[Big G]

Hi All,

I've been mulling over the current market place and the rumours of the future for the DSLR market and what we as consumers want to see in future. There was so much expectation at first for the K7 to be a full frame offering, and there seemed to be some disappointment that it was "only" a 14.6 MP APS-C sensor.

But, i've realised that it all makes sense and I can see why Pentax went down this road (I know, i've been slow to understand!).

Take a look at the competition. Sony's rumoured new A800 will have ~15 MP, the rumoured Nikon D400 will have ~15 MP and the offerings from Canon will be the same.

So, is 15 Megapixels the limit for APS-C?

For the serious enthusiast like me, who wants more? As I revealed over in this thread, the 14.6 MP sensor is plenty of resolution for enthusiast prints on a 30" x 45" canvas - so who needs more? With that in mind, ~15 MP is the perfect MP count for the majority of the market. If I become a pro and need more megapixels, I'll get a Sony A900 down the line.

So for the future, what do we REALLY want out of our DSLRs? How would a full-frame 15 MP camera fair? The increased MP count with increased high ISO / low light / larger photosites would give us an excellent Nikon D700-esque camera with higher resolution.

What is important to you going forward? More FPS? Better noise? Faster, more accurate Autofocus?

[Pål Jensen]

For the serious enthusiast like me, who wants more? As I revealed over in this thread, the 14.6 MP sensor is plenty of resolution for enthusiast prints on a 30" x 45" canvas - so who needs more?

I need more. Take any landscape photo that includes grass and I can see it is shot on digital. I've seen images from the A900 (24mp) and 22mp MF backs and they all have too low resolution. Digital is good for clean images and high acutance (sharpness). In resolution it is far behind film.
Give me 100mp.

[kent]

Get a leica s2 or phase one

[Wheatfield]

Film is still the way to go for very high resolution needs. I still have use for 4x5 in my landscape work, though exposure stitching of digital images to increase resolution is very promising, providing the subject is quite static.
Since 4x5 exposures can lead into the several minute range, exposure stitching can be made to work in many situations.
Now I just need to find a B&W printing solution that I like as much as silver.

[petercrane]

I suppose resolution isn't a big deal for me at the moment. I shoot with the K7 and am completely satisfied with the amount of detail I'm able get from each shot. The most important advancement for me would probably be processor speed (enabling ridiculously fast load times, more advanced in-camera effects/corrections, faster frame rates...) and proper Peaking (for help with focusing even with old manual lens use).

I suppose I can't complain about the performance of the SLR market at current. It's fairly amazing, and besides, who am I to tell someone else that they're making cameras wrong. It's not like I could do a better job.

[deejjjaaaa]

So, is 15 Megapixels the limit for APS-C?

Luminous landscape

Number of pixels of optimal size for different apertures of a diffraction limited lens,
wavelengths of the light and formats, considering 2 pixels per Airy disk diameter

[newarts]

There are some fundamental physics relationships that can't be overcome for a particular exposure time:

The per-pixel Noise to Signal ratio is inversely proportional to pixel spacing (for noise inherent in the image.)

Spacial resolution is also inversely proportional to pixel spacing.

Therefore the limiting Noise/Resolution at the pixel level is independent of pixel size.

I'm not sure what that implies about the future of sensor size and pixel spacing.

Further, the overall signal to noise ratio possible for an image depends only on sensor size; hence sensors must get bigger to decrease noise.

Finally, Spacial resolution is fundamentally limited by pixel spacing; hence pixels will end to get smaller.

On the final point, it might be argued that there's a practical limit of pixel pitch due to optical diffraction effects; that is not true if complex mathematical techniques are used to negate diffraction effects, but that's very costly. Without such mathematics, I think we are at about the useful limit of pixel spacing now.

Dave in Iowa

---

Last edited by newarts; 07-26-2009 at 11:13 AM.

[DAZ]

If one is going to have a discussion about something like this you need to set bounds.

For how many megapixels one needs to set a limit on how big is the print and what viewing distants. I have a 23”x 29” print from a cropped K20D on my wall that the people looking at can’t see the limit of the detail. The print is so big that when people walk up to it the start by stepping up to it at about the same distants as they would for a 8” x 11” but then take a step back because they feel to close to it.

