[Edgar_in_Indy]

I never beleived APSC would go higher than 12 milion pixels after the D300 came out, At least without sacrificing high ISO quality.
There are 24 milion pixels now, but as expected, I wouldnt touch one because awful above just 400 ISO.

I've been looking closely at the 24mp images from the new Sony sensor ever since they showed up on the web, and I think that calling it "awful above just 400 ISO" is overstating the case quite a bit. As far as I can tell, the high ISO performance between the old 16MP sensor and the new 24MP one is fairly close. The 16MP might have a slight edge in terms of noise, but the 24MP is close, and captures a lot more detail.

Here's another demonstration image I made using the sample images from Imaging Resource. Both images are shot at ISO 1600 with the NR set to "LOW". I downsized the 24MP Sony image down to 16MP to match the K-5, but other than that nothing has been done. I don't see where one camera is significantly better or worse in terms of the amount of noise. The K-5 seems to have a finer grain of noise, but the down-sized 24MP image is easily sharper and more detailed than the native 16MP image.

[westmill]

Converting the A77 images in Adobe Camera RAW 6.6 Beta confirms the suspicion that we had when examining the camera's JPEG output. The A77's 24MP sensor is producing significantly higher noise levels than the competition, especially towards the top of its ISO sensitivity range. At ISO 1600 and above the disparity is obvious, and becomes more so as the ISO sensitivity increases. Above ISO 12,800, chroma noise from the A77's raw output is very high indeed.

As always, you should keep in mind that this is a pixel-level judgement and these files contain a lot of pixels. In the real world, with judicious use of post-capture noise reduction you can get much more from the A77's raw output than you might expect from the images and graphs on this page. That said, we've found that at high ISO settings in poor light, it is hard to 'rescue' images taken at ISO 6400 and above, even with careful raw post-processing.

Poss a bit strong, but this is what my opinion was based on ! this I just copied and pasted from depreveiw. Who in my opinion are the best and most trusted testers of camera equipment. Its also simply my opinion too.

[Edgar_in_Indy]

Above ISO 12,800, chroma noise from the A77's raw output is very high indeed...

That said, we've found that at high ISO settings in poor light, it is hard to 'rescue' images taken at ISO 6400 and above, even with careful raw post-processing.

Poss a bit strong, but this is what my opinion was based on ! this I just copied and pasted from depreveiw. Who in my opinion are the best and most trusted testers of camera equipment. Its also simply my opinion too.

I generally don't go above ISO 1600 on my K-x. In a few rare instances I have gone to 3200, but in those cases I never expect to do more than 4x6 prints. And I don't think I'm alone in this. It has been my observation that a lot of people on this forum view ISO 1600 as kind of a self-imposed ceiling.

So talk of how noisy a camera is at ISO 6400 and 12,800 really doesn't concern me. At sane ISO levels, the 24MP sensor does just fine as far as I can tell. I'm at ISO 100, 200, and 400 way more then I'm at ISO 1600, let alone 6400 or 12,800. And at the lower ISO levels, the 24MP is pretty darn impressive.

So tell me, do you not see in the above image how much sharper the down-sized 24MP, ISO 1600 image is compared to the same ISO 1600 image from the 16MP K-5? Reading others' opinions is all well is good, but the proof is in the images.

[westmill]

I generally don't go above ISO 1600 on my K-x. In a few rare instances I have gone to 3200, but in those cases I never expect to do more than 4x6 prints. And I don't think I'm alone in this. It has been my observation that a lot of people on this forum view ISO 1600 as kind of a self-imposed ceiling.

So all this talk of how noisy a camera is at ISO 6400 and 12,800 really doesn't concern me. At sane ISO levels, the 24MP sensor does just fine as far as I can tell. I'm at ISO 100, 200, and 400 way more then I'm at ISO 1600, let alone 6400 or 12,800. And at the lower ISO levels, the 24MP is pretty darn impressive.

So tell me, do you not see in the above image how much sharper and more detailed the 24MP, ISO 1600 image is compared to the same ISO 1600 image from the 16MP K-5? Reading others' opinions is all well is good, but the proof is in the images.

