CFexpress Type B Card Performance
Tested on EOS R5
[See SD Card performance here; and CFexpres Type B card readers review can be found here.]
Pricing developments in the past few months have changed the memory market. Where CFexpress Type B cards were once the super-expensive-but-fast options for flagship bodies, they are now super-fast, almost-as-cheap-as-SD-UHS-II cards. And they got faster still. Wise, Delkin and Angelbird launched new cards and Mark II versions of older cards. Each launched a flagship line designed for more reliable high-bandwidth video recording and an improved large capacity line of cards that is cheaper per gigabyte than before.
Of all the cards tested over the last two years, the top three performers were all launched in the past few months. Launched in March 2022, the Angelbird AV Pro XT Mark II 660 GB card (pictured left) proved the fastest in pushing still photos onto the card. The next two, the Wise 320 GB and the Delkin Black 512 GB were number two and three respectively; as seen in the graph below with the blue outlined boxes showing speed (higher is better).
Just a few weeks later, in May 2022, Delkin hit back, reclaiming the title as fastest card with its Delkin Black 325 GB.
But to put things in perspective, the now-second-place Mark II version of the Angelbird AV Pro XT card’s speed increase is only 5 frames over 30 seconds faster than the Mark I version of the same card; now in seventh place. They’re all pretty spectacular. When you factor in the roomy buffers of the Canon R5 and R3 especially, not many photographers are going to be forced to shoot at the worst-case frame rates these charts suggest.
A note on the Lexar card and company. Our copy of the Lexar card happened to be the only one of dozens that failed since we started testing in 2020. Unfortunately, Lexar told us it is unable to directly replace the card due to its parent company, Longsys, remaining on a US sanctions list. Some people who have bought through a third party retailer report that they have been able to get replacement cards through the retailer, but we were out of luck. In the absence of additional examples, we do not consider the card failure to be more than a fluke, but it should be noted that this manufacturer is limited in its service capacities.
Below are two charts, one for each size class (above 256GB and below), separating out the cards’ pricing per gigabyte by brand and line of card. This allows a more fair comparison of price based on apples-to-apples comparisons of size.
The Upshot: Our original recommendations (Delkin Power and Angelbird Pro cards of >500 GB) still stand as the best bang for buck, but added to this list of recommended cards is the cheaper of the Wise brand cards. They launched two new lines, one simply branded Wise, and another Wise Pro. The Wise branded cards provide the same $/GB as the Delkin Power and Angelbird Pro cards. A normal price of about $0.50 per gigabyte will be seen in the highest-capacity cards of 2,000 gigabytes. Regular deals bring them down to $0.31 to $0.40 per gigabyte. 512 GB cards in those lines will cost about 95 cents per gigabyte. Sabrent provides the best deal in 512 gigabyte cards with $0.66 per gig.
This updated review wouldn’t be possible without the cooperation of the companies mentioned. Not only that, over the course of the last year and a half, representatives from these companies have engaged in an ongoing conversation about the increasing specificity of the CFexpress standards and the card manufacturers’ various responses. Canon has spent significant engineering resources to help these manufacturers gain more consistent compatibility, and the industry as a whole has generally provided photographers with major speed and consistency gains by cooperating together rather than trying to use CFexpress standards as a means to a competitive advantage. It didn’t have to happen this way, and it bodes well for the format’s continued adoption.
Since Camnostic first published the comprehensive review of all available cards in 2020, staff took the intervening 18 months to torture test multiple cards from Delkin and Angelbird, along with Sony, ProGrade, Transcend, and – more recently – Wise and Acer. They all performed consistently without error or problem with excessive use, heat and cold. These cards were put through many camera heat shutdowns, with millions of frames of stills taken. We did not see any difference in reliability between these brands. On the other hand, our own experience and, according to reports on forums, those of others show that the Pergear and Lexar cards did have some problems. The Pergear issue may be more of a firmware problem, but the Lexar issues included hardware failures. Frustrations can be seen on online forums regarding service issues. Outside of those two brands, reliability appears to be very high among the others.
