Problem with RX100 images stabilization

[YiannisPP]

I concur. And I don't even have an RX100! Just thinking of buying one and spent the last hour browsing full scale samples in flickr. And I just noticed blurred edges at shutter speeds of up to 1/100sec, in many different photos and that brought me here...

I know this is just impossible with any Canon unless you're trying to shake the camera. In fact I get away with 0.5sec with my SX220, about half of those come out perfectly sharp. 1/10sec and it's almost a certainty there's no hand shake recorded at all...

And I was ready to place the order...might still do but that's dissappointing.

 

[B1ackhat]

YiannisPP wrote:

I concur. And I don't even have an RX100! Just thinking of buying one and spent the last hour browsing full scale samples in flickr. And I just noticed blurred edges at shutter speeds of up to 1/100sec, in many different photos and that brought me here...

I know this is just impossible with any Canon unless you're trying to shake the camera. In fact I get away with 0.5sec with my SX220, about half of those come out perfectly sharp. 1/10sec and it's almost a certainty there's no hand shake recorded at all...

And I was ready to place the order...might still do but that's dissappointing.

The best I can do handheld with any of my P&S cams (including my Canons) is 1/30. It sounds like I would really struggle with the RX100. :x

 

[Skuto]

For what it's worth, some testing with Stabilization On/Off does show a significant difference, so it's certainly doing something. I think I'll try some more controlled test shots to get an idea how many stops it's worth.


I think the fact that the camera has a 20Mpix resolution means that it's harder to get sharp images if you are going to pixel peep. If you compare with a 5Mpix camera, you can obviously "shake" the latter twice as far to get the same amount of pixels blurred, for a similarly composed shot.

So maybe the formula 1/focal length isn't a good guideline if you're going to look at such a high resolution image at 1:1 pixel level.

 

[tkpatric]

I am very curious about Sony's Steady-Shot image stabilisation. I expect minimum two stops or I buy the G15.

 

[BertIverson]

FWIW. I took a reference shot from a tripod with 2 sec delay at 1/15. Lens at 100mm. I next shot a bunch, of the same frame, hand held at 1/15. Another sequence at 1/30 and a final sequence at 1/60. With my old hands, I can definitely say 1/15 definitely looked somewhat blurry. at 1/30, most of the shots looked OK and the 1/60 shots were all fine. All viewed at 100%. So for me, 1/15 at 100mm is definitely not desirable. I will be happy with the 1/30 shots. I have NOT done a similar test with my A700 DSLR
Bert

 

[Skuto]

Stable 1/30 at 100mm is pretty OK, would be about 1.5 stops? (But I think it's actually more due to the very high resolution, which is why I wanted to test the feature directly)


If you can do stable 1/30 at full zoom I think I need to cut back on the coffee.

 

[BertIverson]

Skuto wrote:

... If you can do stable 1/30 at full zoom I think I need to cut back on the coffee.

Well at 100%, some of the 1/30 shots showed some blur. Others (pure luck) seemed excellent. Of course, if one can fill the frame with optical zoom, there is no worry with slight blur seen at 100%. If my shot will require digital zoom, I will stick to 1/125 or faster shutter. Sunny days seem to be 1/250 at f5.6.

I'm quite happy

Bert

 

[kazkioken]

I find the stabilization to be pretty amazing on this camera, especially in video. As far as shooting f1.8 @ 1/30 I find great results like this:



--

500px

www.500px.com

 

[Skuto]

I attached a note to my wall, zoomed in a little (37mm equiv) and fired a bunch of shots with OS on and off. I tried to pick representative ones, though its hard because of course you don't really shake in a constant manner.

I got usable results with OS up to 1/8s (the 1/15's were perfect). Without OS I started getting good shots at about 1/30s.

So the stabilization from this quick little test seems to be, roughly, 2 stops, which IIRC is similar to what Sony claims.


I'm going to withdraw my criticism based on this. The OS works and obviously provides >1 stop worth. I guess my problem is the combination of a very high resolution (20Mpix) combined with a small, harder to grip body, which makes it harder to get completely (up to 1:1) unblurred pictures.

 

[tkpatric]

Thank you! I wonder if there is a scientific method to measure the quality of image stabilisation...

 

[technic]

Skuto wrote:

This doesn't happen with a DSLR and stabilizing lens. The RX100 isn't that, but it should be pointed out if it's going to be touted as a possible replacement. (And if other compacts have better IS/OS, well...)

Maybe you were lucky with your DSLR then? My Canon 450D with 15-85IS standard zoom is not reliably sharp at 1/30 either, even at 1/60 I still get some slightly blurry shots when working with WA or standard focal length.

 

[technic]

Michael S. wrote:
I have one shot here with f1.8 and 1/125s (S-Mode) used with flash of some young people...I have to say that I don't have the original jpg I made - this has already some sharpening - about 90%, 0.8 and 3 settings in PS Elements.
I don't find it that bad for fully open aperture:

Cheers,
Michael S.

looks soft (or wrongly focused) to me, but that could be because of the lighting. Totally impossible to judge sharpness from such a low resolution shot ...

 

[technic]

YiannisPP wrote:

I concur. And I don't even have an RX100! Just thinking of buying one and spent the last hour browsing full scale samples in flickr. And I just noticed blurred edges at shutter speeds of up to 1/100sec, in many different photos and that brought me here...

I know this is just impossible with any Canon unless you're trying to shake the camera. In fact I get away with 0.5sec with my SX220, about half of those come out perfectly sharp. 1/10sec and it's almost a certainty there's no hand shake recorded at all...

And I was ready to place the order...might still do but that's dissappointing.

yes, a compact (without mirror slap etc.) should be able to do better. This shrinks the advantage of the big RX100 sensor even more in real life (other factors are needing to stop down for sharp corners and 1-2 stops smaller aperture at tele setting compared to some other quality compacts). Maybe the RX100 IS doesn't work well under certain conditions (this occurs with many different cameras ...)?

 

[YiannisPP]

I guess the problem with prospective buyers that have owned a recent Canon (like myself) is that Canon offers a true 4 stop IS advantage. Well not really a problem, but something to take into account.

 

[Docno]

YiannisPP wrote:

I guess the problem with prospective buyers that have owned a recent Canon (like myself) is that Canon offers a true 4 stop IS advantage. Well not really a problem, but something to take into account.

You mean the Canon s100 IS offers a 4-stop advantage??

 

[Skuto]

At least the official S100 specs claim 4 stops. I don't own one but the responses here seem to indicate it's probably not much if any exaggeration.

 

[Skuto]

Your Canon has 2x the pixels my old DSLR had. The RX100 has 4 times as much. This matters and I am now pretty sure its the reason for the discrepancy. When checking blur I'm looking at the zoomed in 1:1, not the zoomed out picture.


While Googling for the official claimed stops on the RX100 I found a 3 month old thread almost identical to this one.

Same conclusions, too: IS worth about 2 stops, being less than some other compacts, and this while the camera has more Mpix. More careful selection of shutter speeds recommended.

 

[John TF]

Cameralabs review:

"To test the RX100's stabilisation, I zoomed the lens to its 100mm position and took a sequence of shots in Shutter priority mode at progressively slower shutter speeds with SteadyShot both on and off. As you can see from the crops below, SteadyShot enables hand holding at speeds down to 1/10, a little over three stops slower than convention would suggest is safe without stabilisation. It's worth noting that the RX100's stabiliation isn't quite as impressive as the Lumix LX7 or PowerShot G1X, and I'm almost certain that's a consequence of its small size. One advantage of the bigger, heavier models is it's easier to hold them steady. "

 

[BertIverson]

John TF wrote:
... and I'm almost certain that's a consequence of its small size. One advantage of the bigger, heavier models is it's easier to hold them steady. "

Or because the RX100 has 20MP and slight movements can be more easily seen when pixel level viewing. That's my theory anyways. -- I guess one could add some lead to the bottom to make it bigger and heavier ;-)
Bert

 

[technic]

Skuto wrote:

Your Canon has 2x the pixels my old DSLR had. The RX100 has 4 times as much. This matters and I am now pretty sure its the reason for the discrepancy. When checking blur I'm looking at the zoomed in 1:1, not the zoomed out picture.

the pixel number matters, but don't be fooled. When I moved from my 6 MP 300D to the newer 12 MP 450D DSLR, I had far more problems with camera shake. In fact, the 450D with an IS lens like the 18-55IS (which has excellent IS) would produce more blur than my 300D with a Sigma 2.8/18-50 non-IS, at similar shutter times.

First I also thought it was because of the higher resolution, but downsampled to 6 MP the problem is still very visible. I'm now convinced this is because the 450D is lighter, has a less solid grip and a faster shutter/mirror mechanism that causes far more vibration than the 300D. I think much of this applies to the RX100 as well: low weight, difficult to hold firmly, relatively big shutter mechanism for the camera size. The IS just needs to do more work compared to 1/1.7"sensor compacts ...