Marsu42 said:
briansquibb said:
Nikon of course is a 1.5 crop
I know, fuzzy description on my side - but 1.5 or 1.6 doesn't really make a difference, does it?
It does factor into spatial resolution, DLA, etc. Those are actually pretty rough numbers, I think Canon's is actually closer to 1.623x, which increases the difference a bit more over Nikon's crop. Someone mentioned we are "well past" the point of diminishing returns. I wouldn't say that at all...you can always gain from increases resolution, and so long as we are not completely outresolving all lenses on the market, we have plenty of returns to gain. We would only begin moving into "diminishing returns" realm once sensors are offering more spatial resolution than the highest resolution lenses that are readily available at "affordable" (read that as professional grade lenses that professionals would buy, since that is where a significant amount of Canon and Nikon revenue comes from) prices.
Even if an image "appears" soft at 100% crop doesn't actually mean it is. Computer screens are 1/3rd to 1/4th as dense as the average print (@300ppi), and if you have fine details that you want to fully reproduce, you need a high resolution print (@600ppi). At 600ppi, computer screens are 1/6th to 1/8th as dense. Printing a 7D photo that might look soft on a computer screen usually results in a beautifully detailed, crisply sharp 13x19" 300ppi print. A 24mp sensor should be able to get you up to around 16x20, etc.
There are other "hidden" benefits that are very hard for consumers to see, understand, and realize, so they are never marketed. One of the most significant, and one of the reasons that I believe more resolution is a good thing despite the negative 100% crop connotations, is that once the bulk of your noise is at a sub-detail level...and by that I mean you need many RGB pixels to comprise the smallest facets of detail in your photo, noise removal becomes significantly more effective. When you remove noise in an image where even a single pixel might comprise a full element of fine detail, you run the high risk of blurring detail. Using Lightroom as an example, you might crank sharpening up to 60, and set the sharpening radius to 1.0, maybe 1.3. With more pixels per finest detail, you could crank the sharpening radius up to 2, or even the max of 3 (and when the smallest element of meaningful detail in a photo requires 5-10 image pixels to render, you could even benefit from larger sharpening radii). A larger sharpening radius helps to eliminate noise in and of itself, reducing the need to use luminance noise correction, or reducing the level to which you might need to push it.
There are plenty of benefits to be gained from pushing sensor resolution "beyond the limit" (i.e. the 173lp/mm spatial resolution of a perfect f/4 lens, say). Its just that those benefits are often complex, hard to understand, and difficult to sell. The inevitable perceptual outcome of continued progress on the sensor resolution front will always eventually be
"all my photographs are soft!" not because they really are in any meaningful context (i.e. downscaled for exhibition online or printed at up to 2x enlargements)...only because they look soft at 100% crop on a relatively low density screen.
(I would actually DIE for a 300ppi screen with a "100%" crop actually set to 33%...that would eliminate the "That camera's pictures are soft!" issue once and for all, with the added benefit that if you zoom in beyond 100%, things don't go pixelated until 400%.
)
