BobYIL: MTF curves indicate that this will be a better one than the 35mm Art. Remarkable!
Do you have a link to the MTF curves?
dstanton: It's a good point the mechanical shutter would be in the way; but...
... I don't think all have mechanical shutters, All the MFT's I have; Pano and Oly -- the CCD is fully exposed. It's the newer models that recently came out where mechanical shutters are now in use.
But... who knows... I thought the same there -- wasn't there a standard for Micro Four Thirds -- for all to follow??
The Blackmagic cameras don't conform to any m4/3 standard (smaller sensor, 16:9 aspect ratio, no shutter, thinner filter stack, etc.). So its reasonable that these specialized Speed Boosters take full advantage of the unique opportunites presented by these cameras. Besides, there are several Speed Boosters that fit standard m4/3 cameras already.
parkmcgraw: The dark photos do not appear to be the result of Pixel Vignetting.
Condition statement, the iris aperture blades are located at an intermediary focal plane, hence not affecting the lens field of view.
The Speed Booster is an optical compressor. Hence, the last element is at a fixed distance from the focal plane. As such the solid angle field of view per pixel is also fixed.
Altering the lens aperture to the maximum open setting only alters the illumination intensity and not the solid angle cone of light at the detector (pixel).
Pixel vignetting, like film vignetting, being a non uniform field effect, the level of vignetting increases with off axis distance. The full field B&W test photos appearing uniform in illumination.
The dark photo and slight color shift looking like coating phase interference, visible as a drop in black levels on axis being minus the chromatic halo seen in the corner of the image field. The phase interference affecting the accuracy of the exposure meter.
No. The Speed Booster is designed to accept a maximum aperture of f/1.2 (actually f/(2^(1/3)) = f/1.2599) from the attached lens and then output a maximum aperture of f/0.90 to the camera.
The only possibility of clipping the on-axis ray bundle is if you attach a lens that is faster than f/1.2. The only SLR lens that meets this requirement is the Canon 50mm f/1.0. If you put that lens on the SB and open up to f/1.0 you won't get f/0.71 but rather only f/0.90 - the same as if you had attached an f/1.2 lens.
Pixel vignetting, which I call sensor-induced apodization, occurs everywhere in the image, including the center of the field. Its all about marginal ray angle and the reduced sensor sensitivity as the angle of incidence increases.
At f/2 the marginal rays on-axis strike the sensor at 14.5 degrees, which most sensors handle very well. At f/1 the marginal ray angle increases to 30 degrees, which causes significant signal reduction with most sensors. At f/0.7 the marginal ray angle is 45 degrees which. Remenber, this is for the center of the image, and has nothing to do with field-dependent falloff.
So, opening up a lens from f/2 to f/1 will not give you four times as much signal from the sensor. This is purely a sensor effect and has nothing to do with the optics.
Salvador Abreu: Interesting, this is in essence what Nikon (& Fuji) did with back in 1995 (!!) with the Nikon E2 and E3 (aka Fujix DS-560 and Fujix DS-565) - see http://en.wikipedia.org/wiki/Nikon_E_series for more details.
Now what I'd really like to see is a Nikon F-mount adapter like this with add-on AF, for use with MF lenses, a bit like the TC-16a. That would be mighty useful on DX bodies...
Salvador:Any similarity between the Nikon E system and the Speed Booster is superficial at best.
The wikipedia article leaves out several critical details that ultimately doomed the Nikon E series. True, it used an image reduction system, but it was a relay type located entirely behind the normal image produced by the objective lens. I call this a Keplerian type reducer. As a result, it added a huge amount of bulk to the system. The lens aperture was limited to f/6.7 in early E versions and f/4.8 in the last version. In other words, even if you used an f/1.4 lens, if you opened it up wider than f/4.8 (or f/6.7 in the earlier versions) the exposure and depth of field would not change.
By contrast, the Speed Booster is a Galilean type reducer that fits between the objective lens and the image plane. The Speed Booster actually reduces the system length and bulk, and is also fully usable with f/1.2 objecttives, which means you can take full advantage of a shallow depth of field.
Krich13: Could you please clarify the testing procedure? When the Canon lens is tested at say f/1.4 with Speed Booster, is the "boosting" effect already taken into account? In other words, is the lens itself stopped down to f/2, and "becomes" f/1.4 using the adapter? Or is it used wide-open at f/1.4 and effectively turns f/1 with the adapter? Later in the review you discuss the 40 mm lens at f/2, so it seems that the effect of the adapter is taken into account, I just want to be sure.
The same question about the f/8 test. Are both _lenses_ set to f/8 or both _systems_ (including the adapter) set to f/8?
Thank you in advance
According to metadata for the wide-open test chart image (as reported in Photoshop) the focal length is 35mm and the aperture is f/1.0. So I assume that at least for this comparison we are looking at f/1.0 vs. f/1.4.
Cal22: You can't ignore the disadvantages this adapter comes along with (more CA, lower contrast, corner softening, poor AF, more weight and bulkiness of FF lens/adapter). That's why it won't be a tool for any photographer in the mirrorless sector. But for some the adapter might prove useful, especially if it comes to wide angle or high aperture FF lenses you possess anyway.
Question: What about the adapter combined with a wide angle shift lens? Will shiftening be increasing optical flaws?
Hi Joe:Was this the review (17mm TS-E + SB): http://verybiglobo.blogspot.com/2013/01/metabones-speed-booster-review-nex-7_31.html
Its a little surprising that shift-lenses can function with the SB since the optical axes of the attached lens and the SB become totally disconnected. But it seems to work just fine.