The newest cameras that are coming out have capabilities that can really tax a memory card. From 4K video to very fast frame rates like 120fps to very long burst shooting, these features need a card that can keep up.
To find out whether the memory card is up to the challenge, you really need to look further to the read and write speeds. Of these, the write speed is by far the most important for most camera uses. That tells you how fast you can get data on to the card. That doesn’t tend to matter as much when using microSD cards in smartphones or GPS units, but it does matter for cameras when you’re shooting in burst mode or recording high definition video. The read speed, which is often used in marketing, tells you how fast you can get data off the card. So in ranking the cards here, I’m focusing on the write speeds and basing them on my own tests.
It’s important to bear in mind in buying a card, though, that it doesn’t always work that you’ll get much benefit from the fastest card. It depends on what you’re putting it in. To take advantage of UHS-II speeds, for instance, you’ll need a camera or reader that’s compatible with UHS-II. If it’s not, the card will still work–they’re designed to be backward compatible in nearly all cases–but you won’t get the highest speeds the card is capable of.
The Current 3 Fastest SD Cards
Lexar Professional 2000x UHS-II U3
With UHS-II and a speed class of U3, Lexar rates this one for a read speed of up to300 MB/s (or 2000x) but, as usual, doesn’t publish a write speed other than “write speeds slower.” But this card boasts impressive sequential write speed as well and is currently the fastest SD card I’ve tested. I was able to get over 241 MB/s of sustained write speed out of it. It’s in its Lexar Professional line and comes in 32GB, 64GB, and 128GB versions.
This one also comes with its own card reader, which happens to be the fastest SD reader I’ve come across yet.
Toshiba Exceria Pro UHS-II U3
Available from: Amazon
Toshiba isn’t often one of the first brands that come to mind with memory cards, but they’ve had some impressive cards in their Exceria Pro line for a while. Performance-wise, I found this card to be very close to the Lexar 2000x. But it falls short in terms of value–it’s consistently significantly more expensive than some of the other lines. It comes in 16GB, 32GB, 64GB, and 128GB versions.
SanDisk Extreme PRO UHS-II U3
Rated for a read speed of up to 280MB/s and write speed of up to 250MB/s, I found that its write speeds fell a bit short of what I expected, and I ended up running the tests a few extra times to make sure. But it’s still a solid performer from a top-notch brand and a very good option. It comes in 16GB, 32GB, and 64GB versions.
The Detailed Test Results
These are sorted by default by descending sequential write speed, but you can click on the column headers to sort by other criteria or use the search bar to filter by brand, etc. You can also scroll the table right to get the other columns.
|Brand||Model #||Class||UHS||Tested Write / Seq||Tested Read / Seq||Tested Read / But||Tested Write / Ran||Rated Read MB/s||Rated Write MB/s|
|Lexar||Professional 2000x LSD32GCRBNA2000R||U3||UHS-II||241.4||278.0||248.3||180.5||300||-|
|Toshiba||Exceria Pro N101||U3||UHS-II||231.7||254.4||236.9||231.7||141.5|
|SanDisk||Extreme Pro SDSDXP||U3||UHS-II||196.8||257.7||222.7||7.2||280||250|
|SanDisk||Extreme Pro SDSDXPA||U3||UHS-I||89.2||94.5||85.3||77.2||95||90|
|Lexar||Professional 1000x LSD32GCRBNA1000||U3||UHS-II||82.2||156.3||141.6||58.2||150||-|
|SanDisk||Extreme PLUS SDSDXSF||U3||UHS-I||64.6||88.7||85.6||59.6||90||60|
|PNY||Elite Performance P-SDX128U395-GE||U3||UHS-I||51.5||95.2||90.2||32.6||95||-|
|Lexar||Professional 633x LSD64GCBNL633||U3||UHS-I||40.1||79.4||75.7||24.2||95||-|
Obviously, this doesn’t include every SD card available. It’s a growing list that’s a work-in-progress, and I’m adding new cards as I get the opportunity to test them. I’m putting the priority on cards that are readily available and from major manufacturers.
Real-World Speed Tests
In conducting these tests, my objective is to test the performance that we can realistically expect using off-the-shelf hardware.
There are two things I am not trying to do. I’m not trying to replicate the manufacturers’ lab tests. And I’m not trying to play gotcha and test whether the speed ratings the manufacturers claim are accurate. There are, after all, several things that can affect the speeds you can get out of cards in practice.
What I am trying to do is find out which cards perform best in real-world conditions and how they compare to each other. Because those are the things that matter to me when I’m trying to decide which card to buy. In short, I’m looking for practical speeds, not theoretical speeds.
So I’m using a real-world computer setup, not some high-end custom rig optimized to squeeze every last bit of bus transfer speed but not much good for actually processing photos and videos. There are dedicated hardware devices that exist only to test the speed of memory cards. Those are ridiculously expensive and not useful for any other purpose. Instead, I’m using readily available standard hardware that photographers might have on hand.
