CFExpress Cards Speed Tests

[Kayaker72]

Thanks for the info, Brant. Like you, I tend to shoot in short bursts so probably a smaller card would be fine. Based on the file size of the R6 (slightly more MP than the R3), even a 64 GB will hold close to 3,000 RAW images. I can’t imaging shooting that many frames in a session, even at 30 fps, and I generally prefer more cards than bigger ones. If I shot video I may feel differently…but I don’t. So given that, I’ll probably pick up a pair of 64 GB CFe cards and see how it goes.

I also need to sort out the SD cards. I have a pair of 64 GB cards that I swap for my EOS R (they came in a bundle with the camera, along with a couple of video accessories that I gave away). However, they’re UHS-I cards and while that wasn’t an issue with the R, I think they’ll be too slow for the R3. I’ll pick up a pair of 64 GB UHS-II SD cards and see how they perform with writing RAW simultaneously to CFe and SD. If that slows it down, I’ll write just to the CFe and periodically Image Copy the files over to the SD during downtime (to mitigate against card failure).

I’ll definitely enjoy the R3, whenever it comes. In the meantime, I have an RF 70-200/2.8 showing up next week.

Regards to the RF 70-200 F/2.8, nice! Everything about it seems good, but the size and ability to pack it upright in particular makes it appealing. I am not sure what my first RF lens will be and when I will start making the transition, but that lens could easily end up in my bag.

After my last post, I took a look at the R3 specs a bit more. Canon has the buffer depth rated at 150 RAW images. Assuming 30 MB per file for the R3, 60 MB/sec UHS-I, 180 MB/sec UHS-II and 500 MB/sec CFE, you are looking at 2 fps, 6 fps and ~17 fps after the buffer has filled and 75 sec, 25 sec, 9 seconds to clear the buffer. If you ever fill it.

As Joel proved, UHS-I cards will really work just fine for most of us even in action instances. Most of us just will not take that many images that quickly. That said, UHS-II does seem to strike a nice balance.

[neuroanatomist]

Regards to the RF 70-200 F/2.8, nice! Everything about it seems good, but the size and ability to pack it upright in particular makes it appealing. I am not sure what my first RF lens will be and when I will start making the transition, but that lens could easily end up in my bag.

My first RF lens was the 24-105/4L, which I bought along with my EOS R. That was before the 24-70/2.8L IS came out, and I now need to decide if I replace my EF 24-70/2.8 II with the RF version. For this sort of thing, I find analyzing the EXIF data from my library to be very helpful. As a Mac user, I used to use Aperture for my library organization, and I had an app called Aperture Inspector that graphed out usage by lens, focal length, aperture, etc. I was disappointed when Apple killed Aperture, and I was forced to use Photos for my library (I've been reluctant to move to LR even though I use PS CC – mostly because I prefer DxO PhotoLab for RAW conversions but it doesn't do well with library management). Photos is bearable. I found a Mac App Store download called PhotoStatistica that performs the same analysis and graphing functions as Aperture Inspector.

IMO, the f/2.8 vs. f/4 choice is more difficult with RF lenses. For a higher-end DSLR, an f/2.8 lens means better AF performance (more cross-type points, more high-precision points) and a brighter viewfinder, and those benefits apply regardless of aperture choice. For the R-series bodies, with DPAF and the EVF, the difference really just comes down to how much benefit one derives from the more light and greater OOF blur of a faster aperture. For the 70-200m focal length, it was almost a no-brainer since over 50% of my shots with the EF 70-200/2.8L IS II were taken wider than f/4.

For the EF 24-70/2.8 II, about 25% of my images are wider than f/4, which is still a pretty substantial fraction. But there the trade off is not only size/weight/cost, it's 70mm vs. 105mm at the long end. A bit deeper analysis shows that with the EF 24-70, about 20% of my images are taken at 70mm, but looking at the resulting image sizes only a total of about 6% of the images with the lens are cropped beyond the 1-2 MP loss from correcting rotation and minor framing adjustments, so most of the time a 70mm long end is fine.

But coming at it from the other direction, with a 24-105mm lens about 25% of my images are in the 70-105mm range (applies to the EF lens when I had it as well as the RF lens). So if I have the extra 35mm, I use them about as often as I use the wider aperture. Does that mean I need both? Probably I'll end up buying the RF 24-70/2.8 and seeing which I use the most. If I had to bet, I'll use the 24-70 more but take the 24-105 when I travel.

After my last post, I took a look at the R3 specs a bit more. Canon has the buffer depth rated at 150 RAW images. Assuming 30 MB per file for the R3, 60 MB/sec UHS-I, 180 MB/sec UHS-II and 500 MB/sec CFE, you are looking at 2 fps, 6 fps and ~17 fps after the buffer has filled and 75 sec, 25 sec, 9 seconds to clear the buffer. If you ever fill it.

As Joel proved, UHS-I cards will really work just fine for most of us even in action instances. Most of us just will not take that many images that quickly. That said, UHS-II does seem to strike a nice balance.

I'd probably be fine with the UHS-I cards I have for my EOS R, but I already ordered a pair of 64 GB CFe and a pair of 64 GB SD UHS-II.

[Kayaker72]

If I had to bet, I'll use the 24-70 more but take the 24-105 when I travel.

That makes a lot of sense to me. In all the analysis, my primary thought beyond where it is used, is what could you live without? The missing 35 mm at the long end of the 24-70 is easy to compensate for by cropping. Sure, lost of pixels, and noise becomes more apparent if at high ISO, but it can be done with minimal penalty. Plus, I often have another lens in the bag that could be used to compensate.

But f/2.8 vs f/4. If you need that stop of light, if you want that bokeh, there is not an easy replacement.