5dM3 dual card write speed results

[msdarkroom]

I did some 5dm3 testing and came up with some write speeds for the SanDisk memory cards. Full size raw for all tests. High speed continuous.

The cards I used
SD: http://amzn.to/IxaUC2
CF: http://amzn.to/KrRKE4


Dual CF/SD
1 through 14: 0.225 seconds each
Image 15: 0.6 seconds
Image 16 and higher: 1.35 seconds

CF only
Images 1 through 20: 0.225 seconds each
Images 21 and higher: 0.55 seconds each

SD only
Images 1 through 15: 0.225 seconds each
Images 16 and higher: 1.3 seconds each

 

[PhilDrinkwater]

Thank you for doing this. I'll be using 64gb cards but the same spec as the ones you're using so I really appreciate the results. They look fairly encouraging to me. Shame the 5d3 doesnt support uhs cards but I doubt it'll ever affect me in reality...

 

[msdarkroom]

Sure thing. I know I had told somebody I would test this in a different post. Better late than never!

 

[Mt Spokane Photography]

How did you conduct the test? Were the cards erased beforehand (Low level format as opposed to in-camera format).

It takes much longer to write to a card if the memory site being written to must first be erased, since a whole block of memory has to be erased to write just one bit.

Test results can be highly skewed if its not done right,

 

[msdarkroom]

Hi Spokane,

IMO most people format their memory cards with their camera. I formatted the cards in camera, did the test, looked at the results, used the camera to format the cards again, and then did the test again. 6 tests in total (2 for each group).
This is how I tested the cards and came up with the write speeds that I believe most people can expect for the cards listed above when using them in real world situations.

-MS

 

[Old Shooter]

Very interesting! Your results agree with Rob Galbraith's!

http://www.robgalbraith.com/bins/camera_wb_multi_page.asp?cid=6007-12452

Basically, the fastest CF cards will write 4x faster than the fastest SD cards...

If you want to maximize your burst rate, Lexar 1000x is the way to go!

Save your money on the SD card; even the Extreme Pro's can't hit 20MB/s on the 5DIII...

Wonder how this would have worked out if Canon had enabled UHS-1 on the SD slot?

 

[Mt Spokane Photography]

Hi Spokane,

IMO most people format their memory cards with their camera. I formatted the cards in camera, did the test, looked at the results, used the camera to format the cards again, and then did the test again. 6 tests in total (2 for each group).
This is how I tested the cards and came up with the write speeds that I believe most people can expect for the cards listed above when using them in real world situations.

-MS

Unfortunately, the results will vary based on how much use the cards have had since they last recieved a complete format.

The CF card has electronics that look for blank unused memory, but the SD card doesn't and can really slow things down. Thats why the in-Camera format offers low level formatting for the SD card.

I prefer to see valid tests done by professionals who understand how to test.

 

[PhilDrinkwater]

Best to leave testing to professionals who know what they are doing.

Harsh!!

 

[PhilDrinkwater]

Very interesting! Your results agree with Rob Galbraith's!

http://www.robgalbraith.com/bins/camera_wb_multi_page.asp?cid=6007-12452

Basically, the fastest CF cards will write 4x faster than the fastest SD cards...

If you want to maximize your burst rate, Lexar 1000x is the way to go!

Save your money on the SD card; even the Extreme Pro's can't hit 20MB/s on the 5DIII...

Wonder how this would have worked out if Canon had enabled UHS-1 on the SD slot?

I ended up with mid level sd cards - well the lowest level 64gb sds I could get. I looked at robs results and decided there wasn't much point getting the fastest ones. Saved me a bunch of money

 

[Mt Spokane Photography]

Best to leave testing to professionals who know what they are doing.

Harsh!!

Yes, I changed it a few seconds later. Having run a test lab, I'm a bit hard on sloppy testing.

There are good websites like the one run by Rob Galbraith who tests almost all of the memory cards and cameras, and does it right. This test could be accurate or way off, since we had a unknown condition of the cards. Erasing or doing a low level format before running the test has the benefit of finding and fixing any card errors as well as starting from a known point.

http://www.robgalbraith.com/bins/multi_page.asp?cid=6007

 

[msdarkroom]

This thread, started by myself, was created with the intent to answer a question I had been asked previously but was slow to respond to. I went ahead and found the original question.

It's here: http://bit.ly/JaBy4p
The question was asked by vWings. I believe it was the 6th or so entry on this thread. My response was about the 10th entry.

I conducted some tests that, in my opinion, would be helpful to people who were looking at buying the SanDisk Extreme cards shown above. I conducted them using what method I believe most people use: Shoot. Import. Format in camera. Shoot.

It's unfortunate that it turned into what it did.


-MS

 

[Ryant]

Darkroom I am glad you made the topic. I did not know about Rob Glbraith's site until people here brought it up. I did some texts with my lexar 600x and I wondered if I could get a bit more out of the lexar 1000x which I found out is a burst of 54 vs 75 so thats a big difference.

The part I dont get is if rob galbraith is testing both the SD and CF cards together. Darkroom gave a test that used both which I have not seen much of. I personally decided to skip using a SD card because I dont want saving to it to slow down shooting or limit shooting in any way.

The other thing that I feel rob galbraiths test skips is how many quick burst pictures you get before the camera memory clogs. I found the lexar 600x 32 gig card let me take a burst of 18 RAW only before it slowed down.

Ryan