Review: Kingston UHS-I U3 64GB SDXC Card (As it Applies to a DSLR)

[hokiealumnus]

I recently(ish) got a higher performance DSLR camera (the Canon 70D), complete with 7FPS shooting capability. One thing that can impact overall FPS, especially is a faster memory card. Previously, the only higher capacity cards I had were 16 GB Class 6 cards, so when Kingston emailed about their new UHS-3 cards, I couldn't resist and jumped on the opportunity to review one. They kindly sent along a 64 GB UHS-I speed class 3 (U3) rated SDXC card.



Rated for a very speedy 80 MB/s write and 90 MB/s read, these cards are billed as being built for 4K Ultra-HD video capture, but they will also help when using SLRs with fast burst speed, which also have to write tremendous amounts of data in short periods. You can see the detailed specifications on Kingston's site.



As I didn't have a USB 3.0 card reader, and because Kingston is nice like that, they also sent along a media card reader so this thing could be tested not only in the camera, but when transferring files to a computer.







It's supremely easy to work. With WindowsXP firmly in the past, all you need to do is plug it in via the included cable and it's instantly recognized as a multiple-drive device. That does clutter up your computer drive display a bit with quite a few new 'drives', but that's typical of any card reader, both internal or external.

Today's competition comes courtesy a Transcend 16 GB Class 6 SDHC card that has served us well for several years now.

[hokiealumnus]

The Transcend card did allow the camera to burst at its full ~7 FPS in all modes (JPEG, RAW & RAW+JPEG), but the buffer always took a long time to clear and after the buffer was full, the burst speed dropped precipitously, to an abysmal 0.5 FPS.



The Kingston card surprisingly didn't add very much to the buffer fill time at all. It increased but not by much.

The two things that DID improve significantly were "FPS after buffer fill" (while the camera is frantically trying to write to the SD card and clear the buffer) and the 'time to clear buffer' metrics.



Yes, you read that right. It takes under one second (so fast I couldn't pretend to accurately measure) to clear the buffer in Fine JPEG mode. While it takes longer to clear RAW and RAW+JPEG buffers, time to clear RAW was measurably less and time to clear RAW+JPEG was cut nearly in half.

The "FPS after buffer fill" also increased substantially, by over three TIMES when shooting RAW only. Though still just under 2 FPS, it's a lot better than half an FPS.

Now we'll put the two side-by-side and you can see how they compare directly.

[hokiealumnus]

Of course, another big advantage to much faster SD cards is transfer speeds, so you can get to work sifting through (and post processing) your photos that much faster.

To measure transfer speed, I used CrystalDiskMark and ATTO Disk Benchmark.

*Note I made a mistake in the text running these benchmarks. The Kingston card is not UHS-3, which does not exist. It's UHS-I U3.





On average, the Kingston card is rated right where it should be. In ATTO, write speeds were just short of the rating but CrystalDiskMark shows writes exceeding the rated 80 MB/s. Both benchmarks show the card exceeding the rated 90 MB/s by a solid 6 MB/s.

The last thing to talk about is price. UHS-I Class 3 cards aren't as cheap as you may be used to in SD cards. The Kingston UHS-I U3 64 GB card, model # SDA3/64GB, runs $108.80 shipped from Newegg. The 32 GB card is a more palatable $57.71 shipped and 16 GB will run you $40.25 shipped. As any early-adopter knows, that kind of markup is what you pay for the newest, fastest hardware.

I'm not here to tell you whether it is worth it for you to buy an ultra fast SD card, but I am here to show you that if you do, Kingston's new UHS-I Speed Class 3 cards will perform right where they says they will, and they will make your ability to get the shot that much stronger - not to mention get you imported and editing that much faster.

Thanks to Kingston for sending this card for evaluation. I'm happy to answer any questions that anyone has and run additional tests if you would like; just ask!

[itznfb]

If the "Time to Fill Buffer" value for JPEG is infinity then the "Time to Clear Buffer" value should be null. Unless you're counting the time to write as the time to clear.

[hokiealumnus]

When you stop shooting in JPEG mode, it takes a second for what is in the buffer to get off the buffer. Very little is on the buffer.

[itznfb]

When you stop shooting in JPEG mode, it takes a second for what is in the buffer to get off the buffer. Very little is on the buffer.

The value for the Kingston card is strange to me. If there is enough data for it to take 6 seconds to clear then it would be safe to assume that the Time to Fill Buffer would not be infinity. It should fill pretty quickly if the card is 6 seconds latent.

It's a small nitpick in the data and it may just be how you're interpreting some of the data flow.... good write ups in both threads using the camera though. Useful set of data. Thanks for taking the time to put it together.

Do you know if they are going to release UHS-II cards any time soon?

[hokiealumnus]

That's fair. I was assuming reasonable amounts of shooting. Basically, I was over 100 shots after 15 seconds of shooting and I assumed (possibly wrongly) that nobody is really going to shoot 15 seconds+ consecutively. Anything that a sports photog would need 7FPS continuous shooting for is going to be over long before 15 seconds is up, therefore for practical purposes, it is infinity. Hope that makes sense. Plus I didn't feel like wasting 1000+ shutter count just to find out when the real limit was.

No clue on Kingston's timeline for UHS-II cards, but I'll try and remember to post up here whenever they announce the series.