now-fraught choice of flagship cameras for sports, the Sony A1 now provides faster mechanical shutter.
The simpler BG-R10 (pictured above) works as one would expect. Not only will it allow for two NH batteries to be used to maximum effect, it will even allow a single Canon LP-e6NH battery couple with a third party LP-e6NH copy and maintain the higher frame rate mode until just the Canon battery gets below its threshold (typically <35 percent). This is the configuration in the image above, with a Neewer battery mated with a Canon NH battery. Because the third party batteries show roughly 90 percent the gross power capacity of the Canon versions, this means having one of each in the grip will almost double the time two third-party batteries will run at 12 FPS, as the firmware runs the batteries down together, not serially. Running two third party batteries would hit the lower FPS mode when they reached about 75 percent of capacity, about half that time. Interestingly, running either one or two older LP-e6 or LP-e6N batteries, however, would set the framerate down to 9 FPS from the get-go, indicating that those battery types are sensed and treated differently.
Using Power from the Grid
Canon includes a high-quality USB-C cord for connecting to a computer or to an external battery, but not a power supply that can connect to a wall. For that, Canon sells the optional $129 PD-E1, which will both charge and power the camera. A battery needs to be in the camera or grip in order for the wall power or external battery power to run the camera.
We tested the use of the PD-E1’s USB connection with just an unconnected dummy battery in the battery compartment to see if the wall power USB interface would work in that situation. It did not, even as it indicated a battery was charging with the little lightning symbol. An actual battery needs to be in the receptacle for the USB power to work. This is further evidence that the camera operates off of the battery even when it is plugged into wall power.
Unfortunately, the highest power output for the power supply is 9 volts at 3 amps (which makes for 27 watts, as you multiply volts and amps to get the wattage), below the threshold for 12 FPS in mechanical shutter. Our test showed an average of 9.3 frames per second with the PD-E1 power adapter. There is another way, though…
Using a High-Watt USB Battery
It isn’t a high proportion of them, but some USB batteries allow for higher wattage output. Using a good USB-C cord that supports the PD standard and a battery that outputs enough wattage will power the camera and charge the battery. As with the PD-E1 wall plug, this arrangement will allow for the 9 FPS capability and not the 12 FPS capability. [Ed note: Reader reports indicate that a 30-watt battery is sufficient for most, although there are cases where some 30-watt batteries do not work, including a couple of our own. The 45-, 60- and 90-watt batteries tested all work without exception so far. We are conducting more tests and welcome reader feedback on their own findings.]
The RavPower 90W 30,000 mAh PD External Battery Charging Away (Note the green light on the BG-R10 indicating which battery is soaking up the juice)
One downside to running the camera this way is that you have a cord sticking out of the left-hand side of your camera, and if someone stumbled or knocked the battery down, it could break USB connection. As Roger Cicala and Aaron Closz found in their tear-down of the R5, the USB and other ports on that side are attached directly to the main PCB – essentially the very expensive brain of the camera. The image below is from their excellent LensRentals blog tear-down post and is used with permission.
One partial protection from this would be to use the cord protector that came with the camera – that little, odd-looking plastic bit with a screw that you probably threw away or lost some months ago. Oddly, there doesn’t appear to be an aftermarket version of the cord protector available that doesn’t involve first nesting the camera in a video rig cage. Some L-brackets do provide some protection if offset such that there is some stand-off room between the bracket and the side of the camera with the ports.
If you use a battery that outputs 18 watts, it will charge the LP-e6NH (and only NH batteries) inside it when the camera is turned off. It will not run the camera, or charge the battery if the camera is on. A battery of less than 18 watt output, connected via the USB cable, would not do anything but to serve as an opportunity to crack your motherboard.
To run the camera with an external battery, it must provide more watts. This story was first published stating it required 45 watts, but at least one reader reports that a 30 watt battery was sufficient. We do know that some 30 watt batteries do not work, but we have not yet determined whether this is as designed, or a sign of wear. [Update, we purchased a 30-watt battery recommended by that user, and can confirm it can indeed run the camera.] We will report further.
The biggest downside to USB batteries – for some people at least – is the fact that they all seem to have firmware that causes them to sleep after a period of time with no draw, and a surprisingly high proportion of them require a button press or a USB port re-insertion to wake up. This makes using regular USB power bricks fraught for applications that require long periods of waiting. Some intervalometer setups, and certainly remote rigs for sports or wildlife can be ruined by this.
