Ooops! Card vs. Port Performance Updated!

If you read “Are you like me?” or “Who the heck is this guy?“, you hopefully read in the very first sentence of the disclaimer: “I’m not an expert, I’m going to make mistakes (and learn from them).”  Well, here we are.

My test results confused me a bit, and I went digging for answers for why I didn’t see as much difference between USB 2.0 and USB 3.0 as I expected.  Looking at the USB Wikipedia page, USB 2.0 should be capable of 60MB/sec. (theoretically) and USB 3.0 should be 625MB/sec. (!). Since the 400X card is rated for 60MB/sec. minimum sustained transfer speed, it should pretty much fill the USB 2.0 bus.  The 1000X card has a whopping 150MB/sec minimum, so clearly you’re going to need USB 3.0 to get the performance the card offers.

So what was the problem?  Drivers.  My ThinkPad T430s was running a generic load of Window 7 that didn’t have the USB 3.0 drivers installed, so I didn’t get the speed benefit on my first run of tests.  I updated everything and re-ran the tests using the same methodology as before, though I had to use a different set of files – I’d uploaded the ones I wanted and deleted the rest.  The new batch was more files but smaller on average, totaling 2.02GB.

So how did it all come out?
Updated Results
400X SD from USB 3.0 Reader: 51:0
1000X CF from USB 3.0 Reader: 11:0

Internal Reader: 1:02

400X SD from USB 2.0 Reader: 2:29.5
1000X CF from USB 2.0 Reader: 1:09.8
The updated drivers didn’t change the USB 2.0 or internal reader results, but saw the USB 3.0 performance crush the previous tests and USB 2.0.  The 1000X card was over six times faster on USB 3.0 where it could run at full speed, just 11 seconds for over 2GB of photos!

So the new takeaways:

  1. If it sounds wrong, it probably is.
  2. Always, always keep your drivers up to date.
  3. If you want maximum performance from newer cards, USB 3.0 rocks!


Card Test: X’s aren’t the only thing that matters…

[UPDATE: If a search engine has brought you here, I found an error in my testing – a quick scan might be useful here, but the final information is found in the “oops!” update.

I recently posted about the benefits of “X’s” – one measure for measuring the performance of our digital film cards, both SDHC and Compact Flash.

Now that I have the 1000X Lexar Professional Compact Flash cards in-hand to go along with my 400X Lexar Professional SD cards and both USB 2.0 and 3.0 card readers, I figured I’d test the combinations and see how things turned out.  The results were surprising…

Here was my methodology: I had 2.03GB of JPEG files from shooting the day before.  I would attach the reader with the card and dragged the DCIM folder from the card to a new folder on the drive.  I timed it from the moment I released the mouse button until the moment the copy window disappeared.  After each test, I deleted the folder so the same amount of space remained on my hard drive, which is a 180GB SSD (which is pretty fast, so your times might vary on spinning hard drives).

First up were the cards on the USB 3.0 reader, a Kingston model with good reviews on Amazon, though sometimes I wish I’d bought the Lexar version, which looks smaller and is less cluttered with other card types I don’t care about.  The 400X SD card finished copying in 2:05.4, where the 1000X Compact Flash card took only 1:09.1, not quite twice as fast.

Next I plugged in the much older USB 2.0 reader, which was also a Kingston product.  I bought the USB 3 reader when I thought one of my kids had lost this one.  Since I knew I was soon getting a ThinkPad T430s with USB 3.0, I went ahead and got the latest thing.  Then my wife found the old reader.  Oh well, at least I have my own now.

Anyway, after running the tests, I was a little dumbfounded.  The SD card took 2:31.8, which gives USB 3.0 a decent 26 second advantage.  Since this was only 1/8 of the card capacity, that can really add up.  However, the 1000X card finished in 1:10.7, giving USB 3 a modest 1.6 second advantage.

I heard back from Lexar when I asked about incompatibility issues, and their reply was they were maintaining compatibility and would work with any customers experiencing issues., so I guess this is good news for people with USB 2.0 readers and PC’s with USB only 2.0 slots.  At some point I’ll try to re-run these tests on a USB 2.0-based machine.

The last test surprised me quite a bit.  I copied the same files from the SD card using my built-in SD card reader on my machine.  The 400X SD card was exactly ten seconds faster than the 1000X Compact Flash card!

I’m sure someone can explain the reasons why, but the lesson I take away from it is this:  just like all cards at the same rated speed don’t deliver the same performance, overall performance isn’t limited to the card.  Your reader makes and the machine itself can make a big difference, too…

Knowing what I know now, I’d probably buy 600X SD cards and slower Compact Flash cards as backup – my goal in buying 1000X cards (which are not-quite-twice the price of the 400X cards) was to get faster transfers off the cards.  My guess is 600X cards off the built-in reader would be faster than the 1000X cards for about half the price.

If your PC doesn’t have a built-in reader, the faster CF cards give you a big benefit, but if you do, the SD cards might be a faster an more affordable alternative…

What do more X’s get you?

