[Agnostic]

Only 1/3 stop? Not almost double?

That's what Sony states. Might be somewhat optimistic under real world conditions.

[Kunzite]

My bad, yes, they're sending ~2/3 of the light to the sensor and ~1/3 to the AF system (70% - 30%).
I was applying some strange math, with very different results

L.E. The error is that we're talking about 1/3 of the usable light, and not 1/3 EV. Am I right?

[blende8]

30% = 3/5 ev

[Edvinas]

DPR data

Useable DR:

K-7 8.6
7D 8.3
300s 8.4
50D 8.3

Canon EOS 7D Review: 17. Photographic tests (DR): Digital Photography Review

And according DxOMark:
K-7 10.6
7D 11.7
300s 12.2
50D 11.4

K-x 12.5

[Agnostic]

My bad, yes, they're sending ~2/3 of the light to the sensor and ~1/3 to the AF system (70% - 30%).
I was applying some strange math, with very different results

L.E. The error is that we're talking about 1/3 of the usable light, and not 1/3 EV. Am I right?

DP review states: "The mirror in the A55 is semi-translucent. According to Sony, it allows 70% of the light that strikes it straight through to the 16.2MP CMOS imaging sensor. The remaining 30% is reflected upwards to a large 15-point phase detection AF array. This equates to a roughly 1/3EV reduction in light transmission for live view/image capture"

[Steelski]

DP review states: "The mirror in the A55 is semi-translucent. According to Sony, it allows 70% of the light that strikes it straight through to the 16.2MP CMOS imaging sensor. The remaining 30% is reflected upwards to a large 15-point phase detection AF array. This equates to a roughly 1/3EV reduction in light transmission for live view/image capture"

So if my maths is correct,
Sony bad Jpegs + only 70% light throughput = Pentax real Iso 25600 and close to D700 performance. ish.


The sensor is good.
I can confirm that after looking at the Sony NEX5 vs the A33 at Imaging Resources that the NEX5 Iso 12800 is almost as good as the 6400 from the A33
http://www.imaging-resource.com/IMCOMP/COMPS01.HTM

A screen capture to prove this.
http://www.dpreview.com/galleries/3600674786/photos/492403/a33-vs-a560?inalbum=for-forum-images

[opiedog]

is that what one pentax dealer meant when he said "pentax will release something that will give full frame a run for its money"?

[knoxploration]

I find it fun that when K-7 came out I read tons of reviews where ppl where overwhelmed. It was so fantastic... Not that much different then the K20D in the specs - I think?

Not sure why you'd think that. The resolution might've been the same, but the list of changes from the K20D was extensive:

* New 4-channel readout sensor w/ on-chip noise reduction
* Burst speed 3fps -> 6fps
* 16-segment -> 77 segment metering
* Two axis sensor shift -> Two axis plus rotational sensor shift
* Sensor shift dust reduction -> Piezoelectric dust reduction
* 1/4000 -> 1/8000 shutter
* 95% -> 100% viewfinder
* 2.7" 230k dot -> 3.0" 920k dot LCD
* Composite -> composite + HDMI video output
* Uprated AF with light color source detection, dedicated AF assist lamp
* More accurate white balance
* New body (smallest in class by a significant margin)
* New features -- in camera HDR, in-camera lens correction, AF assist lamp, auto level, composition adjustment, 30fps HD movie mode
* Etc. etc., by no means a comprehensive list

[Aristophanes]

is that what one pentax dealer meant when he said "pentax will release something that will give full frame a run for its money"?

That assumes FF is standing still.

Not!

[Urkeldaedalus]

is that what one pentax dealer meant when he said "pentax will release something that will give full frame a run for its money"?

I think the quote you're referring to was from a JohnBee post, and that the Pentax official supposedly said the new camera would give the Nikon D700 a run for its money. I assume the newer Nikon version of this camera (D700s, D800, whatever its called) will do even better than the D700 as far as high ISO performance.

Still, if this new camera could get anywhere close to Nikon D700 performance, my interest in full frame would be reduced. Really, if they even get it to K-x performance or slightly better with a 16 mp sensor that will be fantastic.

[rparmar]

I don't care so much about ISO 4 billion. I would just like ISO 100-400 to be cleaner. Pentax, stay with your heritage and remember the photographer!

P.S. Yes, I am a candidate for the 645D. Please send donations through my website.

[pcarfan]

And according DxOMark:
K-7 10.6
7D 11.7
300s 12.2
50D 11.4

K-x 12.5

Fair enough...can we settle on the average of the two then

[pcarfan]

Still, if this new camera could get anywhere close to Nikon D700 performance, my interest in full frame would be reduced. Really, if they even get it to K-x performance or slightly better with a 16 mp sensor that will be fantastic.

+1.......................

[thibs]

And according DxOMark:
K-7 10.6
7D 11.7
300s 12.2
50D 11.4

K-x 12.5

As long as the K7 has lower DR (not that it is any news, really), it isn't as crap as many here do hint at.

It is less good and could be better but nothing to write home about IMO.

[ChristianS]

My biggest Problem with my K20: if i use an iso value above 400 ASA i knew that i will have a big Probelm: NOISE
and so it sounds very good that the K5 may have an ISO-range up to 25000 - and so i think i can use 6400 without any noise busting.
The second Problem: The AF for fast objekts to slow in low light to aggressive so perhaps make the focosing speed swithable: hi speed for action and low speed for high precission or critical light situations
this two things i wish for my next cam.
for lenses: please a 18 - 135 WR as al lens for short trips with less package