One way to mitigate it is to use a Tether Tools Case Relay connector (pictured above) between the USB battery and the camera. The Case Relay product not only provides a connection between a dummy battery and a USB battery, but it also has its own (roughly 1,000 mAh) internal battery that it uses as a buffer, allowing hot-swapping of USB batteries and also allowing it to sent “wake” signals to the USB battery when needed. In practical experience, though, this has proven to be flaky. It is difficult to determine which component causes failure in any given instance, but we do know that there have been times when USB battery firmware has prevented the wake signal from working after an extended period of time, and we also know that there have been times with the Case Relay battery has frozen in the cold much earlier than the larger USB battery, causing a general system failure – which of course fails to recover upon thawing due to Canon’s battery copy protection startup dialog boxes.
A Safer Solution
Some applications – like wildlife remote cameras – require lots of power to last days or weeks, and often in very cold conditions, where batteries can see their power halved just from the temperature. Camnostic will be producing a series of stories and interviews on that use case in the future. In the meantime, we tested out the large video camera batteries that have an established an at-scale market already with batteries that commonly reach 270 Watt Hours, or a little more 16 times the LP-e6NH’s actual measured performance.
Our favorite turned out to be a 98 watt hour battery from IndiProTools.com (Pictured at right…). It has the trifecta of useful interfaces: a D-Tap connection; a 7.4-volt 2.1mm connector compatible with most dummy batteries; and a simple USB port. All three work to power the R5, depending on what connection your dummy battery employs. It costs $99, comes with a charger, and costs only 25 percent more than a single Canon LP-e6NH battery, yet has the power of six of them. The non-D-Tap version is more popular and is more frequently sold out, but the most desirable one is the one that includes the D-Tap connection.
Unfortunately, like the other dummy batteries, the D-Tap-to-dummy-battery connectors available on the market so far act only as LP-e6 batteries, rendering the frame rate the “White Blinking H+” mode, or 7 FPS no matter the voltage. For remote setups, this is often a perfectly acceptable compromise in order to get the large, cheap batteries that provide reliable power without overly-clever sleep routines, but most times with those setups users are employing the e-shutter at 20 FPS anyway, making the power-related framerate irrelevant. The larger battery sizes available also lead to less freezing and generally higher reliability in the field.
For a particularly long remote camera set requiring no human intervention for weeks at a time (bobcat den), we drilled holes into a grip to provide for two dummy batteries to be placed, connected to two large D-Tap batteries, providing for six weeks of power. We used another battery, hacked to a thermostat part, to provide 5 volts to a small heating pad (originally designed to keep lizard tanks toasty) when the temperature in the cooler-enclosed battery pack got too cold. Setups running on an intervalometer turn out not to need the heating pad, as the slight consistent draw from the battery provides its own warmth from the inefficiency of battery-delivered power, provided the battery rests in an adequately-insulated cozy. For adequate insulation, we found that a well-wrapped battery inside a six-pack cooler worked great for this.
A single 98 Watt Hour battery photographed a barred owl nest on an intervalometer for more than 72 hours before finally reaching the last indicator light of power. The heat generated from simply taking a picture each minute kept the uninsulated battery from freezing in nights that reached 15 degrees F below freezing. The same setup running on the same night, but on an infra-red trigger rather than a constant-draw intervalometer froze solid, preventing the chemical reactions involved in the production of electricity.
The upshot is that the D-Tap batteries designed for video use can provide almost limitless power, without the danger of having something stick out of your USB slot. The D-Tap connector shoots over to the camera via a dummy battery in the battery slot. The downside: you don’t get maximum framerate in mechanical shutter because the known dummy batteries are LP-e6 dummy batteries and not LP-e6NH dummy batteries. Perhaps some enterprising engineer will figure out how to make one of those in the future.
Other Reasons for Slower FPS
Canon tells us we’ll get 12 frames per second so long as certain conditions remain true:
- The temperature needs to be warm enough, with the benchmark set at about room temperature
- It indicates that a Canon LP-e6NH needs to have at least about 60 percent juice left (although our tests show this can go consistently down to 35 percent and still perform at the fastest speed, provided it’s a Canon NH battery)
- Shutter speeds must be high enough to not infringe on the FPS rate, of course
Not Going to Cut It…
- Not connected to wifi
- Flicker reduction turned off
- Set to maximum aperture
To this we can add that power must be drawn from the battery compartment, rather than the USB interface.