Well yes of course, but there are plenty of images on depreiveiw that was just a tiny bit of the findings.
And yes... upto 400 ISO its unbeatable and resolves simply stunning detail. Some people just have no reason to ever need go beyond 400 ISO anyway. Im a pro photographer though. I need to produce the goods. to give you an idea, I was at huntingdon race track on boxing day covering
the racing as the official press photographer. There was 6 races. the last race was run at 3.15 and it was going dark. I started the day shooting at
800 ISO by the last race i was up to 3200 ISO. The other wk I was there.... 7 Races.... The last two races I was in a downpour of rain, freezing
cold and ubeleivable strong winds and I ended up using 6400 ISO on the last lap of the last race. The Sony simply could not do this.
http://www.dpreview.com/reviews/sonyslta77/
If you wish to see and read a full and comprehensive review of the Sony 24 milion pixel sensor click on the above. Its Good !

---

Last edited by westmill; 01-07-2012 at 05:09 PM. Reason: added new info

[Edgar_in_Indy]

I can certainly see your point, and understand why the sky-high ISO is important for some people. I myself shoot a lot in the studio with strobe lighting at base ISO, so for that kind of work the 24MP would be awesome. That's why I'm hoping Pentax puts the 24MP sensor in something.

Thanks for the link to the review. It's been a while, but I've already read reviews of the A77. The pictures I used from Imaging Resource are from the NEX-7, which as you probably know doesn't have to deal with the translucent mirror, and therefore has a slight (but real) noise improvement over the A77. I'm hoping that if Pentax does use the sensor, history will repeat itself and Pentax will squeeze better performance out of it, just as they did with the 16MP sensor.

[philbaum]

Converting the A77 images in Adobe Camera RAW 6.6 Beta confirms the suspicion that we had when examining the camera's JPEG output. The A77's 24MP sensor is producing significantly higher noise levels than the competition, especially towards the top of its ISO sensitivity range. At ISO 1600 and above the disparity is obvious, and becomes more so as the ISO sensitivity increases. Above ISO 12,800, chroma noise from the A77's raw output is very high indeed.

As always, you should keep in mind that this is a pixel-level judgement and these files contain a lot of pixels. In the real world, with judicious use of post-capture noise reduction you can get much more from the A77's raw output than you might expect from the images and graphs on this page. That said, we've found that at high ISO settings in poor light, it is hard to 'rescue' images taken at ISO 6400 and above, even with careful raw post-processing.

Poss a bit strong, but this is what my opinion was based on ! this I just copied and pasted from depreveiw. Who in my opinion are the best and most trusted testers of camera equipment. Its also simply my opinion too.

That doesn't sound good. This is a shot that i took at 20,000iso during a dark play (Dracula). It was underexposed even at 20,000, so i had to apply +1.4ev of exposure to get a decent picture. Of course, i had to apply Lightroom NR.




I printed the above at 12"x18" and it looked great. 16"x24" might be a stretch and 20"x30" would not be doable. I eventually ended up giving the picture to the most visible actress in the pic. and she was delighted with it. (ps - they don't allow flash and they don't listen to lighting comments from mere mortal photographers :-))

All i'm saying is that if the A77 can't go higher than 6400 in order to develop a 4x6", than its performance is well below the 16 megapix Sony sensor of the type thats in the K5. IMO

[zapp]

It seems like when Sony came out with the 24M exmor sensor, that folks assumed that the K5 replacement would use that sensor. The neat thing the 16M does for Nikon D4, is to allow a buffer that holds 200 shots compared to K5's 40 jpeg or 15 RAW. D4 also weighs 3 lbs and cost $6K list.

Keeping the same 16M for the K5 successor would also allow more money to be spent on other features. Of course its too late now for Pentax to change whatever they cooked up for a replacement. Just saying... 2012 will be an interesting year.

I came here wanting to ask the same question...
16 megapixel is a good limit for 35mm format cameras. Lenses can typicaly not deliver the resolution required by tiny pixels and it is counterproductive to capture pixels that do not carry any addtional information - anything abover 16 megapixel is really pushing the limit. Dynamic range, high ISO performance ... could need improvements, looks like some companies understood the message. What about a non-Bayer pattern sensor. Same size pixels, but 3 times the information!
Canon and Nikon came from higher resolution cameras, so one may ask whether or not they will place a new camera format in the 16+ megapixel arena. Something Leica did a couple years ago - smart move..?! With faster computers, faster interfaces and increasing storage space, large sensor (large file size) cameras will become popular as well as video from traditional still camera systems.
What does this mean for Pentax. The APS-C system works OK and Pentax is the only manufacturer with a comlete range of APS-C lenses (in full frame land they do not offer anything really tempting...). The K5 does not need more pixels, 16MP may already be too much. Pentax already offers the 645 system with larger pixel counts. So life is looking good.

[Geoff H]

Here's a picture I put together and posted a while back when discussing the merits of more megapixels. For anybody shooting in the studio, more resolution is always a good thing. And as the image below demonstrates, the extra pixels in the new 24MP Sony APS-C sensor are not just empty pixels. There is a real and significant increase in resolution/detail.

Pay particular attention to the words "Pure Brewed", and how the white spaces between the horizontal black lines are rendered. Also look at the seal, and the tiny text above and below it, and how it becomes much more legible as the megapixels increase.


And here's the full image Imaging Resource test-shot that the crops were taken from:

Can't really see anything because we're at screen resolution and obviously the bigger picture will be clearer. DPreview have a nice comparison gadget on their reviews. If you want to see that 'white space between lines' look at the coin at JPG 100 (extra sharpness)on the K5 which really leaps out at you! Pixel peeping on their site is interesting because the Pentax reviews consistently look softer in RAW than cameras using the same sensor (for the K5 read Nex-5, Nikon 5000 etc) and that's got to be down to the lens they use for the Pentax, a 50mm 2.8 - either it's not a good example of the lens or DPR are fiddling the result... Either way the lens looks to be the limit.

On their 'coin' and looking back to back with the new 24 mpx sensor, yes there is marginally more definition, but better to compare the Nex-5 with the Nex-7 where they use the same lens. Not so clear a difference is there? - a tiny bit clearer, but it's marginal. What isn't marginal is the 1 - 2 stop lead in ISO performance the smaller sensor has.

From what I can see even with the camera bolted to the bench, mirror locked up and optimum aperture the advantage of the 24 mpx is marginal to say the least, and I'd bet that with any zoom, hand-held it'd be zero. On the other hand noise differences are clearly visible as are the dobbin' great files produced...

Personally, for the photography I do, I doubt I'll never hit the stops that 16 mpx gives.

Incidentally, I remember some mag doing such a test with 3 sony SLR's a while back - a 10, 12 and 14 mpx. The 12 showed a tiny improvement in resolution over the 10. The 14 didn't because the lens was the limit. The 10 had the best low-noise performance. Overall the 'best' was a toss-up between the 10 and 12 with the 14 some way behind.

Now if they could use the 24, but also an option to set it as a 12 mpx with the sensors doubled for increased dynamic range a la Fuji there might be some sense to it...

[Geoff H]

I've been looking closely at the 24mp images from the new Sony sensor ever since they showed up on the web, and I think that calling it "awful above just 400 ISO" is overstating the case quite a bit. As far as I can tell, the high ISO performance between the old 16MP sensor and the new 24MP one is fairly close. The 16MP might have a slight edge in terms of noise, but the 24MP is close, and captures a lot more detail.

Here's another demonstration image I made using the sample images from Imaging Resource. Both images are shot at ISO 1600 with the NR set to "LOW". I downsized the 24MP Sony image down to 16MP to match the K-5, but other than that nothing has been done. I don't see where one camera is significantly better or worse in terms of the amount of noise. The K-5 seems to have a finer grain of noise, but the down-sized 24MP image is easily sharper and more detailed than the native 16MP image.

Can you please do the same for raw images from the Nex-5 and 7 because I don't think you'll see the same result.