For the reader, I’m using a Lexar Professional Workflow SR2 over USB 3.0. It’s UHS-II compatible. Technically, it’s not actually the fastest SD reader I have. That spot belongs to the reader that comes included with the Lexar 2000x card, but for now the only way to get that reader is to buy that card, so I don’t see that as a practical real-world option for the testing.
For the software, I’m using the benchmarking suite in Digital Media Doctor by LC-Technology. All cards were new or near-new and freshly formatted. Results can vary between each test. So I’m running each set of tests three times and averaging the results.
For the computer, I’m using an iMac Retina 5K late-2014 with an internal SSD. There are faster, more powerful computers that might squeeze out higher transfer rates, but this provides a useful real-world platform that is widely used and available.
The Notes / Definitions
All speed ratings I’m using here are in MB/s (megabytes per second), which is not to be confused with Mb/s (megabits per second). Megabits per second is more commonly used by cameras and in recording video. There are 8 bits in a byte, so to get from megabits per second to megabytes per second you multiple by 8. So 80MB/s is the same as 640Mb/s. Here’s a handy conversion calculator.
X Rating vs MB/s. Some manufacturers use a more cryptic x rating in place of MB/s. Lexar, in particular, has long used this system. It comes from the old way of measuring the speed of CD-ROM drives when the standard speed of a CD-ROM drive was 150KB/s. Each x therefore equals 150KB/s.
Here’s a quick cheat sheet for converting some of the common x ratings to MB/s:
SDHC vs SDXC
SDHC (Secure Digital High Capacity) is a design specification that refers to SD cards that are between 4GB and 32GB in capacity and formatted with the FAT32 filesystem. FAT32 supports individual files up to a maximum of 4GB.
SDXC (Secure Digital eXtended Capacity) refers to SD cards with a capacity larger than 32GB. They’re formatted in the exFAT filesystem.
SD cards are given a speed class rating that refers to its category for writing data, with each category describing a real-world video recording use.
Class 2 supports SD video recording with a minimum serial write speed of 2 MB/s.
Classes 4 and 6 are designed to support from 720p and 1080p video (but not all framerates) with a minimum serial write speed of 4 MB/s and 6 MB/s respectively.
Class 10 is designed to support 1080p recording at minimum (but again, not at all framerates) with a minimum serial write speed of 10 MB/s.
U1 is designed to support real-time broadcasts and HD video (720p and 1080p) with a minimum serial write speed of 10 MB/s.
U3 is designed to support 4K video recording at a sustained video capture rate of 30MB/s.
For practical purposes, the most common currently are Class 10, U1, and U3, with U3 being the current fastest class. It is technically still possible to find older, slower cards that are class 2, 4, or 6, but most modern cameras are likely to work better with at least class 10 cards, especially for recording video.
UHS-I vs UHS-II
Newer SDHC and SDXC cards have a feature called ultra high-speed bus, which refers to the interface. So far, there is UHS-I and UHS-II. UHS-II is the newer, and potentially faster, system but adoption is still not widespread. And a UHS-II isn’t necessarily guaranteed to be faster than every UHS-I card in practice, as you can see from the test results above.
The product labeling for cards with this technology will have either UHS-I or UHS-II, or sometimes just I or II. Technically, it should be Roman numerals, but you’ll sometimes see it list with a number 1, like UHS-1, even by some manufacturers.
You can also tell them apart by looking at the cards themselves. UHS-I cards have a single row of pins on the back. UHS-II have two rows. In this example, both of these are SanDisk Extreme Pro cards, just different generations.
But it’s important to note that taking advantage of the potential speed increases of UHS-II requires that both the card and the host (eg. card reader or camera) support it. Putting a UHS-II card in a camera or host that only supports UHS-I will result in it falling back to UHS-I speeds. Ditto on putting a UHS-I card in a UHS-II-enabled device.
SD vs microSD vs CompactFlash vs CFAST 2.0 vs XQD 2.0
Some cameras come with multiple memory card slots so you can choose which to use (or use both).
The speed of SD cards has come along in leaps and bounds in the past few years. At the moment, the fastest SD cards are faster than the fastest microSD cards, faster than the fastest CompactFlash cards*, and slower than CFAST 2.0 and XQD 2.0 cards.
*This is true of the cards themselves, but some devices can read much more quickly from CompactFlash slots than SD slots.
- There are counterfeit memory cards in the marketplace, especially when you’re buying online. Buying from a reputable retailer helps minimize the risk of getting a fake.
- Memory cards are complicated electronic products. A small percentage of electronic products end up being faulty from the manufacturing process. So it’s good practice to test your card before using it in a mission critical application. Better yet, have spare/s as backup.
- SD cards are remarkably resilient, but they’re not unbreakable. And they’re not designed for long-term archival storage of photos and video. It’s good practice to download the data as soon as practical and get it backed up securely.
- Format the card in the camera rather than with your computer. It reduces the risk of formatting problems.