Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
09-19-2012, 08:32 AM   #31
Pentaxian
JinDesu's Avatar

Join Date: Jun 2011
Location: New York City
Photos: Gallery
Posts: 5,624
Well lit maybe, but you can look around my files where the really dark shadows are (under the cams, where it's black anyways). You can see the fine grain, unlike your images which are splotchy. I can't imagine shoot underexposed and bringing it back will introduce splotchyness to that grain. I don't know - maybe the model I was using had all NR turned off somehow.. but it would be weird.

Heck: http://farm9.staticflickr.com/8283/7585831518_364cc35e10_o.jpg <- ISO 25,600, mildly underexposed (and horribly color balanced, sue me)

Noise grains are heavily evident, I don't really see much noise reduction. Definitely nothing that looks like your pictures.

09-19-2012, 08:54 AM   #32
Junior Member




Join Date: May 2011
Location: Melbourne
Posts: 44
Original Poster
Indeed the grains are there....interesting. I have no idea what's going on...Was this a production unit you used? Maybe they introduced the NR later somehow...here's a 25600 sample from my camera.....not much grain here...I mean you can see clearly NR has been apllied...

09-19-2012, 09:00 AM   #33
Pentaxian
JinDesu's Avatar

Join Date: Jun 2011
Location: New York City
Photos: Gallery
Posts: 5,624
Am away from the computer for now, so I can't view your images in full size. Maybe send me one of your RAW files, and I'll take a look when I am home at night?

I used a floor model at B&H, which usually ends up sold at a cheap price later on, so I doubt it's a "special" model. I can't say that with 100% certainty, so it'd be great if other k-30 users weigh in. I used Lightroom 3 to import the RAW files and output with no editing. Let me see how your RAW files look in my Lightroom, I can send you my RAWs when I am home
09-19-2012, 09:50 AM   #34
Junior Member




Join Date: May 2011
Location: Melbourne
Posts: 44
Original Poster
Here's another comparison ISO 6400. It's more obvious on the right side of the can.


K5


K30


09-19-2012, 10:45 AM   #35
Veteran Member




Join Date: May 2012
Posts: 698
QuoteOriginally posted by MBT74 Quote
The K30 has menu options to set the high ISO NR and Low Shutter Speed NR. Is this a case of these settings being applied to the raw files as well as the JPEGs? I had a quick play with mine with the camera set to RAW only and it did seem to affect the noise visible in the images displayed on the camera screen I think. Didn't have time to look at the files in Lightroom though.

As far as I know the "Low Shutter Speed NR" is a dark frame subtraction rather than actual noise reduction as usually understood. This subtraction has to be applied to the RAW. (Even if you are recording JPEG the subtraction would still have to be performed on the temporary raw before converting it to jpeg.)

What dark frame subtraction does is to take a second picture with the exact same ISO and same exposure but keeping the shutter closed. This second image would be entirely dark if noise did not exist, but since it does it will give an image representing something close to the 'fixed' noise of the sensor, that is the component of noise that does not change much between one shot and another and is due mainly to the slight inconsistency in the performance between individual pixels (mostly the 'dark current').

Dark frame subtraction has been done in the past by the photographer manually shooting that second picture with the lens cap on and then doing the subtraction in post processing, but one only did that because the camera couldn't do it on board. (It is a technique that astrophotographers are very familiar with). Unlike most other post processing procedures there is no advantage in doing dark frame subtraction in post processing rather than in camera since unlike the complex algorithms used for noise reduction it is just a straightforward pixel by pixel arithmetic subtraction, a very simple process that cannot be improved upon.

All this of course hinges on whether my interpretation of "Low Shutter Speed NR" is correct. I am sure that it does the dark frame subtraction because the camera spend a long time apparently doing nothing after each exposure. For instance a 30 second exposure will be followed by 30 seconds of apparent inaction. In reality the camera is 'exposing' the dark frame during that time. The actual subtraction is probably done in a small fraction of a second. The bit I am not sure about is whether or not it is also doing software noise reduction together with the dark frame subtraction.
09-19-2012, 01:51 PM   #36
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,585
QuoteOriginally posted by Nicks Quote
tried both Pentax's Digital Camera Utility and Lightroom 2 with pretty much the same results. ...If you have a darker picture it's pretty much ruined as far as I'm concerned. ...So they chose to ruin the capabilities of a perfectly good camera...
You are going overboard, don't you think?