[UPDATE: At this writing in early fall 2022, it has now been more than two years since the launch of the R5. Much has happened. CFexpress cards have come down in price to be about the same price per gigabyte as SD cards. Super large CFexpress cards have come on the market, with 2 terabyte configurations common and 4 terabyte models available. Lightroom, formerly perennial slowpoke of image organization, got pretty zippy a few revisions back. All of this adds up to pressure on the batteries, as photographers slip into 20 frames per second mode and let loose, without as much worry as before about running out of room or torturing themselves in front of a computer, culling photos.
The LP-e6NH batteries have been holding up as expected, with their performance degradation over time plateauing. Tests are being run to show this in chart form in the near future.
The R3’s launch about a year ago had some Canon-ites move to that low resolution shooter, with its 1-series battery form factor. Those looking to maintain high resolution still find the R5 to be the best Canon camera.
As promised, we tested a series of additional 30w output USB batteries, and all of the new ones we tried successfully powered the R5. Our initial tests were the outlier, with a 30w model that likely wasn’t quite putting out 30w.
Several significant firmware upgrades in, video has been given longer leash, allowing the camera to heat up more than was previously possible. This will burn through the battery faster, but people pushing past the guardrails know what they’re getting into, and likely have a dummy battery leading to a DTAP battery anyway.
Ever optimistic, Canon fans have been pressing the refresh button on rumor sites looking for news of new offerings. It is likely that the R5 will get refreshed in the typical Canon timeframe a four-year target. Will a higher-resolution version come out in the meantime? What features could possibly justify a Mark II version anyway? This problem presented itself on the launch of the R3, which did manage to collect some great ergonomic improvements and a 30 FPS shutter, but which – for many – is still shooting a resolution from the 2010s. The R5 remains the pro workhorse for the brunt of the market. We just need to carry some more batteries.
As ever, field observations by readers prove very valuable. Have an additional tip, correction or another observation that might improve this review of power options? Please contact us here.
Hands-On Evaluation of Sony A1 Versus Canon R5
While the world waits for the February 4 Sony embargo date for reviews that include pictures, video and image quality evaluations, Tony Northrup gave his fairly complete set of impressions of the newly-launched Sony A1 flagship camera. Northrup has the camera in-hand, and was able to compare it directly to the camera that appears to be its most direct competitor, the Canon R5.
As is normal for major new series releases, initial spec sheet marketing gives way to the coverage of limitations and caveats.
The biggies on the positive site:
- Rolling shutter appears to be completely quelled
- EVF is indistinguishable from optical viewfinder
- Flash abilities beat those of any other mirrorless camera
- One of the two key features doesn’t work quite as advertised: the 30 frames per second appears to be possible only when using a lower bit depth; a small set of expensive lenses; and while the tracking and autofocus aren’t being engaged. Real-world conditions give frames per second of about half of the rated 30 FPS, which is less than that achieved by the R5.
- The tilt screen tilts in only one dimension, largely preventing its use shooting in portrait orientation.
- Newly-launched bird eye tracking isn’t as good as Canon’s R5 equivalent
- The heat envelope isn’t so different from that of the R5.
Forum Dwellers Discover Canon R5 Speed Limit
At its fastest, with a speedy CFexpress Type B card, the Canon R5 eats RAW files at a rate of about 10 per second. This effectively gives shooters a 10 FPS camera without any cache limitations, even though the camera can shoot 20 FPS in bursts. But forum dwellers on the internet, in trying to figure out why they were getting different cache clearing speeds, discovered that the operation of the electronic viewfinder (EVF) radically slows down this ingestion rate.
Where the R5’s cache can be read out to the card in about five seconds, using the EVF while the cache is clearing causes it to take just under 20 seconds, a 400 percent slowdown. Whether from an additional power draw, or – perhaps more likely, additional processing needed – the delay is added on for only the time the EVF is activated. As many shooters set the camera to automatically turn on the EFV when their face is near the viewfinder, the cache clearing can be accelerated simply by moving one’s face away and watching the cache clear on the back LCD, which does not slow down the data transfer.
The effect is also negated by setting the EVF to the lower refresh rate. This accomplished by going to the red section in the menus, on the 8th tab, where “Display Performance” will be found.
The CFexpress card review published here, comparing the main cards available on the market, was unaffected by this phenomenon, as the data was collected with the camera on a tripod employing the LCD.