[dylansalt]

Also one has to take into account that with a 24mp aps-c sensor - the absolute perfect technique required to obtain sharp pics

It has been discussed on forums that a shutter speed equal to the focal length used is now history from 16mp upwards - 1.5 - 2x is required depending on how good your technique is.

This is due to the fact that the pixels are so packed together that the slightest movement when taking a photo can induce slight blur effecting the sharpness of the pic.

Also the more mp shifts your optimal f stop(read best mtf result) downwards thus (DIFFRACTION) you will find your basic consumer lens/zoom now yielding soft results on the 24mp even shot at F8.

Thus only pro lenses with constant F STOPS will yield the ultra sharp pics as they hit their sweet spot at F5.6 (where most consumer lenses start)

The D4 I believe will yield fantastic sharpness and acuity with just about any FF lens without the need for absolute critical technique when in the field

[Geoff H]

If I've got it right you should see what I mean below... The 100 asa show the NEX 5 to be fully a match for the already noisy Nex 7, and also show the typical (on DPreview) 'soft' pentax lens. By 6400 asa the Nex 7 is unuseable.

[Ubuntu_user]

That doesn't sound good. This is a shot that i took at 20,000iso during a dark play (Dracula). It was underexposed even at 20,000, so i had to apply +1.4ev of exposure to get a decent picture. Of course, i had to apply Lightroom NR.




I printed the above at 12"x18" and it looked great. 16"x24" might be a stretch and 20"x30" would not be doable. I eventually ended up giving the picture to the most visible actress in the pic. and she was delighted with it. (ps - they don't allow flash and they don't listen to lighting comments from mere mortal photographers :-))

All i'm saying is that if the A77 can't go higher than 6400 in order to develop a 4x6", than its performance is well below the 16 megapix Sony sensor of the type thats in the K5. IMO

Wow. That's with the K-5, right? It seriously looks like a painting...

[Wolfie665]

It seems like when Sony came out with the 24M exmor sensor, that folks assumed that the K5 replacement would use that sensor. The neat thing the 16M does for Nikon D4, is to allow a buffer that holds 200 shots compared to K5's 40 jpeg or 15 RAW. D4 also weighs 3 lbs and cost $6K list.

Keeping the same 16M for the K5 successor would also allow more money to be spent on other features. Of course its too late now for Pentax to change whatever they cooked up for a replacement. Just saying... 2012 will be an interesting year.

for APS-C format there are some physical limitations between pixel ptich size and the resolution: Clarkvision: Does Pixel Size Matter

According to this guy the optimal pixel picth size is 5 microns for APS-C sized sensor.

Although Sony sensor looks impressive I think they have gone already too far in pixel pitch size making it too small 4 microns only.

Based on that there is plenty of room to optimize the APS-C sensors, and to have a new sensors with bigger pizel size and better dynamic ranges and low light performance, not neccessarily bigger resolution then 24M. Why not have a new say 18M sensor packed with bigger pixel 5 microns pitch?

So the pixel resolution war is over at 24M. However the sensor quality war just has started. We may expect many new APS-C sensors with better quality and resolution between 16-24M.

Having said that there is a lot of room for FF sized sensor to go over 33M.

D4 is FF sensor hence it is one small step in the pixel resolution war. Do expect many new FF sensors till the war is over.

[Geoff H]

Wow. That's with the K-5, right? It seriously looks like a painting...

It's beautiful... And with the K5 you can do so much without a flash. For a hand held shot the K5's anti shake gives you say 2 stops. The noise at high ISO is probably 2 stops better than the 24 mpx sensor, so ideally you can shoot 4 stops lower light. I've done shots with campfires, candlelight etc that would be impossible with most cameras and a worth a million extra megapixels. This christmas I did lots of family shots indoors and in dull winter light, I guess the average ISO I used was 3200 with the kit lens, several at over 20,000

[Edgar_in_Indy]

The noise at high ISO is probably 2 stops better than the 24 mpx sensor,

Okay, now things are really starting to get absurd. A 2-stop advantage? Are you seriously saying that the K-5 is less noisy at ISO 12,800 than the NEX-7 is at ISO 3200?