Opticallimits
Tests of the SEL 1855 - Printable Version

+- Opticallimits (https://forum.opticallimits.com)
+-- Forum: Forums (https://forum.opticallimits.com/forumdisplay.php?fid=4)
+--- Forum: Sony (https://forum.opticallimits.com/forumdisplay.php?fid=14)
+--- Thread: Tests of the SEL 1855 (/showthread.php?tid=904)

Pages: 1 2


Tests of the SEL 1855 - karelderaat - 02-09-2016

Thanks Klaus for this info. So maybe all those horror stories about the 16-70 have to be taken with a pinch of salt because te real culprit maybe the sensor. Would be nice to see whether positive stories about this lens are restricted to the 16mp sensor.




Tests of the SEL 1855 - Klaus - 02-10-2016

Quote:Thanks Klaus for this info. So maybe all those horror stories about the 16-70 have to be taken with a pinch of salt because te real culprit maybe the sensor. Would be nice to see whether positive stories about this lens are restricted to the 16mp sensor.
 

The problem is, however, that the 24mp sensor is mainstream today.

The A5000 has a 20mp sensor at the low end of Sony's range but the 16mp one is gone (except among users, of course).

 

Another aspect is that a lower mp sensor is much more friendly to lenses.

 

Higher mp = higher magnification of the flaws and a bigger spread between center (which can still scale) and borders (which struggle). The latter also is easily visible in the current EOS 5Ds R tests as well (stellar center, comparatively mediocre borders). 

 

Finally there are question marks regarding the future of Sony's APS-C camera segment. Sure, they released the A6300 (at an insane price) but unless I am mistaken they didn't bother to release new APS-C format lenses in 2015 (did they even release something in 2014?).  It's also worth to point out that the A6300 roughly costs as much as the A7 mk I so there's a bit of a message there as well .... 



Tests of the SEL 1855 - Brightcolours - 02-10-2016

Quote:The 16-70 was as soft on the A6000 as on the NEX7.

 

At some stage people complained that I should abandon the NEX7 so I bought the A6000.

I did some tests and there was no indication that something changed.

Some people claim that there's a difference but I reckon this is related to different focus settings.

Thus I sold the A6000 again because the viewfinder was worse.

FWIW, initially I even bought a 2nd NEX7 to verify the results but at no effect.

 

Maybe things have changed on the A6300 ... there's always another hope. ;-)
The 16-70mm f4 tests badly on every body, it is the least liked standard zoom across formats. So, it seems that its bad results were due to itself and not the nex7 issue, which does ont turn up with every lens to the same amount.

 

The A6000 does have a different sensor, but you never really wanted to dig into that as far as I remember (you did not have lenses that had a real issue with the nex 7 available when you had the A6000?).

The new A6000 sensor:

http://www.verybiglobo.com/sony-alpha-a6000-iso-comparisons-corner-smearing-color-casting/

Whether the A6000's sensor is without issues, or has just less issues than the nex7 only methodical testing could tell, and I have neither camera.



Tests of the SEL 1855 - karelderaat - 02-10-2016

I have to be careful here, because I am really no expert lens tester (far from that), but it seems to me that the increase of resolution by an increase of mp AS SUCH should not result in less brilliant lens-testing results as regards sharpness/resolution.

 

However, the Imatest does apparently show that the new high mp sensors do result in less brilliant results. This would mean that the only reason why we aim at high mp counts, i.e., higher resolution and by that more detail, did not come through. So, in general, there is an optimum mp count per sensor type??????

 

OR is an increase in mp count gained at a cost in another aspect of sensor quality, resulting in blurring the theoretical gain in resolution away? That would be a very bad thing. And how does that explain the fact that the difference between the edges and the centre increases? I may hope that there is no quality gradient over the sensor. I agree, all speculation, but should we ask Sony to produce an A6000/6300 with 16 mp? Or is the Imatest fooling us? I am puzzled ....

 

I fully agree with your remarks about Sony and it approach of the APS-C NEX/A E-mount range. If they will not come with a serious gamma of lenses very they kill this system. It is no option to let APS-C pay the extra costs of full-frame lenses. I must say that if I had to start again, I would not choose the Sony APS-C E-mount system again. However, to be honest, my practical photo experience is not really suffering from these flaws.



