NEX7 and Rangefinder Wides

[uhoh07]

Wow the C3 does better with the 18 than the M9, to my eye.

 

[headofdestiny]

Wow the C3 does better with the 18 than the M9, to my eye.

Yeah, Sony really seems to have worked some magic with the C3/5N sensor edges...unless it's just a firmware trick.

 

[paulcurtis]

I'd started another thread but it's disappeared so may as well add to this, although it's degenerated a bit.

I've just canned my NEX7 preorder and ordered a 5N to replace my 5. I use a lot of wides and the samples from the 7 are simply unacceptable. Sony got so much right, apart from the sensor it seems. Too high res as well. I wonder how many people really need 24mpix. Especially if you shoot RAW then you have to deal with much bigger files.

When the 7 is around and bedded in, perhaps i will change, but right now i wanted to get in with the 5N in case supplies become limited.

Such a pity

cheers
paul

 

[artur6]

It can't be a firmware trick because, with legacy glass, the NEX won't know which lens is mounted and the vigneting is specific for each lens - i.e a ZM18 will have much more color cast in the edges than a ZM35.

Most Leica M lenses are coded and the camera applies the correct profile for each lens. Nevertheless, a fullframe sensor is more demanding than APS-C and we see from the examples of the former page that a M9 with a non-coded lens performs worse than a NEX-5N, regarding purple edges.

 

[headofdestiny]

It can't be a firmware trick because, with legacy glass, the NEX won't know which lens is mounted and the vigneting is specific for each lens - i.e a ZM18 will have much more color cast in the edges than a ZM35.

Most Leica M lenses are coded and the camera applies the correct profile for each lens. Nevertheless, a fullframe sensor is more demanding than APS-C and we see from the examples of the former page that a M9 with a non-coded lens performs worse than a NEX-5N, regarding purple edges.

Yeah, unless the sensor can somehow detect the amount of shift and correct accordingly, but that's a long shot.

The most interesting thing to me about those M9 shots is, even if you crop out an aps-c portion out of the non-coded M9 samples, the NEX-3C still comes out ahead, so, at least in terms of color shift, the NEX-C3's micro lenses may actually be better tuned than the M9's.

 

[paulcurtis]

It can't be a firmware trick because, with legacy glass, the NEX won't know which lens is mounted and the vigneting is specific for each lens - i.e a ZM18 will have much more color cast in the edges than a ZM35.

Most Leica M lenses are coded and the camera applies the correct profile for each lens. Nevertheless, a fullframe sensor is more demanding than APS-C and we see from the examples of the former page that a M9 with a non-coded lens performs worse than a NEX-5N, regarding purple edges.

The leicas use a coding system on each lens, there are no electronic contacts used. But you can also set the lens on the body. The firmware could allow you to select from presets.

In fact if the NEX firmware could let you dial in a focal length or preset lens to add to the EXIF data, that would be fantastic.

However, the colour shifts show issues with light rays hitting at difficult angles. So there is a finite amount of fudging you can do. I would also imagine the colour shifts would also mean smearing at the edges too. It's a physical characteristic of the light angles with lenses that are very close to the sensor (rangefinder style)

The samples from the 7 are much worse than the ones from the old 5 for example.

cheers
paul

 

[TEBnewyork]

Interesting comments since some of us remember you selling all of it on more than one occasion as well.

You are correct. I owned a M6 for many years and purchased a M7 just before the digital thing really got going. Leica glass was so much cheaper back then and really good Leica glass can cost $5000 a piece.

I am glad I bought all of my new Leica glass when the M8 first came out as year or two later it all started to skyrocket!
--
Greg Gebhardt in
Jacksonville, Florida

--
terry
http://www.terrybanet.com

 

[captura]

no vignetting, covers well, super-cheap.
http://www.flickr.com/photos/stanbury/6026039467/

 

[uhoh07]

The samples from the 7 are much worse than the ones from the old 5 for example.

cheers
paul

Where are you getting that? The really horrible example is from the notorius cv 15, which requires cornerfix on the M9. And in that shot it looks very much like Rock has focused past inifinity, since the center is suspiciously softer than the other samples--or at least it seemed that way to me.

