Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
02-11-2012, 11:36 AM   #1
New Member
gks's Avatar

Join Date: Oct 2011
Location: Evergreen, Colorado
Photos: Gallery
Posts: 8
K20D & high ISO issues

Previously, I had only used my K20D on ISO's of 400 or less. I'm now taking a class that includes a lot of low light, high ISO shooting. With my first batch of images, I found everything above 400 to be seriously bad grainy. I thought I read an article somewhere that gave Pentax fairly high ratings for its high ISO range.

What experience with this have any other forum members had?

thanks.
gk

02-11-2012, 11:50 AM   #2
Loyal Site Supporter
blackcloudbrew's Avatar

Join Date: Mar 2009
Location: Cotati, California USA
Photos: Gallery | Albums
Posts: 3,782
Practically speaking I found that the K20d was at it's acceptable limit at about ISO800 for me. That's factoring in some noise removal in post processing. Don't know what you are taking pictures of in low light but if you can be on a tripod and shoot longer exposures, you can reduce the ISO to acceptable levels regardless. Honestly, after upgrading to the K5 (where I can operate quite nicely up to between ISO3200 to 6400), I just don't bother with trying to get low light performance from the K20d.
02-11-2012, 11:52 AM   #3
Pentaxian
Wheatfield's Avatar

Join Date: Apr 2008
Location: The wheatfields of Canada
Posts: 10,043
QuoteOriginally posted by gks Quote
Previously, I had only used my K20D on ISO's of 400 or less. I'm now taking a class that includes a lot of low light, high ISO shooting. With my first batch of images, I found everything above 400 to be seriously bad grainy. I thought I read an article somewhere that gave Pentax fairly high ratings for its high ISO range.

What experience with this have any other forum members had?

thanks.
gk
The K20 was pretty good for it's time, but remember that a few generations of sensor have come and gone since then. I never found the K20 to be good for much past about ISO 800 without some noise reduction being applied in post processing.
Also, if you are shooting low light, resist the temptation to under expose and push the image in post. You still need correct exposure, no matter what the lighting conditions.
02-11-2012, 12:25 PM   #4
Pentaxian
johnyates's Avatar

Join Date: Aug 2011
Location: Saskatoon, SK
Posts: 1,175
QuoteOriginally posted by Wheatfield Quote
You still need correct exposure, no matter what the lighting conditions.
+1
For best results, ETTR.

02-11-2012, 12:54 PM   #5
klh
Site Supporter
klh's Avatar

Join Date: May 2009
Location: Rochester, New York, USA
Posts: 765
My experience is similar in that I wouldn't go beyond ISO 800 unless I was planning to do noise reduction later. If you need higher than 800, I suggest shooting raw, and then using something like Topaz Denoise. That allowed me to go up to ISO 3200 with good results with the K20D.
02-11-2012, 01:04 PM   #6
Ash
Community Manager
Loyal Site Supporter
Ash's Avatar

Join Date: Mar 2007
Location: Toowoomba, Queensland
Photos: Gallery | Albums
Posts: 22,679
The K20D did improve on the K10D in this department but only gave perhaps 1 to 1.5 stops of higher ISO performance. I had successfully gotten reasonably clean results with the K20D at ISO 1600 as long as I did expose correctly or to the right, as mentioned. But at ISO 3200 and above, results are unacceptably noisy regardless of how well I exposed the shots.
02-11-2012, 01:20 PM   #7
Pentaxian
JohnBee's Avatar

Join Date: Sep 2008
Location: front of computer
Posts: 4,496
I've always found the K20D to be a rather special camera with high ISO. However, getting the most out of it requires a little TLC. For example, I've found the out-of-camera JPG and NR to be quite limiting in comparison to RAW. And so if you're shooting to retain as much detail as possible then RAW will likely be your thing.

One other aspect of high ISO shooting with the K20D is with the use of a dedicated NR software(as klh mentioned), to which I'd add Denoise is likely one of the best software out there at this time.

And finally... if higher ISO(32/6400) is your thing, then you might want to look at: GordonB's Magenta Cast remover, which removes the magenta cast created by the sensor at higher sensitivities(ISO4000 - 6400). And a debanding software wich is built-into Denoise and other NR software such as Niksoft Dfine.