For one, PDCU and Lightroom 2 (!!!) may not be the best tools to use. And you are drawing conclusions from your flawed testing about the K-30's RAWs that are simply not consistent with the data from - for example - DPR's camera testing.



I don't see any 'heavy NR' or 'smearing' there in the RAW's at all. Just fine grain and some chroma noise etc.

Technically the NR in the RAW may be measurable, as DxO report, but it doesn't appear to be visually significant compared to cameras with similar sensors.
09-19-2012, 02:19 PM   #37
Pentaxian
JinDesu's Avatar

Join Date: Jun 2011
Location: New York City
Photos: Gallery
Posts: 5,624
Well, I believe the NR is usually measured by how different the SNR % is from ISO 800 1600, etc. If there is no NR, the % curve is linear.

Again - please post a RAW at ISO 3200 or 6400 for us to see.
09-19-2012, 07:19 PM   #38
Junior Member




Join Date: May 2011
Location: Melbourne
Posts: 44
Original Poster
QuoteOriginally posted by rawr Quote
I don't see any 'heavy NR' or 'smearing' there in the RAW's at all. Just fine grain and some chroma noise etc.
Really? You really can't see it there??? I don't see any grain there. All I see is just a soup of colors....And I use the same software for both K-5 and K30 so I don't think the software can consistently make one camera look better than the other one although I may be wrong....


QuoteOriginally posted by JinDesu Quote
Again - please post a RAW at ISO 3200 or 6400 for us to see.
JinDesu I couldn't find any online service to accept my raw files.....

I'm not trying to convince anyone about this problem. Again, I am comparing K-5's ISO with K30's because I have both cameras and, to my eyes, K-5 looks much better. Of course this is subjective and other people can see thing differently. In rawr's post I see clearly the smoothing, he is seeing fine grain so....I don't want this thread to become an argument between us so I'll just stop here because we're going nowhere.

09-19-2012, 08:04 PM   #39
Junior Member




Join Date: Aug 2010
Posts: 44
Nick i also have notice what you are saying... See here https://www.pentaxforums.com/forums/pentax-k-30/197055-k-30-studio-samples-just-posted-2.html. I really appreciate your time to make this pics, you just confirmed what i saw on k30 raws..
09-19-2012, 08:40 PM   #40
Pentaxian
JinDesu's Avatar

Join Date: Jun 2011
Location: New York City
Photos: Gallery
Posts: 5,624
Nicks, I'm not questioning if you are imagining the problem or not. I can clearly see that in your picture. However, it doesn't match the output of the k-30 I used at B&H. Maybe the one I used just didn't have NR on or something..

I was suspecting that your RAW programs had some odd NR applied to the k-30 files, maybe because they didn't have the right k-30 profile. That's why I asked you to send me your raws (and I can give you my e-mail if you want) so I could compare.
09-19-2012, 08:58 PM   #41
Pentaxian




Join Date: Jun 2009
Location: GMT +10
Photos: Albums
Posts: 10,585
QuoteOriginally posted by Nicks Quote
You really can't see it there??? I don't see any grain there. All I see is just a soup of colors..
Perhaps we are facing a barrier here that relates to what you understand as 'noise', and what 'noise' should look like?

That 'soup of colors' is noise. The smaller the colour dots, the finer the noise 'grain'.

