Stunning AF demo of the R8 by the Fro

R2D2 wrote:

Alastair Norcross wrote:

R2D2 wrote:

Alastair Norcross wrote:

antonio-salieri wrote:

Laqup wrote:

antonio-salieri wrote:

But EFCS is good for situations where you want to avoid shake, but still want full-bit depth images/to avoid some ES downsides, and ES gets you the highest frame rates (though it's better on the newer cameras than on the R5 for rolling shutter). In short... it's still good to have all three options. Having only two is a downside. But it's not a dealbreaker anymore.

The R5 is as good as the R6 II. The R6 is a bit slower and the R7 is really slow.

Ah, yes, it does seem the R5/6ii/8 have similar ES readout. Then the R7/10/50 are slower. The R3 is faster, though.

The R6II is a bit faster than the R5, but the difference probably isn't noticeable. I find even the 'slow' R7 to be more than fast enough for my shooting. Out of about 10,000 shots taken so far with e-shutter, I have noticed rolling shutter distortion in precisely 3 of them. It all depends on what you are shooting, of course, but too many posters here say silly things like 'the R7 in e-shutter is useless for fast-moving subjects'. That's patently false, of course. It all depends on the angle and direction of the movement. I have lots of e-shutter shots of fast moving subjects that show no rolling shutter distortion at all. I love the e-shutter on my R7, and find it very useful indeed. I just wish I could set a burst rate somewhere between the slow 3fps and the (usually) too fast 15fps. The 10fps of my old 7DII was fast enough for most of my action shooting.

I actually get quite a lot of rolling shutter in my R7 action shots (esp seen in the backgrounds while panning). Just something one has to be aware of.

It’s one of the reasons that the R7 just didn’t work out for me as an action shooter. I admit to shooting a lot of fast moving subjects (birds, planes, bikes, skiers, indoor sports, and so-forth). Fortunately I have other bodies that can do the job.

Yes, it all depends on how much panning you do and what is in the background when you pan. I haven't done much (any) panning with prominent verticals in the background, which is mostly what would show up.

Sure. If you don’t ever pan, then you won’t see it in the backgrounds. Only in subjects occasionally.

Also, of course, if the background is blurry enough, and the verticals aren't important, even if they are slanted, it doesn't matter.

I beg to differ. Slanted backgrounds get noticed, whether it’s a stand of trees, a fenceline, gym walls, utility poles, or buildings in the background, even when substantially OOF. This phenomenon is usually very hard (and time consuming) and often impossible to correct in post.

That's why I mentioned the vertical not being important. Your examples are all of verticals that people would expect to be vertical, because they would be recognizable, even when OOF. Not all background elements are familiar to viewers.

Someone recently posted a shot with a blurry signpost in the background, with slanted verticals. He said that the actual sign isn't slanted. But it didn't matter at all to the shot. Who cares or knows whether those verticals are actually slanted? It didn't look at all unnatural.

If you want to simply dismiss it, go ahead. Doesn’t mean it’s not obvious to other viewers. Being free of rolling shutter effects is absolutely more preferable to having them.

I agree, and I'm not dismissing it. I cite that example, because no-one who didn't know about that particular sign would have expected it to be vertical, as opposed to slanting. That's why the poster had to specifically tell us that the actual sign wasn't slanted. We didn't know that in advance. Obviously, buildings and trees are a different matter (though I have seen some dramatically slanted trees on the California coast).

R5 and R6ii are much better, with only occasional rolling shutter effects.

R2

Does anyone have the R8’s sensor readout figures?

If it's the same sensor as in the R6II, as everyone seems to be saying, won't it be the same readout speed?

It’ll be interesting to see how it’ll perform. The Fro’s samples do look promising.

R2