[honey bo bo]

That's an interesting and valid point of view!

Scientifically sound observations of my 35 year old son who is a Technocrat with the local Phone Company. He has been my Photo helper since he was a baby.was given various camera's over the years has even apprenticed in my Lab before the turn of the Century. Last cameras were a Nikon S6400 and a Canon EOS. Last and only picture taken at the age of 8 was the Air intake of a F-16 Fighter with a camera because he didn't have a Cell phone. Since then a million Pics on a iPhone 3,4,5,6. ########## etc. Loves my Pic's but won't touch the Holy Grail.

[dtmateojr]

That does not compute. If they all have 16MP, different size sensors can't have the same pixel density, so what do you mean - same density or same total number of pixels?

I'm also curious about how you obtained your plot. When I go to DxO mark here and click on Measurements -> SNR 18%, the plot looks rather different and shows a pretty sizable advantage of the Canon over the Olympus.

I said saNe not saMe. Anything more than 36Mp is insaNe.

Look at the SCREEN snr in dxomark because that tells you the REAL sensor performance. PRINT snr varies depending on print size and is not constant.

[Ikarus]

Look at the SCREEN snr in dxomark because that tells you the REAL sensor performance. PRINT snr varies depending on print size and is not constant.

Sure, it isn't constant, but what these two plots tell me is that at this particular common print size, despite both having similar sensor technology (having the same sensel SNR), the Canon has an almost 6dB better SNR than the Olympus due to its sheer sensor size.

[mee]

I'm looking forward to the 36MP FF sensor in the K-1 and, maybe, one day the 100MP sensor in the next Pentax 645 body.

No one here can accurately predict 10 years down the road in the camera world. As monochrome said.. soo much can happen in that time (iPhone isn't even 10 years old and it has made so many improvements in its lifetime).

The better idea is to get a feel for what is out there now, what will be out there tomorrow (near term) and use what best suits you. I would find it odd to try to play the camera market like the stock market.

[dtmateojr]

Sure, it isn't constant, but what these two plots tell me is that at this particular common print size, despite both having similar sensor technology (having the same sensel SNR), the Canon has an almost 6dB better SNR than the Olympus due to its sheer sensor size.

Correct.

You could improve the SNR of the print from the smaller sensor though simply by printing smaller or viewing it from farther away. Billboards for example require less than 2Mp. Both approaches do not really tell you about the performance of the sensor, no? Because print SNR has absolutely got nothing to do with light gathering or hardware performance in general. And yet the FF fan boys will keep selling you the print SNR graphs to justify spending three times more money for essentially the same camera.

When dxomark releases the SNR results for the K1, I can assure you that it will be no different from the ancient K5. It's predictable because it's very simple math and physics.

[Ikarus]

Correct.

You could improve the SNR of the print from the smaller sensor though simply by printing smaller or viewing it from farther away. Billboards for example require less than 2Mp. Both approaches do not really tell you about the performance of the sensor, no?

The fact remains that for a given print size, the FF sensor gives you a higher SNR, a whopping 6 dB in the specific case selected by DxO Mark. Clearly, this would have to be attributed to sensor performance. Of course, whether you consider this relevant for your own purposes is a different question. Some people add noise to their shots for artistic expression.

---------- Post added 03-16-16 at 09:53 PM ----------

When dxomark releases the SNR results for the K1, I can assure you that it will be no different from the ancient K5.

I wouldn't be surprised either, as far as the sensel noise is concerned.

It's predictable because it's very simple math and physics.

... as is the fact that the K-1 will outperform the K-5 handily in terms of noise at a given print size.

[quant2325]

Getting back to the OP:
1) Smartphones will continue to take over and eventually dominate the industry. Modular add-on lenses and even sensors will improve and become available.
2) APS-C sales will slowly decline, but the cameras will be better and better due to sensor improvements, which will improve in a way similar to Moore's Law.
3) For most serious amateur photographers an APS-C camera is all that is needed. At some point the MP in each sensor will be "good enough" and the prices will significantly drop.

[wildman]

What do you think will happen to APS-C over the next 10 years?

For all practical purposes and in terms of markets and possible future tech developments...

...FF will become the new "MF"

APS will become the new "FF"

4/3 will more or less disappear as a market - with future tech developments in other sensor formats it will prove to be both too big and too small.

The mass mainstream dedicated camera market will be dominated by 1" sensor cameras in one form or the other - both small enough and good enough. The new mainstream norm. This, of course, assumes there will be a large mass market for dedicated cameras in ten years - a big assumption at this point in time.