The CFexpress Type B standard matured in the past year. (Sony continues to push the CFexpress Type A card, which is smaller and slower, but adequate for their high-throughput bodies.) In this past year, the uncertainties about how camera makers would handle CFexpress firmware and other plumbing issues have been resolved. Through a series of both card and camera firmware upgrades, a single, consistent interface has arisen, leading to high reliability. Canon drove this process with the release and ongoing firmware development of its EOS R5 camera. It actively cooperated with card manufacturers to get most vendors all on the same boat of reliable performance.
Delkin came out with an even-faster Black version of its cards in 2021. No card was faster at the time ingesting 56 MB RAW files in 30 seconds while shooting in 20 frames per second – our comparison standard. It is an 8K video-swilling memory monster, but is pricier than Delkin’s Power series, which ties for the win on bang-for-buck. It remained on top until…
Wise launched its Pro line. Its movement into the competition is notable in that its quite expensive Pro cards are as fast as the very fastest cards, and the normal Wise cards provide the same value-to-performance ratio of the best-in-class Angelbird and Delkin Power cards.
But then, in late March of 2022, Angelbird refreshed its lines with new cards that just barely edged the others in performance. Just two months later, Delkin hit back with its new Black 325 GB, which remains – for now – the fastest we’ve tested.
In the meantime, Acer and Sabrent came out with very respectable value-for-money cards.
This “comprehensive review” is still highly-focused on stills rather than video. While there were component and firmware issues with keeping up with the minimum throughput on early card models, the higher-end cards all handled high bandwidth 4K video pretty well. Newer models of cards appear to be designed specifically for this performance and brag about their ability to keep up with 8K recording. The newer cards released are able to record the full 30 minutes an R5 can record 8K internally – even the cheaper Sabrent ones.
When choosing a card for video, the primary factor is inevitably the reliability of that minimum write speed. Shoots are interrupted when a camera drops a frame, or worse, stops shooting due to a card hiccup. This single factor makes us recommend the following cards over the others: Angelbird AV Pro XT; Delkin Black; Prograde Cobalt; Sony Tough; Wise Pro. None of those cards ever dropped a frame as we shot 4K HQ for more than a year. Of those cards, the Angelbird Pro XT 660 GB card is significantly cheaper than the others at the time of this writing, with its Mark II version a couple percent faster and a bit more expensive.
The other big issue for CFexpress cards and video is the heat they create. For the first time, we’ve tested all of the cards available to us for how long they’ll shoot 8K video on a Canon R5. Three stand out as being able to reach the 30 minute recording limit; the most recent Angelbird and Delkin releases, along with the two new Sabrent offerings.
Aside from the Sabrent cards, these lines of cards are the more expensive ones, unfortunately. In some cases, the same brands have cheaper lines that work almost as well. For instance, we spent a great amount of time shooting Delkin Power cards because those are the cards one of the two review authors purchased after initial publication. There were times in that year where he saw throughput stop a video recording when shooting 4K HQ. The Delkin Black cards would not have had that issue. But, given that the Black cards are 40 percent higher in price, and most of that shooting was for personal projects rather than professional work, he indicated he likely would not have changed anything had he the opportunity to do it all over.
Cards optimized to maximize the minimum throughput consistency also appear to have better stills shooting cadence – the consistency with which pictures are taken at a constant rate, even when shooting with a full memory cache.
WHY CADENCE MATTERS:
A “good cadence” means that the pictures that are taken are spaced out relatively evenly. When you have frames taken at 12 per second, that’ll cover most needs. Certainly 20 FPS covers almost all needs. But if you’re trying to shoot after you’ve already filled your buffer, some CFexpress Type B cards will be better than others at being ready for the next shot. If the camera starts-and-stops, you might not be able to time critical action.
A good example showing the difference is the two different lines of Delkin cards. The Delkin Power cards provide 99 percent of the same throughput as the Delkin Black cards, but it is the Black cards that will have a consistent frame rate while shooting when the buffer is full. The Power cards will shoot at full FPS for a few seconds, and then stop for a few seconds. This doesn’t make for great sports photography.
