[derekkite]

OK' here are a few taken inside at ISO 1000, 1600, 2000
Same lens (yeah ! even inside!)
No scientific testing of course. The calendar used is curved on top and bottom, so of course that will show as OOF at the borders.
Under tungsten light but no camera setting done for that sort of lighting.
No flash, camera on a bench (no tripod) no remote, no delay.
Distance was about 12 feet.




---------- Post added 02-20-16 at 04:58 PM ----------


Yep ! RAW (DNG).
I'll just have to start with the slight front focus ... not too great using those charts .... better idea?
What is the "falloff" ?

---------- Post added 02-20-16 at 05:16 PM ----------

@ Penumbra:
Yes, the first one is cropped a bit (was showing at around 16% originally and I cropped it to 32%)

Falloff in quality. You will find that it is either stunningly sharp, clear and nice, or garbage, no in between.

[NZ_Ross]

Now I think about it - Ed from Photouniverse did a big rant on how bad the K-3 image quality was when he first got it coming off a K-5.

Then he adjusted the default factory settings and seemed to end up happy enough with the image quality. From memory he also has a video on the settings he uses for a K-3 which is basically how I have my K-3 set up. Just do a search on K-3 on his Youtube site and you will find them. There might be some helpful stuff there for you.

[jpzk]

I faced the same dilemma before, and because I have so many photos (thousands) to process each time after a gig, I had thought about that before. I didn't like the previous versions of Silkypix (prior to 6). I just use the standard version which surprisingly is cheaper (currently on special price) than Photoshop elements and LR. I am a very happy user of Silkypix now, of course, YMMV.

Thanks again.
I will certainly consider Silkypix 6 once I go through the paces with what I presently have.

---------- Post added 02-20-16 at 07:27 PM ----------

I don't know if CS5 applies anything by default on import so I cannot help there. Lightroom is 'script' driven so on import it will apply automatically a set of adjustments if you want, which I do.

What I mean by develop for the sensor is simply forget everything you learned or did with your k-5 images. The K-3 is different, the sensor does not produce the same image as the k-5. If you develop a k-3 image the same as a k-5 image you will be disappointed. I am definitely not the only one to notice that. There were a lot of complaints initially as people used k-5 develop settings on k-3 images with poor results. Several initial reviews were very negative, I suspect because of this reason.

So, basically forget everything you know about developing an image and start over. K-3 sensor is different. My reaction, at first, was the same as NZ_Ross "what have I done". Seriously considered sending the k-3 back! But glad I did not, the image quality is there you just have to learn how the sensor reacts. Think of it as learning to develop a different brand of film. Things are mostly the same, but not completely and that gets you in trouble.

Well, I suppose this is like coming to a new system altogether.
I must have put at least 90,000 clicks with the K5 and I have been so used to it that it became "intuitive" .... automatic!
I also did some tweaking in the camera settings for Hi ISO NR: I've set it up to my taste in the "Custom" NR pull down sub-menu, after watching the video suggested by NZ_Ross - Ed from Photouniverse .
That should help.

---------- Post added 02-20-16 at 07:28 PM ----------

Falloff in quality. You will find that it is either stunningly sharp, clear and nice, or garbage, no in between.

Alright, and thanks again for the help.


---------- Post added 02-20-16 at 07:29 PM ----------

Now I think about it - Ed from Photouniverse did a big rant on how bad the K-3 image quality was when he first got it coming off a K-5.

Then he adjusted the default factory settings and seemed to end up happy enough with the image quality. From memory he also has a video on the settings he uses for a K-3 which is basically how I have my K-3 set up. Just do a search on K-3 on his Youtube site and you will find them. There might be some helpful stuff there for you.

Good thing!
Just viewed the video and applied some of the settings he suggested, especially as they relate to Hi ISO NR settings.
After a little tweaking, I'll go back out tomorrow and retest this K3 again.

Thanks for the help!

