Pages: [1] 2   Go Down

Author Topic: Let's discuss some DR (Dynamic Range) samples  (Read 11198 times)

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Let's discuss some DR (Dynamic Range) samples
« on: November 15, 2014, 01:47:13 am »

Update -- samples added:
Here is the present set:
http://echophoto.dnsalias.net/ekr/Articles/DRArticle/DS4.html

Files, see below:
A 99 SLT 2.5s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC4758.ARWWindows not clipped
A 99 ALT 8s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC475.ARWWindows clipped
P45+ 7.5s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/CF045286.IIQWindows clipped
P45+ Lumariver HDR
multiple exposures
fused
http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/20140617_lumariver.dngHDR, Windows not clipped

Hi,

Lot of talk about Dynamic Range recently. In these discussion I often state that I seldom see DR as a real obstacle. I feel it may be good to discuss some real world images. I would say that these images may be one of those where DR does matter.

What we have here are two images shot at Lockenhaus castle in Austria. One is shot on Sony Alpha 99 SLT, probably using the same sensor as the Nikon D600 the other is on shot on a P45+ using a Kodak sensor. According to DxO Mark the SLT99 has about 1EV advantage in "print mode" (pixel count taken into account) and 1.4 EV in pixel peeping mode. The P45+ has higher exposure which doesn't allow recovery of outside landscape, while the Alpha 99 image allows for decent recovery. Well, these are real world shots, testing is better done in lab conditions….

The images are from a real world shoot and they are not really equivalent. Both are taken from a series intended for HDR rendition.

Anyway, original raw images are here: http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/

Here is my best conversion of the P45+ image:


And my best conversion of the Sony Alpha SLT image:


On the P45+ I did also a conversion in Capture One V7.3 using "HDR" sliders and no curve adjustments. My processing skills on Capture One are quite limited.


The links below go to my usual "full disclosure" indexes. The first one refers to the original DNG files (and does exclude the Capture One conversion as it was made from the IIQ file). The second one shows the JPEGs but does not refer to the raw files. These are limitations in my program generating those links.

http://echophoto.dnsalias.net/ekr/Articles/DRArticle/DS1.html
http://echophoto.dnsalias.net/ekr/Articles/DRArticle/DS2.html

This is the raw histogram from the P45:


And this one from the SLT 99:


Above the piano there is a rectangular dark area. I made a small selection in that area, selection histograms:

P45+:


SLT 99:


I hope this samples may be a starting point for a good discussion.

All photographic images:

Best regards
Erik
« Last Edit: November 16, 2014, 06:50:26 am by ErikKaffehr »
Logged
Erik Kaffehr
 

Abe R. Ration

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
    • Abe R. Ration's lens and camera blog
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #1 on: November 15, 2014, 02:55:42 pm »

By a happy coincidence I just took a couple of test frames on the subject of DR. I will someday, though not anytime soon, create a better test procedure, but these two are what I took today.

The first one way taken with Sony A7 and CV 50/1,5 Nokton.

The other one I took with old Olympus Zuiko 50/1,8 equipped with an old multi element fisheye adaptor (with optical flaws as well) to see how much reflections can influence at worst.

I had two papers with some patterns - one pattern with darkening background, the other with white background and brightening details. One paper was behind a solid plate and was under 2 lux illumination, while the other was under a halogen lamp with upto 3050 lux on the paper.

Plenty of curver etc. in both images to make the details as visible as possible.
Sorry that this is such a quick and dirty experiement.


Logged
Abe R. Ration
amateur photographer, amateur scientist, amaterur camera buff
http://aberration43mm.wordpress.com/

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4066
    • Photos of Arkansas
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #2 on: November 15, 2014, 04:09:23 pm »

Hello Eric:

I worked with both files, and to me, the P45+ is the better file overall.  The only problem with the P45+ is as you noted, the light coming in the windows is too harsh to really capture much of information outside, whereas with the Sony, you can pull down the windows and still hold some image.

However.

1.  I thought that C1 worked much better on the overall image of the p45+, especially shadow recovery, both overall and on the roof.  The Sony file I added an adjustment layer to pull up the roof, I did not need this on the P45+
2.  I did a keystone adjustment to the Sony just for looks, I left the P45+ as is was.
3.  P45+ image had a good bit of CA around the windows, but that cleaned up in C1, Sony image did not
4.  Sony image loaded too red to me especially on the bricks, but no knowing what the interior really looked like, I sided with the P45+ as it gave a much better tone to the bricks
5.  P45+ easily wins on details, I did add a bit of structure to the P45+ as the image seemed just a bit dulled down, I saw it was 7 seconds.  When you look at the seating in the background, wood details stand out much better in the P45+ file

Both files are clean, and I assume both were at base iso, I saw the P45+ was at 50.  It would be very interesting to try the same test, at a shorter shutter speed with higher iso, and see if the same holds true.  In this test, you have proven to me that the CCD can do excellent work, when given enough light and time.  

