Crop factor of NEX 5

Found it, I think it's 1.5X. More importanltly, can you use non u4/3 lenses ultra wide ange lenses (not telecentric optical design) without smeary image quality in the outer field, the biggest drawback to the u4/3 format?
 
Initial results show much better corner/edge performance than that of the smaller sensor micro 4/3. The RD1 has the same size sensor as the NEX and to my knowledge doesn't show poor corner performance, even though it doesn't have any offset micro lenses on the sensor, ala M8 or M9 (correct me if I'm wrong.)
 
digitalintrigue said:
Initial results show much better corner/edge performance than that of the smaller sensor micro 4/3. The RD1 has the same size sensor as the NEX and to my knowledge doesn't show poor corner performance, even though it doesn't have any offset micro lenses on the sensor, ala M8 or M9 (correct me if I'm wrong.)
Click to expand...

My simplistic understanding of the corner-smearing issue is that it's less related to sensor size and more related to:
1. flange>sensor distance
2. angle of light exiting rear element
3. special sensor mojo for corner sharpness a'la M8/M9

I would think closer flange>sensor distance means greater possibility of corner smearing, with more corner smearing from near-symmetrical rf wides than from mostly-retrofocus SLR wides. So I'd expect less smearing from NEX than m4/3, and maybe less from R-D1 than NEX, light exit angle and sensor mojo being equal. I'd also expect less from std. 4/3 than m4/3, with flange>sensor distance mattering more than sensor size.

The R-D1/M8-M9 question is interesting, I'm not sure if I've heard anything definitive. Are corners better on the M8/M9 than the R-D1 if using non-retrofocus wides? Have there been corner smearing complaints with ultrawides on the R-D1?
 
I think you're right

I think you're right

I haven't used, or seen many images with the 12 and RD1, but the CV 15 and 21, and multiple 28s on the RD1 are all as good on corners/edges, or better, than on the M8, and as on legacy 24x36 systems.

The Nex is looking very interesting as a backup M body to haul around. Hmmm.

digitalintrigue said:
Initial results show much better corner/edge performance than that of the smaller sensor micro 4/3. The RD1 has the same size sensor as the NEX and to my knowledge doesn't show poor corner performance, even though it doesn't have any offset micro lenses on the sensor, ala M8 or M9 (correct me if I'm wrong.)
Click to expand...
 
Yes, I am hoping someone will show some side-by-sides with the same lenses on NEX and RD1. From what I've seen, things look pretty good so far...
 
PCB_RF said:
My simplistic understanding of the corner-smearing issue is that it's less related to sensor size and more related to:
1. flange>sensor distance
2. angle of light exiting rear element
3. special sensor mojo for corner sharpness a'la M8/M9

I would think closer flange>sensor distance means greater possibility of corner smearing, with more corner smearing from near-symmetrical rf wides than from mostly-retrofocus SLR wides. So I'd expect less smearing from NEX than m4/3, and maybe less from R-D1 than NEX, light exit angle and sensor mojo being equal. I'd also expect less from std. 4/3 than m4/3, with flange>sensor distance mattering more than sensor size.

The R-D1/M8-M9 question is interesting, I'm not sure if I've heard anything definitive. Are corners better on the M8/M9 than the R-D1 if using non-retrofocus wides? Have there been corner smearing complaints with ultrawides on the R-D1?
Click to expand...


Of course, the flange-to-sensor distance of a rangefinder lens manually adapted to a u4/3 body is the same as that of any other rangefinder-compatible body (M8,M9,RD-1,Nex, etc.) otherwise infinity focus would be off. This is what the M-adapter's length is for.

And since the flange distance is the same, then so is the angle of light exiting the rear element.

Me thinks the corner blurring issue has more to do with the sensor itself and how it receives the incoming light.

~Joe
 
finally a very compact camera with a crop factor that allows rf users to use their lenses without turning them all into teles. really excited about reading these nex threads.
 
I handled an NEX5 in a camera store in Kyoto today. Dang, that thing is tiny. Much smaller in the hand than a GF1.
 
Image quality with wide angle M lenses are likely better that my Oly E-P2 and the crop factor of 1.5 and less edge smear have some nice advantages for users of legacy wide angles. However, the lack of an EVF and it's tiny size make Leica teles like a 90/2 or any 135 more difficult to use.
 