[Fauxton]

I have both the K5 and K3. I had the same reaction at first. Many of us did. The K3 has the same noise levels at the K5 at the same size. However if you view the K5 at 100% and the K3 at 100% - pixel peep in other words - you will find only about the same level of noise in K3 compared to the K5. For example have a look here Pentax K-3 vs Pentax K-5 II vs Pentax K-5 .
So what can be done? Well, in the K3 (and K3II) you have better processing and more pixels over the K5. The K3 is natively sharper as it also does not have an AA filter. So at similar print sizes you will have better resolution and more information with the K3 than K5. Also you will have a bit better noise characteristics with the K3 - it has different grain characteristics. So its your choice to apply NR and trade off a bit of the resolution/information should you wish against the noise. As a few have mentioned you don't need to go far with the NR in post to clean the image up nicely. I find I can do this almost without penalty to 1600, quite well up to 3200 generally, and with a bit more trade off and depending on the subject up to 6400. You might get more or less depending on your subject, size of print, resolution requirements and your level of acceptance for noise. My NR of choice is Nik when using CS/ACR or the NR in DXO Optics 10. The NR in DXO Optics 10 is outstanding. Another program that has excellent noise handling for the K3 is Rawtherapee (probably the best - IMHO - there is for its demosaicing and noise balance - it is complicated though). C1 V9 is very good as well. I don't bother with DCU, Silkypix or Adobe's NR (tried them and I don't get as good results as with the others above).

Search for the posts either here or on DPReview. There is a fair amount of info. It likely will take you a couple of hundred shots, maybe a few thousand to really see what the K3 can do and dial it in and adjust your post process workflow. The results are well worth your time. One last point - its also worth your time to dial in the JPEG settings. Once you've done that SOOC can be pretty amazing. But the default setting are not particularly well chosen - Pentax doesn't do itself any favours with its downright goofy default settings (magenta bias for example, over saturation in the default Bright setting for another, heavy handed NR defaults for another....)

[jpzk]

I have both the K5 and K3. I had the same reaction at first. Many of us did. The K3 has the same noise levels at the K5 at the same size. However if you view the K5 at 100% and the K3 at 100% - pixel peep in other words - you will find only about the same level of noise in K3 compared to the K5. For example have a look here Pentax K-3 vs Pentax K-5 II vs Pentax K-5 .
So what can be done? Well, in the K3 (and K3II) you have better processing and more pixels over the K5. The K3 is natively sharper as it also does not have an AA filter. So at similar print sizes you will have better resolution and more information with the K3 than K5. Also you will have a bit better noise characteristics with the K3 - it has different grain characteristics. So its your choice to apply NR and trade off a bit of the resolution/information should you wish against the noise. As a few have mentioned you don't need to go far with the NR in post to clean the image up nicely. I find I can do this almost without penalty to 1600, quite well up to 3200 generally, and with a bit more trade off and depending on the subject up to 6400. You might get more or less depending on your subject, size of print, resolution requirements and your level of acceptance for noise. My NR of choice is Nik when using CS/ACR or the NR in DXO Optics 10. The NR in DXO Optics 10 is outstanding. Another program that has excellent noise handling for the K3 is Rawtherapee (probably the best - IMHO - there is for its demosaicing and noise balance - it is complicated though). C1 V9 is very good as well. I don't bother with DCU, Silkypix or Adobe's NR (tried them and I don't get as good results as with the others above).

Search for the posts either here or on DPReview. There is a fair amount of info. It likely will take you a couple of hundred shots, maybe a few thousand to really see what the K3 can do and dial it in and adjust your post process workflow. The results are well worth your time. One last point - its also worth your time to dial in the JPEG settings. Once you've done that SOOC can be pretty amazing. But the default setting are not particularly well chosen - Pentax doesn't do itself any favours with its downright goofy default settings (magenta bias for example, over saturation in the default Bright setting for another, heavy handed NR defaults for another....)

