The primary issue is that digital sensors are more sensitive to the light path than film. As the angle of incidence increases, diffraction around the edges of photosite wells: chromatic aberration and moire effects become a problem, as does total illumination at the photosite causing darkening at corners and edges. This effect is less pronounced with SLR lenses than with RF camera lenses because SLR lenses have already been formulated to accommodate the mechanical obstruction of a swinging mirror, moving the nodal point forwards relative to the focusing plane, but it remains an issue as you get to the ultrawide range shorter than 24mm in focal length.
To solve this really requires a different approach to lens design for the digital sensor with short focal length lenses in particular, overall a wider lens mount diameter relative to the format size helps a lot as it gives more leeway for lens design to produce optimized lenses. A shorter mount register, relative to format size, also helps as it allows room in the lens design for additional collimating elements to the rear of the primary lens groups. Both Canon and Leica, in producing 24x36mm sensors and in accommodating RF lens designs respectively, have both opted for a compromise sensor with offset microlenses to help reduce the diffraction/lightloss/CA issues, and with in-camera image processing to massage the data as well. Olympus, with Kodak and other vendors, designed the 4/3 system specifications from the ground up with these notions in mind. Note the diameter and register of the lens mount relative to the format in that system: this is a near-ideal, optimized relationship of these components: regardless of whether they've achieved other performance goals, they suffer very little moire, chromatic aberration and corner light falloff. But of course they also started afresh with lens designs and ignored backwards compatibility with OM system lenses other than through adaptation with reduced functionality. The smaller format 16x24 sensor ameliorates a good number of these issues for legacy lens systems while still maintaining a great deal of quality and compatibility. The secondary issue is that Pentax has elected to go with in-body image stabilization rather than in-lens stabilization. This addresses three concerns: first was the notion of the cost/bulk/weight compromise in lens design, second was the issue of lens quality and durability with constantly moving elements, and third was the notion of compatibility/applicability for older lenses. To achieve good image stabilization this way with long lenses requires movement of up to 5mm off the null position center, which means that lenses which barely cover the 24x36mm image circle would no longer be able to take advantage of image stabilization with this schema. --- For my uses, the 16x24mm sensor format has worked well, allows suitable quality and sensitivity as well as adequate wide angle coverage with exceptional quality, such that I am happy to have the smaller size optics required for my desired field of view range as well as adequate/satisfactory compatibility with older lens designs. Good noise and quality at ISO 1600 and 10Mpixel is good enough for my work. I also look forward to the larger, higher quality results with medium format sized sensors ... and I expect prices in that market to remain pretty steep for a while to come. 24x36mm sensors ... well, for me it's mostly irrelevant now as I have tailored my kit to the 16x24mm format, but if they do come out with a price/quality sensible alternative in that size that provides compelling advantage AND is compatible with all my current lenses, it really requires that I buy just one more slightly longer focal length lens and I'll be ready for it. Godfrey On Dec 21, 2006, at 6:01 AM, Tom Simpson wrote: > Im kinda' new here, but one of the things I picked up in passing is > that > the K-mount might be fundamentally incompatible with a full-frame > DSLR? > Or did I miss something? If I didnt, please explain to me what the > issue > is, as I was under the impression that if you put a sensor the size > of a > 35mm frame in the same place in reference to the lens as a 35mm > frame of > lens, it would all work out just peachy. Why not? I hope not, as I am > hoping that the bigger sensor will give the quantum physicists more to > work with to give us all the resolution we can eat AND bring dynamic > range more in line with film. From my understanding, this is likely an > either/or proposition with an APC sensor. Cant have both. -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