To account for this factor, we’ve developed a new measurement to help compare cards. We call it the maximum choke (MC). This is the largest number of seconds the camera stops taking pictures while shooting with a full cache over a 30 second period. To arrive at this figure, we first shoot for the period of time needed to fill the cache (typically about 4 seconds in electronic shutter or 12 seconds in one of the mechanical modes for the Canon R5), and then start timing the interruptions that occur in the next 30 seconds. The cards tend to be consistent with their maximum interruptions, and range from 1.5 second to 3 seconds of interrupted shooting.
PANASONIC AND NIKON PERFORMANCE:
Camnostic published a review of CFexpress cards as they perform in Panasonic S1R and S1H bodies. The take-away: those cameras have throughput limitations that makes CFexpress card choice less important, the best providing only 1/7th better throughput than the XQD cards those cameras supported natively.
A similar performance test was applied to the newly-firmware-updated Nikon Z7 II, which can be found here. The firmware for these older-design Nikons needs to be updated before the XQD port can do double duty as a CFexpress card port. As it is, the Nikon’s new application of the standard provides only about 1/3rd additional speed over the old XQD method. The Nikon Z9 has not yet been tested, but we expect it will shed many of these limitations.
We also recently published a similar review of SD cards here.
– If you want the best price per GB while still shooting almost as many pictures over time, you’ll have one of the larger Delkin Power or Angelbird cards. The gargantuan 2TB versions of those cards cost less than 40 cents per gigabyte sometimes (deals come and go). We’ve found that pricing for those cards tends to sit at a higher price for part of the time, and then go through weeks or months at a time at significantly lower prices. The more-reasonably-sized 512GB cards cost just below a buck a gigabyte, with the Wise card being the cheapest of them at $0.82 per GB.
– If you want the most number of shots in 30 seconds, you’ll have the Delkin Black card, with the Angelbird Pro cards, Wise Pro 320 GB or the Delkin Black 512 GB following closely behind. These cards barely eclipse their competition by just a shot or two.
– The card that clears the buffer the fastest is the SanDisk Extreme Pro 512GB, clearing it more than half a second faster than the other cards of its capacity, closely followed by the new Wise Pro card.
– The best all-around card is arguably the ProGrade Cobalt, which doesn’t win outright in any category, but essentially ties for first in several of them.
– The card that takes the most shots before the buffer kicks in while shooting in mechanical shutter is the Delkin Black – but by just a smidge.
– The cards that perform best as hard drives hooked up to the computer are the Sony cards. They’ll offload images and video faster than the others.
Biologists call a flock of CFexpress cards a “kidney” of cards
Some of the tests below were conducted when some cards were unavailable, so some measures do not show data for the complete list of cards. With that said, let’s get to the measurements….
It has to be said that the CFexpress format is a leap forward. The cards are all so fast that the difference between the top performer in clearing a full buffer and the card that came in second from last is just 1.5 seconds. The buffer is that on-board, super-fast memory that fills up, and then you’re just left with the raw speed of your memory card, chewing through what the buffer can upload to it. That the memory cards can eat up the full buffer so quickly now has implications for people thinking about which factors are most important. If clearing the buffer is the difference between 4.5 and 6 seconds, it might not be as big a priority as it use to be. In the old days (2019), the difference between cache clearing in one card to another could easily have been more than 15 seconds. When shooting SD cards on the Sony A9, it seemed you could do your laundry and have a little time left for a snack while the buffer unloaded to the memory card. With CFexpress, buffer clearing times are just less important. All that said, the SanDisk 512GB card is tops at clearing the cache, as the chart to the right shows. (The Lexar card was eliminated from consideration because, while speedy, it eventually failed. More on that later.) This is the time it takes to completely move the contents of the cache to the card. In the R5, this appears to be about 2 GB of data, or about 40 full RAW files. (The cache will appear to be about 60 shots deep because fast cards will allow you to fill up about half again as much data during the time it is uploading to the memory card. This number – reported in the viewfinder for those who have it set so – will vary mostly based on the ISO setting, as that will affect the size of the files.)
Since this comparison review of CFexpress cards was first published, we’ve added data from additional cards. The Mark II of the Delkin card is a new firmware version. The Transcend 512 GB card was added (and much liked) and the Lexar card was added (and, later, subtracted as it failed and the company could not replace it). Also added were the two types of Wise cards and the Delkin Black line.