If the point is sensor sizes then cost it an issue. Much to the chagrin of some, bigger sensor cost more and will likely continue to cost more no mater what. Lenses that can cover bigger sensor and resolve more resolution have always cost more. So the point becomes at a price point how much is good enough.

Digital and film are not the same thing so will never be apples to apples but one needs to keep the discussion to at least the same points. No point in bringing up 4x5 film if the sensor sizes are going to be APS-C or even a 35mm sensor size. Those is like saying a 4x5 film has more resolution then 35mm film but complain that 4x5 is so big and needs to be the same as 35mm film. Then I come back with 12x12 film blows away 4x5 film for resolution. There really is no point to this.

If dynamic range then keep it to color not some special ultra slow B&W film using a special custom hand developing method. Keep it to what most people could and would use.

These questions and the discussions that follow can be extremely informative as long as most everyone is discussing the same thing.

DAZ

[Torphoto]

I don't understand how some one could be disappointed in a 14.6 or 15 mp aps c dslr, the resolution is far greater than the 6 and 8 mp cams a few years ago when everyone was hailing film is dead.

I just enlarged a few rolls of 6x6 from an old tlr, the lens is soft and the film grainy but it could enlarge better than an 15 mp digial image. This is where resolution takes an advantage over sharp and clean, you just have so much more to work with.

Now with that and now that digital has finally equaled 35mm film ( current ff dslrs ) we are begining to hitthe limit of the lenses, in apsc there is no point going to more mp as the lenses will not be able to resolve more, same goes for noise and pixel spacing, there only so small you can make a pixel before you get to the point of loss on returns mp vs pixel size vs optics. Look at those12mp point and shoots, a good 7-10 mp point and shoot looks so much btter than these new 12 and now 15 mp ps cams. Do we want the same for aps-c sensors? You might beable to squueze say 15-16 mp out of apsc with the best optics but you'll then hit a wall, a wall we have already hit! beyond 16 mp why bother, same with ff sensors, with 21 and 24 mp they are great! maybe you can push them to 28 but that to will be pointless as now it's hard to get the full use from these sensors with out going to premium optics and even then the sensor is out resolving the lens.

Film will be aroud as simple brute size ( 4x5 8x10 ) will compensate for all of this, and remain as long as we need resolution, that is until some one stitches 15 or 20 35mm type ff sensors and cools them enough in an array to make a 4x5 or 8x10 view cam back at an affordable price

[pingflood]

Now with that and now that digital has finally equaled 35mm film ( current ff dslrs ) we are begining to hitthe limit of the lenses, in apsc there is no point going to more mp as the lenses will not be able to resolve more, same goes for noise and pixel spacing, there only so small you can make a pixel before you get to the point of loss on returns mp vs pixel size vs optics. Look at those12mp point and shoots, a good 7-10 mp point and shoot looks so much btter than these new 12 and now 15 mp ps cams. Do we want the same for aps-c sensors? You might beable to squueze say 15-16 mp out of apsc with the best optics but you'll then hit a wall, a wall we have already hit! beyond 16 mp why bother, same with ff sensors, with 21 and 24 mp they are great! maybe you can push them to 28 but that to will be pointless as now it's hard to get the full use from these sensors with out going to premium optics and even then the sensor is out resolving the lens.

Not following your math here. A 1.6 crop sensor (e.g. 50D) at 15mp would have the same pixel density as a full frame sensor of around 39mp. So why would you not be able to push the FF to the same density as APS-C?

[GLXLR]

Hi All,

I've been mulling over the current market place and the rumours of the future for the DSLR market and what we as consumers want to see in future. There was so much expectation at first for the K7 to be a full frame offering, and there seemed to be some disappointment that it was "only" a 14.6 MP APS-C sensor.

But, i've realised that it all makes sense and I can see why Pentax went down this road (I know, i've been slow to understand!).

Take a look at the competition. Sony's rumoured new A800 will have ~15 MP, the rumoured Nikon D400 will have ~15 MP and the offerings from Canon will be the same.

