[audiobomber]

Please re-read my comments. I didn't say APS-C will max out at 16-18 Mpix, period. I said "with current technology".

The Sony 24mp sensor is current technology and has the same noise performance as the Sony 16mp sensor therefore your statement, that 16-18 maxes out current technology, is incorrect.

Of course technology will continue to improve, that's why I believe we're well below the highest pixel count for APS-C. Lots of people bemoaned sensors moving beyond 10-12mp, once considered to be the sweet spot for APS-C. A few years ago, it was widely thought that more pixels would mean higher noise. Sony has proved that it ain't so with its 16 and 24mp sensors.

[normhead]

The Sony 24mp sensor is current technology and has the same noise performance as the Sony 16mp sensor therefore your statement, that 16-18 maxes out current technology, is incorrect.

It's bizarre that so many years after experts said man would never fly faster than the speed of sound, or go into space, we still have folks that think they can make a name for themselves imposing limits on technology. Even more bizarre are those that try to impose limits on a class of technology, like APS-c while pretending that FF or another system don't suffer from the same limitations. In the case of this board, we have folks who use 5 year old questionable data to make their case. At this point, we've reached the stage where other considerations besides the sensor technology used are the defining factors in photography. I know it upsets the geeks but we have entered the 400 horsepower car stage of photography. We don't need more MP, and as printer technology improves we'll need even less. Camera makers will continue to produce high Mp cameras, the way auto makers continue to make 400 horse power cars, not for people who need them because almost no one needs them, but for people with small *****es who are trying to compensate.

Dynamic range , getting to a full 16 bit palette and increased sensitivity are all more important at this point than increased Mp. Everyone is impressed by more MP, but no one can post examples clearly showing why they need it. While it's possible it might be necessary to the few, those doing museum type prints at sizes up to 60 inches, that isn't a market most of us have access to, and is hardly worth even discussing. People love to pretend they need big. But using what you've got efficiently is almost always better than acquiring the biggest, no matter what the product.

I don't listen to the people driving 400 Hp gas guzzlers either. They are off in their own little out to lunch world. They resent reality, and have their little arguments they use to justify their excesses. Most of which make sense as excuses to buy something new rather than a verifiable solution to a list of real needs.

[Anvh]

Dynamic range , getting to a full 16 bit palette and increased sensitivity are all more important at this point than increased Mp.

Imaging that maybe over some years the pixels are so small that 4 pixels can easily be binned together, then you can throw away demosaicing that's normal these days.
Resulting in more accurate images.

Depending on the colour we could be working not with 16mp but with only 4mp at some area's within the photo with the K5.

[audiobomber]

I'm looking forward to 24mp, for the same reason I moved to 14.6; more megapixels means more ability to crop. I've been waiting for a 1.4X TC, but now I'm thinking a 300mm prime and more megapixels might be preferable. Cropping means no lens change, one stop faster, no extra cost, wider or narrower FOV.

[normhead]

That's a good point. Twice as many Mp. All other things being equal, mean a longer effective focal length for each lens. So if you consider a wildlife shoot, you can get the same size image with 36 Mp camera and a 300mm lens and a 50% crop, as you could with an 18 MP camera and a 600mm lens. The issue with this theory, is that the further you are from an image the "flatter" it looks, and despite using high resolution lenses, the more detail you lose, especially in shooting less than ideal conditions, like a tripod that isn't completely locked down or hand held. As someone said," if you don't like your pictures move closer." I'm always thinking about achieving the best possible image. And that will always be an image taken as close as possible to the subject, with an appropriate lens for your subject. Shooting further away with a longer lens and more MP can give you shots you won't get any other way, but could also have you settling for less than the perfect image, because you can. But, if you're shooting from closer to your subject, you have the advantage of more detail and a more pleasing and natural look to the final photograph.

That being said, if you're photographing Grizzly bears you want to be using at least 600mm on an FF camera (400mm APS-c) because of the distance you should be away from them for safety. Trying to pretend your 200mm APS-c is the equivalent of 600mm FF because it's 36 Mp could get you into some grizzly trouble. The whole area discussion is fraught with unknown factors and I'm guessing only field tests will resolve many of the issues. Sometimes 600 mm is 600mm.

