• News about Products and Latest Company Direction • Summaries of Reported Problems and Potential Issues • Technical Articles on Photographic Gear and Technologies
Friday, October 11, 2013
Worth Re-posting: How Meaningful is a 24MP APS-C Sensor?
fact is, everybody wants a full frame. if Pentax is not going to do it, at least the k3 should able to do sweep panorama. This way you can at least fake a full frame frame look with the Brenizer method with ease instead in Post. The Sony Nex however, it has sweep panorama builtin and the full frame A7r is going to be out soon.
Now I use the cheapo Sony A3000 + Mitakon lens turbo + FA lenses to "make" photos. It's not that bad with focus peaking. And meanwhile I will continue to do point and shoot with the k5 and as a backup. I'd be better off spending my money on full frame lenses than to upgrade to a gimmick. If the Chinese manufacturers are smart enough, like Mitakon , they should team up with electronic toys maker and software developers. There is a huge market for autofocus lens adapters. Also, If sigma can come out with unlicensed PK mount autofocus lenses, I bet they can do a better version of the speed booster. Remember, Sigma already has an autofocus version of the 2x teleconverter.
From what I see in resolution tests on various websites is that the 24mpix cameras do indeed have significantly more resolving power than the 16mpix ones
I agree with Rice, 24mp APC is a stretch unless using the best glass which really doesn't exist, to bad Zeiss discontinued there Pentax mount lenses. Ricoh is now going to have to put out a Technical Bulletin like Nikon did when the D800/E came out explaining what lenses work the best, proper technique etc... But I am sure they won't.
In regards to the D800 there is really only one lens that is really up to 36MP, that is the new Zeiss 55/1.4 The new Sony's coming out soon, I would stay away from the 36mp FF I would have like to have seen the 24mp FF version have no AA but I don't think that will be the case.
Contrary to Rice's opinion, there are many lenses that are pretty good on 24MPix APS-C and on 36Pix FF as well. 24MPix FF with no AA filter would be pretty prone to more as it has the pixel density of ~10MPix APS-C...
24mp is already far more than real pictures can chew. AND it slows down processing AND simply gobbles up space like never before. All these disadvantages for the occasional aerial shoot with a fast prime on infinity focus? THAT is where you might benefit from that resolution if your subject matter actually requires any sort of depth.
14mp..16mp was already diffraction-limited beyond ~ f/7. So hopefully the next time I see a bump in resolution in Pentax-land, it will concern the 645D-sucessor. There simply is no point in pumping up the the number of sub-pixel dots any further. It's time to focus on their light-sensing capabilites and/or find a way to get rid of the Bayer-matrix. When is Sigma going to license their foveon-tech / supply Pentax-Ricoh with a nice full-colour-pixel based sensor? That moment - even if that sensor only consists of 10..12 million dots: we all win! Unfortunately "Foveon" hasn't been ready for prime-time, yet. Especially considering it still is much too power-hungry, with limited usable sensitivity and lacking any video read-out.
Alright: case in point: if the 24mp-sensor of the K3 can be pushed beyond the capabilities of its 16mp-brother seen in the K-5 (in the sense of colour-differentiation and dynamic range) -> alright with me. It just is time to stop the meaningless megapixel-mania. Do other stuff. Sensors have become "cheap" nowadays. Plenty of other things to improve. Like some new exciting lenses. How about some FA28/FA31-sucessor with wr?
More pixels with the same noise performance means better quality in the end and more post processing possibilities, so event if you don't get more information, you have more pixels to work with and better quality picture as a final result on a same size print.
The more dense the sensor is, the less chance to get moire in real life situations you have, first of all, because of patterns that cause it become smaller and less common, second of all, because the lens resolution and diffraction limit the "moire prone window" to pretty narrow aperture settings range.
Completely wrong concept! As I have said many time, Nyquist limit is NOT dependent of pixel density! You should really find some elementary digital sampling theory articles to read and understand the basics.
As for the diffraction limit, now you start to agree with me, i.e., the denser the pixels, the easier the diffraction limit to be reached. So, a dense sensor is no good again!
