A77 -- no ghosting!

[agorabasta]

how about the rest: AF speed and accuracy, contrast, etc

Speed is exactly the same as with the a55 for all lenses I tried, both srew-drive and motorised HSM/SAM (HSM means Sigma, both the 17-50 and 18-250 work OK). Low light focusing without lamp assistance is definitely better. The bad thing is that it has more trouble than a55 had while focusing on simple ordered textures; the 55 was already much worse than a700 in that regard; the 700 had no problems focusing on a uniform grey LCD screen catching the subpixels that projected on main sensor at about the sensel size.
Single-press release happens when the lens has really finished adjusting focus, quite unlike some a500 that often released the shutter before lens stopped. Single-press accuracy is very good even with that super-slow SAL18250.

All my A-mount lenses required no AF adjustments on my older bodies, and neither do they with this adapter. But the 5N AF adjustment option is activated with this adapter even for the Sigma HSM lenses I have.

The most important thing to me is that this adapter preserves the infinity where it is supposed to be. That was a major problem with LA-EA1 that really killed the close focus ability for any lens attached.

As of now, the adapter is fully functional only with Nex5N. Both my 5 and C3 don't even see any lens attached through it. Waiting for the promised FW updates...

[agorabasta]

Here are two 100% crops, one with an SLT adapter, another with older 'empty' one.

Shot at ISO6400 1/60s handheld so framing is slightly different, very same detail/NR and other processing on both.

Now guess which is what!

[agorabasta]

P.S. One more thing - there are no CA/distortion correction for the A-mount lenses with Nex5N, at least not for the SAL18250 and 35f18 SAM.

[agorabasta]

My educated guess would be that mirror in LA-EA2 is the same new mirror as found in the 77. I think it's great news.

And now we really may evaluate the true effect of light loss on the noise levels registered as we don't need to compare different cameras.

So here are two more crops from the darkest part of the image, enlarged about 1.5x and with the Lr detail/NR module switched off. 'Standard' sharpening at JPEG export.

[[SiC]]

Second one is with adapter?

[agorabasta]

The 9647 is with the SLT adapter. And the first pair the EA1 is LA-EA1, the EA2 is the LA-EA2 SLT adapter.

[[SiC]]

The 9647 is with the SLT adapter. And the first pair the EA1 is LA-EA1, the EA2 is the LA-EA2 SLT adapter.

Oh!

[agorabasta]

Did some further testing for 'light loss'.

Assuming older Nexes don't 'understand' the SLT adapter yet, they may be used to measure the actual loss.
So I used spot metering, far defocused fully manual Samyang lens at f/11 and pointed it at a white uniform monitor screen.

The Nex5 metered exactly 1/15s independent in rather wide range of directions and distances to the monitor screen when used with the 'empty' LA-EA1 adapter.
Switching to the SLT LA-EA2 returned metering switching between 1/13s and 1/15s. Averaging it to 1/14s would seem realistic.

So it's either the adapter still tells something to the non-supporting camera and it adjusts the gain, or the loss at that mirror is mere 7% in that artificial 6500K light.

[alphaomega]

Agorabasta

Single-press accuracy is very good even with that super-slow SAL18250.

Funny that the SAL18250 should be called "super slow". I own one of these and always find, when fitted on to my A580, that it is normally in focus when taken to my eyes or moving to another object. I find that lens OK for focus, but then it is not used for sport or other fast moving objects. I don't think it is bad for a screw drive.
Grateful for the info. on the new SLT LA-EA2 adapter. I am also waiting for the firmware upgrade to my NEX-5 as I fancy that adapter for use until some more useful lenses are released dedicated to the NEX system. I now have the SEL55200 and find it slow to focus and particularly when aiming at low contrast objects. If you try and photograph a forest at or around the 200mm setting the lens will not react. Must be a combination of the contrast detect and the slow lens (4.5-6.3).

[agorabasta]

Once again tried to check the light loss at the mirror, this time using AEL function on 5N.

Got a very steady value of (0.36-0.38)EV in a pretty long series.

This time it was measured with the 35f18 SAM. This lens has a pretty large rear element, and I expect fast lenses with large rear elements to exhibit relatively larger light loss when used with SLT as there is more light hitting the mirror at oblique angles.