Home Search Gallery How-To Books Links Workshops About Contact
I tested these seven memory cards.
I
get my goodies at Ritz, Amazon and Adorama. Pro Tip: Always format every card in your camera (not your computer) every time you use it. This prevents the possibility of data errors and corruption. See also separate SanDisk Extreme IV Review and Lexar 300x UDMA Review. INTRODUCTION Download Play Format Shoot Compute Do We Care? I care about speed for downloading. I don't care about it for shooting, since camera memory buffers eliminate any shot-to-shot delay even with a slow card. Modern cameras have internal buffer memory. Buffer memory stores the photos you just took before they get written to your card. The buffer writes your photos to your card while you are taking more pictures. You do not have to wait for the card, unless you overflow the buffer. It is next to impossible to overflow the buffer unless you're doing something silly like shooting numerous, continuous 15 shot bursts in uncompressed raw mode. I've never been able to fill a buffer on any of my cameras, ever. Therefore I don't care about write speed in my camera. I use both fast and slow cards and have never noticed any differences in shooting. I do care about download speed. This is how long I need to wait around to get the images into my computer. Which Cards I Tested My friend, Humanitarian Photojournalist Karl Grobl, loaned me his new SanDisk Extreme IV combo, so I decided to compare it to the other cards I own. These cards are: Compact Flash: SanDisk 2GB Extreme IV (borrowed September 2006 review) SD: SanDisk Extreme III 2GB (bought September 2006, $112) The Microtech is the pig at this party. It's the only card I've ever used which has ever lost data. It's why I no longer consider cards made by anyone other than SanDisk and Lexar, the two top brands. PERFORMANCE back to top Recommendations Download Play Format Shoot Compute Download Play Format Shoot Compute This is the most important difference between cards. Downloading a card takes long enough that I wait around for it. It's not negligible, like shooting times, and it's not so long that I walk away or can concentrate on anything else. Download speeds matter. Test Procedure I copied 111 images, 168 MB, originally shot on a Canon 30D, to each card. I then timed the download from each combination. These times are on my Quad 2.5 GHz G5, bought January 2006. One test was on my older 800MHz 12" iBook G4 laptop, bought February 2004 . I dragged and dropped the files in Finder. For the Canons, which don't appear as drives when connected, I used Apple's Image Capture to transfer the images. The card reader is SanDisk's Extreme IV Firewire 800 card reader. I also talk about the card there.
Card Reader, Firewire 800 Download Times, Seconds (111 files, 168MB)
Holy Cow! The Extreme IV completely smokes any other card I've used, by at least a factor of two!
Card Reader, Firewire 400 Download Times, Seconds (111 files, 168MB) Same card reader as above, except connected to a Firewire 400 port.
Good news! It's just as fast with Firewire 400. The Extreme IV is still over twice as fast as anything else I've tried.
Card Reader, Firewire 400 (iBook) Download Times, Seconds (111 files, 168MB) Same card reader as above, except connected to a Firewire 400 port on my 12" iBook.
More good news! It's almost as fast on my laptop's Firewire 400 port.
Nikon D200 via USB Download Times, Seconds (111 files, 168MB)
The party's over. Over the USB connection from my D200, everything is slower. The speed advantage of the Extreme IV is lost - it's no faster than the Extreme III using my camera's USB connection.
Canon 30D via USB Download Times, Seconds (111 files, 168MB)
The 30D is a little slower than the D200 for downloads.
Canon 20D via USB Download Times, Seconds (111 files, 168MB)
The 20D is much slower than the 30D. This is a good reason to get the 30D.
Nikon D70 via USB Download Times, Seconds (111 files, 168MB)
Forget using the D70 for downloads. What took under eight seconds in the Firewire card reader takes over three-and-a-half minutes via my D70's old USB 1.1 connection.
Nikon D80 via USB Download Times, Seconds (111 files, 168MB)
Good news! The D80's USB connection is as fast as the card reader with Firewire 400. The Extreme IV isn't available in SD (yet.)
Canon SD700 via USB Download Times, Seconds (111 files, 168MB)
More good news: Although not as fast as the D80, the SD700 pocket camera is faster than any DSLR tested using CF cards.
Playback, in-Camera back to top Recommendations Download Play Format Shoot Compute How fast do images flip forward and back, how fast do they come up, scroll around, and zoom? Nikon D200 No difference among cards. Nikon D70 No difference among cards. Canon 30D No difference among cards. Canon 20D No difference among cards. Nikon D80 (SD cards) No difference among cards. Canon SD700 (SD cards) No difference among cards. There appeared to be differences when playback was zoomed and I used a trick mode to move between images at the same zoom, but these turned out to be differences in files sizes. In zoom mode the SD700 has to read the entire image from the card, and on Canon some JPGs are much larger than others. Tested this with the same images and you'll get the same results. Playback Summary: Cameras aren't slowed by reading from the cards. They are slowed by their own internal processing. I can't see any differences among these cards for playback functions.
