Zim, be more practical.
The 5D III has a rather large memory buffer (bigger than most cameras). When you shoot, it's always sending data to the buffer... then writiing from the buffer to the card(s).
For "JPEG" shots, to the SD card slot, the shooting is, for all practical purposes, "unlimited" (you will fill the card capacity before it runs out of buffer space).
If you shoot "RAW" shots, then the camera WILL eventually slow down. But that wont happen until it shoots at least a dozen continuous frames (I think Canon officially claims 13 frames before the buffer slows it down -- but I think in personal usage I think I counted 18 (that's practical because RAW images do "technically" compress... they just don't use a "lossy" compression method like JPEG. In other words you'll seem some variability on this number depending on how well the images compress.)
I have BOTH an SD and CF card inserted in the camera -- generally at all times. I still get the performance equal to, or higher than, the numbers Canon claims even with both cards in the camera.
Not having the 3rd party information to directly scrutinize what they say, my inclination is to disagree (but I'd need to see EXACTLY what they claim to be certain.)
My point about the practicality is this... what are you shooting and how many frames per second do you actually require before the camera slows down. Experienced photographers who aren't shooting action photography will tend to prefer to shoot in RAW, but they usually aren't bursting off shots like crazy. Sports photographers are bursting off shots like crazy... but those who need to know that they can hold the button down for many seconds at a time will set the camera to JPEG mode specifically because the shooting is, for all practical purposes, unlimited (even when using the SD card.)
If you REALLY need "action" photography camera performance... a 7D II or 1D X would be on my list of cameras you should look at.
EDIT: I followed the link to the BorrowLenses website to see what they "discovered" (not exactly a "discovery" - anyone who understands technology would have known this even without performing a test.)
What they claim is that if you not only insert both cards, but also USE both cards concurrently (e.g. tell the camera to write your image to both cards after each shot - e.g. one card is used as a "backup" of the other) then this means the camera can go no faster than the speed of the slowest card. Sure... that's true.
HOWEVER... you wont be able to measure any difference at all... until you fill the camera's internal buffer. Then the delay is based on how long it takes to clear the buffer by writing to the cards. Just writing a single image twice (once to each card) is going to slow things down -- you're making the camera do double the work. This slow down would be percevied even if both cards were operating at precisely the same speed. I suspect this has more to do with the "slow down" than the fact that the SD card is operating in a UDMA 7 slot.
Tim Campbell
5D III, 5D IV, 60Da