The m4/3 smearing issue probably has a lot to do with sensor topping thickness (AA filter, CFA, etc.) Vivek, over on getdpi, claims to have measured the thickness on m4/3 cameras, and it is thicker than anything else out there. Like their 4/3 brethren, m4/3 sensors perform better with telecentric lenses, and the sensor microlenses have issues with angled light rays at the sensor edges.

If the Sony NEX proves to handle these angled light rays better, that, combined with the 1.5x crop, gives them a big advantage in M lens adaptation.
 
Duuuhhh----hey!

Duuuhhh----hey!

JoeV said:
Of course, the flange-to-sensor distance of a rangefinder lens manually adapted to a u4/3 body is the same as that of any other rangefinder-compatible body (M8,M9,RD-1,Nex, etc.) otherwise infinity focus would be off. This is what the M-adapter's length is for.

And since the flange distance is the same, then so is the angle of light exiting the rear element.

Me thinks the corner blurring issue has more to do with the sensor itself and how it receives the incoming light.

~Joe
Click to expand...

You've run circles around me, logic-wise! Monty Python fans?

The first thing that occurs to me is that you're right, of course. I completely forgot/ignored the fact that with legacy lenses on adapters the mount and rear elements are always the same distance from the film/sensor plane, regardless of mount. Mea stupida.

The second thing that occurs to me is that I shouldn't be posting in the morning before I've had my 3rd cup of coffee.

The third thing that occurs to me is that, with the m4/3 sensor being smaller than APS/FF, the sensor corner is seeing a more-centric part of the image circle than the bigger sensor would. That makes sensor construction and microlens/light angle interface really critical. Imagine how bad the corners would be on a full-frame sensor using the same construction as current m4/3.

Could you build an m4/3 sensor with less corner smearing using legacy wides without compromising performance with native telecentric m4/3-4/3 lenses? Just lessen the filter pack?

So maybe the only valid thing from my post is that SLR wides, being retrofocus, are more telecentric than RF wides, so should have less corner smearing, in general, on m4/3?

BTW, I've had no coffee yet this morning, so please forgive me, and feel free to correct me, if all of this is drivel.
 
PCB_RF said:
Could you build an m4/3 sensor with less corner smearing using legacy wides without compromising performance with native telecentric m4/3-4/3 lenses? Just lessen the filter pack?

So maybe the only valid thing from my post is that SLR wides, being retrofocus, are more telecentric than RF wides, so should have less corner smearing, in general, on m4/3?
Click to expand...

That seems to be what some are implying. Leica did quite a trick with the microlenses of the M9, and I'm sure that m4/3 sensors could be improved in this regard. It seems that NEX is somewhere between the two, and, although the sensor is bigger than m4/3, early tests seem to show that NEX performs better at the edges with M lenses. I'm waiting to see more tests before I dive in.

As for retrofocus, you've got the right idea. Wide M lenses tend to only be slightly retrofocal, so the exit pupil is close to the sensor, causing the corner issues on digital sensors. The highly retrofocus wides of DSLR systems shouldn't be an issue on m4/3 or NEX at the edges.
 
The absence of micro-lenses does not cause "smearing", i.e. loss of resolution in the corners. That can only be caused by diffraction on additional material on top of the sensor. M8 and M9 micro-lenses primarily reduce vignetting and some color shift.

So I'm assuming the NEX 5 sensor has less "topping", as Douglas suggests. And let's not forget that both M8 and M9 use in-camera soft/firmware (even for "RAW") and lens coding for further lens correction. With LTM or M lenses generically adapted to APS-C or u4/3, this software correction will always be missing. It's clearly being used for native lenses though.

Roland.
 
Last edited:
correction

correction

Not necessarily. Only if you turn it on. Leica's acknowledged it's not always perfect or ideal.

I don't know what the X1 does, or if it's lens is a standard M lens, but the only other camera I know that takes M lenses easily with an APS-C sensor is the Epson RD1.

There have never been corner or smearing issues I've ever seen, and I've owned 4 of these.

u4/3 is a different beast though, since the native lenses for it have high levels of distortion that need correction in firmware to produced decent images.


ferider said:
...And let's not forget that both M8 and M9 use in-camera soft/firmware (even for "RAW") and lens coding for further lens correction. With LTM or M lenses generically adapted to APS-C or u4/3, this software correction will always be missing. It's clearly being used for native lenses though.

Roland.
Click to expand...
 
You must log in or register to reply here.
Back
Top Bottom