If you’re shopping for memory lately, you’ll see a dizzying and probably confusing array of performance indicators on the label or in the name of the product.  I guess one way to know which is faster is to just buy the more expensive ones within the same brand, but that isn’t the best way to shop or get the performance you want.

One increasingly common way of denoting card speed is “X” (e.g. 133X, 400X, etc.) – which is a carry-over of the way we used to measure the speed of optical drives.  Ultimately, it isn’t a guaranteed way to measure speed, and even “speed” is a reference to two different things that usually have different performance.  So what do you you want to look at?  Read Speed and Write Speed.  What are the benefits of each one?

Write Speed:  There are a couple of related advantages to a card with a fast write speed.  If you’re shooting at high speed (maximum frames per second), you’ll be able to take more shots before you have to stop due to the camera’s buffer hitting its memory limit.  This is because the camera is able to write the shots from the buffer to the card more quickly.  When you do hit the wall, you’ll wait less time for the buffer to clear so you can shoot more photos.  You’ll notice the little light that means your camera is writing images to the card goes out more quickly.  At the end of the day, if you don’t shoot a lot of continuous shots, you won’t see the benefit of fast write speeds.

Read Speed:  Do you get tired of waiting for your photos to download from your card to your PC?  I definitely do.  I have three spawn and a wife who I constantly compete with for time on the family PC (the one with the big screen).  The longer I wait for my shots to import the more people I’ve had to defend against for the rights to my chair.  The benefit of read speed is easy.  Faster is always better – it helps eliminate the “watched pot” syndrome.  You’ll get the benefit every time you transfer photographs off your card.  A caution:  I’ve heard that 1000X cards (currently the fastest available speed, only available on CompactFlash) require a USB 3.0 port and a USB 3.0-based card reader as well.  It is unclear to me whether that is a true compatibility issue or just getting the full performance of the card.  I’ll try to get some clarification.  I haven’t heard similar issues with speeds under 1000X.  When looking at cards, the premium brands tended to have better performance here.  For example, the 16GB Lexar Professional 1000X CompactFlash card is rated for 95MB/sec. write and 150MB/sec. read where Delkin 1000X cards (the only other 1000X cards currently available at B&H) are rated for 150 MB/sec. read but only 50-80MB/sec. write depending on which capacity you choose.

Even those numbers don’t tell the whole story.  Rob Galbraith has a great collection of speed testing on a huge variety of cards.  See the results here.  What you’ll find is the premium cards like Lexar Professional and SANDisk Extreme are the fastest in their class (pretty much in that order) and then you see the less-expensive guys.  Net: when it comes to speed, you get what you pay for.  Or maybe it is you pay for what you get.  Whatever.  🙂

On the subject of compatibility, you might have questions about whether your camera is compatible with the latest cards.  We’re kind of left on our own here: camera manufacturers don’t spend too much time testing new cards on old cameras, and the information isn’t always complete or easy to find from the guys who make the cards.  So what do you do?

The good news is that for the most part, the cameras use nearly any card that fits the slot.  You might not get the highest in-camera performance possible from the card, but it will probably work.  That said, if you have an older camera – let’s say more than 5 years old – I wouldn’t bother investing money in the latest card technology or the biggest card.  My D300s, the flower of 2009 technology, works fine with the 1000X UDMA 7 CompactFlash cards from Lexar.

Is it worth it?  It really depends.  I don’t typically fill my buffer – I can shoot several seconds of RAW shots at 7 frames per second before filling up my buffer, and in JPEG mode I’ve got even more than that.  I don’t know if I’m getting any benefit while shooting from UDMA 7 (I doubt it).  But I noticed a huge bump in speed pulling my shots off the card when I moved from my 133X SD cards to my current 400X.  I’m waiting for my 1000X CompactFlash cards, and I’m expecting even faster transfers.  That is worth it to me.  If the price is too high for your budget, the good news is there are usually several tiers of performance available with lower prices to match.  Pay attention to the write and read speeds specifications, and buy the best you can afford.  For the same money, I’d buy smaller cards of better quality/speed than bigger cards from the cheapo guys.

As far as brands, I’m personally a fan of Lexar, and all my current cards are from their Lexar Professional series.  I like them because they make their own chips and have direct control of virtualy all aspects of their components and manufacturing.  Even SANDisk can’t say that.  Lexar’s warranty and support are excellent.  I wouldn’t shy away from SANDisk (who also makes some of the fastest cards along with Lexar, in their SANDisk Extreme line).  Kingston and Transcend both seem to have loyal followers.  I buy Lexar cards because I think memory is a poor place to get cheap.  If your card is bad, your photos are likely gone forever (though Lexar does have some software that can sometimes help if you have a corruption issue, formatted the card, etc.).  From my work in the technology industry, I know a little about the world of component sourcing, and the cheap guys are cheap for a reason – they are often buying chips that don’t meet the quality and/or performance specifications of the big guys, among other things.  The price differences between the premium brands isn’t much, and I think it is worth it not to worry about it.

I’d love to hear your comments – what brands do you use?  Any horror stories?  Got a good experience?  Do you prefer several smaller cards or one big one?