Formatting back to top Recommendations Download Play Format Shoot Compute Pro Tip: Always format every card in your camera (not your computer) every time you use it. This prevents the possibility of data errors and corruption. These are the times to format in-camera. They will vary depending on what's already on the card. Nikon D80 In my D80 every card formatted in less than a second. Nikon D200 In my D200 they formatted in less than two seconds. Canon SD700 My Canon SD700 can do a more thorough "low level" format that takes a little longer. Here's what it took, in seconds:
Write Speeds, Shooting back to top Download Play Format Shoot Compute Recommendations As I explained at the top, write speeds aren't important unless you set your camera to something silly, like shooting sports in uncompressed raw. Any of these cards, even my ancient ones, are fast enough for responsible shooting. If you're shooting so fast and with files so huge that you need fast cards, you probably lose far more time transferring and processing it all later. Test Procedure I made six shots with each card and camera combination. I measured the time from when I first pressed the shutter to when the CF light extinguished. I turned image review off. Don't compare speeds between cameras: each camera made different file sizes.
Nikon D200 Write Time, Seconds
(1MB Large Basic Optimal Quality JPGs, 15.2MB uncompressed raw.)
Nikon D70 Write Time, Seconds My D70 buffer is only 4 shots in raw, so my D70 paused before shot 5 and shot 6. No big deal, it was writing the whole time anyway.
(1.3 MB Large Normal JPGs, 5.4MB compressed raw.)
Canon 30D Write Time, Seconds
(1.5MB JPG Large Normal JPGs, 7.5MB raw.)
Nikon D80 Write Time, Seconds
(2.5MB JPG Large Normal, 8.3 MB raw)
Canon SD700 My Canon SD700 writes so fast to its card that I can't time it. It writes as fast as it shoots for normal JPGs. At my normal settings, ISO Auto, Large Normal JPG (950 kB files) every card ran for at least 20 shots at 2 FPS, no problem. SD700 Buffer depth at 2FPS, Typical 1MB JPGs
SD700 Buffer depth at 2FPS, ridiculous 4.75MB JPGs Trying to break something, I set my SD700 to make the most ridiculously huge files I could: 4.75MB to see if I could fill its buffer. I did this at ISO 800 and superfine JPGs. Here's how many shots I could make with each card at 2 FPS before the buffer got full:
Even after I filled the buffer, it still ran at about 1 FPS. Write Speeds, in Computer back to top Download Play Format Shoot Compute Recommendations This is if you move files from your computer back onto the card. This isn't important, but I timed it since I put the same set of files onto each card to test download speed. Test Procedure I copied 111 images (168 MB) originally shot in a Canon 30D, by dragging them in my Mac's Finder. I used the SanDisk Extreme IV Firewire 800 CF Card Reader. I uploaded to the SD cards in my Nikon D80. Upload Times, Seconds (111 files, 168 MB in FW800 card reader)
*96.68 via Firewire 400 port
Upload Times, Seconds (111 files, 168 MB via a Nikon D80 and USB)
WHERE THEY'RE MADE The Lexar are made in USA. The SanDisk are all made in China. The Microtech is made in Japan. This is where the cards I have are marked. This changes weekly, so don't worry about it. RECOMMENDATIONS back to top Performance Download Play Format Shoot Compute All cards are fast enough for DSLR shooting, unless you're doing something bone-headed like shooting long continuous action sequences in uncompressed raw. Download speeds do vary, but only if you use the fastest cards and external readers. By the time you've yanked a card out of your camera and jammed it in the reader, you've probably fallen behind just reading it from your camera. To get the fastest speeds from the fastest card, the SanDisk Extreme IV, you need to use its special reader. If you're part of a pro shooting operation with runners grabbing your cards and downloading them for you, the SanDisk Extreme IV will save your assistants a lot of time. If you're like me and download from the camera, there is little benefit to the fastest cards. Get the Extreme IV if you want the best. It's so inexpensive today there's no reason not to. I paid more for a no-name, unreliable 256MB card in 2002 than the 2GB Extreme IV card costs today. Go ahead, live it up. If you want to save money, you'll probably never miss anything with a slower card. PLUG If you find this as helpful as a book you might have had to buy or a workshop you may have had to take, feel free to help me write more with a donation. Thanks for reading! Ken |
Home Search Gallery How-To Books Links Workshops About Contact