[Zygonyx]

I'd prefer 16 MP FF CCD than 24 MP APS-C CMOS. As for me - DXO tests are like 3D Mark for photo amateurs.

I wouldn't have such an extreme jugement, but for sure the K-5's dynamic range seems to be out of reach of the new Sony 24 Mp sensor...
So having the 645D, i shall pass my way with the K-3 if it is bound to get such a sensor.

[gazonk]

I wouldn't have such an extreme jugement, but for sure the K-5's dynamic range seems to be out of reach of the new Sony 24 Mp sensor...

I wouldn't say that. IF (as Falk cautioned...) Pentax is able to do the same to this sensor that they did to the 16mp sensor, they may actually achieve even better DR. This makes waiting for the K-3 quite exiting!

[Rondec]

A77 is rather noisy in RAW (see various reviews) in spite of DXO tests. I don't think that 24 MP sensor will be strong in Pentax. Maybe not worse than K-5.
But it's more correct to compare K-5 and A77. Anyway, 24 MP APS-C sensor is VERY NOISY in RAW, and it's noisy at ALL ISO.
It looks like old D2X sensor with new NR technologies. It's MP race and trickery of users in modern consumer society.

It sure looks like it is at least as good. Sure the individual pixels aren't as strong, but the question really is if you print at an equivalent size, which photo will look better. Looks to me like the 24 megapixel sensor will be slightly better than the 16 megapixel sensor.

It isn't correct to compare the A77, rather than the Nex 7. The whole point is that SLT cameras do worse at high iso because their pellicle mirror soaks up light and forces them to shoot at higher sensitivity for a given light level. The mirrorless cameras don't have this issue and so they give a better indication of how the sensor would perform in a Pentax or Nikon camera with a standard SLR setup. That said, Sony doesn't seem to do very well at getting the most out of their sensors. Both Pentax and Nikon outdistanced them with the 16 megapixel sensor that is in the D7000 and K5. Hard to say if they could do the same thing with the 24 megapixel sensor.

[Zygonyx]

I wouldn't say that. IF (as Falk cautioned...) Pentax is able to do the same to this sensor that they did to the 16mp sensor, they may actually achieve even better DR. This makes waiting for the K-3 quite exiting!

Well, let's look at figures :

1- 16 Mpix sensor case

-Sony A35 takes 12,7 EV in "landscape" protocol

-Sony NEX5N the same (which sounds "bizarre" because it has no translucent)

-Pentax K-5 takes 14,1 EV thus a difference of 1,4EV in favour of Pentax, considering the 16Mpix sensor

2- 24 Mpix sensor case

-Sony A65 takes 12,6 EV, whereas A77 takes 13,2 (an other bizarre thing)

-Sony NEX7 takes 13,4 EV (wich is a significant step forward)

-Pentax K-3 should take less than 14 EV in any case imho, so at best it would equal K-5. The reasons :
* Sony is making progress in image processing
* Dynamic range is not a linear function in relation with same size sensor at varying technology (if i am not mistaking - but perhaps i hope i am... )

[gazonk]

Well, let's look at figures :

Yes, let's look at the figures, and they say that Sony has improved DR with 0.3EV in their 16mp (e.g. from A55 to A35), but the NEX-5n is still 0,7 below the NEX-7. So I maintain my statement that IF Pentax is able to do the same to the 24mp sensor as they did to the 16mp sensor, the "K-3" will beat the K-5 in DR - and by a good margin. This is pure speculation of course, but your pessimistic outlook is IMHO even purer speculation

[falconeye]

Well, let's look at figures :
Sony A35 takes 12,7 EV in "landscape" protocol
Sony NEX5N the same (which sounds "bizarre" because it has no translucent)

A77 takes 13,2 (an other bizarre thing)
NEX7 takes 13,4 EV (wich is a significant step forward)

Please, no jumping to conclusions here.

First, DR is (theoretically) unaffected by the SLT's semitransparent mirror. The low iso score is, the landscape score isn't. The semitransparent mirror reduces effective iso (moves DR[ISO] curve left, not down), that's all.

Second, at DR as high as delivered by the K-5 (and only if as high) is negatively affected by a 12 Bit A/D converter. In your comparison, not all A/D converters are 14 Bit, the K-5's is.

Third, 14EV DR is hard to measure accurately. We talk about 3 electrons remaining read-out noise! Don't put to much interpretation in differences in the order of +/- 0.25EV.

IMHO, all in all, to say that the 24MP and 16MP sensors perform very similiarly is the best we can currently do.