Tests of the SEL 1855 - Brightcolours - 02-10-2016

Quote:I have to be careful here, because I am really no expert lens tester (far from that), but it seems to me that the increase of resolution by an increase of mp AS SUCH should not result in less brilliant lens-testing results as regards sharpness/resolution.

 

However, the Imatest does apparently show that the new high mp sensors do result in less brilliant results. This would mean that the only reason why we aim at high mp counts, i.e., higher resolution and by that more detail, did not come through. So, in general, there is an optimum mp count per sensor type??????

 

OR is an increase in mp count gained at a cost in another aspect of sensor quality, resulting in blurring the theoretical gain in resolution away? That would be a very bad thing. And how does that explain the fact that the difference between the edges and the centre increases? I may hope that there is no quality gradient over the sensor. I agree, all speculation, but should we ask Sony to produce an A6000/6300 with 16 mp? Or is the Imatest fooling us? I am puzzled ....

 

I fully agree with your remarks about Sony and it approach of the APS-C NEX/A E-mount range. If they will not come with a serious gamma of lenses very they kill this system. It is no option to let APS-C pay the extra costs of full-frame lenses. I must say that if I had to start again, I would not choose the Sony APS-C E-mount system again. However, to be honest, my practical photo experience is not really suffering from these flaws.
You don't get it yet. The nex7 sensor has a flaw, be it the sensor glass, the micro lenses or both, which makes it perform badly with lenses which let light fall on the borders at certain angles. Normally, one would never see lower numbers when testing with higher resolution sensors. Especially with mirrorless with short flange distances and bigger angles of light with certain lens designs, in combination with careless sensor design, we can see problems.

 

Problem is, it depends on the lens tested whether such an issue shows up. The NEX5 and Nex7 both have issues, the Nex 5R had it resolved and the A6000  doe snot seem as affected either. 

 

If the camera designers take care designing the micro lenses and sensor stack, there is no limit to the sensor resolution (meaning, a high res. sensor will not show lower MTF figures than a lower res. sensor).

 

By the way, Sony is not the only one making such errors. Leica had a similar issue with one of the M series (forgot which one) where corners would get wrong colours. And there are a few reports that the Canon EOS M3 can show some green discolouration in corners (although so faint that many users report not seeing any issue):

http://www.cameralabs.com/reviews/Canon_EOS_M3/

 

But as badly messed up as Sony did with the nex7, that is special.



Tests of the SEL 1855 - chrismiller - 02-10-2016

karelderaat you need to bear in mind that imatest numbers are not absolute, they are relative to the resolution of the sensor (line widths per picture height). With an increase of resolution of the sensor the test score of a particular lens that is not out resolving the sensor will drop. Of course the amount of detail captured would be the same

 

Quote:OR is an increase in mp count gained at a cost in another aspect of sensor quality, resulting in blurring the theoretical gain in resolution away? That would be a very bad thing. 
Increased noise from smaller pixels - see sony a7s / nikon d4 etc have low resolution sensors (12MPix & 16MPix) but the best low light performance.

 

color casts and edge smearing also apply to sony 35mm sensors with wide angle lenses - eg a7r + voigtlander 15mm. Voigtlander have actually released a new version that solves the issue (with an increase in size) now.

 

Chris




Tests of the SEL 1855 - karelderaat - 02-10-2016

Thanks Brightcolours, the angle argument makes me 'getting it'. 




Tests of the SEL 1855 - Klaus - 02-10-2016

Quote:The 16-70mm f4 tests badly on every body, it is the least liked standard zoom across formats. So, it seems that its bad results were due to itself and not the nex7 issue, which does ont turn up with every lens to the same amount.

 

The A6000 does have a different sensor, but you never really wanted to dig into that as far as I remember (you did not have lenses that had a real issue with the nex 7 available when you had the A6000?).

The new A6000 sensor:

http://www.verybiglobo.com/sony-alpha-a6000-iso-comparisons-corner-smearing-color-casting/

Whether the A6000's sensor is without issues, or has just less issues than the nex7 only methodical testing could tell, and I have neither camera.
 

 

I am aware of this blog - this  is inconclusive. The blog was actually the reason to purchase the A6000 and check what's going on.

 

The edge coloration can be compensated via post processing the camera. It doesn't necessarily originate in a better sensor.

 

I did some tests with the A6000 vs NEX7 at the time. There was no indication that the results weren't any better on the A6000 except for a weaker AA filter - as far as I remember - it has been a while. This is a linear effect. The factor has been taken into account in the 16-70 review.