[Camera lucida]

I looked at the photo of the two ladies shot with the 645D.
I've suddenly realised that the old pro's were right. My eyes were instinctively drawn to the lower left point were the lines of composition intersect.
Isn't photographic theory wonderful!

[er1kksen]

When you open a file in Camera Raw, there is an option at the bottom to select image resolution. My theory is/was that if you raise the resolution, that ACR will effectively reduce the pixel size which in turn would allow the noise reduction software to treat the image differently.

Do you know if the camera raw module in PS elements should be capable of this as well?

[kiwi_jono]

I looked at the photo of the two ladies shot with the 645D.
I've suddenly realised that the old pro's were right. My eyes were instinctively drawn to the lower left point were the lines of composition intersect.
Isn't photographic theory wonderful!

LOL!
I noticed the same thing but there may be other forces at play here

[Atindra]

Details in John Bee's sample with early ACR conversion are too good for that range of ISO. One would expect hairs should look like a wig but here still one can see hairs are finely resolved.

One thing is sure that 51200 will be usable with some training in Noise removal.

Atindra

[fs999]

Thank you all !

jpzk : the door of the three K-5s were sealed or glued, but one did open...
I changed the SD card in the morning, but didn't see that the photos were saved as PEF, so I went back in the afternoon and saw that the original 2GB SD card was stolen !
Sorry for the server reset, I forgot that my site provider makes a backup at that time...

Excellent work JohnBee !

[Jonson PL]

Great shot's always difficult to compare, because of not yet firmware version 1.0 but stil as a trial-version.
I had the chance on wednesday to try the 645D. Took some shots at iso-1000 didn't change this.

I’m in awe at this being Iso 1000 from a CCD sensor. I know that sensor size correlates to noise, but with this quality, you could easily push one stop, and then have actual Iso 3200. Didn’t know Medium Format could do this well at higher Iso. And this is just from JPEG.
What aperture was it ?


Regarding the K5, all the numbers we've gone over the last weeks, are one thing. But I’m impressed now that I see actual photos. Iso 3.200 looks very good. And Iso 12.800 would be useable too, considering it is from JPEG and early firmware. Thanks a lot for taking the shots and posting them.
The Raw 51.200 Iso shot, this could easily be converted to B&W, if out on night safari stalking wildlife in Africa.

Thanks
I too tried Noiseware Pro and a host of other NR tools. Mainly to see how they would response to the image grain. However, after a few trials, I settled on a blend of ACR Chroma/Luma and finished off with Denoise afterward. So a little of both.

One of the main issues I've found with reference to the grain in this particular sample is that external noise reduction systems don't seem to recognize the noise patterns as actual noise. If I had to guess, I'd say that the grain(cross hatch noise) is too large to be effectively identified as noise by the actual software.

So to test this theory, I interpolated the image to 26MP in ACR, to effectively shrink image grain size and ended-up with the following:

K-5, ISO51200 (ISO102400 equivalent), Early ACR sampling:


It's not a huge leap over the previous sample, though it is an improvement.
So at least we will have that much to look forward too, when better samples do come around

I think this one is way better, than the first one you posted.
ISO 102.400 equivalent, this is almost difficult to come to terms with.
If you added 2 EV, isn’t that two stops ? So that it would be more like Iso 152.600 ?

[Christopher M.W.T]

Thank you all !

jpzk : the door of the three K-5s were sealed or glued, but one did open...
I changed the SD card in the morning, but didn't see that the photos were saved as PEF, so I went back in the afternoon and saw that the original 2GB SD card was stolen !
Sorry for the server reset, I forgot that my site provider makes a backup at that time...

Excellent work JohnBee !

Great work!

What else can you share about the AF performance and overall responsiveness of the K-5 in comparison to the K-7?

[Christopher M.W.T]

Great shot's always difficult to compare, because of not yet firmware version 1.0 but stil as a trial-version.

