Canon's 50mm prime lenses are all non-IS (Image Stabilization) for some reason. Considering the popularity of these range primes, shouldn't Canon be interested in shipping a new, optically improved IS version?
-
\$\begingroup\$ The reason for asking this question is this: I take a lot of impromptu/adhoc shots. And I have noticed significant shake problems with EF 50mm f/1.4 (it is a gem for calculated, take-your-time shots though). As a noob who have tasted the smoothness of 70-200mm f/4L IS USM or a kit lens for that matter, I started wondering why not IS on a 50mm? \$\endgroup\$– mixdevFeb 1, 2011 at 20:08
-
1\$\begingroup\$ See also: Is it true that there are no stabilized prime lenses (and if so, why)? \$\endgroup\$– jristaNov 24, 2011 at 4:21
10 Answers
There are two main reason why there's little chance we're going to see 50mm IS lens in the near future:
50mm lenses tend to be very simple and therefore cheap. Complicating its design with an image stabilizer group would push the price significantly higher, while the added benefit of IS wouldn't be that high at 50mm focal length.
If you want an expensive and light-hungry 50mm, there's already the huge lump of glass that is EF 50mm f/1.2L. There's not much point in having IS in such fast (and relatively wide) lenses.
-
16\$\begingroup\$ I've used a 35/1.4 with stabilization, and been really happy for it. Far from being pointless, it's actually extremely useful. \$\endgroup\$ Feb 1, 2011 at 18:13
-
2\$\begingroup\$ Same as Jerry, I use my fastest lens (like the Canon 50mm f/1.4) in very low light (say, concerts) where I push my 5DMII as far as I can in terms of ISO and noise. At this point, shutter speed is the last parameter I can tweak, and I'm nowhere at a speed that would not benefit from IS. My suspicion on this question is that it would make the lens much bulkier, and 50mm lenses are traditionally supposed to be small/portable. \$\endgroup\$ Feb 1, 2011 at 18:47
-
4\$\begingroup\$ Have you tried a monopod? It'll get you more time than IS will. \$\endgroup\$– user2719Feb 1, 2011 at 19:06
-
3\$\begingroup\$ Jerry: Which manufacturer makes a stabilized 35/1.4? \$\endgroup\$ Feb 2, 2011 at 18:58
-
1\$\begingroup\$ @Jon.Griffen on a Pentax body, all lenses are stabilized; so you could either pick a Pentax-M, a Zeiss, or a Samyang/Rokinon lens. Or you could go with Sony body+lens. \$\endgroup\$– ImreJun 17, 2011 at 15:58
That is a question for Canon. Honestly, I would not expect them to answer anything about their reasons or plans for not doing something.
What you may note is that there are very few prime lenses with stabilization. If you look across ALL major brands, there are only 20 of them among 213!
One would assume the base logic is that prime lenses have wider apertures than zooms and so stabilization is less needed. This is actually one reason I primarily shoot with Pentax and I enjoy using a F/1.8 lens with stabilization for extreme low-light shooting.
There may be a technical answer, but I would just be guessing. Perhaps if you start moving lens elements, you may eat away at some of the quality advantage, or perhaps it is harder to stabilize at wider apertures? I'm curious too but don't really know.
Disclosure: I am the owner of the site linked above.
-
2\$\begingroup\$ I just took a quick look at the link that was posted here. As I guessed most of the lenses here are over 200mm, with about half of 300mm +. Many have max apertures of f/4 or f/5.6. As for a use case, I'd imagine mounting one on a monopod rather than a tripod (you usually turn IS off for tripod mounting). \$\endgroup\$– MikeFeb 2, 2011 at 6:07
One of the oldest rules of thumb in photography is the "Sunny 16" rule. Essentially, it says that you can get a proper exposure using 1/(ISO linear rating) as the shutter speed at an aperture of f/16 in bright daylight conditions.
At an ISO setting of 100, then, a midday exposure would be 1/100 s (or, more practically on a film camera, 1/125 s for slides or 1/60 s for reversal films).
Another rule of thumb for th 35mm format is that you can safely hand-hold a camera and get reasonable sharpness at a shutter speed of 1/(focal length of lens) in seconds. Anything from 1/60 s down is in the hand-held range for a 50mm lens on a full/frame camera, and 1/80 s and faster ought to be okay on a crop-sensor camera.