[dtmateojr]

The fact remains that for a given print size, the FF sensor gives you a higher SNR, a whopping 6 dB in the specific case selected by DxO Mark. Clearly, this would have to be attributed to sensor performance. Of course, whether you consider this relevant for your own purposes is a different question. Some people add noise to their shots for artistic expression.[COLOR="Silver"]

Nope. That is entirely due to downsampling. They are interpolating the SNR when a 50Mp Canon is printed at 8x10. The interpolated graph is derived from the real SNR measurements which is that graph I posted here.

Any image when downsampled will result in apparent increase in SNR. Downsampling is no different from increasing the viewing distance. That's why drive-in movies look fine even if they are standard definition videos. That's why billboards are only 2Mp.

[assa1]

Nothing will be the same when QIS will finally ready for production. It could turn the photo industry upside down and my imagination isn't able to imagine myself whan may happen.

[dtmateojr]

Here's why relying on dxomark PRINT snr leads to wrong impressions with regard to the true performance of a sensor.

If you compare print snr of D4 and D800 you would think that you got a bargain for going with the half-priced D800



However, looking at the screen (real) snr reveals the truth that the D4's sensor is superior to the semi-pro D800



So, do not rely on the print snr because it does not give you the real picture (pun intended).

[Rondec]

These are really tough things to predict. The question really lies more with the brands themselves than with consumers. As long as Canon, Nikon and Sony continue to sell APS-C ILCs for less than they sell full frame cameras, I would say there will be plenty of folks who are interested in purchasing them. Point and shoots are in the process of dying.

Enthusiasts on the forum are always looking for a little bit of improvement in image quality -- whether that means getting a faster lens, a bigger sensor, or a newer generation camera. But in the real world, we have reached the point where all of these cameras -- micro four thirds, APS-C, and full frame are good enough and this is what has hurt ILC sales the most. People are shooting most day to day snap shots with their phones and then break out their SLRs or mirrorless cameras for low light situations where they need it. And even a camera from two or three years ago is really good enough. If you own a K30 or a 70D or a D7200, it is hard to say (short of you dropping your camera and breaking it) what features a camera company would have to stick into their camera to get you to want to buy a new camera. 4K video is supposed to be it, but I don't really buy that people are going to buy a new camera based on that.

Monochrome is probably right that in the long run EVFs will replace OVFs -- at least on mid to entry level cameras. As to whether or not that means new lens mounts, I am a little questionable on the subject. Canon and Nikon's biggest tool for selling cameras is their amazing lens portfolio. If they ditch their current lens mounts to go with a shorter registration distance, even if they have a nice adapter, and they risk angering their current users and losing the leverage they have over Sony in the marketplace. Feels like they won't change till they have to.

[beachgardener]

in the long run EVFs will replace OVFs

I would certainly object strongly to EVF replacing OVF in APS-C DSLR. Long live OVF!!!!!

[BigMackCam]

Why does it bother you so greatly that a smaller format has IQ advantages over larger formats in certain situations?

Firstly you were the one who started imposing unrealistic restrictions to justify FF and now I'm giving you very realistic situations and you can't accept it? Your bias is showing inspite of your claims to be "fair" and only here for "discussion purposes".

Oh, you misunderstand - it doesn't bother me at all, nor am I biased. I have FF, APS-C, 2/3" and smaller sensors in my cameras and I enjoy and make use of the benefits from them all (in fact, I lean more towards APS-C than FF for my own purposes - it seems to suit me best overall). I certainly don't believe one is better than the other, or that they are all capable of exactly the same results with the exception of IQ. On the contrary; it's you who believes the only advantage of larger formats over smaller is sensor performance, and you're coming across as argumentative and defensive, which I think is a shame. Anyway, good luck to you.

[Ikarus]

Nope. That is entirely due to downsampling. They are interpolating the SNR when a 50Mp Canon is printed at 8x10.

The requirement for downsampling by a certain amount is not arbitrary. It is what you have to do when doing an apples-to-apples comparison at the same print size, and the fact remains that the Canon's print is cleaner. I have a hard time seeing how this bolsters your case that m43 is as good as FF by that metric.

---------- Post added 03-17-16 at 09:05 AM ----------

Here's why relying on dxomark PRINT snr leads to wrong impressions with regard to the true performance of a sensor.

If you compare print snr of D4 and D800 you would think that you got a bargain for going with the half-priced D800

The numbers don't lie. If getting clean prints at that size is all you care about, you did get a bargain with the D800. I don't want to get in an argument with you about why that is. Suffice it to say that bringing up a case where two FF cameras match each others performance by that metric does not help your case claiming that m43 is just as good. What would help is you bringing up an example where an m43 camera actually gets close to an FF in this department, but I suspect you'll have a very hard time digging that up.