A77: why long lenses seem to be so out resolved?

Started 5 months ago | Discussions thread
VirtualMirage
Senior MemberPosts: 2,235Gear list
Like?
Re: A77: why long lenses seem to be so out resolved?
In reply to Michel J, 5 months ago

Michel J wrote:

VirtualMirage wrote:

DxO typically only tests with the RAW data. Here is a quote from their site on lens testing:

  • All lens scores pertain only to RAW imaging performance. They do not take into account other criteria such as image processing efficiency, mechanical reliability (autofocus, robustness), value for money, weight, etc.

Link for reference:

http://www.dxomark.com/About/Lens-scores

-- hide signature --

Paul

DxO, but NOT dPr.

But the person questioning this was asking about DxO, not DPR. So my response still stands.

Anyway, I would love to agree with your points, but your statements about the fact that some sensor "outresolve lens/es" is contentious matter to a conclusion by Allan Olesen arguments >>>

"Moiré, lenses definition VS resolution power" >>>

I reserve my position on that, but it sounds good to me.

So you linked (twice) to a previous discussion that you were having and referred to comments that clearly state that your conclusion comes from your failed attempts to properly understand Allan's explanations. And on top of that, you linked to someone's own personal opinion that has no documented evidence that proves it.  That doesn't bode well for your argument here.

I also don't completely agree with the idea that if a sensor starts to moire that it means that the lens is outresolving the sensor.  Moire is an artifact that is introduced due to sensors using a bayer color filter array and the demosaicing algorithms used to piece together the color information.  Because of this, an OLPF filter is used to reduce the chance of moire.  Replace the bayer method sensor with a Foveon, and you will get no moire.  If all sensors were Foveon type or black and white only, then there would be no moire.  If there was no moire, then you guys would be trying to find some other reason to be pointing at to try and make your case with.

In the other hand, I think we can compare lenses with Nikon mount yet, every time is a Sony sensor with the same pixel count (and of course with the same sensor generation)! And between A77 vs D7100 it seems to be we can extrapolate (and with the Nex 7 as well)? Because now with today 24MP/APS-C sensors, the AA filter effect, is progressively confused with the s/n ratio. And if the Nikon one resolve 3000LPH when the A77 resolve 3400LPH, I assume that the AA filter of the Sony is not noticeable, if it's not present.

No, it just shows how contradictory DPR is and the flaws in their testing.  And if you read the review closely, they stated that although the A77 appeared to have captured more detail in JPEG it did so at the expense of haloing (which is a sign of too much in camera sharpening).  The reason why they didn't label the D7100 as being higher in detail is they most likely stopped counting when moire was becoming present.  And remember, this is a JPEG comparison.  There is too much cooking going on to really determine its resolution capabilities by looking at the JPEG.  You are much better off looking at the RAW.  Unfortunately, while DPR mentions that the D7100 produces even more detail in RAW, but doesn't mention how much their measurements came to.

Anyway, it does not reflect a tangible explanation of tests from the same camera (A65): why the Sony DT 35mm F1.8 SAM have the same score >>> of the Sony 35mm F1.4 Gwhereas in reality, there is no noticeable diffrence with the Sony DT 18-135mm f/3.5-5.6 SAM at 35 mm. And when the kit-lens itself DT 18-55mm is rated higher than the last one!

Sony 35mm F1.4 G 13 p-MP (DxO T score: 21)

Sony DT 35mm F1.8 SAM 13 p-MP (DxO T score: 21)

Sony DT 18-135mm f/3.5-5.6 SAM 7 p-MP (DxO T score: 12)

Sony DT 18-55mm F3.5-5.6 SAM 9 p-MP (DxO T score: 11)

First, You can't compare the score of a zoom with a prime. A zoom has it's score weighted across its whole focal length, there is no comparison

Secondly, the scoring system is based around a fixed scenario:

  • The DxOMark Score is measured for defined exposure conditions corresponding to low-light scene with 150 lux illumination and an exposure time of 1/60s. These conditions were chosen as we believe low-light performances are very important for today’s photography and it is also important for photographers to know how well lenses perform at the widest aperture.

Lastly, did you even look at the charted data of all four lenses?  Or did you only look at the final numbers?  At 35mm, there is a clear difference between the primes and the zooms.  You also can't compare the score of an 18-135mm to an 18-55mm.  While the 18-55mm performs better in its focal length when compare to the same range on the 18-135mm, the 18-135mm improves itself in the range the 18-55mm can't reach.  This is why it achieves a higher score.  Not because it is better than the 18-55 in that range, but how it performs throughout the 18-135mm range and averaged.

Can anyone tell to us which is which from the lenses above (courtesy of enemj!!):

f5.6

Compare what?

Both images use different settings and one appears to be out of focus. We also don't know if it was a JPEG (then what settings) or RAW (what post processing is done, etc.)  It is not an accurate comparison and does not do one of the lens any favors.

If the one that is out of focus is actually in focus then I would say it is either a bad copy of the 35mm F/1.8 or it is the 35mm F/1.4. My 35mm F/1.8 is sharper than that and doesn't produce the fringing and hazing that is present in that image.

And if you go start comparing all lenses at F/5.6 or F/8, you would find that a lot of them would me the demands at that aperture for most people's needs.  But that isn't the only point of having a more expensive lens, or better lens.  The point can range from how fast the lens is (aperture), how does it render colors and bokeh, as well as durability.  Focusing on a fairly small aperture and using that as the judgement of a lens is short sighted and proves no point.

-- hide signature --

Paul

 VirtualMirage's gear list:VirtualMirage's gear list
Sony RX100 Sony SLT-A77 Sony a77 II Sony 50mm F1.4 Tokina AT-X Pro 11-16mm f/2.8 DX +23 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow