Diffraction is a physical limitation to resolution, and even though I usually agree with you, TTJ, I think the tech is getting close to where it makes a difference in some practical situations. Without having looked at the A77 images and just based on the increase in lateral resolution, I would think that f/5.6 is probably not where the physical limitation is on DX for 24 MP (my guess is closer to f/8, but someone could crunch the numbers).

Though f/8 is a sweet place to be for a lot of photographers, it's not that the technology isn't useful to anyone; there will still be people who want those 24 MP and are going to shoot at f/4 or f/2 or some other large aperture most of the time. For them, diffraction is a non-issue and they can print their wicked huge prints (or just gush about how much data there is in their pictures—look at those nose hairs! I've never seen anything like this before). I think more photographers are in this group (less the nose hairs) than the group for which diffraction is starting to matter, which is why manufacturers just keep increasing resolution like low res is going out of syle.

It's the landscape photographers that are probably going to realize that 24 MP (or 32, or whatever) is not going to help them on DX because they're already running into diffraction at the apertures they set. They can always go to larger format, though.

Niko, shouldn't we start a 'DX is dead' thread, just for old times' sake?

TaoTeJared said:
I'm not a believer that diffraction is an issue yet. There are too many variables that manufacturers can change to counteract it that all of the equations ignore. Changing the algorithms or even the pixel pattern can easily change where diffraction comes in. Not a single diffraction equation makes any variable available for that or aspherical elements for that matter of fact. If Diffraction was a real hindrance, why is every company still moving quickly forward without hesitation? Because they have found ways either physical or software to move around it.

I saw with my own eyes, the diffraction limit when comparing samples of the A77 images. It hit's hard at about F/5.6. Sharpness is drastically decreased, even in the center. This might be exaggerated due to Sony's inferior glass, but it was still there and it was terrible. I don't know how they could work around this issue, really. Beyond that, the sensor is nothing short of amazing(DR and SnR wise). I really couldn't believe my eyes, it really suprised me. The images were really close to the D7K in IQ with the slight edge still going to the D7K sensor(which is very hard to beat right now). Who knows, after seeing what I saw, maybe Nikon will tweak this sensor and use it in the D400.

I'm not a believer that diffraction is an issue yet. There are too many variables that manufacturers can change to counteract it that all of the equations ignore. Changing the algorithms or even the pixel pattern can easily change where diffraction comes in. Not a single diffraction equation makes any variable available for that or aspherical elements for that matter of fact. If Diffraction was a real hindrance, why is every company still moving quickly forward without hesitation? Because they have found ways either physical or software to move around it.

I took a serious look at the new 24.5mp Sony sensor from the A77. The sample images were very impressive in some ways, but had serious diffraction issues. I was most impressed with the SnR of the images, despite the crazy pixel pitch of that sensor. I bet Nikon could really make this sensor shine, but don't think they'll go there due to diffraction problems. I really think we'll see a even better tweaked version of the D7K sensor in a D400. This would be my best guess.

Beyond diffraction, I am pretty impressed with the A77 sensor. Nikon's Noise Reduction algorithms are so much better than everybody else's, I would actually love to see what could be achieved, even though I think the pixel pitch is overkill and would much rather see Nikon use the amazing D7K sensor for the D400.

There's a lot of excellent discussion on this thread. Can I add something very simple?

I've just had some of my D90 images printed at 24x16. And here's the point - they look good.

Do I, personally, want prints bigger than this? Well, no, not really.

So do I want more pixels? Well, it would be nice but, no, it's not important to me.

Like so many things in life, there comes a time when what you are are happy with what you've got.

Take music CD technology, for example. It cannot be improved because CDs already exceed the limits of human hearing. That's not true of camera sensors (yet!), but for SOME photographic uses we have already surpassed what is necessary.

PB PM said:
Production isn't just a problem on Nikon's end though. I doubt it takes any longer to make a D700 than say, a D300. Part of the issue is yields of FX sensors. The number of FX sensors that you can get from a single wafer of silicon is much lower than that of DX sensors.

Production is a problem as well as the added tolerances for Pro cameras. SS has it right, the point is to transfer DX lines to FX lines. The number of sensors from a wafer has no bearing how fast they can pump them out. On the production side they only will make what they need and expect demand to be. It is really surprising how much demand there really is still for the D700/D3s/D3x - 3years out of release and they still can't keep up? That is nuts. It was getting steady before the quake but that put the kibosh on that hope. An educated guess is that it is the assembly that back logs the cameras (when the parts were available.)

There is actually a good deal of difference between the D700 and the D300.

Super Shooter said:
I don't think 12mp sensors on a new DSLR would sell well up against all these new 20 plus megapixel cameras which now include point & shoots. Same reason nobody wanted a new D2h with 4mp back in the day.

Realistically, only serious photogs and pros pay $2k for the high end cameras and most know what the capabilities are and are less prone to drink the marketing hype. They also have a slew of lenses and are not going to switch brands. 4mp D2H couldn't create a 24x36 print from an 3200 iso image either. I think 12mp sensors would sell like crazy for another 2-3 years easily. Bump the bit rate to 16 and a bit cleaner on the ISOs and there would be none left on the shelves.

Oddly enough I was hungry. ;-)

Production isn't just a problem on Nikon's end though. I doubt it takes any longer to make a D700 than say, a D300. Part of the issue is yields of FX sensors. The number of FX sensors that you can get from a single wafer of silicon is much lower than that of DX sensors.

PB PM said:
TacoTeJared, the sounds good in theory, but could be troublesome for actual production. Nikon already has enough trouble keeping FX bodies in stock, do you really think having more models will help with that situation? I sure don't.

Taco Te Jared? Ha ha! PB PM must be hungry!

Obviously if Nikon had more FX models they would plan for a higher production run. Current DX factories could be switched over to FX production since the sensors would be made by someone else anyway. If they were to have trouble keeping up it would be because they'd be ultra hot sellers. Which is true of just about everything Nikon sells now.

I don't think 12mp sensors on a new DSLR would sell well up against all these new 20 plus megapixel cameras which now include point & shoots. Same reason nobody wanted a new D2h with 4mp back in the day.

TacoTeJared, the sounds good in theory, but could be troublesome for actual production. Nikon already has enough trouble keeping FX bodies in stock, do you really think having more models will help with that situation? I sure don't.

DaveyJ said:
Increased file size is a problem for all of us.

I agree. Those who say hard drives are cheap really don't realize what the real issues are and will be. If you have ever tried playing with 80mb MF RAW files you will realize real quick an upgrade in a camera can mean many, many more expensive upgrades.

Being a Daytime analyst, I tend to see the Economics of building cameras VS. Consumers perceived needs (self induced intoxication on marketing juju). Camera companies are just the referees between the two ends of the spectrum.

Example: Economically the more of a particular part you make, the cheaper it gets. Consumers want affordable FX products and want Nikon to be on top of the tech wars. My suggestion to the referees? Continue to use the 12mp sensor from the D3s in the two bodies below it (D700/D300) and stop developing things like video for people who do not utilize them. Leave the video upgrades to lower bodies with a faster refresh rate. Than leave a High end pro body with the newest tech for the pros who really need it just as you do now. Then flow that exact tech downward. Continue this for the next two refresh cycles.

Cost would drop for the next 3 years where money spent on development two cameras could be put back into "real" tech improvement. We could all get FX with phenomenal IQ, decent video for under $2k.

So - who could live with a D3s sensor in a D300/D700 for the next 3-4 years or do you really thing you need 24mp?

One clear problem with this topic is speculation about future technology. I am not against large accurate file size of valuable subjects. Realize that I have seen the final images shot from 45 second digital large format backs and they are inaccurate. I just don't have much reverence for false images that do not represent what the scene really was.I should add the images were not very attractive=marketable. I am very interested in what the D4, D800, and D400 file sizes will be. A higher process speed will also greatly (I hope) reduce video artifacts in DSLR applications.

I think 16mp is perfect for most things. By now, Nikon should be able to provide a 16-20mp FX format sensor with better signal-to-noise ratio than the amazing D3s sensor, while increasing DR. Just look at what the D7000 was capable of with a pixel pitch of 4.9. I'm almost positive the new pro DSLR cameras from Nikon be capable of 16bit color depth, which will Increase IQ even more. Hopefully the hardware advances enough to be able to keep frame-rates high while in 16bit mode. Look at the processing power in garden variety smartphones(like dual core processor Android phones). There should be no reason they can't implement this processor tech into their DSLR's.

I'm all in for a D4 and I do hope it's no more than 16-20mp with a stop better DR than the D7000(the DR king) and a stop better ISO handling than the D3s(the clean high-iso king). I am not willing to sacrifice IQ for more pixels, however. If I wanted that, I would be shooting Canon.

Throw me in the group that wants more mp but not anything crazy. I want the best image I can get and I believe that nikon can still give me the amazing low light ability and a few more mp at the same time while keeping image quality great. This is also why I love new products coming out. Its neat to see where nikon is going.

TaoTeJared said:
no one wants to go back to a camera that shoots 1 frame per sec.

Or less like the Pentax 645D!!!!

SquamishPhoto said:
You clearly do not speak for everyone here.

+1

I guess that in a few years from now, when technology advances, we will have faster cameras, faster disks, with bigger capacity that the 16MB file generated by the D7000 will be small.

I remember 10 years ago, when the first digital cameras were coming out, people were complaining that files above 5MB are hard to deal with.. now most of the phones have that capabilities...

Only now, the digital technology is crossing the limits Film photography had in pixel count...

It is true that the more dense is the sensor, more sensitivity is affected due to noise... but I guess that is a technological problem that may be solved in the future.

DaveyJ said:
Increased file size is a problem for all of us.

You clearly do not speak for everyone here.

Some of those large format backs are scanning backs, which gives you an effect like the video jello effect but for stills. I don't like it no matter how big the image can print.

About the pixel pitch estimate, to get vertical pitch you divide sensor height (16mm on DX, 24mm on FX, etc.) by number of vertical pixels. Same thing with horizontal pitch only you use sensor width and horizontal pixels. There's about a 2% variation in the different sensors, but who cares? (You can look that up if you do.) This gives pixel pitch in mm, so you need to multiply by 1000 to get the answer in μm.

Sample calculation (D7000):
V pitch: 16mm / 3264 pix * (1000 μm / 1 mm) = 4.9 μm / pix
H pitch: 24mm / 4928 pix * (1000 μm / 1 mm) = 4.9 μm / pix

Increased file size is a problem for all of us. I hope I (and you) don't lose some really valuable image because it got lost in the shuffle. I myself think that file size with larger digital formats can be very self defeating. Awhile back I was in the Blue Ridge and stopped at a scenic vista i knew well. Got out of my vehicle and here is a guy shooting the scene with a digital large format. 45 second exposure. The clouds were shifting as he photographed. His resulting file size was being handled by a powerful laptop right beside him.

At what point is increased file size overkill?? For one I think the D700 and D7000 are somewhere in the hunt and having owned some very serious medium formats, I can say no thanks.In fact the primary reason I am on Nikon Rumors is to share the thought that larger formats are very expensive and I don't think they are worth the trouble. Huge file sizes are trouble. Beyond say 24 MP I do not want that camera. Last point I shoot in JPEG mostly and RAW when it is important partly due to file size and partly because it is not worth the difference. Even with 11X17 prints.

I use 2 x 16 in my D7000 and have the settings for overflow.

1 16 carried as a backup, but never had to use it.

I am selective with what i keep and I do discard those that are not keepers.

2 x 1 tera externals

those teras I back up unto cds ever so often.

depending on what I am doing, I sometimes shoot jpeg and sometimes raw.

and yes I do agree with you, it is something that has to be managed and bigger files mean more managing.

I would love the option though(we can already toggle between 12 bit and 14 bit). I would be happy with a camera that had a 16 bit option even if it shot 1 or 2 fps (2fps at 16bit, 6fps at 14 bit and 10fps at 12 bit). I don't know what the buffer on the a77 (probably small but you never know) is but 24mpx at 10/12fps is fast and possible (and the fact that it is a SLT doesn't matter because the shutter mechanism has nothing to do with bit rate)

It isn't that I don't want more mpx it is just that I am happy at 12mpx. When I got my D7000 I found out that I couldn't fit anywhere as much on an 8gig memory card and being a photog that shoots between 700 to 3000 shots a day it really set me off. I already have to get a new hard drive every other month. The issue with telling the camera to shoot smaller is that I may need to crop as well as print large (think 20x30in) so I can't shoot at 6mpx as it is an obvious drop in quality. Also I need the clean high Iso as I shoot at night frequently...

Dave_Robo said:
The problem with a lot of pixels on a small sensor is that you lose contrast. The number of different variations in color gets smaller.

I don't believe that is accurate as what you are describing has to do with color bit depth.

The size of the pixles has little to nothing to do with contrast or color. It is the bit depth of the sensor that captures this. Digi cams/ cell phones are usually 6-10bits per color, lower DSLRs are usually 10-12bits, Higher DSLRs are 14bit and Medium Format is usually 16bit. The added colors create smoother shifts, contrast, sharpness, etc.

Personally I believe if Nikon or Canon moved their cameras to 16bit they will really start to rival MF cameras. The issue is that is huge amount of data to crunch fast and no one wants to go back to a camera that shoots 1 frame per sec.

bjrichus said:
Nice table... were you doing the math or did you get this info from somewhere like manufacturer web sites?

It is sometimes hard to get this info from marketing departments, as they usually don't understand things like pixel density/pitch etc, so don't include it in anything.

If they actually realized HOW to work out what makes a great image, rather than just focus on a nice page layout, they'd usually have a better story, but then, that's life...

It is listed on DxO. I'm not sure if they calculate it or for which ones or if they rely on the manufactures stats. Press releases rarely state these but all brands publish these if you go deep into their specs.

kanuck said:
...... Nikon Spokesman are saying the public isn't even asking Nikon for one but I know they have been working on one. .......

May be they have been listening to Steve Jobs . The man credited with providing us with products we didn't know we wanted

On the subject of pixels and cost, just why is the D3x so much more expensive than a D3s

It would be interesting to see how the mirrorless system factors into all this higher MP debate if that camera ever is released. Nikon Spokesman are saying the public isn't even asking Nikon for one but I know they have been working on one. Different crop factor so again another physics twist....

It is my hope that the New Nikon bodies will have both high pixel count for high IQ in good light and pixel "binning" (combining two or more pixels) for high IQ in low light. In the end I just want astonishing Image Quality no matter how many pixels or what technology gets me there. Certainly there will be trade offs. We could all go buy a Nikon D3x now for more pixels or a D3s for high IQ in low light but very few will because the negative trade off is cost. When more pixels become available at a lower cost, such as in a 24 megapixel DX sensor, the negative trade off will be whatever reduction in IQ comes with it and perhaps more time processing or purchasing a faster computer. Everything is going to have its pros and its cons. I doubt Nikon will give us more pixels without a corresponding increase in IQ from those pixels. But we just have to wait and see. We cannot evaluate the trade offs until we see what the new Nikon bodies can do.