01-10-2016, 12:55 AM
Doing something for a long time doesn't make you good at it 
GM has been making cars for a long time but they still managed to go bankrupt, and they still put some real pieces of shit on the road (which isn't to say everything they make is truly terrible). Same with Toyota and their now constant recalls for the frames of their vehicles.
Good lens designs have low angles of incidence on all optical surfaces; high AoIs produce higher order aberrations, and very sensitive to alignment. To reduce the AoI one method is to "expand" the lens design and make it larger.
In the case of telephoto lenses, it would be trivial to scale an existing double gauss lens up to 200mm or 300mm and then slow it down to f/4. The result would be functionally the same as a sonnar-type design, a telephoto, a tessar, etc.
There is a rather large formula for the telephoto ratio of a lens... it is a first order property akin to focal length or aperture. Certainly a lower telephoto ratio (and thus shorter lens) increases the AoI, which reduces performance. If the AoI cannot be sufficiently controlled in an all-spherical design and the aberrations cannot be canceled through clever design, newer technologies can be adopted (DO, aspherics, freeforms, newly "BR") to reduce the aberrations to an acceptable (i.e. near-zero) level.
The great white lenses often have relatively poor telephoto ratios (near 1x) as their designs accommodate many features (such as super-fast autofocus, teleconverter compatibility, image stabilization, reduced focus breathing, and exceptional correction), but the 300/4 PF sets a record in the consumer space at ~0.48x and is I would say is nearly at the limit for such a large detector.
Assuming Olympus lacks the skill to properly implement diffractive optics, they could use aspheres to solve the same issues. However, even if they simply remove the DO aspect and accept the reduced performance, the detector is much smaller so coma, astigmatism, petzval, distortion, and lateral color are already reduced greatly. Re-optimizing with that in mind would yield a good performance.
I mean, the Canon 300/4 is a hair shorter and it covers 4x the detector while being well over a decade old now.
Regarding the ability of the various consumer optics firms, I would rate them something like the following:
1. Zeiss - EUV litho and medical division for their aggressive utilization of freeform and aspheric optics
2. ARRI - same reasons
3. Canon - they seem to be miracle workers re: price/performance (10-18 stm etc) and also capable of very radical designs (11-24)
4. Leica - Excellent use of aspheres, very good to excellent performance in extremely small sizes
5/6. Zeiss (ZE/Milvus/FE/etc) - not as good as they used to be but still offering top-notch lenses
5/6. Sigma - very good lenses of late
7. Nikon - They're this far down for by and large being too conservative in the last decade+. Simple lenses, average/good performance, expensive, not particularly small.
8. Schneider - very good lenses, but extremely long development times. Give a man 4 years to design a lens and almost anyone could produce something great by chance.
9. Tamron - I may be being overly punitive because of sloppy construction on many/most lenses, but they have some excellent models like the new 35 and 45mm lenses in that regard.
10-xx Sony, Olympus, Panasonic, Fuji et al - I would never purchase a lens from any of these companies. 16% distortion in some models is simply inexcusable, very bad assembly quality, poor mechanics, and not particularly great performance out of any of them.
GM has been making cars for a long time but they still managed to go bankrupt, and they still put some real pieces of shit on the road (which isn't to say everything they make is truly terrible). Same with Toyota and their now constant recalls for the frames of their vehicles.Good lens designs have low angles of incidence on all optical surfaces; high AoIs produce higher order aberrations, and very sensitive to alignment. To reduce the AoI one method is to "expand" the lens design and make it larger.
In the case of telephoto lenses, it would be trivial to scale an existing double gauss lens up to 200mm or 300mm and then slow it down to f/4. The result would be functionally the same as a sonnar-type design, a telephoto, a tessar, etc.
There is a rather large formula for the telephoto ratio of a lens... it is a first order property akin to focal length or aperture. Certainly a lower telephoto ratio (and thus shorter lens) increases the AoI, which reduces performance. If the AoI cannot be sufficiently controlled in an all-spherical design and the aberrations cannot be canceled through clever design, newer technologies can be adopted (DO, aspherics, freeforms, newly "BR") to reduce the aberrations to an acceptable (i.e. near-zero) level.
The great white lenses often have relatively poor telephoto ratios (near 1x) as their designs accommodate many features (such as super-fast autofocus, teleconverter compatibility, image stabilization, reduced focus breathing, and exceptional correction), but the 300/4 PF sets a record in the consumer space at ~0.48x and is I would say is nearly at the limit for such a large detector.
Assuming Olympus lacks the skill to properly implement diffractive optics, they could use aspheres to solve the same issues. However, even if they simply remove the DO aspect and accept the reduced performance, the detector is much smaller so coma, astigmatism, petzval, distortion, and lateral color are already reduced greatly. Re-optimizing with that in mind would yield a good performance.
I mean, the Canon 300/4 is a hair shorter and it covers 4x the detector while being well over a decade old now.
Regarding the ability of the various consumer optics firms, I would rate them something like the following:
1. Zeiss - EUV litho and medical division for their aggressive utilization of freeform and aspheric optics
2. ARRI - same reasons
3. Canon - they seem to be miracle workers re: price/performance (10-18 stm etc) and also capable of very radical designs (11-24)
4. Leica - Excellent use of aspheres, very good to excellent performance in extremely small sizes
5/6. Zeiss (ZE/Milvus/FE/etc) - not as good as they used to be but still offering top-notch lenses
5/6. Sigma - very good lenses of late
7. Nikon - They're this far down for by and large being too conservative in the last decade+. Simple lenses, average/good performance, expensive, not particularly small.
8. Schneider - very good lenses, but extremely long development times. Give a man 4 years to design a lens and almost anyone could produce something great by chance.
9. Tamron - I may be being overly punitive because of sloppy construction on many/most lenses, but they have some excellent models like the new 35 and 45mm lenses in that regard.
10-xx Sony, Olympus, Panasonic, Fuji et al - I would never purchase a lens from any of these companies. 16% distortion in some models is simply inexcusable, very bad assembly quality, poor mechanics, and not particularly great performance out of any of them.