Go Back   Steve's Digicams Forums > Digital SLR and Interchangeable Lens Cameras > Pentax / Samsung dSLR, K Mount Mirrorless

Reply
 
Thread Tools Search this Thread
Old Feb 16, 2008, 9:08 PM   #1
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,888
Default

iso 800













iso 1600






You can pixel peep by clicking original size of the images above. Both image with no NR and no flash of course

I can only speculate that to deal with noise, it is better to have more on the right side (histogram)+ iso 1600 than iso800 with the histogram more on the left side.

Daniel
danielchtong is offline   Reply With Quote
Sponsored Links
Old Feb 17, 2008, 3:08 PM   #2
Senior Member
 
thkn777's Avatar
 
Join Date: Jun 2004
Location: Germany
Posts: 1,831
Default

Hm,
I noticed that some concert shots in similar situations came out surprisingly well with ISO1600 - ok. But for a closer look on that behaviour you can't use just these two pictures...

Do some tests at the same subject, with the same setup. Shoot RAW. And include some over- and underexposed shots in the series. It IS a difference if you step up/down while taking the photo and then correct in the RAW converter. So this maybe already an explanation or at least a start to research further: in the concert ISO1600 shot you have very high dynamic range and nothing in the histogram middle, where things look completely different in #1

Since I do a lot of low light shots I'd be interested in your findings.


thkn777 is offline   Reply With Quote
Old Feb 17, 2008, 4:39 PM   #3
Senior Member
 
snostorm's Avatar
 
Join Date: Jan 2004
Location: Chicago Suburb, IL, USA
Posts: 2,770
Default

Hi Daniel,

A good illustration of why we should not jump to conclusions about IQ from any single shot, and that the "rules" have exceptions that aren't often considered by many. "Higher ISO = increased noise" is one example of this, and many seem to take this "rule" at face value.

What is generally forgotten is the condition -- "given equally accurate exposure" -- with "exposing to the right" as a good technique to get the best result from a possibly less than optimal situation.

I hope to see the "greater pixel density = more noise" "rule" challenged by the new Samsung sensor. I don't know how many times I've read "you can't go against the rules of physics", but the mitigating factor here is -- "given the same level of technology", and from what I've seen, we might get the high ISO performance that we enjoyed with the 6MP Pentaxes from the new 14MP K20D sensors -- and we'll apparently be given the choice of applying different degrees of in-camera NR or not, so we'll be able to choose between noise and detail.

Scott
snostorm is offline   Reply With Quote
Old Feb 17, 2008, 4:50 PM   #4
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,888
Default

thkn777 wrote:
Quote:
Hm,
I noticed that some concert shots in similar situations came out surprisingly well with ISO1600 - ok. But for a closer look on that behaviour you can't use just these two pictures...
Do some tests at the same subject, with the same setup. Shoot RAW. And include some over- and underexposed shots in the series. It IS a difference if you step up/down while taking the photo and then correct in the RAW converter. So this maybe already an explanation or at least a start to research further: in the concert ISO1600 shot you have very high dynamic range and nothing in the histogram middle, where things look completely different in #1
Since I do a lot of low light shots I'd be interested in your findings.*
Hmm I just have more questions
The subject is intriguing.

These are the unknown:

1. While shooting, the histogram of the first image (iso800) was more promising. There was no way that I would have set the iso to 1600 on the spot.

2. What cause the 2nd one to have lower noise level – aperture or shutter speed? That is unheard of. By accident, there were two subjects in the 2nd pict. I had to step down to get bigger depth of field F5.6 instead of F4 which necessitated the use of iso 1600. Well I did that by accident. Also I did not have SR in DL. So I rested my arm on the seatback in front of me and well I did that in 1/45s (for an FA135mm!)



Now these are known facts:
1. We all agree noise is a huge deal for indoor low light shots
2. What started all these thought came from me reading about the noise level of the new Nikon D300 even at 6400 - supposedly to be very good as proved time and time again. But there are D300 shooters reporting that there is quite a lot of noise in the shadow (like the BG of my first image) at speed below 1/3sec even at ISO 200.

This reinforces my suspicion that Nikon did a great job in tweaking the firmware or software to lower the noise level of the image (with the same Sony sensor - same size as K100D & K10D physically). Hope Pentax does not take too long to replicate.

Any comments? If we all know how to tweak this (factors like shutter speed, iso & aperture) we all could be better off.

So we do not know what to choose. We will just shoot like crazy. Take more shots and pray that we have more keepers than otherwise.

Daniel






p.s.