My NR of choice is Nik when using CS/ACR or the NR in DXO Optics 10. The NR in DXO Optics 10 is outstanding.
That's the Google software? As in here ( Nik for NR): https://www.google.com/nikcollection/products/dfine/# at $169.00 CDN
And the DXO Optics .... $129.00 USD
Both quite expensive though.

[jpzk]

UPDATE:

I am slowly getting used with the K3 as far as "noise" is concerned ... quite a departure from the K5 indeed !
Here are a few AF "test" shots (not at all scientific) that I did today.
There are too many images to post them all here so I am providing a link.
I am not sure, but I think comments can also be made in the Flickr images.
Any comments would be welcomed !

The Flickr album containing the images is here: https://www.flickr.com/gp/rdlphotos/P042EX
and I am hoping you will be able to view this album because I have made it "private".
Edit: I changed the link. It wasn't working.

[boriscleto]

UPDATE:

I am slowly getting used with the K3 as far as "noise" is concerned ... quite a departure from the K5 indeed !
Here are a few AF "test" shots (not at all scientific) that I did today.
There are too many images to post them all here so I am providing a link.
I am not sure, but I think comments can also be made in the Flickr images.
Any comments would be welcomed !

The Flickr album containing the images is here: https://www.flickr.com/photos/rdlphotos/albums/72157664405926250
and I am hoping you will be able to view this album because I have made it "private".

Private means no one but you can see it. The link just gives a page not found error.

[jpzk]

Private means no one but you can see it. The link just gives a page not found error.

OK, I'll correct this right away.
I can always put in back "private" later.
Thanks for the warning.

Edit: I've just made the album "available with people you will share it with": https://www.flickr.com/gp/rdlphotos/EE29y6
The exact wording was : "Any private photos in this album will remain private. Only people you share this link with will be able to view them. Everyone will be able to download these photos."

Let me know if that works.

[boriscleto]

Let me know if that works.

Yes, that works.

[jpzk]

Yes, that works.

Yeah !
Thanks !

[jpzk]

For those who had a "page not found" error clicking on the original link,
See Post No. 23 .... that link works.
Cheers!

JP

[howieb101]

Hi.

Here is a link to a forum post on hi-iso concert photo's with the K3 and use of noise reduction.
https://www.pentaxforums.com/forums/172-pentax-k-3/269926-concert-shot-k-3-sigma-18-35-a.html

If you have read of this you can see that what can be achieved with a hi-res APS-C sensor is pretty remarkable. It involves post production work but if you compare it too buying a brand new full frame camera, buying some computer software may make the $US170 seem like small fry.

That's until you give in and buy the full frame anyway. Lol.

Howie Be

[jpzk]

Hi.

Here is a link to a forum post on hi-iso concert photo's with the K3 and use of noise reduction.
https://www.pentaxforums.com/forums/172-pentax-k-3/269926-concert-shot-k-3-sigma-18-35-a.html

If you have read of this you can see that what can be achieved with a hi-res APS-C sensor is pretty remarkable. It involves post production work but if you compare it too buying a brand new full frame camera, buying some computer software may make the $US170 seem like small fry.

That's until you give in and buy the full frame anyway. Lol.

Howie Be

Thanks, Howie!
I've looked at the images from Stagnant with the bands .... Impressive!!
Now, about this DxO ... well worth it, you say?
What is this DxO 9 using Prime noise reduction engine ?
Is that part of DxO or an add-on ?

Edit: just went to DxO site: $199.00 USD for the "Elite" version which is required if you want to have the "Prime" denoising engine.
Quite a bit of $$$ !

[howieb101]

Hi jpzk.

Well, I didn't actually say well worth it. More that for the money it is a lot cheaper than buying a full frame camera.

Also, if you didn't know, the sensors used by Pentax, Sony and Nikon are nearly all made by Sony (there are a few exceptions every now and then).