I had the chance on wednesday to try the 645D. Took some shots at iso-1000 didn't change this.

Your friend terrifies me, but that's an amazingly clear shot for 1000iso.

As for the 2nd photo, all I can say is "EUROPEAN ass " and I absolutely miss it!

They should hold Photokina in Belgrade Serbia if you want the best looking women

[raz]

Iso 800 and 1600 - looks very good, better than my K100D. But...maybe I overreacting, we need to see some samples in the dark.

Iso 3200 - the subject is out of focus ? I cannot tell how many details are preserved.

Iso 12.800 - looks very promising. Th NR kicks in and I don't like how it does to the bokeh, but the face looks good.

Iso 51.600 - very noisy but again, lots of details are kept and this is impressive. His hair looks very good, after de-noising details can still be visible on his face, I think this can be printed on decent sizes and still would look very decent.

I think the results are better (even a lot better) than the D7000 I seen from another thread. Let's hope Nikon doesn't see this thread then tweaks their camera before the dpreview comparative tests

[melander]

After a quick look at the samples, I have come across a few early observations.

The first is that the scene is(at least...) 2 stops underexposed. Which is good in some way because it leaves a little headroom for a slightly better noise performance that what is seen here.

The bad news is however, that I wasn't able to extract much out of the image. Though I am hopeful the K-5 will have the capacity to provide a little more in terms of usable output at this range, I'm leaning toward the idea that getting it will require a fairly stringent workflow with respect to camera settings and conditions in which images will be recorded. ie. something along the lines if a a blend of camera NR and post processing.

Anyways, here's a very quick ACR sample with +2 eV(stops) of exposure added to balance scene. And though it does seem pretty far from ideal at, I think its safe to say that we can expect a little more in terms of detail with better exposure and some carefully chosen settings. (time will tell)

K-5, (early) ISO51200 Sample from RAW w/ +2 eV in ACR:


Original with Default Exposure:


So clearly the +2 EV in ACR took its toll on the final image(ISO102400?).
So all is not lost just yet

So this is equal iso 204800? Iso 51k pushed 2 stops.

[Douglas_of_Sweden]

Hopefully we can turn off NR for jpeg at 3200 and do our own processing. I can shoot RAW if needed...just never learned to process them correctly!

Think I read somewhere that you are even supposed to be able to set the NR individually for each ISO.

[loule]

Merci Fred et RonHendricks pour ces échantillons très très prometteurs.
51200 iso sur le k5 = 800 iso sur mon K10.....

[RonHendriks1966]

Here two other shots made with the 645D and the D-FA 55mm. All are made at the Pentax boot, so there is some snapshot to it. I would have take it for a walk, but then there was only one 645D left over for all the other visitors that day

All three picture's where taken at iso1000 and f2.8 I couldn't find the setting to go to iso1600 at the time that I had the 645D in my hands. I had about 15 minutes, so that's nice but not enought to see everything. In it I had a class 4 card so writing was slow. I made 10 RAW+Jpg pictures and that made up for 809 MB data. To my surprise when I put that card in my K-7 I could see the jpg that's in the RAW file on my screen, so it's excactly the same as in the K-7, but no possibility to enlarge or RAW conversion.




This last picture was put bhere before, but the other was from Jpg and this is from RAW and a little more detail in the dark parts.

When I looked true the viewfinder it's a lovely sight. No Idea how big it is in comparisson with the K-7 but it looks good. But then I was at the stand of Hasselblad and looked true the 65 megapixel camera and that was like looking out a window so large. No idea how they compare viewfinder-wize.

[fs999]

Great work!

What else can you share about the AF performance and overall responsiveness of the K-5 in comparison to the K-7?

Thank you !

The AF is very impressive also with SDM !
The trigger makes less noise than K-7 !
7 i/s is very fast !
Most other features are the same as the K-7...

[RonHendriks1966]

Thank you !

The trigger makes less noise than K-7 !

Thank you I had that exact idea, but wasn't shore.