I have attached a side by side of both images in C1, I used Capture 1 vr8 as it's much better than 7 in so many areas.  

Edit:  I could not have gotten such a clean transition with my P45+, the roof area would have looked much worse and the seats.  After looking that this image the first one you posted from the dock scene, I am quite sure that Phase One made some changes to the P45+, in the later models, and or Kodak was able to change something in the chip fab (which I doubt).  I have been of the belief that most camera companies do this, (under cover improvements), and looking at your P45+ images really convinces me of that.  Keep that back as it's a winner for sure.

Sincerely
Paul

« Last Edit: November 15, 2014, 04:22:05 pm by Paul2660 »
Logged
Paul Caldwell
Little Rock, Arkansas U.S.
www.photosofarkansas.com

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #3 on: November 15, 2014, 05:39:00 pm »

I am looking at the images on an old laptop which I know has a very poor screen relative to my regular system. All the images seem to fit well in the capability of the screen. There is no banding. All I have to do is look away from the screen to see things much darker and much brighter in the room. The colors seem fairly limited in variety. The main thing changing is lighter or darker.

My point is, in terms of current output technology, our mind can, and does, make it seem realistic, when we know that putting the laptop down in the actual real scene will show instantly all the shortcomings.

I think we are at the analogy of looking at mp3 music, saying it looks realistic, when we have not seen CD quality music.
Logged

Ray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 10365
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #4 on: November 15, 2014, 06:56:15 pm »

The first point I would make here, Erik, is the importance of using equal ETTR exposures when comparing the dynamic range of two images.
In the examples you've shown, the P45+ image shows no detail through the small windows. That part of the scene through the windows is blown. If you were comparing 2 shots from the same camera, one would expect to see better shadow detail in the shot with the greater exposure, that had blown highlights.

The P45+ shot shows better shadow detail than the A99 as one would expect, because of that greater exposure. Attached is a 100% crop comparison, after interpolating the A99 shot to the same size. Notice the blotchy noise in the A99 image on the left, the one that has been exposed for the highlights.

The second point I would make is, this is not an ideal scene to demonstrate the importance of using a camera with a high DR. The details of the scene out of the small windows are not relevant to the composition. The windows are simply a light source. If the hall were lit by artificial lighting, you wouldn't expose for the electric light bulbs so that you could see the filaments inside the bulbs, would you?  ;)

What these shots of yours demonstrate is a principle of ETTR that is sometimes overlooked, that is, one should not hesitate to blow certain highlights in a scene if detail in such highlights is largely irrelevant to the general appeal and impact of the composition. In other words, one should not sacrifice image quality in the main body of the composition in order to preserve irrelevant detail in small areas of brightness.

Another example which would illustrate this point would be to expose for the small patches of sky visible through the branches and foliage of a tree when photographing a bird, or other creature, sitting in the shade of the tree. Which is more important, the bird or the small patches of sky? If the patches of sky were large patches with interesting cloud formations, that would be another matter.

Logged

JV

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1013
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #5 on: November 15, 2014, 09:01:54 pm »

one should not sacrifice image quality in the main body of the composition in order to preserve irrelevant detail in small areas of brightness.

+1.  Very well phrased.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #6 on: November 15, 2014, 09:25:13 pm »

Hi Paul,

Thanks for your processing suggestions. I understand that C1 V8 has introduced some "tone mapping" functionality, that was one of the functions I was missing in V7. I would suggest it is needed to bring out details without making for boring image.

Thanks for suggesting that I keep the P45+. I am thinking about selling the whole kit, so an alternative viewpoint is welcome.

These images were not intended as a test or comparison of the P45+ and the SLT 99, the intention was more to supply raw images for discussion.

One thing I observed is that roof reproduces well in both images, although photon count is in all probability quite low.

Best regards
Erik




Hello Eric:

I worked with both files, and to me, the P45+ is the better file overall.  The only problem with the P45+ is as you noted, the light coming in the windows is too harsh to really capture much of information outside, whereas with the Sony, you can pull down the windows and still hold some image.

However.