Anyway, what this means is that if you were looking at an APS-C 24 Mpx camera by Pentax, Sony and Nikon they are likely to have similar hi-iso performance (on the assumption they are within the same generation). Note: Sony and sometimes Nikon sometimes have first use rights to a sensor so it can take 6-12 months for say Nikon or Pentax to get the same or equivalent sensor.

Anyway, after all of that it is out of the way, my main point would be that if the APS-C K3 image quality doesn't take your fancy it is unlikely any other 24 Mpx camera by Sony or Nikon is likely to be any better. As a further aside, Canon sensors have been known to be lagging behind Sony manufactured sensors for approx. the last 5 years. Google, Sony vs Canon sensors or dxo comparisons and you'll find out what I mean.

So, yes, the software cost of DXO is quite high relative to the camera body cost, especially if buying in the USA. Only you can decide if its worth it.

By the way, if you get proficient with Photoshop or equivalent, you can get away with your own noise mitigation. It will take some work but if it is only on selective photographs you can save yourself some $$$.

I should add, I have bought DXO myself when their was a sight discount here at Pentax Forums. Whether it was worth it for me is probably marginal but I am pretty good at Photoshop. For a person who just wants to click on a few buttons, DXO is definitely easier.

Howie Be

[jpzk]

Hi jpzk.

Well, I didn't actually say well worth it. More that for the money it is a lot cheaper than buying a full frame camera.

Also, if you didn't know, the sensors used by Pentax, Sony and Nikon are nearly all made by Sony (there are a few exceptions every now and then).

Anyway, what this means is that if you were looking at an APS-C 24 Mpx camera by Pentax, Sony and Nikon they are likely to have similar hi-iso performance (on the assumption they are within the same generation). Note: Sony and sometimes Nikon sometimes have first use rights to a sensor so it can take 6-12 months for say Nikon or Pentax to get the same or equivalent sensor.

Anyway, after all of that it is out of the way, my main point would be that if the APS-C K3 image quality doesn't take your fancy it is unlikely any other 24 Mpx camera by Sony or Nikon is likely to be any better. As a further aside, Canon sensors have been known to be lagging behind Sony manufactured sensors for approx. the last 5 years. Google, Sony vs Canon sensors or dxo comparisons and you'll find out what I mean.
I think this is going to get better, after being acquainted with the K3. It does need more work but probably well worth it.
So, yes, the software cost of DXO is quite high relative to the camera body cost, especially if buying in the USA. Only you can decide if its worth it.

By the way, if you get proficient with Photoshop or equivalent, you can get away with your own noise mitigation. It will take some work but if it is only on selective photographs you can save yourself some $$$.
I am not at all proficient with Photoshop, except for some basic PP such as exposure, sharpening, etc ... nothing fancy.
The ACR RAW developer has never been my forte.
I should add, I have bought DXO myself when their was a sight discount here at Pentax Forums. Whether it was worth it for me is probably marginal but I am pretty good at Photoshop. For a person who just wants to click on a few buttons, DXO is definitely easier.
I downloaded the trial version of the DxO Elite, with the Prime denoising engine included, and I must admit that it seems to be much easier than PS/CS5/ACR and I have actually (already) processed some files which I will put in an album at Flickr tomorrow. Looks like DxO does a better job.
Saty tuned for an update and let me know what you think when all is ready to be viewed!
Howie Be

Thanks again for the help !

---------- Post added 02-23-16 at 11:00 PM ----------

Here is another Flickr album in which I posted some shots taken tonight (very quickly and certainly not under the best circumstances) for comparing DxO "Prime" denoising engine (DxO 10 Elite version - trial) vs. Photoshop CS5 ACR noise reduction.
All shots were taken RAW (DNG) and shown here (Flickr album) in JPEG's.
Note: I am NOT very good at anything "software" or "post processing" of any kind, so please bear with me.
Let me know if this link works:
https://www.flickr.com/gp/rdlphotos/80bwu6

Thank you !
JP