Metabones Speed Booster Review (NEX 7) Part II - Canon TS-E 17 f4 L

Canon TS-E 17mm f4 L USM, Metabones Speed Booster (Canon EF to Sony E-mount) and Sony NEX 7.

There is ongoing debate on the forums between the "I like NEX because of the small size majority" and me. (Well, maybe few others :-))

The reason why I choose NEX had nothing to do with the size or weight. (It is welcomed addition, I admit).
I choose NEX because of the shortest flange distance, that allows use of almost all lenses ever produced and very good sensor especially in terms of DR. Later with the NEX 7, it comes also ultimate resolution bellow MF. NEX 7 has higher pixel density than my Nikon D800E.

My problem with this "Size is all that matter majority" is only one - Sony. Because of the mass market consumers pressure on the size and weight, Sony involved those extensive software corrections in their .arw files, that used to be RAW files, not so long ago.

Beside the fact, that I like to be in control of the things, there is yet another consequence. (And that one is also on purpose, I am sure.)

Lens testing is loosing its sense, as we can only test certain combinations of camera and its firmware, lens and its firmware and RAW development software (in a current edition). That makes too many variables and  combinations to allow objective qualification of the lenses and their optical performance.

And this apply to the lens adapters too of course. So, while the following shots are taken with Sony NEX 7 and firmware 1.01, and developed in LR 4.3, results with any other camera, or with NEX 7 with the original firmware (1.00) might be completely different.

Is there any reason to reed this review than? I don't know... But if you are all about the size, I am sure you can spend your time much better, because this combination will not fit in your pocket.

(Recently I had been reading one dicscussion on DPR NEX forum, about the reason for buying new Sony SEL 20mm f2.8 pancake, for 360 USD instead of Sigma 19 f2.8 EX DN E for recently 139 USD (or 99 USD if bought in bundle with Sigma 30 f2.8).
One of the members, after deep analyze that included - nicer color finish of the barrel, and overall more "cool" look, if I understood well, said that his jacket pocket is made for pancake lenses, and therefore he will certainly buy new Sony. The size difference between the lenses is some 25 mm shorter, BTW.)

For me (and those few others) for whom the size is not everything, or the bigger the better if you like, Sony NEX 7 with Canon TS-E 17 f4L and Speed Booster, sounds great. Landscape but even more architecture photographers, that are after high resolution, optical correction and convenient stitching capabilities of TS-E 17 f4L, probably should pay attention. If it deliver, what Metabones SB promise.

Before I post first image, let me say, that I have and will make much more with this combination. I was impatient to see how it performs, and I wanted to see, how does it compare to Canon 5D Mark II, but unfortunately my battery in Canon was ready to die, and without a charger at work, I could only made few snaps before full dark-out.

Here is the direct comparison. Image are somewhat reduced, but if you enlarge them to the original size, I believe that you would still be able to see the differences.

Comparing Sony NEX 7 with Metabones SB vs Canon 5D MII, both with Canon TS-E 17 f4 L at f5.6

I should notice, that because of fast change of the lighting conditions, I was not able to calculate real exposure difference (or Speed Boost), but I will do that when I get to the studio and could control lighting.
The FL lens referred to the Sony NEX 7 was 11.9mm, and maximum aperture was listed as f2.8. As you can see, TS-E 17 f4L is a bit tighter than on Canon 5D MII. As the Metabones SB magnification is listed to 0.71, it should refer to 12.07mm on Sony APS-C, or to 18.1mm FF equivalent. I would say, that difference in FL is slightly bigger than 1,1mm but that can be also due to the different physical size of the systems. All in all, it is close enough.

LR 4.3 basic info.

Sony NEX 7 with Metabones SB and Canon TS-E 17 f4 L at f2.8

Let's look at some 100% crops

Comparing Sony NEX 7 with Metabones SB vs Canon 5D MII, both with Canon TS-E 17 f4 L at f5.6
100% Crops

This crops certainly shows one thing - Metabones SB does what it promised to do. All crops from Sony, shows slightly higher resolution than those from Canon. There is few Mpx more on Sony too, but I would say, that what we can see here, is a real lens performance booster effect. (Just a reminder, none of those shots was sharpened in post processing, beside standard LR 4.3 sharpening).

So far so good. Great. Fantastic. There is 1K USD APS-C camera that can do better than 2K + DSLR in terms of IQ. Where is the hook, you might ask?

If almost 700 USD is not enough big hook for you, there are few more...

The biggest one is overall more CA especially purple fringing.

Here are crops before and after correction at f2.8 and f8:

Sony NEX 7 with Metabones SB and Canon TS-E 17 f4 L at f2.8 and f8, CA and PF correction

While stopping down doesn't help to eliminate CA and PF, correction in LR 4.3 for both, is really trivial. 

The second trade off is heavy vignetting at maximum shifts.

Comparing Sony NEX 7 with Metabones SB vs Canon 5D MII, both with Canon TS-E 17 f4 L at f8
Maximum shift problem

This finding indicates, that designers pushed probably a bit hard on the projection circle at this FL. (Remember - one correction lens, that rule them all?)
Maybe Brien can give as more reasons behind (if he ever read this article), but it could be that they had to go over, because of lack of space related to highest possible refractive index (before they lost control over CA and PF). That would also correspond to the vignetting at longer focal lengths, where the projection circle is also smaller than it should be.

The smaller projection circle could result in the corner smearing, but with a lens set to 0 shift, I certainly couldn't see any more smearing compared to Canon output.

To my eyes, there is also increased barrel distortion. TS-E 17 is a lens with a minimum distortion for UWA, and I will have to make more intensive test, but so far it seems that distortion might be also boosted with SB.

Finally, let's see if stopping down the lens, improves resolution.

Sony NEX 7 with Metabones SB and Canon TS-E 17 f4 L at f2.8, f4 and f5.6. 100% crops.

As you can see, stopping down result in better overall result. At f5.6 image has better resolution across the frame and light fall off is significantly reduced. But, image at f2.8 is still very good, and with some post processing, it is more than usable.

I have to make studio resolution, exposure test, fall off test, color cast test and NEX 7 firmware 1.00 if needed, but so far, Metabones SB didn't disappoint. (Once and if, I manage to adjust infinity focus with my favorite lenses).

TS-E 17 combination comes out very attractive. One of the best Canon UWA lenses gets
a new body to shines on, and could be used at almost one f-stop faster resulting shutter speed.

This, in combination with very good high ISO performance of NEX 7 (and all other NEXes), makes it kind of revelation for the architecture photographers.
How's about Astro photography? Instead of f 4, there is now welcomed f2.8 high image quality 17mm lens.

Once uploaded, you should be able to download RAW (well in Sony case, something alike) files from this link - Original Files, but be patient, as I don't know if it will work. If you succeed, feel free to do with images whatever you need for your evaluation, but don't publish them without my permission, please.

Link to the part I - Metabones Speed Booster Review (NEX 7) Part I Introduction 
Link to the part II - Metabones Speed Booster Review (NEX 7) Part III - Canon EF 50 f1.2 L USM

Next lens to jump on booster will be Canon EF 50 f1.2L. Meanwhile, here is one Bokehlicious shot from 94.5 f1.4 :-)