Hope this helps.

PS. here are a few examples of some K20D ISO6400 output using the above mentioned tools:
META: ISO6400, 1/60s

click for larger image

META: ISO4000, 1/60

click for larger image

META: ISO6400, 1/10

click for larger image

META: ISO6400, 1/40

click for larger image

Granted some of these were processe3d with earlier NR software(beta) and in many cases would yield better results today. Though I think for the most part, that they should give you an idea of what can be expected from high ISO files coming from the K20D.

Hope this helps,
JohnBee
02-11-2012, 01:42 PM   #8
gks
New Member
gks's Avatar

Join Date: Oct 2011
Location: Evergreen, Colorado
Photos: Gallery
Posts: 8
Original Poster
ISO issues

Thanks, everyone, for your input. I was relieved that others had experienced the same things! I do shoot in raw, but like to spend as little time as possible on the computer, so I'm not very interested in utilizing additional software to clean up. I had been thinking about a new camera (K5!) so maybe it's time.

thanks again.
gk

02-11-2012, 01:44 PM   #9
gks
New Member
gks's Avatar

Join Date: Oct 2011
Location: Evergreen, Colorado
Photos: Gallery
Posts: 8
Original Poster
more

OH, and JohnBee, those are absolutely lovely images. Especially the girl in candlelight!

gk
02-11-2012, 05:34 PM   #10
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
It takes all but ten seconds to apply NR in your PP program, but if you don't want to spend those ten seconds, then you can spend those same ten seconds turning on the NR in your camera. Could just be you have unreasonably high expectations and are doing too mich pixel peeping rather than judging your images from actual prints or displays that actually fit on the screen. It might help to post sample images you are not happy with. The K-5 is better, but it's hardly night and day.
02-11-2012, 06:29 PM   #11
Pentaxian
JohnBee's Avatar

Join Date: Sep 2008
Location: front of computer
Posts: 4,496
QuoteOriginally posted by Marc Sabatella Quote
The K-5 is better, but it's hardly night and day.
If we consider all of the high ISO IQ advantage bewteen the K-5 and K20D, I think it might very well be night and day.

Here's what I found at ISO6400:

Higher dynamic range
Much less noise
No banding
No magenta cast
Higher resolution

Put together we get this: (K-5, ISO6400)

Click image to see full size

Having said, I don't think we'll ever see anything remotely close from a K20D(not even 1/2 resolution), and so I think this qualifies as night and day.
JohnBee

Last edited by JohnBee; 02-11-2012 at 06:34 PM.
02-11-2012, 06:41 PM   #12
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
At 6400, maybe, but even so, perhaps only once you've applied your very sophisticated NR techniques to get the most out of the K-5. Straight from camera, the differences aren't that noticeable unless pixel peeping. And up to ISO 1600 or 3200 in a typical print or viewed on a computer screen other than at 100%, I doubt most could tell the difference at all. At 100%, the difference might be more apparent, but if you need to go to such extremes of pixel peeping to see them, that hardly qualifies as night and day.
02-11-2012, 07:13 PM   #13
Pentaxian
JohnBee's Avatar

Join Date: Sep 2008
Location: front of computer
Posts: 4,496
QuoteOriginally posted by Marc Sabatella Quote
At 6400, maybe, but even so, perhaps only once you've applied your very sophisticated NR techniques to get the most out of the K-5. Straight from camera, the differences aren't that noticeable unless pixel peeping. And up to ISO 1600 or 3200 in a typical print or viewed on a computer screen other than at 100%, I doubt most could tell the difference at all. At 100%, the difference might be more apparent, but if you need to go to such extremes of pixel peeping to see them, that hardly qualifies as night and day.
I find the K-5 images to be much cleaner from ISO1600 straight from the camera. However, what I think stands-out the most with the K-5 in contrast to the K20D, is in the level of detail present in the final image. For example, since the K20D already has significant chroma noise at ISO1600, the OOC files will show signs of detail loss due to NR very early on. Whereas the K-5 on the other hand has no notable chroma at this same sensitivity. Likewise... the K-5 SNR curve is far less abrupt as the sensitivities rise and so the effects become far more pronounced than with the K-5. And more importantly, where the K-5 to be far ahead of the K20D insofar as JPG and in camera NR processing also.