[cbope]

Do you've any numbers how many photons they collect now, so a percentage.
I thought they pretty much nailed it these days.

Sorry, I don't have numbers for camera sensors. No doubt the amount is reasonably good, otherwise we would have noisy images. Some light will be lost via the AA filter and Bayer arrangement in front of the sensor though.

I would almost predict a minor breakthrough when AA filters are no longer needed as sensor technology and designs improve. There is no doubt that resolution will increase at that point, both native pixel resolution and actual image quality, once cost is not a significant barrier.

[cbope]

Dynamic range , getting to a full 16 bit palette and increased sensitivity are all more important at this point than increased Mp. Everyone is impressed by more MP, but no one can post examples clearly showing why they need it. While it's possible it might be necessary to the few, those doing museum type prints at sizes up to 60 inches, that isn't a market most of us have access to, and is hardly worth even discussing. People love to pretend they need big. But using what you've got efficiently is almost always better than acquiring the biggest, no matter what the product.

I couldn't agree more, there are more important issues to improve than raw MPix numbers for the vast majority of us.

[Anvh]

Sorry, I don't have numbers for camera sensors. No doubt the amount is reasonably good, otherwise we would have noisy images. Some light will be lost via the AA filter and Bayer arrangement in front of the sensor though.

I would almost predict a minor breakthrough when AA filters are no longer needed as sensor technology and designs improve. There is no doubt that resolution will increase at that point, both native pixel resolution and actual image quality, once cost is not a significant barrier.

Funny enough the higher the pixel count the less need there is for an AA filter...

But is there an increase lost in photons if the pixel count increase if not then more or less pixels shouldn't matter for the average.

[salgueiros]

It's bizarre that so many years after experts said man would never fly faster than the speed of sound, or go into space, we still have folks that think they can make a name for themselves imposing limits on technology. Even more bizarre are those that try to impose limits on a class of technology, like APS-c while pretending that FF or another system don't suffer from the same limitations. In the case of this board, we have folks who use 5 year old questionable data to make their case. At this point, we've reached the stage where other considerations besides the sensor technology used are the defining factors in photography. I know it upsets the geeks but we have entered the 400 horsepower car stage of photography. We don't need more MP, and as printer technology improves we'll need even less. Camera makers will continue to produce high Mp cameras, the way auto makers continue to make 400 horse power cars, not for people who need them because almost no one needs them, but for people with small *****es who are trying to compensate.

Dynamic range , getting to a full 16 bit palette and increased sensitivity are all more important at this point than increased Mp. Everyone is impressed by more MP, but no one can post examples clearly showing why they need it. While it's possible it might be necessary to the few, those doing museum type prints at sizes up to 60 inches, that isn't a market most of us have access to, and is hardly worth even discussing. People love to pretend they need big. But using what you've got efficiently is almost always better than acquiring the biggest, no matter what the product.

I don't listen to the people driving 400 Hp gas guzzlers either. They are off in their own little out to lunch world. They resent reality, and have their little arguments they use to justify their excesses. Most of which make sense as excuses to buy something new rather than a verifiable solution to a list of real needs.

This is sooo true....

[stevedh]

So if you consider a wildlife shoot, you can get the same size image with 36 Mp camera and a 300mm lens and a 50% crop, as you could with an 18 MP camera and a 600mm lens.

This as far as I can tell is wrong and I thing, please correct me if I've made some wrong assumptions.

18mp is a resolution of about 5194x3464 and 36mp is about 7350x4900

50%crop of 7350x4900 is 3675x2450 or about 9MP.

The thing to bare in mind is that if you want to print something at double the size at the same dpi, you need 4x as many pixels.
so ignoring all other things going from 16mp to 24mp will only allow you to increase the image by about 20%

[ENicolas]

The sensor is half of the story - the software is the other half. Demosaicing and noise-correction algorithms are in no way maxxed-out as far as efficiency goes. And designers are playing new games with how photosites work on the sensor - some variant of the Foveon scheme, for one example, could ultimately increase resolution markedly.

Remember that lenses are really lagging behind as they are the purely analog component and are fairly passive devices. Imagine a lens loaded with prisms that can divide a scene into quadrants, and direct the light from one quadrant at a time to the sensor. Then the software stitches the four pieces together into one image and you have doubled the resolution of your sensor's output.