Having shot ad infinitum with the colorshifting skopars, from 21 to 35, on my old 5, I've seen no evidence the 7 is worse--But you may have seen shots I haven't so please post your links when you have a chance

There is the strange one with the ZM 25 vs ZE24, where the edges come in gradually and are sort of OK at F8. But

But is the ZM sharp across the frame now on the 5n wide open or at say f/4?
samples or links?

I know of none on other than:
http://www.flickr.com/photos/sportsphotorob/6232392106/in/photostream/

These show the zeiss is soft on the edges wide open but acceptable at f/8. Certainly not worse than 5.

 

[uhoh07]

here is the 5n fuzzing out on the edges with the contax 28
http://www.flickr.com/photos/8399398@N06/6284252934/in/pool-nex5n
but it seems great in lots of others with heliar wides

and in rob's shots the 18 and 21 are better than the 25 or 28 on the edges--or it seems to me they are. At f/8

Since the sensors peak at f/4 shame he didn't do many there.

6000x it's like the hubble

 

[headofdestiny]

here is the 5n fuzzing out on the edges with the contax 28
http://www.flickr.com/photos/8399398@N06/6284252934/in/pool-nex5n
but it seems great in lots of others with heliar wides

and in rob's shots the 18 and 21 are better than the 25 or 28 on the edges--or it seems to me they are. At f/8

Since the sensors peak at f/4 shame he didn't do many there.

6000x it's like the hubble

I'm a little concerned about the detail in the top corners of his NEX-7 photo with the Zm 35/2 at f4: http://www.flickr.com/photos/sportsphotorob/6231917737/sizes/o/in/photostream/

Either way, if the NEX-7 is only as good as the NEX-5 in the corners, I'll have to bite the bullet and go NEX-5N.

 

[Jonas B]

I'm waiting on a NEX7 and I really don't understand why it matters if third party glass works on the NEX bodies.
(...)

That's a strange start. Further down the thread you seem to know everything?

 

[paulcurtis]

Where are you getting that? The really horrible example is from the notorius cv 15, which requires cornerfix on the M9. And in that shot it looks very much like Rock has focused past inifinity, since the center is suspiciously softer than the other samples--or at least it seemed that way to me.

I was referring to Steve Huffs photos specifically. I did ask on another thread to see if there were more CV/wide samples.

I'm comparing those with my CV12 on my 5, chalk and cheese. The 5 is much better (and the CV12 i had always understood to show the effect more than the 15).

I've recently just got a 5N so i will be able to compare the 5 and 5N. I thought i'd get one before the stocks dry up, then if the 7 turns out okay i can swap.

Have you seen other samples?

If corner fix was built into aperture then it wouldn't be so bad, but i'm not about to add another set of steps when processing the photos i take.

Also to my mind if there is colour fringing then the sharpness of those areas would also be affected. It's a physical light interaction thing...

cheers
paul

 

[sroute]

Also to my mind if there is colour fringing then the sharpness of those areas would also be affected. It's a physical light interaction thing...

Paul,

Do you believe the $7,000 Leica M9 has problems with sharpness in "those areas" due to colour shift? Generally speaking, it does not.

But the Leica M9 does in fact have problems with edge color casts, just like the NEX-7 is showing, or the 5 and 3 showed, and what the uncorrected Ricoh GXR A12 M mount module will show.

In the case of the M9 and GXR we know for a fact that these colour casts are corrected by software. Turn off the correction software and you'll get magenta and vignetting a-plenty with both those cameras, both of which were designed for M mount lenses. Yet their edges can be very very sharp, tinted magenta or not.

I'm personally not willing to guess how the casts are corrected in the NEX-C3 and 5N unless someone can point to an official statement by Sony on the matter.

For all we know when "no lens" is detected the firmware looks for an increase in magenta on the sides and automatically applies a fix. Or maybe it does that for every scene. Quick, someone shoot a backdrop lit with light magenta to see if there is reverse vignetting!

