4 ms·
What's the point of all those megapixels? Can lenses work at such a crazy sampling frequency? 40 mpix at APS-C means 2.9 micron pixels, or for Bayer array senso
by anovikov 4y ago
What's the point of all those megapixels? Can lenses work at such a crazy sampling frequency? 40 mpix at APS-C means 2.9 micron pixels, or for Bayer array sensor, 86 lp/mm. By comparison, "resolution" line on the MTF chart is at 30 lp/mm sample rate, 2.9x more coarse. Contrast at such extreme frequency will be virtually zero for any real life lens. These pixels are a waste of photons, and storage space.
- bayindirh 4y agoActually no. Modern lenses coupled with stabilization systems get pretty sharp. I have a 24MP A7-III. The very lenses I use can resolve at 60MP siblings, too. Modern glass foundries of Nikon, Canon, Sony and Fuji can create pretty impressive glass nowadays.
- hef19898 4y agoEven with older lenses the question is whether you can see the "lacking" resolution on the final product or not.
- bayindirh 4y agoIt depends on the printing/display medium. At every case, downscaling a higher resolution image pays off. Especially, if it’s a nice monitor for wallpaper use or a big print for wall display.
- hef19898 4y agoAbsolutely! That is actually the only "issue" I have with my cameras, both are 12 MP. Don't het me wrong, they make great pictures, small file size has a bunch of advantages and I prefer to get composition right in camera anyway. But sometimes, usually for wildlife, you need to crop because even a 720mm equivalent lense only gets you that much of a bird in far distance inframe. And then low MPs only give you that much MPs and details to work with. But hey, I still get some nice shots, and most of the time I do landscapes and architecture anyway.
- bambax 4y agoI'm wondering about this too. Storage of raw images esp. is already a difficult problem (not unsolvable, but expensive). I wonder if the future would not be the opposite? Store low-res images and upscale those you need with the help of AI. Maybe in the future we can have just one super high resolution image for a shoot, store all the other images as super low res, and use the high-res image as a reference when upscaling. That would be very cool I think.
- bayindirh 4y agoLossless compression of raw images solve much of these problems. Also flash memory is getting cheaper and faster than these cameras getting resolution. The problem for the serious photographer is nonexistent. A 2TB drive is very sufficient for any amateur photographer using a camera in 24MP-40MP range. We store raw images to archive what we got. AI interpolates. It's not acceptable.
- bambax 4y ago> A 2TB drive is very sufficient for any amateur photographer No. Maybe for you, but not for me. I have a 4TB drive and it's almost full. I need to upgrade to 6 or 8 TB and it's a pain.
- dannyw 4y agoFairly irrelevant, but it's not a Bayer sensor. It's X-Trans. That's why I love Fuji.
- CarVac 4y agoX-trans… no actual advantages, a big pain. Source: I develop raw processing software.
- kruxigt 4y ago
- Kon5ole 4y agoX-Trans has a color sampling pattern of 3x3 pixels instead of 2x2 for bayer. This means that color for X-Trans is sampled from a 50% larger area, which 1) reduces noise 2) makes noise look slightly "better" than bayer noise (according to many including me). It also has a downside in that it reduces fine detail somewhat compared to Bayer. In practical terms it also has a downside in being less supported by software, maybe because raw processing software developers consider it a big pain. ;) I'm happy they made X-trans also for the 40mpx, it should improve the details (which is a weak spot due to the 3x3) significantly and still maintain the benefits of a 50% larger color sampling area.
- CarVac 4y ago> This means that color for X-Trans is sampled from a 50% larger area, which 1) reduces noise 2) makes noise look slightly "better" than bayer noise X-Trans has less color sampling per unit area, so you get more chroma noise. So the demosaicing either gets lots of color artifacts, or it has lots of baked in chroma noise reduction. On Bayer, you can simply choose a different demosaic method (IGV or LMMSE, for example), or apply additional chroma noise reduction, and the playing field is leveled noise-wise.
- Kon5ole 4y agoHave a look at the color patches on dpreview's studio shot comparison and compare any X-trans sensor with a bayer sensor with the same size and pixel count. Use RAW and high ISO to clearly see the difference. X-trans is clearly better for noise, and also clearly worse in detail.
- sbaiddn 4y agoAt f1.8 maybe. At f22 not a chance in hell. Diffractiom blurs everything. All should not be lost though. A lot of noise in a picture is independent of a pixel's neighbor. So, if you have more pixels, you can average out a neighboring pixel to get better low light sensitivity.
- twehrou32423f 4y agoThey certainly do for full-size sensors; I've heard APS-C lenses are harder to make, but don't see why not.In any case, even if they don't it helps bring down the noise when down-sampling. That said, all of us do love new toys : D
- wyager 4y agoIn theory, we're still well below the diffraction limit even at f/4. Aberration, depends on the lens. Plus, superior lowpass filtering in software.
- smilekzs 4y agoI think it's the same reason audio is processed at 192 kHz --- basically oversampling to make anti-aliasing less of an issue.