[maxfield_photo]

So having the 645D, i shall pass my way with the K-3 if it is bound to get such a sensor

I own neither the K5 nor the 645D, my K20D is due for a replacement, yet I will also pass on a 24Mp APS-C sensor, there's just no reason for that many megapickles unless you are making movie poster sized prints. 16Mp is plenty, 18Mp is a bit more than most people need but not without it's uses in certain situations. By 20Mp its just a bloated file unless we're talking about a larger sensor.

[gazonk]

I own neither the K5 nor the 645D, my K20D is due for a replacement, yet I will also pass on a 24Mp APS-C sensor, there's just no reason for that many megapickles unless you are making movie poster sized prints. 16Mp is plenty, 18Mp is a bit more than most people need but not without it's uses in certain situations. By 20Mp its just a bloated file unless we're talking about a larger sensor.

16 mp images look better than my 10mp K10D images on my current computer screen (2560x1440). If Ultra High Definition TVs suddenly take off, you're going to view your pictures on a 33mp screen. And then you may want to have 4x that number of pixels for really optimal quality. So in a couple of years, 130mp cameras may make sense even if you don't print three story high posters.

And then there's cropping... just a little cropping and a 24 mp image suddenly lacks resolution for that 2560x1440 screen...

[eddie1960]

The big advantage i see is the cropping room 24MP gives. if it can maintain or improve on K5 performance while giving me the extra MP to print LARGE or Crop and still print Large I see it as a boon and would likely skip over the k5 and go to the K3 (of course this assumes there is no FF 24 or 36 MP model coming which is where I'd rather go)

[Winder]

The advantage that I see in the 24MP sensor is that if everything else is equal (DR/Color/noise) then the noise will appear as finer grain and will be more appealing in final output. The technical charts might not change, but the actual appearance of prints will be better. The ability to crop is improved and for people like me who don't like 3/2 that is a great thing.

I think the K-5 replacement with the 24MP sensor will be a huge success. You have to think that Pentax has improved all of the support technologies of the K-5 (image processing, AF, SR, shutter induces blur). Add those thing to the 24MP sensor and it becomes a pretty compelling camera.

K-3:
24MP APS-C
21-point AF (wider spread & smaller more accurate points)
Larger OVF
SR that comes faster than the AF.
Improved 14-bit image processing
1/250 flash sync
2 SDXC card slots

Everyone is using the same sensors, so Pentax is going to have improve the features and support technologies in-order to separate their products.

[Zygonyx]

Please, no jumping to conclusions here.

First, DR is (theoretically) unaffected by the SLT's semitransparent mirror. The low iso score is, the landscape score isn't. The semitransparent mirror reduces effective iso (moves DR[ISO] curve left, not down), that's all.

Second, at DR as high as delivered by the K-5 (and only if as high) is negatively affected by a 12 Bit A/D converter. In your comparison, not all A/D converters are 14 Bit, the K-5's is.

Third, 14EV DR is hard to measure accurately. We talk about 3 electrons remaining read-out noise! Don't put to much interpretation in differences in the order of +/- 0.25EV.

IMHO, all in all, to say that the 24MP and 16MP sensors perform very similiarly is the best we can currently do.

Thanks for these complements.

[MrPetkus]

24MP is a step forward, not back. More megapixels = more detail, more cropping latitude. Sure, this may impact noise but this can be mitigated to a degree by the sensor implementation. To deny the advantages of a higher resolution sensor to consumers because there is some more noise or <<you>> don't need to print or view on large media is foolish.

[Winder]

24MP is a step forward, not back. More megapixels = more detail, more cropping latitude. Sure, this may impact noise but this can be mitigated to a degree by the sensor implementation. To deny the advantages of a higher resolution sensor to consumers because there is some more noise or <<you>> don't need to print or view on large media is foolish.

But there really isn't more noise with the new sensor. S/N ratio is almost identical. The end result is better IQ.

[audiobomber]

The advantage that I see in the 24MP sensor is that if everything else is equal (DR/Color/noise) then the noise will appear as finer grain and will be more appealing in final output. The technical charts might not change, but the actual appearance of prints will be better.

I believe the DXOMark SNR score is normalized to 8mp. If two cameras score the same, they will show the same amount of noise when printed.

[JohnBee]

I believe the DXOMark SNR score is normalized to 8mp. If two cameras score the same, they will show the same amount of noise when printed.

Wow I didn't know that!
Hmmmm that doesn't seem very logical though does it?
I mean... I often crop my prints and when I do, I often try to maintain a 1:1 output.