For now I'm going to remain uneasy not because I fear edge detail is being ruthlessly destroyed on the NEX-7 with wide angle rangefinder glass, but because we do not truly know the answer and won't until someone does an authoritative look at this issue, considering not just colour casts but more importantly edge sharpness and detail preservation.

Units are supposed to ship to us regular folks in 7 to 14 days if Sony Style's estimated ship dates mean anything at all. Sooner or later some regular Joe or Jill is going to take delivery and shoot the test shots all these big name bloggers could have done but haven't.

 

[uhoh07]

I was scratchning my head over a number of shots, and then I started looking at every long sample he had and seeing some blur in the corners to one degree or another: including the sony 24.

Whoa, I thought, what's up with that?

Then I checked Huff's giraffes with the leica 35. Wicked sharp right to lower right edge.

I stopped worrying.

That said, the 5n may be the smarter option for us. The EVF which tilts is way better. I worry about the pop up flash being vulnerable to my bashing and thrashing.

But.....I'd sure like to see for myself

I'm a little concerned about the detail in the top corners of his NEX-7 photo with the Zm 35/2 at f4: http://www.flickr.com/photos/sportsphotorob/6231917737/sizes/o/in/photostream/

Either way, if the NEX-7 is only as good as the NEX-5 in the corners, I'll have to bite the bullet and go NEX-5N.

 

[SQLGuy]

The really cool solution would be a firmware update that allows you to shoot a grey card (or white card) and create an in-camera vignetting/color-shift correction profile from that result. Ideally, the camera would allow you to tag and store quite a few of these profiles, rather than making you re-shoot the card each time you switch lenses.

This should be possible to do in-camera, as it's what CornerFix does OOC, but whether it's work that Sony is interested in taking on....?

It can't be a firmware trick because, with legacy glass, the NEX won't know which lens is mounted and the vigneting is specific for each lens - i.e a ZM18 will have much more color cast in the edges than a ZM35.

Most Leica M lenses are coded and the camera applies the correct profile for each lens. Nevertheless, a fullframe sensor is more demanding than APS-C and we see from the examples of the former page that a M9 with a non-coded lens performs worse than a NEX-5N, regarding purple edges.

Yeah, unless the sensor can somehow detect the amount of shift and correct accordingly, but that's a long shot.

The most interesting thing to me about those M9 shots is, even if you crop out an aps-c portion out of the non-coded M9 samples, the NEX-3C still comes out ahead, so, at least in terms of color shift, the NEX-C3's micro lenses may actually be better tuned than the M9's.

--

Nex-5 with kit lenses, Contax G 35, and a number of legacy lenses (mostly Canon FD)

 

[uhoh07]

It hard to believe the 5n does this 100% in the sensor and the firmware plays no part. This would imply the sensor is optimised for very short registers---unless it somehow is just sensitive to all light angles. But up to now it seems in general sensors have been tuned for optimal results from a given register.

Now the 16mm rear element is closer to the sensor than my CV 21mm. Did Sony really just forget about performance with their native 16 on the Nex-7?

I see the 5n still has softness deep in the corners on the 16
http://www.flickr.com/photos/2416images/6236018517/sizes/o/in/photostream/

http://www.flickr.com/photos/bairid/6219253894/sizes/o/in/photostream/

Do we find these far worse than Rob's 16 samples?

http://www.flickr.com/photos/sportsphotorob/6232002288/sizes/o/in/photostream/

Look at the grass in the foreground. Looks nearly the same as the 5n, no?

The really cool solution would be a firmware update that allows you to shoot a grey card (or white card) and create an in-camera vignetting/color-shift correction profile from that result. Ideally, the camera would allow you to tag and store quite a few of these profiles, rather than making you re-shoot the card each time you switch lenses.

This should be possible to do in-camera, as it's what CornerFix does OOC, but whether it's work that Sony is interested in taking on....?

It can't be a firmware trick because, with legacy glass, the NEX won't know which lens is mounted and the vigneting is specific for each lens - i.e a ZM18 will have much more color cast in the edges than a ZM35.