This is Walt the blue-winged teal, our test subject. We have MORE THAN 500K shots (And Growing) of him.
A critical factor – perhaps the most important for sports and wildlife shooters – is the number of shots you can rip before you get the unpredictable stutter of a full cache. In mechanical shutter at 12 FPS, the R5 will give you between 10 and 15 seconds of glorious, uninterrupted shooting. This is the sum of the buffer and the number of shots a card can manage to ingest while the buffer is filling. In our original analysis more than a year ago, the Delkin Power card consistently took the lead here. Since then, the Delkin Black card beat it, as did the Wise Pro card.
Of interest here is the fact that Sandisk’s smaller card sizes – which sport very similar performance metrics on the label – perform terribly relative to the 512 GB card. Unfortunately, this seems to be the norm. Sony’s smaller sizes also perform poorly in the other tests relative to the 512 GB, although they are keeping up with their bigger sibling in this test of initial speed capacity.
An early version of our cache depth test showed much more variation among the cards, but we discovered that running tests very quickly, one after the other, would increase a card’s temperature, and this write performance would be throttled. It appears from talking to reps at multiple manufacturers that this is something that will be true for all CFexpress cards. It also appears to be the reason you don’t see people brandishing minimum write speeds, because it is very temperature dependent. It also may be because an “effective minimum write capacity” number that indicated a speed below 40 degrees celsius would be embarrassingly low – likely around 550 MB/sec. for the best cards, which wouldn’t look cool emblazoned on a card label.
The good news is that there are very few people who will use their cards more intensively than we did for these tests, and the cards very rarely got hot enough to affect performance. After shooting about 3,500 stills in a row, formatting the card after about every 300, we saw some performance degradation that may have been heat related. This is likely an area that video shooters will be more concerned about. We did keep track of temperatures after this, monitoring how much the temperatures rose between each run of 30 seconds of shooting. It was so consistent among the cards, it’s not worth reporting other than to say that we’d see a roughly 2 degree celsius increase per run, and simply plugging 5 numbers into a computer between runs and formatting the card would provide enough time for the camera’s card slot to lose 1 degree. We never saw odd performance changes below 50 degrees C, and stills shooters are unlikely to see those temps. Even after several thousand shots in a few minutes, the cards were typically measured to be only in the mid-40s. Stills photographers won’t need to worry, but video people may find some cards lower writing capacity below the bitrate required for 8k shooting once a certain heat level is reached. More likely, the R5 will stop recording due to its own heat detection firmware routines. Video users of the R5 will likely be using an external recorder for high bitrate clips longer than a few minutes, so the CFexpress performance won’t be relevant to them. That also has the benefit of removing recording time limits. The Ninja V external recorder today costs roughly the same as some of the 512 GB cards in this review.
Another alternative to using CFexpress cards and putting that heat internally in the camera can be found with the just-released ZITAY XFexpress to SSD Converter, reviewed by Matthew Allard over at Newsshooter. Camnostic also tested an external heat-sink-plus-fan device from Tilta, which we reviewed here.
Shots Per Half-Minute
Here is what happens when you set a camera on electronic shutter (20 frames per second) and mash that trigger for half a minute. The scale of this chart starts at 250 images, which exaggerates the differences between the cards. All the cards perform decently, even the pokey SanDisk ones of 256GB and below.
Here is where you see the smaller cards in the Sony system also start to falter, even though their labels claim equivalent performance with the 512 GB version.
The affordable Delkin Power card surprises here by coming in 8th after winning the buffer test that showed how many uninterrupted frames it could take before the buffer caused skipping. It appears to be a sprinter of a card, perhaps with firmware that is more cautious on heat throttling, but that is just speculation. The Transcend card, too, sprinted just as fast, and wound up with the poorest performance in the remaining seconds after the buffer was filled.
The chart above shows the performance of the cards in mechanical (first curtain) shutter mode. This is favored by many with the R5 because its maximum of 12 frames per second is a reasonable rate when you’re not trying to capture the fastest movements. It’s also not very loud for a mechanical shutter. The R5 forces electronic shutter users into the 20 fps rate, which often is a drain of time later when culling pictures. By and large, the cards keep their relative rankings. The scale of this chart changes to a minimum of 200 frames, so the two charts of how-many-frames-in-30-seconds aren’t as different from one another as might initially seem.
