(Yesterday I wrote some quick thoughts about the A1. Now I’m adding more details, especially about the global shutter sensor.)
I guess I’ll have to repeat myself again. When R5/R6 was announced I said:
“I have no doubt that Sony can make an equally powerful or even more powerful camera. I’m worrying about the overall user experience. There are other features I desire, like Olympus’s ND filter emulation, Sony’s good old smooth reflection, focus stacking needless to say, internal RAW processing and a lot more. In the era of excessive performance, these are the things that matter. And Sony has been doing a really really bad job in the past years.”
Yet here we are, exactly what I’ve expected. You can’t say that Sony has done nothing, for example they have in-body image cropping and the long overdue lossless compression now. You can also downscale your JPEG/HEIF to 21MP but somehow it doesn’t work for RAW (I mean, how difficult can it be?). But it’s far from enough, especially when they’ve done the outrageous crime of removing the flip-out screen. This one really caught me off guard and I’m a little bit scared about what’s gonna happen for A7IV now. I though the flip-out screen is 100% guaranteed but now I have some doubts.
Some people are arguing that a tilt screen is actually better than a flip-out screen from a photographer’s point of view. First of all, I disagree:
Second of all, Sony also disagrees. They know perfectly well which screen is more popular, and that’s exactly why they crippled it. This is the equivalency of removing SLOG from the A9 series. If you think it’s actually for your benefits then I’m sorry to tell you: sir, you’re a little bit retarded. We also know that there are screens that combine the strength of a tilt screen and a flip-out screen. But hey, I guess these components are too expensive for a $6500 camera.
What’s even funnier is that the buffer size is only 155 compressed RAWs, which runs out in 5 seconds. I don’t know how well the CFE-A card performs but the label says maximum write speed = 700MB/S, so in the best scenario you’re getting 14FPS with lossy compression after the buffer is fully loaded. I guess now you see the point of a downscaled RAW format.
It’s the typical old Sony that brings out a “slap your face” sensor that crushes their competitions, so they think they’re entitled to cut some corners to make sure that you can’t have the perfect camera. This approach used to work well in the past but not anymore. There is something about the EOS R5 that I think I’ve never shared before:
EOS R5’s readout speed is actually 60 fucking FPS
at the expense of the dynamic range for sure:
But I was expecting a 30FPS sensor with reduced DR. Oh boy, how wrong was I. Of course this doesn’t put R5 in the same ballpark as A1. The readout speed of A1 is at least 1/200s thanks to the integrated memory (it seems to be 1/240s so this sensor is 3.1x faster than the one in A9/A9II). But R5’s 20FPS e-shutter mode is not completely unusable either. It’s actually much much better than any A7x cameras out there (maybe except the A7SIII). Now, consider that R5 also has an excellent IBIS and a flip-out screen, which are probably going to be much more useful for most enthusiast or even some professional users (like wedding photographers), I would be hesitant to choose A1 over R5 even if the prices are the same!
OK to be fair, if you’re more demanding than most people then at least A1 can easily obliterate Canon 1DX3, which is also a $6500 camera with… a fixed screen. I think A1 does mark the death of sport/journalist DSLRs. Canon has tried every mechanical approach to extend the lifespan but 30FPS is just out of DSLR’s reach. Sony wants the sport camera market and it looks like they are going to get it. The A9II is already doing a fantastic job stealing Canon’s contract with those big clients like the Associated Press, and A1 is surely going to be the final nail in the coffin. Canon can make a 45MP30FPS EOS R1 using R5’s sensor but the target clients aren’t going to be as tolerant as me with the 1/60s rolling shutter. Unless Canon has another secret sensor coming there is no way they can stop Sony in this niche market in the future. But this has nothing to do with us mainstream users, unfortunately.
There is a more realistic approach for Canon though: the global shutter sensors. Since GS sensors have more components in each pixel, the image quality (both the dynamic range and the low-light performance) is going to take a hit, and there could be some weird artifacts due to the parasitic light. Camera makers usually try to avoid it because of the IQ issues, but if there is no way for Canon to match Sony’s integrated memory sensor then GS sensors actually make tons of senses:
- The rolling shutter is 0. It doesn’t make it any more useful than A1’s 1/240s sensor for high speed shooting, but it’s gonna be a GAME CHANGER for any strobe light users. There is probably going to be new flash lights designed specifically for GS cameras. While Sony is using some interesting technology to increase the sync speed of the mechanical shutter to 1/400s, it’s still pretty useless during daytime. Tons of people are going to pay for a global shutter camera just because of the infinite flash sync speed.
- Again… while Sony is working hard on their mechanical shutter, a GS sensor does not need it AT ALL as long as you can achieve a good parasitic light sensitivity. This is going to reduce the size/weight/cost of the camera and bring some interesting changes to the autofocus.
- Since mechanical shutters are useless for a GS sensor, Canon can abandon their dual-pixel AF technology and use Sony’s PDAF, so that the AF pixels can still work while the sensor is capturing the image. Therefore, a GS sensor doesn’t have to be super fast to reduce the AF blind time. You can make a 30FPS sensor for a 30FPS camera. This is going to level the playground for Canon because Sony’s excessive speed is going to be meaningless, except for the video specs. However, those professional users don’t really need video specs that much, and I’d say 4K120P is pretty much all we need right now.
- Many people believe that GS sensors are much more expensive than a normal sensor. This is not true. The cost is more or less similar. So it’s gonna be significantly cheaper than Sony’s super fancy sensors.
Well I think these advantages are good enough and I’ve wrote about some of them in the 索兰经. So the major downside is the image quality, but if Canon can make sure that the low-light performance is still OK maybe with the help of the gapless on-chip micro-lens then I’m pretty sure that there are tons of ways to fix the dynamic range. Even if the low-light performance can’t be fixed, desperate times require desperate measures, if Canon is desperate at all. I’ve been constantly underestimating Canon recently so who knows.
Anyway… I’d be lying to say I don’t want a 50MP30FPS camera with A9’s AF capability, but there is no way I’m going to pay $6500 for it when you can buy the R5 for almost half the price. For a normal user like me (and I’m on the more demanding side) the benefit is so limited. The ironic thing is that, at the moment it’s really the E-mount lenses that have kept me from switching. Let’s hope that Sony won’t fuck with the A7IV, and A7RV can come a little bit sooner since A7RIV is obviously a lost cause now. On the other hand it does look like that Canon has some supply issues and it’s not entirely Kung Flu’s fault. I guess their new sensors still have a relatively low yields so Sony’s not in hurry right now.
Photo Engineering 