Angelbird CFexpress Cards First To Get User Firmware Upgrade
Angelbird released an installer to upgrade its AV Pro XT CFexpress cards‘ firmware – an apparent first in the new CFexpress card category. The new firmware version 1.20 eliminates the a startup delay seen with the Canon R5. A mix of interactions between the original R5 firmware and the older card firmware caused a four-second each time the camera woke up, but subsequent RF firmware upgrades mostly relegated the issue to a single delay each time the camera was first turned on, or the Angelbird card was inserted into the slot. The new card firmware eliminates even that initial delay.
Other firms have released new firmware for CFexpress cards, but have asked users to return cards for replacements when necessary, such as Delkin’s proactive release in November 2020 of a new version fixing an issue affecting some people.
The Angelbird cards are among the cheapest and fastest among cards tested, and easily among the few cards that spawn the most interest and questions in regard to the Camnostic review of CF Cards published back in October.
The card firmware upgrade installer requires the use of an Angelbird CFexpress card reader and a Windows computer. A representative from Angelbird indicated a Mac version should be coming soon. A review of card readers appears here, where Angelbird card performs decently its interface type (USB C, non-Thunderbolt). That reader costs about $65 and is one of only two that come in a very portable size.
Panasonic S1R CFexpress Upgrade Gives Only 1/7th Boost
Upgrading the S1R cameras to use CFexpress cards improves card intake speeds by about 1/7th. Its, frankly, a little disappointing, but not surprising. When Panasonic allowed for the use of CFexpress cards in its flagship cameras via a firmware update, many had visions of multiplying data throughput – which indeed is theoretically possible when looking at the CFexpress standard. But there are many, many variables that affect actual throughput, including the camera’s firmware and the hardware and software built-in to components of the bodies that usher the data onto the card. There are so many variables, that after talking to engineers and executives at several CFexpress manufacturers, we stopped keeping track of them and opted instead to just acquire and test the cards to show real-world performance.
The Panasonic S1R and its husky brother, the video-oriented S1H, are limited to just below 10 frames per second in terms of image acquisition, and that won’t change with a faster card, but what will change is the behavior of the camera once the cache is filled up. The cameras start a significant slowdown of framerate after six seconds into a burst. The excellent camera site Imaging-resource.com provides some independent testing of these sorts of things, as we corroborated their findings both with a recent rental, as well the author of this research piece having used the S1R as his primary shooter for most of 2019. With those 62 shots taken before the cache fills, the cameras take about 17 1/2 seconds to finish packing the data stored in the cache onto an XQD card. With upgraded firmware and the addition of any of the fastest CFexpress cards available, that time to pack the data away lowers to 13 1/2 seconds.
In the days of XQD cards, trying to continue to shoot with a full buffer would only offer a photographer 2.5 frames per second. With the CFexpress card enabled, that buffered shooting rate now becomes 2.8 frames per second.
It appears that there are hardware and/or software limitations still in the Panasonic data path separate from the speed of the cards. The cards have been extensively tested with the Canon EOS R5 and R6 to show greater absolute capability, as the link above and chart below show. The five fastest-rated cards tested all took more than 300 shots in a 30-second period on the Canon R5, and showed a framerate with a full cache of 9.9 frames per second (versus the 20 frames per second the camera was set to do in electronic shutter mode during that test).
That 9.9 frames per second (FPS) moving onto the CFexpress card in the R5 compares directly with the 2.8 frames per second moving onto the same card in the Panasonic S1R, a more than three-fold difference. Below is a chart of available CFexpress cards showing how many seconds each takes to clear the R5 buffer, which is a similar size to that of the Panasonic S1R. For one of the fast cards, the average is about 4 seconds, compared to the S1R’s CFexpress buffer clearing of 17.3 seconds using XQD and 13.5 seconds using CFexpress.
We have seen this card format software dance before. When the UHS-II standard became available, a few cameras received upgrades to firmware allowing the new cards’ use. In no case did those upgraded cameras fully exploit the new speeds available with the new card standard, as some combination of other components provided data bottlenecks.
Which Card is Best?
The ProGrade Cobalt cards appear to show the best performance and decent price, but the differences in performance on the Panasonic bodies are very slight versus the differences seen on the Canon (and even Nikon) bodies, where there appear to be fewer bottlenecks. Within the margin of error at the top are also the Angelbird Pro XT card and the Delkin Power card. Sony’s Tough card performs just as well, but is more expensive. Transcend and the ProGrade Gold cards performed measurably, but slightly less well.