Below, is a table of the claimed maximum write and read speeds of some the cards. As discussed earlier, dynamic heat reactions prevent there being a real minimum write speed figure, but even so, these maximum speeds show the manufacturers being pretty generous with themselves. These numbers are an industry convention that doesn’t have a great deal to do with actual card performance.
What about Video?
Video people really need a binary answer of yes or no… Does this card write the desired video format and not skip frames? With all of the cards in the first tier of performance, they can write 4k video until they are full; or more likely, until the R5 gives them an overheating timeout. Generally, the more expensive of the newer cards are advertised as 8K-friendly, while the bulk storage cards that are cheaper for the gigabyte do not make such claims. That said, none of the Angelbird, Delkin, Wise, largest Sony, Prograde Cobalt, or largest SanDisk cards every failed to deliver at least 10 minutes of 4K HQ video before stopping. That cannot be said for the Transcend, ProGrade Gold or smaller sizes of the Sony and SanDisk cards.
Of course, if you’re using the R5, and you’re at all serious about video, you’re recording to an external recorder, and the memory card is precisely the thing you’ve removed and put aside. There is some thought that some CFexpress cards may generate more heat than others. At the request of a forum dweller over on FredMiranda.com, we measured before and after temperatures during a test run of all of the cards, making sure the camera’s slot had cooled to at least a common maximum temperature (45 C). Keeping the temperature to that or lower, we saw no measurable degradation in stills performance at higher temps.
Price mattered to start, of course, but when you figure out that the performance on these cards is pretty consistently good, price takes on an even greater importance as a differentiator. More interesting to most people are the developments at the other low end of the speed graph. Delkin and Angelbird have been reducing the price of their very large capacity cards. While these cards may sit in the middle of the speed rankings, they are still pushing 90 percent as many photos onto the card per second as the top performing card. And they cost about a quarter the price per gigabyte.
The average price per gigabyte for CFexpress Type B cards lowered from a bit over $1.20 back when the R5 launched the demand for the format to around $1.10 in the fall of 2021, and now in 2022 sits below $1.05. This is taken from data collected periodically from major online retailer pricing for the same cards over time.
Compared to other formats, the CFexpress Type B card prices are getting better faster. Pricing for SD UHS-II cards has been stagnant over the past two years. CFexpress Type A, popularized with some major Sony body releases, has only recently seen competition come into the market. ProGrade and Delkin joined Sony in producing the cards, but the new competitors didn’t underprice Sony by much. And the CFexpress Type A cards – both intrinsically slower and smaller capacity – started off at about twice the price per gigabyte as CFexpress Type B cards.
Also new to the market is a do-it-yourself kit sold by Sintech (below), which provides a CFexpress Type B case into which a separately-sourced SSD drive (2230 nVME chips removed from game consoles work well) can be inserted. It requires specialized – also not included – electronics heat paste, and might not be for those uncomfortable with the entrails of electronic devices. But it does provide the cheapest CFexpress Type B option. After sourcing parts from Turkey, the US and China, we put together a 512 GB card that cost us in all about $96.
The surprising thing was how fast the card is. While on the lower end, it is still faster than ProGrade Gold cards and costs less than one fifth the amount per gigabyte.
The card has been used for several weeks without incident. It does appear to get hotter than other cards, and overheated an R5 shooting 8K about twice as fast as most other cards. The internal drive is a 512 GB Kioxia (formerly Toshiba) chip taken out of a game console that was being upgraded.
CFexpress Card Reading/Writing
- There is some correlation between drive performance and actual performance in terms of throughput when used in a camera. The chart here shows a comparison of the drive performance (orange) versus the number of frames that can be blasted through the card in 30 seconds on an R5. These numbers are presented as indices, not actual performance metrics.
- Potential performance as a drive is much higher than actual delivered performance in a camera – by roughly a factor of 2. Even more interesting, that drive performance is only about half as fast as the claimed maximum performance. This means that the R5 is currently getting about a quarter of the performance promised by the maximum write figures.