The first one
http://www.pbase.com/danieltong/image/70731924/original

Focal Length 135 mm
Exposure Time 1/60 sec
Aperture f/4
ISO Equivalent 800





The 2nd one
http://www.pbase.com/danieltong/image/71445971/original

Focal Length 135 mm
Exposure Time 1/45 sec
Aperture f/6.7
ISO Equivalent 1600
danielchtong is offline   Reply With Quote
Old Feb 17, 2008, 5:00 PM   #5
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,888
Default

snostorm wrote:
Quote:
Hi Daniel,

A good illustration of why we should not jump to conclusions about IQ from any single shot, and that the "rules" have exceptions that aren't often considered by many. "Higher ISO = increased noise" is one example of this, and many seem to take this "rule" at face value.

Scott
Scott

Not just one.
I do have a whole pile of shots in the two occasions (as shown above) to confirm that the iso/noise performance.

"Higher ISO = increased noise" is correct IF all factors remain the same. Nobody can dispute that . I think we have the consensus?? But since when shutter speed and aperture have anything to do with noise? And what are their co-relation?

Meanwhile we rely on luck in taking more pict in the hope of getting more keepers?

Daniel
danielchtong is offline   Reply With Quote
Old Feb 17, 2008, 5:12 PM   #6
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,888
Default

This Nikon D300 fellow reported






Well, your observation is interesting to say the least.

I took this image at ISO 200, 1/1.3sec at F16. I am seeing noise in the dark area, more than I expect.


Arthur (in another forum in Toronto)
danielchtong is offline   Reply With Quote
Old Feb 17, 2008, 6:01 PM   #7
Senior Member
 
Monza76's Avatar
 
Join Date: Jul 2004
Posts: 4,093
Default

"Visible light" corresponds to a wavelength range of 0.4 - 0.7 micrometers. Digicams are now down to pixel sizes of less than 2 micrometers, this is getting very close to the physical limit since pixel size is only slightly greater than wavelength. The 6MP Pentax models have a pixel size of about 8 micrometers, considerably more than the accepted optimum of >6.5 micrometers (a figure which will no doubt change with technology). The K10D was between 6 and 6.5 thus it was not quite as sensitive as the 6MP models. The K20D will be less than 6 micrometers but will benefit from improved software and, as Scott pointed out above, it will give you the option of less noise or more detail by allowing adjustment of the NR program.

As a film photographer I accept noise as long as it is grain like, personally I think we sometimes overlook the fact that noise can actually add to the "feel" of an image (which is probably why Adobe allows us to "add noise" in their editing programs).

Ira
Monza76 is offline   Reply With Quote
Old Feb 17, 2008, 6:47 PM   #8
Senior Member
 
VTphotog's Avatar
 
Join Date: Mar 2005
Location: Extreme Northeastern Vermont, USA
Posts: 4,214
Default

danielchtong wrote:
Quote:
Hmm I just have more questions
The subject is intriguing.

These are the unknown:

1. While shooting, the histogram of the first image (iso800) was more promising. There was no way that I would have set the iso to 1600 on the spot.

2. What cause the 2nd one to have lower noise level – aperture or shutter speed? That is unheard of. By accident, there were two subjects in the 2nd pict. I had to step down to get bigger depth of field F5.6 instead of F4 which necessitated the use of iso 1600. Well I did that by accident. Also I did not have SR in DL. So I rested my arm on the seatback in front of me and well I did that in 1/45s (for an FA135mm!)
Quote:
In this case, the histogram tells the story. You have much more information in the lower third of your histogram in the first picture, which equates to more dark detail. Along with the detail, you get noise. The higher contrast in the second shot forces much of the dark detail and noise below the visible threshold. (it's there, but you can't see it) The details of interest are in the upper part of the histogram where the noise is least.
Quote:
Any comments? If we all know how to tweak this (factors like shutter speed, iso & aperture) we all could be better off.
Quote:
With your permission, I will attempt to adjust the first picture's lighting and contrast (without doing noise reduction) to see if I can get some improvement.
Quote:
brian
VTphotog is offline   Reply With Quote
Old Feb 17, 2008, 7:51 PM   #9
Senior Member
 
danielchtong's Avatar
 
Join Date: Dec 2005
Posts: 2,888
Default

VTphotog wrote:
Quote:
With your permission, I will attempt to adjust the first picture's lighting and contrast (without doing noise reduction) to see if I can get some improvement.
brian

Brian

Go with the original
http://www.pbase.com/danieltong/image/70731924/original
I know very little in tweaking lighting and contrast to improve thing
But if I am correct, not much can be done with so much noise there already

Daniel
danielchtong is offline   Reply With Quote
Old Feb 18, 2008, 2:35 AM   #10
Senior Member
 
bilybianca's Avatar
 
Join Date: Nov 2005
Location: Hassleholm, Sweden
Posts: 3,435
Default

danielchtong wrote:
Quote:
But since when shutter speed and aperture have anything to do with noise? And what are their co-relation?
I have observed that longer exposure causes more noise. I think I have read somewhere that it has to do with the pixels getting heated during exposure. If this is right, you're balancing on the point where lower ISO stands against shorter shutter time.

Kjell
bilybianca is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 8:28 AM.