So, is 15 Megapixels the limit for APS-C?

For the serious enthusiast like me, who wants more? As I revealed over in this thread, the 14.6 MP sensor is plenty of resolution for enthusiast prints on a 30" x 45" canvas - so who needs more? With that in mind, ~15 MP is the perfect MP count for the majority of the market. If I become a pro and need more megapixels, I'll get a Sony A900 down the line.

So for the future, what do we REALLY want out of our DSLRs? How would a full-frame 15 MP camera fair? The increased MP count with increased high ISO / low light / larger photosites would give us an excellent Nikon D700-esque camera with higher resolution.

What is important to you going forward? More FPS? Better noise? Faster, more accurate Autofocus?

All of them are very important. MP aren't exactly at a limit right now. Just like a few years ago, many dSLRs were stuck with a 6MP sensor from Sony, such as from the *ist D all the way to the K110D, about 3 years of time difference. The time frame from the K-7 to the K20D has only been two years. My guess is that the next dSLR generation (the K-!@#!@%, Nikon D400, and Canon 70D) will all have 17+MP, although I don't think it matters. More MP just introduces more noise and the 10MP sensor from my GX-10 still produces nice top quality photos with 11X14 and the K-7 probably will have 15X19 with top quality prints. I just want to see better noise performance. If Pentax were to make a camera right now with the same ISO performance as the D700, it would have to be a 6MP sensor, so we just have to wait and see what happens. Even though the K-7 barely produces cleaner images than the GX-10, it has upgrades far more desirable.

[pingflood]

All of them are very important. MP aren't exactly at a limit right now. Just like a few years ago, many dSLRs were stuck with a 6MP sensor from Sony, such as from the *ist D all the way to the K110D, about 3 years of time difference. The time frame from the K-7 to the K20D has only been two years. My guess is that the next dSLR generation (the K-!@#!@%, Nikon D400, and Canon 70D) will all have 17+MP, although I don't think it matters. More MP just introduces more noise and the 10MP sensor from my GX-10 still produces nice top quality photos with 11X14 and the K-7 probably will have 15X19 with top quality prints. I just want to see better noise performance. If Pentax were to make a camera right now with the same ISO performance as the D700, it would have to be a 6MP sensor, so we just have to wait and see what happens. Even though the K-7 barely produces cleaner images than the GX-10, it has upgrades far more desirable.

Well, there is more to the story. Once you start reaching limits determined by diffraction, you won't be able to take any advantage of the extra resolution. Would you pay a premium for a 30 mpix camera that can only theoretically reach full resolution at f/2 or faster? Where are you going to get a lens that's sharp enough near or at wide open?

[Torphoto]

Not following your math here. A 1.6 crop sensor (e.g. 50D) at 15mp would have the same pixel density as a full frame sensor of around 39mp. So why would you not be able to push the FF to the same density as APS-C?

If you squeeze that many pixels into ff, yes you'll get more rez but now loose on the lower noise they benefited from, also your going to really push the limits of the optics.

Remember most lenses that work so well on aps-c are ff ones that are using the sweet spot, now imagine using all of the lenses ability to try and illuminate a 39 mp ff sensor, I bet most of that sensors resolution will be wasted. Now they will have to redevelop new optics, ironically probably larger optics or just innovative designs ( like some of the apsc only lenses ). You will still hit a brick wall in optics with physics. Light is of a certain wave length, there is a point ( that apsc is hitting now ) that ff is also close to where you just get diminishing returns of the usefulness for the extra resolution when your lens can not resolve.

I guess my ramble was not making sense but you'll get the idea.

[pingflood]

Well, I can see part of your point, but whether the outer edges will get outresolved by the sensor is kind of pointless -- this happens already with many lenses and more moderately dense sensors. If say APS-C "maxes out" at 15mp, then you'd want the same density on the FF sensor in order to get the same center resolution as the crop camera.

Now with that and now that digital has finally equaled 35mm film ( current ff dslrs )

You may want to rethink this statement. Film still has more resolution and depth of color.