Why Card Performance Matters
There are 4 primary aspects to be considered when choosing a memory card:
- Reliability – this is discussed extensively in our Choosing
the Right Memory Card, and Protecting your Data article.
- How long it takes to upload your pictures to your computer – manufacturers claim read/write speeds, which they obtain
under optimum conditions using the fastest card readers and computers, individual results may vary significantly depending
on the computer and reader used.
- Card size, making sure you have enough to cover your shooting needs, and one big card verses multiple small cards is
also discussed in our Choosing the Right Memory Card, and
Protecting your Data article.
- How the card performs in the camera. For the landscape shooter or other users that carefully compose each shot, card
speed is unimportant. For anyone that shoots action, card speed can be critical.
This section focuses on how different cards perform in different bodies.
Testing In-Camera Card Performance
All modern DSLR's have onboard buffers – basically RAM memory, just like your computer. When pictures are taken they are
written directly to the buffer, the camera applies any processing (sharpening, noise reduction, saturation etc), and then
the image is written to the memory card. Memory cards are much slower than the buffer, so if you are shooting at a high
frame rate, you will more than likely fill the buffer at some point. The camera continues to move the images from the buffer
to the memory card as quickly as it can, even while you are shooting.
In the real world, how important is the in-camera card speed? Twenty years ago, I was shooting a Formula 1 race at
Silverstone in England with Nikon F Photomic's and manual focus lenses. I'd pre-focus the lens on a certain part of the
corner, and shoot cars as they went through. A second or so later, I'd have wound onto the next frame, steadied the camera
and would be ready for the next car (and possibly missed a car or two that passed in the mean time). If a car slid mid
corner, chances are I'd miss it. If it crashed, it would take a second or so to refocus manually and you'd miss some of the
action. With today's DSLR's with their advanced autofocus systems, the camera's can easily track cars and refocus on the
next car in a split second, and the ability to shoot high frame rates significantly increases your chance of getting that
shot of the car sliding, the overtaking maneuver, sparks flying from under the car and so on.
While you are unlikely to just hold down the shutter and keep shooting for extended periods, taking a burst of 6 shots,
followed a few seconds later by an 9 shot burst, then another burst shortly afterwards is not uncommon when there is a lot
of action. If you are shooting RAW, your camera probably gives you less than 20 shots before the buffer fills. By the third,
fourth or fifth burst, your buffer will likely fill. This is where card speed comes into play – when the buffer fills, and
it invariably will, what becomes important is (a) what kind of frame rate you can expect while the buffer is full, and (b)
how long it takes the camera to empty the buffer to the card. This is what we've set out to test in this article.
Compact Flash (CF) Card Performance Results
The following cards were tested:
 
The SanDisk Ducati, Sandisk Extreme III and Lexar 300x are all UDMA cards. UDMA (Ultra Direct Memory Access) is an
interface that allows a higher transfer rate of date to the camera body, providing the body supports the UDMA interface.
UDMA cards will work in non-UDMA cameras (like the Nikon D200), although they won't be able to take advantage of the UDMA
feature. Cameras like the Nikon D300 that supports the UDMA interface can still use non-UDMA cards.
The following table highlights the summary of results of these cards with different bodies:
| Card |
Card Size |
Actual Size* |
Card Speed** |
UDMA |
Nikon D300*** |
Nikon D200*** |
| SanDisk Ducati |
8 GB |
7.62 GB |
45 MB/s |
Yes |
27.62 MB/sec
2.09 fps
6.87 sec
More |
9.69 MB/sec
0.68 fps
29.82 sec
More |
| SanDisk Extreme IV |
8 GB |
7.62 GB |
40 MB/s |
Yes |
27.13 MB/sec
2.08 fps
7.17 sec
More |
9.57 MB/sec
0.67 fps
30.29 sec
More |
| Lexar Professional 300x |
8 GB |
7.45 GB |
45 MB/s |
Yes |
22.83 MB/sec
2.01 fps
9.28 sec
More |
8.54 MB/sec
0.58 fps
35.21 sec
More |
| SanDisk Extreme III |
8 GB |
7.62 GB |
20 MB/s |
No |
10.90 MB/sec
0.80 fps
18.89 sec
More |
7.74 MB/sec
0.51 fps
37.95 sec
More |
| SanDisk Ultra II |
8 GB |
7.62 GB |
15 MB/s |
No |
5.86 MB/sec
0.33 fps
34.82 sec
More |
5.31 MB/sec
0.35 fps
56.98 sec
More |
| * | The actual available space on the card
after formatting. |
| ** | Card write speed as claimed by the
manufacturer. |
| *** | The MB/sec number is the maximum measured
Megabytes per Second transfer rate from camera to card, the fps number is the maximum number of frames per second achieved
with a full buffer shooting RAW, and the time in seconds is how long it took to clear a full buffer after shooting stopped.
Click on the link in the table for full details and explanation. |
As expected, the UDMA cards are faster than the non-UDMA. The SanDisk Ducati was the fastest card in just about every
test. Despite being rated at the same speed as the SanDisk Ducati, the Lexar 300x was actually slower than the 266x Sandisk
Extreme IV in most cases. All the cards are rated at 8Gb, however the Lexar also had almost 200MB less space available for
images than the SanDisks.
To see the full report and detailed results for a specific body, please click a link below:
Nikon D300
Nikon D200
Secure Digital (SD) Card Performance Results
Coming Soon for the Nikon D90.
| 
