Greg,
I'll get this test done by the weekend and post the results.

What shutter speeds did you try at, if you try slower speeds 1/24 for example does that make any difference?

thanks steven

paul

From my experience it doesn't matter much if flashes were shot with CMOS or CCD. Once you compress them for delivery on DVD you end up with a mess. OTA DVB doesn't fare any better either. I've noticed several movies where the red carpet scenes have serious macroblocking.

I wouldn't cast that example as being fake quite so fast, I did manage to get a similar effect from my EX1 too. In terms of technology limitations though this isn't one I'm personally bothered by. Compared to having a camera with grainy images or requiring a lot more lighting I'd go this route.

There may also be a way to reduce the effect in camera settings, if I still had the EX1 I'd have a play with shutter speeds etc, unfortunately though I'm still awaiting a replacement.


Paul

With the flashes the issue is they split a frame which looks like frame tearing on playback. As you say it depends how this will be delivered because compression down the line can make them worse but then not all footage will be compressed downstream, so for some it is an issue. But these aren't too bad to fix in post.

As Paul says i wouldn't dismiss the footage quite so quickly and for me personally i would just like to see where the limitations start appearing as it will affect *my* use, all i know right now is that there *are* limitations!

Having said that i wouldn't have thought shutter speeds as such should affect this because we're talking about the time taken to *read* the cmos in one go *after* the shutter has been open for X amount of time. Given a high enough read speed we should see the same skew at all speeds so if this footage is real under what settings do we see it. Same thing for Pauls footage (i only saw a frame of the fence but it was very pronounced).

Am i wrong about read speed vs shutter speed?

cheers
paul

I really don't understand so much fuss about the rolling shutter. I've been using my V1E with all sorts of shutter speed (even as high as 1/300th just for experiment sake), and frankly do not find it a problem at all - neither with the pans, nor the flashes. Yes I'm getting partially exposed frames when flashes are around, but watching the video rather than analysing single frames makes it non-issue. And Bob is right that most of "red carpet scenes" on TV do fall apart for split seconds due to the flashes, while I think most of them are shot with CCD cameras...So if anything can really spoil the picture, it's the compression and not the rolling shutter!

Cannot imagine the EX1 being any worse at this than the V1.

Indeed I am; happy to host EX1 clips but am traveling quite a bit right now and thus a little behind on email.

Here's the test I did, I'm not 100% sure of the settings I used, I think it was 720/25p with shutter at 50 (speed).

The camera was mounted on a tripod. I didn't consider it to be a problem, in fact I quite liked the way it accentuated the feeling of speed. It kind of looks the same if you do the same thing with your head!

regards

Paul.

Oops... if it really was at 1/50th, than it's definitely worse than on the V1 (and I only mean the first part, with moderate pan speed).

Cheers Paul.

It clearly does wobble, but not so that it would really show in 99.9% of circumstances.

Worth bearing in mind.

Well, that's a reasonable test to me.
Thanks Paul.
It definately has wobble.

The soccer footage we shot did not really show any issues, at least not obvious . We were panning quickly using 720 60P. So I'm not worried.

For those who are, you may wish to look the other way until Sony offers another cam with these specs that does not have these artifacts to this extent.

Also, for those who tend to use their camera for 24P motion, fast hard pans are what's avoided in the first place.

I fired my camera up last night and tried some tests and really did not see anything bad. Maybe I did not pan it fast enough.
Since I would never keep these shots for display, panning the camera with earthquake movement and showing this type of artifact does not bother me.

And if I were intentionally going for this type of movement, It would be for effect anyways.

Thanks for that Paul.

i wonder what will happen in the real world, for example hand held, where you get the skew but also the jello effect up and down. think tracking motion out of a car or from an aerial shot.

I wonder if there's some combination of settings that is making it worse though. It seems pretty obvious and yet there are plenty of people who are quite loudly claiming no rolling shutter effects.

Maybe there is more at play here than the shutter speed (which i don't think should affect it anyway)

paul

I shot some car footage using fast pans and didn't notice it at all, it's only really going backwards and forwards that highlights it and I can't think why anyone would want to do that anyway unless your covering the latest table tennis world championships maybe!

Paul
Gun and run scenes could force that 'flaw' out. I think the point is that for many(perhaps most), it shouldn't matter since it takes extreme circumstances for it to manifest. However, letting people know about this issue (and the vignetting and whatever others) allows the readers to make informed choices.
Nothing could be worse than if someone buys a camera, shoots an entire production day and finds these problems in the dailies.
The flash/strobe disruption isn't isolated to macroblocking but it also induces partial exposure of at least 2 frames (depending on the strobe timing). One can edit out 1 frame without too much problem but 2 consecutive frames will affect playback.

You can see a vertical shift effect in the V/F on 2/3" broadcast CCD cameras if you pan them quickly. You also get strange tearing effects on computer screens when you pan, so it's an issue that been around in the CCD cameras.