- The correlation between drive performance and camera performance is inconsistent for some cards. This may have to do with firmware differences. Some smaller-than-512 GB cards seem to have similar performance between functioning as a drive on a computer and functioning as a memory card in the camera, where the others show a much better performance when used as a drive.
In conducting the tests for the cards hooked up to a computer, we saw some very strange results, eventually figuring out that major differences would be introduced not only by using different card readers, but also even which port we chose to hook it up to on the computer, and also what sort of cord was used. These results retested, standardized on a Sonnet dual CFexpress card reader using a Thunderbolt 3 port and cord.
Our experience was that on a USB port, the Sonnet was slower than the other two, but on the Thunderbolt port, the Sonnet was fastest yet. Combine that with different cords providing different bandwidth, and some computers having different levels of USB support on different ports, this is rife for confusion and points out that people can compare their own bandwidth results to ours to see if perhaps they could get significantly better download rates by tinkering with those factors.
[Since this review was first published, we’ve come out with a more comprehensive review of card readers.]
Unless indicated otherwise, shots were taken as RAW files at 1000/th of a second in electronic shutter. Two Canon R5 cameras were used to check consistency of results. Both mechanical and electronic shutter were used in a significant number of tests, just to ensure that there wasn’t a relative difference in performance among the cards (there generally wasn’t).
Two sets of data were taken at 250 ISO and at 3200 ISO to see if there were relative performance differences with the different sized files (there weren’t any observable). Some of the data averages are rounded to the nearest ten so we don’t mistakenly imply that our precision is better than it actually is.
Relative to one another, the cards in the same class as one another showed only about 10 percent variance in performance. There certainly is shown a different set of classes, though, when looking at the 128 GB and 256 GB cards versus the higher capacity models. Those smaller cards get an additional roughly 20 percent performance hit on average. So, with everything performing at a very, very high level, and little variation among them, factors like price and service weigh heavily. My own assessment is that Delkin Power and the Angelbird cards win here because the 512 GB versions are less than $1 per GB, and they sell larger versions that dip below $0.31 and $0.40 per GB.
But, really, you can’t go wrong with the Angelbird XT 660GB, or the Prograde Cobalt 325GB, or the Sony Tough 512 GB, or the SanDisk Pro Extreme 512 GB, or – now – the Wise 512 GB and the new Delkin series of Black cards. I’d just stay away from the sizes below those. Some of those larger cards will be a bit better in one area or another. If money isn’t an object, pick your most limiting factor and look at the relevant chart above to pick the winner. But for everyone else, just go get a big Delkin or Wise or Angelbird card. And, while you’re at it, get one of those little Angelbird CFexpress readers, so you’ll never worry about which reader can read what card. Angelbird also allows for user-executed card firmware updates if you use their card reader.
Even the people coming from the Sony A9 bodies or the Canon 1DX Mark III will find the throughput of the Canon R5 a bit mind-bending when using CFexpress cards. The SanDisk 512 card is pushing 560 MB per second to clear the cache. That’s 10 shots per second of 45 megapixel RAW files. That said, even the fastest performing card isn’t close to the claims on the labels. It is about 1/4 the capacity of the CFexpress standard (~2TB) and about 1/3 the claimed maximum write capability (~1.5TB). The implication is that better is still to come; and, indeed, over the 18 months of testing as new models came out, that’s precisely what we’ve seen. To give a sense of scale, a card functioning at the bandwidth claimed on these cards’ labels (typically about 1700 MB per second) wouldn’t even need a cache. The card would suck in the data as quickly as the camera can produce it. As it stands today, current performance gives us about half the bandwidth required to make the cache irrelevant. I expect that will happen in a camera generation or two. We’ll be watching and reviewing.
Special thanks to a few people who helped push this review along. Two Canon R5 shooters who wish to remain anonymous lent multiple CFexpress cards. LensRentals.com was able to provide a couple of the brands on short notice. Jeff, over at cameramemoryspeed.com gave great advice on figuring out methodology. Forum dwellers over at FredMiranda.com also helped figure out what tests would be most useful and helped inspire the effort in the first place. Card manufacturers Angelbird, Delkin, Sony, Sabrent, Acer and Wise lent cards, as well as very useful technical information.
– 30 –