Having said all that... I've printed a great number of high ISO images from the K20D, and I can definitely see the differences between both systems even without heavy processing. But above all, I'd say it's the practical use of the higher sensitivities in the K-5 that would place it in the night and day category first and foremost against the K20D. For example... I've found the K-5 strives where the K2DD would struggle. Such as, in the skating rink or music recitals. At which point the K20D would always require a considerable amount of care and work in order to get to print. Whereas the K-5 does not. Which I think really stands-out as the biggest example of night and day between the two.

Hope this helps.
JohnBee
02-12-2012, 08:11 PM   #14
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
I notice you posted K20D and K-5 pictures, but no back to back comparisons? Why not back to back K20D and K-5 photos, both ISO 1600, straight from the camera, same scene, same lens, same exposure? I still suspect the OP is simply doing too much pixel peeping and will see that at "real" sizes, there's really nothing worth worrying about. I've seen K20D and K-5 pictures taken back to back and see virtually no difference at screen size up to ISO 1600, and barely noticeable at 3200.
02-12-2012, 10:49 PM   #15
Pentaxian
JohnBee's Avatar

Join Date: Sep 2008
Location: front of computer
Posts: 4,496
QuoteOriginally posted by Marc Sabatella Quote
Why not back to back K20D and K-5 photos, both ISO 1600, straight from the camera, same scene, same lens, same exposure?
That's probably because I'm too lazy.

QuoteQuote:
I still suspect the OP is simply doing too much pixel peeping and will see that at "real" sizes, there's really nothing worth worrying about.
I'm not sure if this is a wise approach given that we haven't encompassed such aspects as cropping and the various other aspects which come into play in shooting; ie. pushing contrast, saturation and sharpness etc. which all work to amplify noise, banding, detail loss etc.

QuoteQuote:
I've seen K20D and K-5 pictures taken back to back and see virtually no difference at screen size up to ISO 1600, and barely noticeable at 3200.
I don't think I've ever printed an image at screen size myself(personally). However... I still think were left short in terms of processing both in camera and off(SEE: cropping, IC adjustments etc). Though on the topic of low light, I think the performance terms change dramatically between both units. For example... the K20D can and does exhibit banding as early as ISO1600 in low light. And so I think there's a great deal of overhead between the K20D and K-5 when we get right down to it.

Taking the following scene for example:

META: K20D, 1/20s, ISO1600

click for larger image

META: K-5, 1/20s, ISO1600

click for larger image

As you can see, at this sensitivity, the K20D is already showing signs of banding, chroma noise, and color detail degradation, whereas the K-5 is still very good. Which is quite apparent even at 1/3 resolution. Having said that, if we enabled NR on both systems and pushed for detail, the K-5 would pull ahead by a significant margin given it's superior base IQ at this particular sensitivity.

I also think it's worth noting that since the OP announced that she was not interested in any type of external post-processing, then I believe the K-5 would present itself as a significant upgrade in terms of low light shooting against the K20D in this particular case.

Hope this helps.
JohnBee

Last edited by JohnBee; 02-13-2012 at 08:22 AM.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
camera, dslr, iso, k20d, photography
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
K20d - Night photography and high ISO. emtor Pentax DSLR Discussion 10 03-11-2011 01:43 PM
K20D blueish high iso noise nopps Pentax DSLR Discussion 5 12-20-2010 03:02 PM
m4/3 high ISO better than K20d ?? arbib Photographic Technique 12 09-10-2010 06:30 PM
K20D vs K-x : high ISO & processing skb0 Pentax DSLR Discussion 8 07-16-2010 08:36 AM
K-7 high ISO vs K20D high ISO supa007 Pentax DSLR Discussion 72 05-10-2010 04:24 PM



All times are GMT -7. The time now is 01:59 PM. | See also: NikonForums.com, part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top