Yes, the constant press of technology just keeps making you have to do things that you don’t want to. In today's case study: give up all those CompactFlash cards you’ve been acquiring over the digital DSLR era.
As pixel counts and frame rates went up and video streams demanded higher and higher minimum guaranteed transfer rates, CompactFlash eventually hit its ceiling. UDMA-7 basically maxes out at a data rate of 145-167MB/s, but that’s only attainable through a fast interface such as USB 3.0. Internally, the parts that control writing to the card in the camera tend to max out at 33MB/s, or sometimes 66MB/s for a higher cost part.
CFast uses the CompactFlash physical body, but replaces the internal ATA bus construct of CompactFlash with SATA, which allows up to 600MB/s but requires different write controllers in the camera. XQD uses a PCIe bus construct, and the original version allowed 500MB/s; XQD version 2 now allows 1000MB/s.
It’s not a surprise that as the pro cameras got more sophisticated and demanded more bandwidth out of their buffer-to-card systems that the original CompactFlash design started showing strains. Nikon appears to have chosen XQD for their future, Canon appears to have chosen CFast. It seems pretty obvious that the old CompactFlash card is going to die off soon. Secure Digital (SD) has more than enough horsepower for the lower end cameras for the time being. SD with UHS-II allows up to 312MB/s (via a four lane bus), which is better than CompactFlash can do without moving to CFast or changing something significant in the CompactFlash specification, thus creating another medium.
I really don’t expect to see CompactFlash in future Nikon designs. The high end seems now focused on XQD, while the rest of the models should do fine with SD (and probably UHS-II as cameras require more bandwidth). As Nikon has demonstrated with the Nikon 1 line, for the truly compact cameras we’re going to get microSD cards.
It’s not like we haven’t had card changes in the past. CompactFlash originally came in two sizes, Type 1 and Type 2. Some interesting things came about in the slightly fatter Type 2, including the IBM Microdrive, a really small spinning hard drive. But even if you had bought CompactFlash Type 1 cards back in the early days for your D1 series models, by the time we got to the D3 you needed new cards to keep up with the camera (UDMA). We’ve seen similar moves in the SD world, too.
You may remember an article I wrote long ago (Tip of the Iceberg). The cost of a camera isn’t just the camera. It’s all the things that you pile on top of that acquisition. In the case of many of the latest cameras, that means:
- New cards (e.g. faster, or even of a different type)
- New software (to support the new raw formats)
- Possibly new computer (to support the new software and larger file sizes)
- and so on
All this is the price for admission, folks. From the camera makers’ standpoint, it’s sort of a no win situation for them. They have to improve their product to get people to update, since updates are becoming more of a percentage of their sales than new user acquisition. But in improving the product they risk having some just decide to continue to use their existing perfectly fine product for longer periods, partly due to these added costs of upgrading.
So, if you think you’re going to stay current in the Nikon ecosystem in the future:
- Pro camera users should probably start buying XQD cards
- Consumer camera users should probably start acquiring UHS-II SD cards
There’s no hurry to do either of these things unless you’re buying a D5 or D500. Nevertheless, adding new CompactFlash—or SD cards that aren’t at least UHS-I—is probably something you don’t want to do unless you absolutely know that you’ll still have cameras around by the end of the year that use them.