[ogl]

O my, I don't want to go back that road!

Yes...JPEGs are not bad in terms of noise. But...
Is it new ways to use APS-C camera?
1. Don't shoot in RAW
2. If you use RAW, use good NR software.
3. Never use ISO higher than ISO400 in RAW....
4. If you want to see clear picture, use downsize till 1-2 MP.

hmmm....

K-5 and K-r offer very good balance resolution/noise.

By the way, new Samsung 20 MP APS-C sensor (IMO) is better choice than Sony...

Pentax can improve IQ from this 24 MP sensor a bit...But...I don't think that it will be close to 12 or 16 MP APS-C.

[falconeye]

Pentax can improve IQ from this 24 MP sensor a bit...But...I don't think that it will be close to 12 or 16 MP APS-C.

According to DxO, the new 24MP probably outperforms the 16MP sensor, after taking into account: scaling to print size, light loss by mirror, Sony processing.

[Rondec]

Yes...JPEGs are not bad in terms of noise. But...
Is it new ways to use APS-C camera?
1. Don't shoot in RAW
2. If you use RAW, use good NR software.
3. Never use ISO higher than ISO400 in RAW....
4. If you want to see clear picture, use downsize till 1-2 MP.

hmmm....

K-5 and K-r offer very good balance resolution/noise.

By the way, new Samsung 20 MP APS-C sensor (IMO) is better choice than Sony...

Pentax can improve IQ from this 24 MP sensor a bit...But...I don't think that it will be close to 12 or 16 MP APS-C.

I would say that iso 1600 is max on this sensor, where as I shoot to iso 6400 on the K5. Maybe Pentax could get an extra stop out of this sensor compared to Sony, but I don't know really see the benefit (other than selling to consumers who don't know better who want higher megapixel cameras).

[ogl]

According to DxO, the new 24MP probably outperforms the 16MP sensor, after taking into account: scaling to print size, light loss by mirror, Sony processing.

We see real test studio shot.
Do we make pictures for DxO measurements?
By the way, K-5 has better DR and low-light ISO under DxO data.
Color Depth is almost the same.

DxOMark - Compare cameras side by side

Honest - I don't need 24 MP at APS-C. Yes:The resolution shines. But, 16 MP is my limit for APS-C.

[rawr]

K-5 and K-r offer very good balance resolution/noise

Don't forget the K-x!
Studio shot comparison: Digital Photography Review



It always looks really good in those dpreview comparisons in JPG and RAW. Especially when you use the watch face as the test image. K-x does even better than the 5D2.

In fact K-x and A77 are the only ones to resolve the split in the 'fork' of the 9 o'clock marker, and here the K-x does better than even the A77.

The high resolution potential of the A77 seems to get severely degraded by the noisiness of the sensor.

[ogl]

Some comparisions

[ogl]

One more

Attached Images

 

[falconeye]

Some comparisions

I really don't like misleading comparisons ...

The A-77 has a translucent mirror eating about 1/3 of light according to DxO tests for the A580 vs. SLT A35.

This means that the A77's ISO 1600 really is ISO 2400 for normal cameras (if one wants to compare the sensor...).

Then, Sony underperforms with the same sensor, compared to Pentax and Nikon, as the comparison of the A580 vs K-5 or D7000 shows.

Finally, images must be scaled to match one size.

In the following comparison, I did the necessary work in order not to jump to conclusions like anybody else.

It is all images scaled to 8MP, and A-77 ISO 1600 compared to ISO 3200 from models w/o a translucent mirror. ISO 2400 isn't available in the comparisons.

As you can see, the A-77 and A580 perform almost the same and I see no reason to assume that the sensor embedded in the A-77 has worse low light or dynamic range performance than its 16MP predecessor.

[ManuH]

In fact K-x and A77 are the only ones to resolve the split in the 'fork' of the 9 o'clock marker, and here the K-x does better than even the A77.

It's an exposure issue rather than a noise problem. Which makes me think that the comparison is not quite as controlled as one would like.

Judging by the pictures alone, it's clear that the 5D mk. II is better, follow by K-5 then A77 (I didn't check K-r). The difference in resolution is not that much visible. The 5D is noticeably sharper at any ISO. But that gain could be due to other factors: lens, aperture, light, raw converter, etc. In conclusion it just tells me that the difference are not that big.

[ogl]

I really don't like misleading comparisons ...


It is all images scaled to 8MP, and A-77 ISO 1600 compared to ISO 3200 from models w/o a translucent mirror. ISO 2400 isn't available in the comparisons.

As you can see, the A-77 and A580 perform almost the same and I see no reason to assume that the sensor embedded in the A-77 has worse low light or dynamic range performance than its 16MP predecessor.

I see A77 is worse even at 8 MP. Not the same.
But it's 24 MP camera.


I think it will be better to compare real and indicated ISO by DxO data...

A77 - ISO50 = real ISO63
K-5 - ISO80 = ISO70


A77 - ISO800=ISO635
K-5 - ISO800=ISO717

A77 - ISO1600=ISO1302
K-5 - ISO1600=ISO1613

A77 - ISO3200=ISO2612
K-5 - ISO3200=ISO2864

It's not 1/3....

By the way, from ISO200 - K-5 has better colour sensivity. Better tonal range at all ISO.
Better DR.

[rawr]

I look forward to the DxO assessment of the [non-translucent mirror] NEX-7 for more interesting/ less crippled comparisons. This is fun.

[RonHendriks1966]

I look forward to the DxO assessment of the [non-translucent mirror] NEX-7 for more interesting/ less crippled comparisons. This is fun.

The a77 is 12-bit, and would we get 14-bit in a Pentax model and what would be the difference?

[ogl]

The a77 is 12-bit, and would we get 14-bit in a Pentax model and what would be the difference?

maybe in DR...

[falconeye]

I think it will be better to compare real and indicated ISO by DxO data...

A77 - ISO50 = real ISO63
K-5 - ISO80 = ISO70

IMHO, there is a misunderstanding on your part here.

The real ISO measurement only calibrates the camera exposure meter with a standard.

If two cameras would both have identical real ISO, then it would only mean that identical amounts of light enter the lens.

The amount of light absorbed by the mirror, AA filter, micro lenses, the Bayer filter etc. is not measured and still differ from camera to camera.

Which is why cameras with identical sensors and identical processing can differ in performance.


In order to find out, how much light is absorbed by the translucent mirror, one must compare the two luminance noise curves of SLR A580 and SLT A35 (sharing the same 16MP sensor) and compute the shift along the horizontal axis to make both curves match. This shift is the light loss. This shift is almost exactly 1/2 stop. Which in turn corresponds to 70% transmission and 30% reflection. Because the effect is like a permanant ND filter, Sony can offer true ISO 56 if Pentax can offer true ISO 80

DxO made the same reasoning and wrote it up here. In their own words: "we lose between 1/3 and 2/3 stop, which -unless a more translucent material is found- is the best Sony could come up with"

[Zygonyx]

According to DxO, the new 24MP probably outperforms the 16MP sensor, after taking into account: scaling to print size, light loss by mirror, Sony processing.

I doubt it can be better "dynamic range" wise : -0,9 EV as measured by DxO.