Now, not all pictures are taken at high noon on a sunny day. But ask yourself how many pictures you tend to take with your 50mm lens stopped down to f/16. Unless you keep your camera pegged at the hyperfocal distance for Cartier-Bresson-style "decisive moment" snaps, I'd be willing to bet that most of your small-aperture shots are either landscapes or architectural, in which case a tripod is hardly an annoyance. You probably use the lens opened up to f/8 or wider most of the time. And you've got a further couple of stops of ISO sensitivity to hand before picture noise begins to be noticeable to anyone besides pixel peepers.
What all of this boils down to is that there would be a vanishingly small market for an IS 50mm once the added bulk and expense are taken into account. The only people who would really benefit from IS on a 50 are people like me (I have a chorea not unlike Huntington's disease) and a handful of photojournalists who want to work in available darkness but can't stand grain. Oh, and a few people who have money to spend and can convince themselves that the IS lens must be "better" because it's more expensive and has the letters "IS".
-
\$\begingroup\$ I am not sure that you need to apply crop factor to the 1/focalLength rule, since the projected image circle is the same...you are simply cropping to the center of it. The amount of blur due to shake is the same on FF or APS-C. \$\endgroup\$– jristaFeb 1, 2011 at 19:05
-
7\$\begingroup\$ But the degree of enlargement in the final work will be larger. Motion blur that is imperceptable at size X may be unacceptable at size 1.5 X or 1.6X. The same thing applies to circles of confusion -- they may be the same absolute size on the sensor, but they won't be the same size in the final work. \$\endgroup\$– user2719Feb 1, 2011 at 19:08
-
2\$\begingroup\$ @jrista APS-C has smaller pixels in their sensor. It is like cropping them blowing it up, magnifying the blur. A sub-pixel vibration will not be visible on full frame, however the same vibration may not be sub-pixel on an APS-C due to the smaller pixel size. In fact, when I shoot with my 60D in Av mode and my 50mm lens with Auto ISO, the camera will give priority to shutter speed, keeping it above 1/80 (not 1/50), and raise the ISO instead of lowering shutter speed to 1/focal length. So the camera knows it is an APS-C and applied the 1/(focal length x 1.6) equation instead. \$\endgroup\$– GaptonNov 24, 2011 at 1:48
Purely economic. Both Canon's and Nikon's kit zooms manage to include stabilisation and still cost almost nothing, so it's not the cost of including the mechanism, it's the cost of redesigning the lens. They'd do it if they thought there was a market, but they don't.
Personally I think they're missing a trick. I'd absolutely love a 35mm or 50mm prime with IS, even if it was f/2 as long as it was sharp wide open. I think we'll have to wait for Sony (ie Konica/Minolta) or maybe Sigma to do it. Nikon and especially Canon are pretty unresponsive to user demand. But if Sigma did it and it sold well, then they might be woken from their slumbers.
-
1\$\begingroup\$ I agree with you on the price statement. The 18-55 IS is the second cheapest lens From Canon (I guess?). Probably the expensive lenses may have better IS. But even the kit lens' IS should have been good for the 50mm. \$\endgroup\$– mixdevFeb 3, 2011 at 19:29
-
\$\begingroup\$ @mixdev A maximum aperture of f/3.5 at 18mm (5.14mm entrance pupil) and f/5.6 at 55mm (9.8mm e.p.) leaves a lot more room inside the lens for an IS assembly than a maximum aperture of f/1.8 at 50mm (28mm e.p.) does. \$\endgroup\$ Oct 18, 2018 at 3:38
There is a technical reason for not incorporating image stabilisation in a typical high speed Gauss type lens. Optical IS requires a moving lens component that displaces the image laterally without defocusing. This can be done in most multi-component asymmetric constructions, by moving some component controlled by a stabilising sensor.
However the heavier the moving part is, the more power is needed and more inertia will be fed into the whole system of camera and lens, counter-acting the intention of stabilising. A wide aperture Gauss lens has very heavy components. Moreover, moving any of its component parts will have little effect on where the image falls on the sensor and more effect in defocusing the lens, inducing optical errors in the light path.
It is simply not feasible to stabilise this type of lens optically. Therefore no symmetric construction lens is ever stabilised optically. Only in cameras with sensor stabilisation will you get the very fast Gauss lenses or other symmetric designs stabilised.
