As I mentioned in another thread, I recently bought dp2m and have been exploring its capabilities. One of questions I had was - which class of SD card would be sufficient to cover my needs? The test is not very scientific but I tried to test the cards I had laying around, just to see if it was worth it, or would buying another one slightly faster would result in no improvement at all.
Set up: The camera was put in a static position (should not really matter, but just in case). It was in S, 1/30, is100, f2.8, WB sunlight, only RAW, MF, 0EV.
Methodology:
Scenario: 1 single photo. As I press the trigger, I start the timer as well.
I measured: - how long it takes to clear the buffer (T0), - time between after the buffer cleared (returned to count 7 from 6) and red LED starting to blink (T1), - the duration of flashing LED (T2).
Scenario: 3 photos in continues mode. As I press the trigger, I start the timer. Once 3 shutter clicks happened. I release the trigger. When the light stops flashing and the entire buffer is ready to get 7 images, I stop the timer.
Scenario: 5 continues mode photos - same as 3 continues, but with 5.
Scenario: Fill the buffer (same as 3 but hold the button until the camera stops taking pictures due to filled buffer). I press and hold the trigger until the camera runs out of buffer and stops taking images. As I begin to press the trigger, I start the timer as well. I stop the timer when the light stops flashing and the entire buffer is ready to get 7 images again.
Subjects:
Sandisk SDHC Class 4 (8GB) micro
Transcend SDHC Class 6 (8GB)
Sandisk Ultra Class 10 (64GB) SDXC I micro
Half-baked attempt at analyzing: The DP2M takes 7.3 + 3 seconds to read, and places one image into the buffer + possibly does some processing on it in those 10.3 sec. After that it writes it to the card. This is why 10.3 sec (T0+T1) is constant for all memory cards, but T2 changes in accordance with a given SD card speed class. Hence, the only time that can be shaved off by using faster Sd card is up to T2 (if SD write were to happen instantaneously). T0+T1 is still gonna be there. Also for subsequent images T0+T1 overlaps with T2, so for 7 shots the formula of total time is gonna be this:
total time = T0 + T1 + N * T2,
where N is a number of shots taken (works up to 7 images. Once the buffer runs out, the formula gets more complicated, so I ignored that case. This post got tedious enough already).
The law of diminishing returns is on display here. This is old class 10 card, and it does very well against class 4. But class 6 is marginally worse than class 10. Class 6 makes it just 16% longer than class 10.
upd: i also tested iso 6400 in the case they do more processing at high iso, but the times were the identical, so I discontinued testing for that setting. Running same test 2 times showed numbers did not vary significantly so I only recorded 1 run of each card for each scenario
Set up: The camera was put in a static position (should not really matter, but just in case). It was in S, 1/30, is100, f2.8, WB sunlight, only RAW, MF, 0EV.
Methodology:
Scenario: 1 single photo. As I press the trigger, I start the timer as well.
I measured: - how long it takes to clear the buffer (T0), - time between after the buffer cleared (returned to count 7 from 6) and red LED starting to blink (T1), - the duration of flashing LED (T2).
Scenario: 3 photos in continues mode. As I press the trigger, I start the timer. Once 3 shutter clicks happened. I release the trigger. When the light stops flashing and the entire buffer is ready to get 7 images, I stop the timer.
Scenario: 5 continues mode photos - same as 3 continues, but with 5.
Scenario: Fill the buffer (same as 3 but hold the button until the camera stops taking pictures due to filled buffer). I press and hold the trigger until the camera runs out of buffer and stops taking images. As I begin to press the trigger, I start the timer as well. I stop the timer when the light stops flashing and the entire buffer is ready to get 7 images again.
Subjects:
Sandisk SDHC Class 4 (8GB) micro
Transcend SDHC Class 6 (8GB)
Sandisk Ultra Class 10 (64GB) SDXC I micro
Half-baked attempt at analyzing: The DP2M takes 7.3 + 3 seconds to read, and places one image into the buffer + possibly does some processing on it in those 10.3 sec. After that it writes it to the card. This is why 10.3 sec (T0+T1) is constant for all memory cards, but T2 changes in accordance with a given SD card speed class. Hence, the only time that can be shaved off by using faster Sd card is up to T2 (if SD write were to happen instantaneously). T0+T1 is still gonna be there. Also for subsequent images T0+T1 overlaps with T2, so for 7 shots the formula of total time is gonna be this:
total time = T0 + T1 + N * T2,
where N is a number of shots taken (works up to 7 images. Once the buffer runs out, the formula gets more complicated, so I ignored that case. This post got tedious enough already).
The law of diminishing returns is on display here. This is old class 10 card, and it does very well against class 4. But class 6 is marginally worse than class 10. Class 6 makes it just 16% longer than class 10.
upd: i also tested iso 6400 in the case they do more processing at high iso, but the times were the identical, so I discontinued testing for that setting. Running same test 2 times showed numbers did not vary significantly so I only recorded 1 run of each card for each scenario
Last edited:
