RX100 iii lens issue

[stever]

resolution test today against my old RX100 and GX7 +12-35.  Started at 70mm eff and the RX100 is not great - but better than the iii which also acted like the autofocus was inconsistent.  next 35 eff where both were good - as good in the center as the GX7 but not in the corners (but still respectable for 3x2 format) although again the iii was less consistent than the RX100.  Finally tried 50 eff and the RX100 had 30% better center resolution at f4 than the iii.  I didn't test wider as my Imatest target is really not large enough for me to trust the results.

guess i'll have to send it back and try again in a couple months - that should teach me to pre-order.

otherwise, the evf is great, but I'd be happier with a thinner, more robust camera without the selfie lcd

[Ray]

Thanks for reporting that, Steve. There's been a lot of talk about the benefits of the wider maximum aperture of the RX100-3 at 70mm, and other benefits such as its built-in ND filter option, but I've seen no test comparisons of the resolution of the lens.

DXOMark have tested the RX100-3 sensor, but haven't yet tested the lens. However, they have tested the lens in the RX100-2 and have given it a P-Mpix sharpness rating of just 6, which is the same as the P-Mpix rating for the old 8mp Canon 20D used with the EF 17-40mm/F4.

I was hoping that the RX100-3 lens might get a higher P-Mpix rating of at least 7, which would put it on a par with the 15mp Canon 50D used with the EF-S 17-55/F2.8 lens. I'll have to wait and see. I'm in no hurry.

[stever]

decided to go ahead and test at 28mm although only 2ft from the target (I have tested other WA in the past and the results seem to be meaningful for comparison purposes).  wide open the iii substantially out-resolved the RX100 (wide open is f2.2 at 28mm for the iii) - from f4 to f8 they were essentially equal.  Both had a bit higher resolution than the GX7 with 12-35 or 7-14 up to f8 where the GX7 was a bit better as might be expected from sensor/pixel size.

I take DXO's results as general guidance but they sometimes vary significantly from me measurements and experience.  my 20d with 17-40 or any wide angel lens couldn't come close to the wide angle resolution of the Sonys

if Sony can get the lens adjustment under control (I doubt that poor resolution at 50mm and above is a design problem) the RX100iii will be worthy of the hype.  Keep in mind that the aperture advantage is really only at 70mm (f2.8 vs f4) - both cameras are f2.8 wide open at 35 and 50mm, and as mentioned, the iii is already down to f2.2 at 28mm. 

[eronald]

I had an RX 100-wonderful, broke after a week. Replacement, pretty bad, sold.
The problem  with Sony seems getting a decent sample. I guess DxO and dPreview get superstars.
The other issue is that Jpegs and even Raw conversions are heavily dependent on digital corrections, these break if your sample lens is off-center etc.

Edmund

[Ray]

Lack of good quality control has always been a problem. When DSLRs had low pixel count, like 6 or 8 or 10 mp, I would test lenses after buying them, and return them if they weren't up to my expectations.

I would much prefer the manufacturers to do their own testing immediately after assembling the lens, or camera with in-built lens. They could then divide their results into 3 segments, and charge accordingly. Highest price for the best performers; medium price for the average performers, and the cheapest price for the 'almost' rejects.

I guess the reason they don't do this is because they don't want to advertise the fact there is a quality control problem.

[Peter_DL]

I'll stay far away from the iii
due to the very negative experience I made with the ii
in terms of chromatic aberration / color fringing.

Peter

--

[Dale_Cotton2]

DXOMark have tested the RX100-3 sensor, but haven't yet tested the lens. However, they have tested the lens in the RX100-2 and have given it a P-Mpix sharpness rating of just 6, which is the same as the P-Mpix rating for the old 8mp Canon 20D used with the EF 17-40mm/F4.

DxOMark lens testing on the RX100 and Mk2 only reports on the lens without distortion correction applied. Given OOC JPEGs are corrected, as is LR and ACR output, it's rather difficult to access the uncorrected output, so not sure how relevant their results are. 6 MP certainly doesn't reflect my experience with Mk2 output via Lightroom. Sure: there are some soft corners and some softness at both extremes of focal length, but to me not worse than some mythical average lens. And we already know from test image samples that the Mk3 lens is noticeably better all around.

[Ray]

DxOMark lens testing on the RX100 and Mk2 only reports on the lens without distortion correction applied. Given OOC JPEGs are corrected, as is LR and ACR output, it's rather difficult to access the uncorrected output, so not sure how relevant their results are. 6 MP certainly doesn't reflect my experience with Mk2 output via Lightroom. Sure: there are some soft corners and some softness at both extremes of focal length, but to me not worse than some mythical average lens. And we already know from test image samples that the Mk3 lens is noticeably better all around.

Dale,
As long as their methodology is consistent, then DXO results are useful. However, if one doesn't have a means of adequately correcting the chromatic aberrations etc, due to deficiencies in the RAW converter, then there may be a problem.

Last time I checked, DXOMark have not produced results for either the RX100, or the Mk 3. Only the Mk 2.

The main difficulty I see with their P-Mpix ratings is that one cannot be sure if the same lens has been used on all models of the same brand of camera.

For example, the P-Mpix rating for a specific model of lens on an 8 mp Canon 20D is sometimes no different from the P-Mpix rating for the same lens on a 15 mp Canon 50D, which seems a bit odd since the 50D has almost double the pixel count of the 20D. I can only presume that it was not the same lens that was used in both tests, just the same model of lens.