easy sd card speed test for dslr video

I picked up a canon t2i the other day to save some timelapse-related wear and tear from my 7D. Unlike the 7D, the t2i uses sd cards. I was really happy to find out that sd cards in general are far cheaper than CF cards, but I quickly realized that some cards are better than others, even when they have the same specs. I was inspired by this post over on peta pixel to run a quick test.  This isn’t a very broad test, but I tested all the cards I could scrounge up.

In order to do this, I just popped in an sd card and fired off 15 shots. I recorded audio of the camera firing and then compared the waveforms in final cut.. The quick cards bogged down significantly less than the slow cards. You can see the in-camera buffer get filled up after 7 shots, and then the real test begins. The gap between shots is now due to the write speed of the card. The end of the green bars represents the total amount of time that the card spent writing buffered data. Since I didn’t line them all up starting at time=zero the numbers to the right of the green bar represent the total time to shoot 15 shots.


Sandisk extreme class 10 45mb/sec  32 Gig – 9 Seconds

PNY professional 20mb/s  32 Gig – 15 Seconds

Sandisk Ultra II 2 Gig – 17.75 Seconds

Sandisk 1 Gig – 26.20 Seconds


What does this mean? Well, it shows the comparative speed of different cards I have available to me. You may get drastically different numbers based on your cards and cameras, but you will probably still show the same trend. I wouldn’t even think of using the slower smaller cards for video, but this certainly puts some data behind that feeling. For stills the speed really doesnt effect me very much, but it is crucial for video. If I’m shooting hd with a high iso (lots of detail make a hard to compress image) then I absolutely need to have a fast card. Of all of the cards tested, the Sandisk extreme class 10 45mb/sec is a clear winner for dslr video. It is 166% the speed of the pny card, despite only being 120% the price. If I were a stills-only shooter, then I would consider trading volume for speed.

In the future  I’m interested in testing out different priced cards in the same class to see what the speed / price curve looks like. Lastly, it brings up the interesting idea that a slow enough card can almost replace an intervalometer. My super old sandisk 1Gig card took almost 3 seconds between shots! That could make for an acceptable time lapse interval in a pinch. Do you have a brand that you love or hate? post it in the comments and I’ll try to see if I can test it.

Note – I know that there are much more precise ways you can actually test cards using a computer. I don’t want that. I wanted to test both my camera and sd cards simultaneously.

~ by ross on July 10, 2012.