This article may be useful to see what different types of noise look like:

http://www.cambridgeincolour.com/tutorials/image-noise-2.htm

Last edited by rawr; 09-19-2012 at 09:09 PM.
09-19-2012, 09:18 PM   #42
Junior Member




Join Date: May 2011
Location: Melbourne
Posts: 44
Original Poster
The "fine" grain you are seeing is the smoothed luminance noise. That's the reason they are applying NR so that the usual, "healthy" grain looks like "fine grain" . Yes the chroma noise is more visible but not the luminance noise which makes me believe the NR is targeting luminance noise. Unfortunately killing the luminance noise kills the detail as well if overdone, which seems to be the case for K30.
09-19-2012, 10:27 PM   #43
Site Supporter




Join Date: May 2012
Location: Adelaide
Photos: Gallery | Albums
Posts: 1,081
Has the question been asked as to whether the K-5 and K-30 are metering the scene identically? If the K-30 was programmed to err a little more to darker images (for example to better protect highlights) then the shadows are darker because they have received less light (which is equivalent to the shadows being shot at a higher ISO) so a higher level of noise to signal would be expected, all other things equal. As K-30 is 12 bit depth vs K5 14 bit, it may be that the exposure metering decisions made by the firmware developers may also be a little different.
09-19-2012, 11:13 PM   #44
Junior Member




Join Date: May 2011
Location: Melbourne
Posts: 44
Original Poster
QuoteOriginally posted by southlander Quote
Has the question been asked as to whether the K-5 and K-30 are metering the scene identically? If the K-30 was programmed to err a little more to darker images (for example to better protect highlights) then the shadows are darker because they have received less light (which is equivalent to the shadows being shot at a higher ISO) so a higher level of noise to signal would be expected, all other things equal. As K-30 is 12 bit depth vs K5 14 bit, it may be that the exposure metering decisions made by the firmware developers may also be a little different.
QuoteOriginally posted by JonPB Quote
The K-5 has 14-bit DNGs while the K-30 has 12-bit DNGs. The complaints here have been about shadow detail, which is where the K-5 seems to put those extra two bits. My guess is that we're seeing the same approach to processing between the cameras, but the K-30--in order to cut costs--simply uses simpler analog to digital converters, which shows up in the shadow detail. It isn't noise reduction per se, but rather a lesser amount of data to work from, which might appear to result in a lower signal to noise ratio.
The 2 bit difference has been brought up by JonPB. The shadow detail and excellent DR seems to pull K-5 in front which is a shame considering K30 is the much newer technology...Also, I don't understand why a pentax engineer would think something like this: OK, so our new camera is not great when it comes to shadow detail, let's completely kill that already low detail by applying some NR to it.....

Last edited by Nicks; 09-20-2012 at 12:56 AM.
09-20-2012, 12:31 AM   #45
Forum Member
Richdog's Avatar

Join Date: Dec 2006
Location: Zurich
Posts: 54
QuoteOriginally posted by Nicks Quote
You can set your camera as you like, Richdog. I'm happy with my settings right now but the default settings are probably made to appeal to the entry level users. Very rich, unnatural colors, high sharpness that will give you a lot of artifacts and high contrast that will kill the DR. You can tune down all these setting so nothing to worry about. I definitely recommend the camera unless you like shooting in low light/high ISO situations in which case I would recommend waiting for the upcoming K-5II. K30 is really not too bad in the low light department but it is worse compared to K-5. BUT don't forget that K-5 is the best APS-C camera when it comes to high ISO resolution!
But the K5 II is double the price and in a different class... the point is we shouldn't have to suffer bad NR techniques on a new camera that uses the same sensor as a 2 year old camera. It seem like wilful deception, mild sabotage even, from the Pentax engineers. Why would they do that when they know it affects IQ on RAW files of all things?
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!
bit, iso, k-30, k-5, k-50, k30, nr, pentax, pentax k30, pentax k50
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
Shadow correction applied to raw files? outsider Pentax K-5 21 10-18-2013 11:54 PM
Microsoft releases RAW Codec with support for Pentax PEF raw files NKK Digital Processing, Software, and Printing 16 08-27-2012 06:00 AM
Here are some K30 raw files to play with. Good for 7days only Riv Pentax K-30 & K-50 2 07-28-2012 12:07 PM
Pentax RAW files Kendrick Pereira Troubleshooting and Beginner Help 16 07-27-2012 10:56 PM
Does the Photoshop RAW converter work with Pentax RAW Files? mconder Pentax DSLR Discussion 3 05-15-2008 12:14 PM



All times are GMT -7. The time now is 04:23 AM. | 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