Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
01-23-2009, 05:07 PM   #1
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
K200D + DA40 at the equivalent of ISO 6400

The picture I am posting below did not strike me as especially remarkable, until I started thinking about all the ways in which conventional wisdom would have suggested it should not have come out *nearly* as well as it did. It is a very quick snap shot taken in very low light, *two stops underexposed* on the K200D with my DA40 at f/2.8, 1/20", ISO 1600 and then pushed two stops in PP. This is the equivalent of shooting at ISO 6400.

Of course, everyone "knows" Pentax AF is not quick enough to "capture the moment" in low light. And, if not for SR, 1/20" would be too slow for handheld shots at 40mm, especially when shooting from the hip as I was. But more importantly, consider:

- The K200D uses the Sony 10MP sensor, not exactly famous for its high ISO performance.

- Unlike the K-m/K2000, the K200D does not feature ISO 3200, further limiting its apparent usefulness in low light.

- Shooting underexposed exacerbates noise.

- The DA40 is limited to f/2.8, which many consider insufficient for low light candid photography because it forces you into slower shutter speeds, higher ISO, and/or underexposure as compared with "faster" lenses.

And yet, with a two-step push and just a *little* NR in ACDSee Pro:



I have seen enough ISO 6400 samples from the K20D to suggest that this is not a ton worse. Of course, this isn't great either, but really, perfectly usable for the web or a 4x6 print.

I think most people would agree the image would have been better captured by a K20D and FA35, so it could have been shot at f/2 and ISO 3200 and not needed pushing. Or maybe even shot at higher ISO with a faster shutter speed. And I would be honestly curious to see some samples in these kinds of conditions with these kinds of settings for comparison. But as I said, I've seen enough samples to not be expecting to be blown away by the difference the FA35 or K20D (or your favorite Canikon equivalents) could have made. I know I've tried and failed many times to do as well with my DS and A50/1.7.

No fair posting shots taken in better light or processed with heavy NR from high end noise reduction packages like Neat Image, BTW...

01-23-2009, 08:20 PM   #2
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
QuoteOriginally posted by Marc Sabatella Quote
I have seen enough ISO 6400 samples from the K20D to suggest that this is not a ton worse. Of course, this isn't great either, but really, perfectly usable for the web or a 4x6 print.
This thread may be of interest to you. LOTS of k20d high-iso shots:
https://www.pentaxforums.com/forums/pentax-dslr-discussion/27927-pentax-k20d-...o-gallery.html

K20d in low light at ISO 6400, no PP other than raw to jpg conversion, no noise reduction. k20d is good at high ISO...


Last edited by PentaxPoke; 01-23-2009 at 08:29 PM.
01-23-2009, 08:33 PM   #3
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,642
Sorry Marc, that noise level does not look acceptable to me.

I've come to the conclusion that f/2.8 is too slow, so I'm going to buy an FA 35.

Last edited by audiobomber; 01-24-2009 at 06:01 AM.
01-23-2009, 08:35 PM   #4
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,642
QuoteOriginally posted by PentaxPoke Quote
K20d in low light at ISO 6400, no PP other than raw to jpg conversion, no noise reduction. k20d is good at high ISO...
That's amazing to me, even though I have a K20D. What's your secret?

01-23-2009, 08:35 PM   #5
Veteran Member




Join Date: Jun 2008
Location: Borås, Sweden
Photos: Gallery
Posts: 3,165
The noise in the first picture is really disturbing and far more than I would personally accept. (Frankly, it looks like it was choked down to about a 4 bit range.)

Here's an ISO 5000 shot from the K20D.

01-23-2009, 09:35 PM   #6
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
QuoteOriginally posted by audiobomber Quote
That's amazing to me, even though I have a K20D. What's your secret?
My secret? I shoot PENTAX !

Thanks though. There's really no secret other than exposing to the right and using a k20d.

Nice example Ping. You should add that to the k20d High ISO thread!
01-23-2009, 09:37 PM   #7
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,642
QuoteOriginally posted by PentaxPoke Quote
There's really no secret other than exposing to the right and using a k20d.
So you use some positive EV? How much, +.3?
01-23-2009, 09:42 PM   #8
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
QuoteOriginally posted by audiobomber Quote
So you use some positive EV? How much, +.3?
It kind of depends what I meter on. I would say that on average, I bump up the EV half a stop. Whatever it takes to get the histogram to the right without blowing highlights.

01-23-2009, 11:08 PM   #9
Veteran Member
creampuff's Avatar

Join Date: May 2007
Location: Singapore
Photos: Gallery
Posts: 3,955
Was it ever a contest?
K20D wins easily. The sensor is that good...
01-24-2009, 09:46 AM   #10
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
Original Poster
OK, you've definitely convinced me the K20D is capable of far better. Although I'd note none of the posted samples - including the ones I saw in the thread PentaxPoke mentioned - are anything like the same sort of lighting. On the other hand, I've also seen samples from the K20D at ISO 6400 even in pretty decent lighting that looked pretty much exactly like my image. So while I am of course impressed with what you've shown me, I still wonder what the K20D would have done in my situation.