Most Leica M lenses are coded and the camera applies the correct profile for each lens. Nevertheless, a fullframe sensor is more demanding than APS-C and we see from the examples of the former page that a M9 with a non-coded lens performs worse than a NEX-5N, regarding purple edges.

Yeah, unless the sensor can somehow detect the amount of shift and correct accordingly, but that's a long shot.

The most interesting thing to me about those M9 shots is, even if you crop out an aps-c portion out of the non-coded M9 samples, the NEX-3C still comes out ahead, so, at least in terms of color shift, the NEX-C3's micro lenses may actually be better tuned than the M9's.

--

Nex-5 with kit lenses, Contax G 35, and a number of legacy lenses (mostly Canon FD)

 

[captura]

News is an adapter for the 5C and 7 for this 110 lens.

Thought I should post something about the Pentax 110 lens, the widest of which is 18mm. PERFECT for m43 as it is the exact cover for it's sensor. Works quite well with NEX, some vignetting on some pics, but better than cine lenses by far.
That is the OLD news.

http://forums.dpreview.com/forums/readflat.asp?forum=1041&message=34465828&changemode=1

http://forums.dpreview.com/forums/readflat.asp?forum=1041&message=36492215&changemode=1

The NEW news is that the troublesome old adapter appears to be updated by a better one from China which appears to be modified top also suit NEX7, NEXC5 and presumably 3C. Still too expensive, I'm looking for a cheaper adapter alternative. Any ideas? (for my NEX3.)
.
Pentax 110 Lens to Sony NEX-3 NEX-5 NEX7 NEXC5 Adapter

http://www.ebay.ca/itm/Pentax-110-Lens-Sony-NEX-3-NEX-5-NEX7-NEXC5-Adapter-/120771249906?pt=LH_DefaultDomain_0&hash=item1c1e8706f2

 

[headofdestiny]

It hard to believe the 5n does this 100% in the sensor and the firmware plays no part. This would imply the sensor is optimised for very short registers---unless it somehow is just sensitive to all light angles. But up to now it seems in general sensors have been tuned for optimal results from a given register.

Now the 16mm rear element is closer to the sensor than my CV 21mm. Did Sony really just forget about performance with their native 16 on the Nex-7?

I see the 5n still has softness deep in the corners on the 16
http://www.flickr.com/photos/2416images/6236018517/sizes/o/in/photostream/

http://www.flickr.com/photos/bairid/6219253894/sizes/o/in/photostream/

Do we find these far worse than Rob's 16 samples?

http://www.flickr.com/photos/sportsphotorob/6232002288/sizes/o/in/photostream/

Look at the grass in the foreground. Looks nearly the same as the 5n, no?

There isn't much of a penalty by optimizing micro lenses for short registration distances. Maybe a bit more vignetting on telephotos, but nothing much to worry about. However, forum member Agorabasta doesn't believe that it's a microlens change at all. He claims that Sony sensors often have a small gap in between the filter pack and sensor, in order to aid in not seeing dust on the sensor, and Sony could have just removed that gap on the C3/5N. Who knows?

Agorabasta posted a 16mm corner comparison of the NEX-5 and NEX-C3 on photo club alpha, and the corners of the latter seem better.

NEX-5 top right corner:

http://www.photoclubalpha.com/forum/download/file.php?id=3930&sid=4a48142d6df43644bc43b2fcd9719dec

NEX-C3 top right corner:

http://www.photoclubalpha.com/forum/download/file.php?id=3931&sid=4a48142d6df43644bc43b2fcd9719dec

Also, there is a Xitek thread showing all kinds of NEX-5 vs. C3 comparisons, and, not only is there much less color shift, but resolution also seems to be improved.

see thread here: http://forum.xitek.com/showthread.php?threadid=902861&pagenumber=1

ZM 21 at f8, whole frame:



NEX-5, NEX-C3 center crop:



NEX-5, NEX-C3 lower left corner crop:



NEX-5, NEX-C3 lower right corner crop:



The center crops look pretty similar, but the corner crops of the C3 look to be resolving more.