Of course it is - the Nyquist limit is the maximum theoretical resolution a sensor can resolve, so it is a function of a sensor density as well, different density sensors of the same size have different Nyquist limits. And as for the diffraction limit - when you reach it, it's the lens which becomes a natural "AA" filter, you don't get more "real" information, but you still have it described by more pixels, so it looks better printed or downsampled to a web resolution (oversampling). Basically, denser the sensor with the same "per pixel" noise level, better the overall picture quality on the final output.
I'm a engineer and often do work in signal processing so I'm very familiar with nyquist. jaad75 above has a point. I think it is you who needs to do a little reading up. The lens focusing an image onto a sensor produces a given frequency, and Nyquist says you need at least 2 physical samples for each cycle to avoid aliases. So yes, at a basic level, the higher the pixel density the less likely you are to see alias effects with all other things being equal.
We all have been talking about a more packed sensor against a less densed one, with the same pixel count. So, if you still think this is relevant to Nyquist, you would better go back to the school to study engineering maths 101 again, if you really had an engineering degree!
Logging you in...
Anonimity is Great · 599 weeks ago
RiceHigh 110p · 599 weeks ago
jaad75 · 599 weeks ago
Sony camera do mundo · 599 weeks ago
at least fake a full frame frame look with the Brenizer method with ease instead in Post. The Sony Nex however, it has sweep panorama builtin
and the full frame A7r is going to be out soon.
Sony camera do mundo · 599 weeks ago
Mike · 599 weeks ago
RiceHigh 110p · 599 weeks ago
JohnC · 599 weeks ago
Walt · 599 weeks ago
Steven k · 599 weeks ago
Ricoh is now going to have to put out a Technical Bulletin like Nikon did when the D800/E came out explaining what lenses work the best, proper technique etc... But I am sure they won't.
In regards to the D800 there is really only one lens that is really up to 36MP, that is the new Zeiss 55/1.4
The new Sony's coming out soon, I would stay away from the 36mp FF I would have like to have seen the 24mp FF version have no AA but I don't think that will be the case.
jaad75 · 599 weeks ago
Heiko · 599 weeks ago
24mp is already far more than real pictures can chew. AND it slows down processing AND simply gobbles up space like never before.
All these disadvantages for the occasional aerial shoot with a fast prime on infinity focus? THAT is where you might benefit from that resolution if your subject matter actually requires any sort of depth.
14mp..16mp was already diffraction-limited beyond ~ f/7. So hopefully the next time I see a bump in resolution in Pentax-land, it will concern the 645D-sucessor. There simply is no point in pumping up the the number of sub-pixel dots any further. It's time to focus on their light-sensing capabilites and/or find a way to get rid of the Bayer-matrix. When is Sigma going to license their foveon-tech / supply Pentax-Ricoh with a nice full-colour-pixel based sensor? That moment - even if that sensor only consists of 10..12 million dots: we all win! Unfortunately "Foveon" hasn't been ready for prime-time, yet. Especially considering it still is much too power-hungry, with limited usable sensitivity and lacking any video read-out.
Alright: case in point: if the 24mp-sensor of the K3 can be pushed beyond the capabilities of its 16mp-brother seen in the K-5 (in the sense of colour-differentiation and dynamic range) -> alright with me. It just is time to stop the meaningless megapixel-mania. Do other stuff. Sensors have become "cheap" nowadays. Plenty of other things to improve.
Like some new exciting lenses. How about some FA28/FA31-sucessor with wr?
jaad75 · 598 weeks ago
RiceHigh 110p · 598 weeks ago
jaad75 · 598 weeks ago
RiceHigh 110p · 598 weeks ago
As for the diffraction limit, now you start to agree with me, i.e., the denser the pixels, the easier the diffraction limit to be reached. So, a dense sensor is no good again!
jaad75 · 598 weeks ago
RiceHigh 110p · 598 weeks ago
Joe · 598 weeks ago
RiceHigh 110p · 598 weeks ago