BTW, I personally don't mind noise so much when eliminating it messes with detail. But FWIW, here's what Neat Image has to say about my shot:

01-24-2009, 10:38 AM   #11
Veteran Member
PentaxPoke's Avatar

Join Date: Jan 2008
Location: Oklahoma
Posts: 1,411
QuoteOriginally posted by Marc Sabatella Quote
OK, you've definitely convinced me the K20D is capable of far better. Although I'd note none of the posted samples - including the ones I saw in the thread PentaxPoke mentioned - are anything like the same sort of lighting. On the other hand, I've also seen samples from the K20D at ISO 6400 even in pretty decent lighting that looked pretty much exactly like my image. So while I am of course impressed with what you've shown me, I still wonder what the K20D would have done in my situation.
Marc,

OK, if you don't think any of those pictures were in dark enough conditions, how about this one? The room was so dark that I could not see the cottage on the table with my own eyes. I took this quite a while ago to show the high ISO capabilities of the k20d at retaining fine detail. I had to focus with the lights on, turn them off, and then take the picture. That "blast" of light you see in the lower left is a dim little night light in the kitchen. No PP, no NR at all of this picture. This was ISO 6400 with a 4 second exposure. Your photo had about 80 times the shutter speed. k20d is good in low light...

Last edited by PentaxPoke; 01-24-2009 at 11:26 AM.
01-24-2009, 02:56 PM   #12
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
Original Poster
QuoteOriginally posted by PentaxPoke Quote
OK, if you don't think any of those pictures were in dark enough conditions, how about this one? The room was so dark that I could not see the cottage on the table with my own eyes.
My concern wasn't just with the *level* of light, but also with the *nature* of the lighting in the scene. My scene had considerably higher dynamic range and the light was more strongly colored than what you are dealing with here. In my experience, both factors contribute heavily to the appearance of noise.

The relative lack of light and shadow in particular strikes me first about the shot you just posted. It's actually quite noisy - closer, IMHO, to my first posted image than to the other K20D samples posted here. But you have to examine the images more carefully to see this, in large part because the shot is more flat.

Strongly colored light affects the appearance of noise because you are really dealing with only one channel that has any significant amount information. At least, that's my assumption - one I may try testing at some point.

This particular shot is also not a apples-to-apples comparison since the K20D would have used dark frame subtraction on a shutter speed that slow. Actually, according to GordonBGood (who is pretty expert on these matters, having written his own RAW conversion routines to examine output from a variety of different cameras), the K20D applies some NR on *all* shots at ISO 1600 and above - yes, even for RAW, and even if you have high ISO NR turned off.

So really, it's kind of unfair to ever compare ISO 6400 form the K20D against the equivalent from the K200D unless you apply similar NR to the K200D image. Unfortunately, I have no idea how aggressive the K20D's NR is in ISO 6400 RAW images and have no way of reproducing it. But with that in mind, maybe the more apt comparison would be to my processed image, and then see to what extent the K20D wins on detail.

Anyhow, at least with regard to the other images posted, I don't deny they look a heck of a lot better than my initial posted image in terms of noise. But I also know that the K20D images posted here are not necessarily typical - they were specially chosen to demonstrate the advantages of the K20D. The sample pictures posted in various reviews I've seen tell a somewhat different story. And indeed, one such review did a controlled test of the K20D versus K10D at ISO 3200 (equivalent via push processing on the 10D, of course) and also concluded there was little discernible difference demonstrated by their test.

It is these sort of images I had seen elsewhere that gave me the impression the K20D's ISO 6400 was not particularly impressive. But some of images posted here (including on the thread you referenced) are obviously more so. Clearly, there are factors that affect the appearance of noise - and the K20D is not immune to these factors. I'm trying to get a handle on what these factors are.

QuoteQuote:
k20d is good in low light...
I am not questioning this. But I am trying to reconcile the differences between images like yours of the lens on the focus chart with ones like these, from an examination of the K20D by Dale Cotton:





Those are probably the worst I've seen, so I'm not claiming those are typical either. But for anyone wanting to read more on the subject of comparing K200D and K20D with regards to high ISO noise, see Gordon's contributions to this thread:

k20d vs k200d [Page 1]: News Discussion Forum: Digital Photography Review

The subject is not as black and white as it is being made out to be.

Last edited by Marc Sabatella; 01-24-2009 at 06:42 PM.
01-24-2009, 03:09 PM   #13
Pentaxian
Marc Sabatella's Avatar

Join Date: Sep 2006
Location: Denver, CO
Photos: Gallery
Posts: 10,686
Original Poster
QuoteOriginally posted by audiobomber Quote
I've come to the conclusion that f/2.8 is too slow, so I'm going to buy an FA 35.
That's a perfectly reasonable decision, and I'm probably going to get one for my wife at some point, too. But I should point out that the sort of light I was shooting in isn't typical *at all*. I've shot probably thousands of similar pictures at f/2.8 with ISO 800, 1100, or 1600, with shutter speeds of 1/30" or faster, and am usually within a half stop of "correct" exposure. It's extremely rare to have to shoot at 1/20" and then still have to push exposure two full stops. I didn't post the picture I did because it was the best sample I could come up with out of hundreds of attempts, but rather, because it is the first time in a *long* time I can recall having to underexpose so badly, and I was surprised to get anything at all.

Which is to say, I don't normally feel very limited by f/2.8 (or by the K200D). But of course, f/2 or better is nice to have when you *do* need it, if you can live with the shallower DOF.
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, da40, dslr, image, iso, k200d, k20d, light, photography, samples, shot
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
K-5 vs K-7 First ISO performance testing (ISO-6400) starscream Pentax News and Rumors 95 09-25-2010 07:02 PM
K-5 ISO 6400 developed sirluke Pentax DSLR Discussion 24 09-23-2010 10:57 AM
People Iso 6400 jct us101 Post Your Photos! 1 06-27-2010 06:03 PM
K7 + ISO 6400 cpk Pentax DSLR Discussion 1 02-26-2010 07:46 PM
Iso 6400 on the Kx is a Homerun Igilligan Pentax DSLR Discussion 23 01-02-2010 03:52 PM



All times are GMT -7. The time now is 06:19 PM. | See also: NikonForums.com, CanonForums.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