APS-C + speedbooster vs full frame

Question

I own a Nikon in DX-format (Nikon name for APS-C sensor, crop factor ~1.5) and I'm considering buying a 50mm prime lens soon for it. (To be complete, my camera is a D5200, although I don't think the precise model matter so much.)

Of course, because of the crop factor, such a 50mm lens would give me a 75mm effective focal length (by which I mean when compared to a full frame body). I have two choices, as I see it:

1. go for a 35mm lens instead to end up with a ~50mm effective focal length in the end,
2. go for the 50mm together with a 0.6 or 0.7 speedbooster.

The second option has the advantage that I'll still have a 50mm lens if I decide to go for a FF body in the future. What are the drawbacks of a speedbooster really? From what I gather online, the only thing that could be bothering is that it shifts your aperture range towards low values (f4 becomes roughly f1.2 in my imaginary setup). And I would actually consider that a plus considering what I usually shoot.

I'm looking for the opinion of people who experienced one or both setups before. Would you recommend option 1 or option 2, and why?

---

Edit. By speedbooster, I mean something along the lines of what metabones has done since a few years for mirrorless cameras. Maybe focal reducer is more accurate.

This is new territory for me, and I surely don't understand everything optic-wise. Please do tell if this post is not making a lot of sense (for example if option 1 and 2 are widely different or not comparable in any way).

Answer 1

Speed Boosters

The primary design intent of speed boosters is exactly what the name suggests. They make the lens "faster" (i.e. provide more light to the image plane).

• They are primarily an alternative to increasing ISO and secondarily an alternative to spending lots of money expensiver faster lenses. Being an alternative to higher ISO's sets the bar for optical quality...potential image degradation via the optical path is traded off for potential image degradation due to the noise of higher ISO.

• The changes to field of view is a side effect (and often a desirable one).

• The primary market for speed boosters is video/film making. There is more latitude for overall optical quality because images move. And noise reduction in post is a computationally expensive calculation which has to be done 24 or more times per second of video. Also variations in noise artifacts between frames can be problematic.

• Keep in mind that slow shutter speeds are far less of an option for video due to frame rates, and even less of an option for movie making where a consistent 1/50 second is desirable.

• The standard for video/film making is Super 35. This is approximately APS-C. It is not full frame. One of the motivations for speed boosters is to bring Micro 4/3's cameras closer to Super 35. This simplifies thinking about lenses. Also the smaller Micro 4/3's sensors tend to have weaker low light performance due to smaller size.

• Speed boosters make more sense when the one stop of speed is the difference between a $4000 cinema lens and a $8,000 cinema lens. In part due to cost, in part because cinema lenses come in series with consistent controls to make setting up camera rigs easier.

• The wider image provided by a speed booster is less important for traditional video and film making because the visual language tends to favor tighter crops. There's less time to ponder details at the edge of the frame when frames are moving 24 or more times per second. Also, tighter crops mean smaller sets...less to build, less to light, more space for everything that has to happen off camera.

A Little Gear Philosophy

1. When in doubt, get the simplest thing that might work. If that's a 35mm lens, then that's simpler than a 50mm lens + a speed booster.
2. Avoid buying gear for cameras you don't have. Between now and the time you get a full frame camera, you probably will have learned something and often learned enough that you will want to make pictures that you haven't thought of yet...in the future you might find that 50mm equivalent is not your preferred focal length or that you actually prefer zooms or even that your preferred upgrade is to medium format or Micro four-thirds.

Remarks on Normal Lenses

• Traditionally the normal lens is a focal length equal to the diagonal of the image plane.
• The normal lens of 35mm or full frame is ~43mm.
• 50mm is "normal" for 35mm film in the sense that there is a visual language built up around it because Oscar Barnak used a 50mm on the first Leica designs and everyone copied that.
• The closest match normal lens for APSC with 1.5 crop is 28mm (42mm equivalent).
• A 35mm lens is about as close the 35mm film normal of 43mm as 50mm is. Or to put it another way, a 35mm equivalent lens (24mm on APS-C) is about as normal as a 50mm equivalent lens. Just in the other direction.

Answer 2

The 35mm f1.8 DX lens is sharp as a pin, distortion-free and cheap as chips. If you want a 'normal' lens (eg for candids or small-group portraits) you can't do better. As others have said, no speedbooster made for Nikon F-mount bodies.

Answer 3

My question would be... why do you want a 50mm lens FoV? My opinion is that about the worst thing you can do in photography is to create "normal" images. And for many things the 75mm EFL is arguably better.

Non-opinion wise.

• The speed boosters function by inserting additional optics into the light path... in my experience that always has some negative impact (sometimes minimal, but it's still there).
• The speed boosters cost as much as either of the Nikon 35mm/50mm f/1.8's, and many other fast primes (AKA "thrifty 50").
• There is no speed booster available for F-mount to F-mount... probably because of the increased complexity required; and because the cost/demand isn't justified (see above).

BTW, the speed booster's F# decrease is only relative to the exposure/light density. In terms of depth of field it increases; because the speed booster reduces the size of the image circle (effective shorter FL). And FL/distance have more impact on DoF than aperture does... for the same recorded image the DoF probably remains about the same, but I haven't compared nor done the math.