So while it is rather easy to stabilise a zoom lens, a prime lens in many cases cannot be stabilised, particularly lenses of symmetric or near symmetric design. Typical wide angle lenses for DSLR cameras as well as tele lenses are however not symmetric designs, and thus in most cases they can be optically stabilised.
It's an interesting idea. But I suppose image stabilization pays off the most on lenses that are slow (F/2.8 is slow-ish by prime standards) or long (100mm or longer) or both. A 50mm prime (which is usually F/2.0 or faster) just isn't a top priority for the IS treatment.
While 50mm primes may be popular among photo enthusiasts, they may still be among the slowest-selling, least frequently revised lenses in a big manufacturer's line-up. Cheap kit zooms are certainly the best-selling lenses Canon or Nikon makes, and so are updated pretty often. Canon's current F/1.4 50mm lens design, on the other hand, is nearly 20 years old (circa 1993).
There are economic arguments for sure (prime lenses tend to be older designs and don't sell in as high volumes as consumer zooms) but here's another way to look at it.
The shutter speed required to prevent camera motion blur is heavily dependant on the focal length. The shutter speed required to prevent subject blur is independent of focal length if the subject fills the same amount of the frame.
For shooting people in a non posed situation I've found that 1/50s is about the limit to avoid motion blur in most cases. The speed you need for shooting people and the speed you need to prevent camera shake are about the same.
Now consider a 50mm lens with a three stop stabilizer that works (going on test results three stops is realistic these days). That takes you to a speed of 1/6s which is going to cause motion blur in a lot of subjects. It's true that sometimes you're not shooting moving subject, but in these cases it's not unreasonable to use a tripod, since you're able to take more time over the shot.
I know some people like to use fast wide lenses with cameras featuring sensor-shift IS, and there are cases where it's useful, however these situations are rare compared to shooting with a telephoto lenses where the speed required to freeze action doesn't prevent camera shake.
It is fairly hard to make a stabilizer in a lens with a large maximum aperture because the standard designs do not have a place where a relatively small/light movable optical element could be placed so that it can compensate vibration/shake. So, you need a "heavy duty" stabilizer unit, like you have in a 70-200/2.8 or even a 200/2.0 lens for many thousand dollars. There is a patent from Canon for a 50/1.8 out there - look at the picture, that kind of shows the problem (for example here: http://thenewcamera.com/canon-50mm-f1-8-is-lens-patent/).
My 5 cents: Would be usable for some (like me!) but fairly expensive to make. If the price was - just for IS, same optical quality otherwise - close to $1000, would you be interested? I might, but not sure...
I have a theory: because you don't need it.
In a 18-55mm, you get f/5.6 at 50mm with image stabilization. In a 50mm f/1.8, you get f/1.8 at 50mm without image stabilization.
The difference in light-collecting ability is 9.68x. If this is not enough for you, get the f/1.4 (difference 16x) or f/1.2 (difference 21.78x).
Now, I have made as test of image stabilization of a 18-55mm lens. It doesn't help at shutter speeds slower than 1/10 s. So, IS helps only for two stops, or gives 4x benefit.
If you really need an IS at 50mm, you can purchase 17-55mm f/2.8 IS, although a 50mm f/1.4 without IS would probably be better in most situations, being 4x faster.
-
\$\begingroup\$ Today's night photographer needs to stay ahead of the unwashed mobile phone crowd as many stops of handheld ability as possible ;) \$\endgroup\$ Mar 13, 2019 at 9:42
It is interesting doing a comparison of fashion shots of models in stidio, with the Canon 24-105mm f4 L and the Canon 50mm 1.4, both at the 50mm focal length, both on a 5D-iii. If the model is still, they are equivalent, with somewhat finer detail in the prime. If the model is moving, the 50mm prime suffers. I can only attribute this to lack of IS in the prime.
As flash sync is limited to 1/200 sec, you sometimes cannot go to higher shutter speeds in the studio to resolve movement in the target.
-
1\$\begingroup\$ The 50mm wide open is obviously more sensitive to change in focus (due to movement) than the slower zoom. Did you use the same aperture on both? Did you make several shots with both to account for the odd failures? Flash duration is much quicker than 1/200 ms (especially at less than full power), and should freeze movement regardless the shutter speed. \$\endgroup\$– ImreOct 27, 2012 at 8:15
-
10\$\begingroup\$ IS does absolutely nothing to help with subject movement, unless you're panning the camera along with the subject! \$\endgroup\$ Oct 27, 2012 at 9:45