We're taking baby steps with this stuff now. Even the next two or three years are going to bring revolutionary new ideas to working commercial devices. Our DSLRs with old-school glass are soon to be dinosaurs.

[Anvh]

The sensor is half of the story - the software is the other half. Demosaicing and noise-correction algorithms are in no way maxxed-out as far as efficiency goes. And designers are playing new games with how photosites work on the sensor - some variant of the Foveon scheme, for one example, could ultimately increase resolution markedly.

Remember that lenses are really lagging behind as they are the purely analog component and are fairly passive devices. Imagine a lens loaded with prisms that can divide a scene into quadrants, and direct the light from one quadrant at a time to the sensor. Then the software stitches the four pieces together into one image and you have doubled the resolution of your sensor's output.

We're taking baby steps with this stuff now. Even the next two or three years are going to bring revolutionary new ideas to working commercial devices. Our DSLRs with old-school glass are soon to be dinosaurs.

Shoot in RAW and most of the limitation you mention won't effect the picture so much.
This is why RAW yields more precise and sharper results then what can be done in camera.



But to go on about high megapixels.
36 mp in the K3 might not be that bad.
They could maybe remove the anti aliasing filter at this resolution meaning more light will hit the sensor and the image from the lens won't get blurred.
Beside that because the pixels are smaller a RGBW bayer filter might be more effective then RGB filter at these kind of resolutions.
The filter was a patent off Kodak but since they don't make sensors anymore it seems Sony has the patent now so who knows.
Sony Global - News Releases - Sony Develops New “RGBW Coding” and “HDR Movie” Functions

[Bestzoom]

But to go on about high megapixels.
36 mp in the K3 might not be that bad.

Ideally yes, even though we owners of K5 agree 16 or 24 Mpx are plenty for day to day use.

Is the K5 upgrade going to target those looking for DSLR with APSc or entry level FF or is it able to keep up with Canikon flagship DSLR?

[normhead]

16 or 24 Mpx are plenty for day to day use.

You haven't been paying attention. benjikan shot the cover of Harper's Bazarre with a K20D, 14 mp, many have pointed out that if you want to go full frame a Nikon D700 with 12 MP is more than adequate. These things can be used for things that go way beyond "every day" use. In fact, for everyday use, some have suggested a 6 mp *ist is adequate.. IN my experience 12 Mp is fine for prints of up to 20x30 inches. I would consider that up in the extraordinary range. Nothing day to day about it.

Someone really needs to define why they need a 36 Mp image. SO far, no one has defined why they need a 16 Mp image. There's kill and overkill. There is makes a difference (in which case the difference can be defined, and should be, if claims are being made) or there's doesn't make a difference.

I'm guessing there are some uses for which big files are necessary, however, no one knows what they are. I've suggested previously that images over 50 inches wide probably need larger than 16 Mp... but I'm guessing. I would consider that to be an extra-ordinary circumstance. There is a long way between that and "day to day"

A more accurate statement would be, "most K-5 users agree 16 mp is overkill for everyday use." I personally rarely use more than 1024 of my 5000 pixels wide. That is overkill on a grand scale.

[Dave L]

This as far as I can tell is wrong and I thing, please correct me if I've made some wrong assumptions.

18mp is a resolution of about 5194x3464 and 36mp is about 7350x4900

50%crop of 7350x4900 is 3675x2450 or about 9MP.

The thing to bare in mind is that if you want to print something at double the size at the same dpi, you need 4x as many pixels.
so ignoring all other things going from 16mp to 24mp will only allow you to increase the image by about 20%

Spot on.

So a new 24MP sensor would have twice the resolution of an original 6MP Pentax DSLR such as my *ist. It's taking a long time to get there!

To be honest, I have enough trouble trying to max out my existing K5 sensor's performance without having to worry about maxing out one that has a 50% greater resolution. It is somewhat easier to max out my older 6MP Pentaxes! Also, K5 is more than halfway from the 6MP *ist to a 24MP APS-C model, so I'm happy with what I've got.

I wish sensor/camera makers would put more effort into improving low light performance and dynamic range, something that all lenses and all photographers would automatically benefit from. You can't say that about higher (than 16MP) pixel counts.