1.  I thought that C1 worked much better on the overall image of the p45+, especially shadow recovery, both overall and on the roof.  The Sony file I added an adjustment layer to pull up the roof, I did not need this on the P45+
2.  I did a keystone adjustment to the Sony just for looks, I left the P45+ as is was.
3.  P45+ image had a good bit of CA around the windows, but that cleaned up in C1, Sony image did not
4.  Sony image loaded too red to me especially on the bricks, but no knowing what the interior really looked like, I sided with the P45+ as it gave a much better tone to the bricks
5.  P45+ easily wins on details, I did add a bit of structure to the P45+ as the image seemed just a bit dulled down, I saw it was 7 seconds.  When you look at the seating in the background, wood details stand out much better in the P45+ file

Both files are clean, and I assume both were at base iso, I saw the P45+ was at 50.  It would be very interesting to try the same test, at a shorter shutter speed with higher iso, and see if the same holds true.  In this test, you have proven to me that the CCD can do excellent work, when given enough light and time.  

I have attached a side by side of both images in C1, I used Capture 1 vr8 as it's much better than 7 in so many areas.  

Edit:  I could not have gotten such a clean transition with my P45+, the roof area would have looked much worse and the seats.  After looking that this image the first one you posted from the dock scene, I am quite sure that Phase One made some changes to the P45+, in the later models, and or Kodak was able to change something in the chip fab (which I doubt).  I have been of the belief that most camera companies do this, (under cover improvements), and looking at your P45+ images really convinces me of that.  Keep that back as it's a winner for sure.

Sincerely
Paul


« Last Edit: November 15, 2014, 10:00:30 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #7 on: November 15, 2014, 09:51:53 pm »

Hi Ray,

I stated quite in the original post that this is not really a comparison or test. Exposures are not identical or similar. I have something like half a dozen exposures from each camera but they are not really comparable. Exposures were based on camera histograms. Intention was to make HDR, but I preferred the images from a single exposure.

Compositions/crops are different as the 24-70/2.8 zoom is wider than the 40 mm on the Hasselblad.

Regarding your observations I agree with what you say.

Best regards
Erik



The first point I would make here, Erik, is the importance of using equal ETTR exposures when comparing the dynamic range of two images.
In the examples you've shown, the P45+ image shows no detail through the small windows. That part of the scene through the windows is blown. If you were comparing 2 shots from the same camera, one would expect to see better shadow detail in the shot with the greater exposure, that had blown highlights.

The P45+ shot shows better shadow detail than the A99 as one would expect, because of that greater exposure. Attached is a 100% crop comparison, after interpolating the A99 shot to the same size. Notice the blotchy noise in the A99 image on the left, the one that has been exposed for the highlights.

The second point I would make is, this is not an ideal scene to demonstrate the importance of using a camera with a high DR. The details of the scene out of the small windows are not relevant to the composition. The windows are simply a light source. If the hall were lit by artificial lighting, you wouldn't expose for the electric light bulbs so that you could see the filaments inside the bulbs, would you?  ;)

What these shots of yours demonstrate is a principle of ETTR that is sometimes overlooked, that is, one should not hesitate to blow certain highlights in a scene if detail in such highlights is largely irrelevant to the general appeal and impact of the composition. In other words, one should not sacrifice image quality in the main body of the composition in order to preserve irrelevant detail in small areas of brightness.

Another example which would illustrate this point would be to expose for the small patches of sky visible through the branches and foliage of a tree when photographing a bird, or other creature, sitting in the shade of the tree. Which is more important, the bird or the small patches of sky? If the patches of sky were large patches with interesting cloud formations, that would be another matter.


Logged
Erik Kaffehr
 

deejjjaaaa

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1170
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #8 on: November 15, 2014, 10:15:12 pm »

I understand that C1 V8 has introduced some "tone mapping" functionality, that was one of the functions I was missing in V7. I would suggest it is needed to bring out details without making for boring image.
and also a new C1 OEM camera profile for your A99
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples LumaRiver added
« Reply #9 on: November 15, 2014, 10:39:02 pm »

Hi,

Anders Torger who often posts on thee forums has developed a program called Lumariver HDR. It does HDR fusion. A great feature of Lumariver HDR is that it can do HDR on undemosaiced raw files and output raw. Here is a a DNG file from Lumariver HDR:

http://echophoto.dnsalias.net/ekr/Articles/DRArticle/20140617_lumariver.dng






Anders Torger has also written code to export IIQ files.

Best regards
Erik
Logged
Erik Kaffehr
 

allegretto

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 660
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #10 on: November 15, 2014, 11:44:54 pm »

Erik,

I understand that you are coming to this in a technical and graphic way, so please don't take my impressions the wrong way...

This post, taken with your post on a previous thread about tech aspects you showed both cameras and concluded from the graphs that it could be "lens flare" causing what you see in the graph. I think I understand the graph but I'm not the Engineer you are...obviously

However I can see...

Whatever you see in the graphs, the +45 clearly takes MUCH better, MUCH richer files and colors than the Sony. It's not even close to my eye and I'll bet some others here too. The other post showed crystals that appeared to be sprinkled ON the photo with the +45 and a pale comparison in the Sony. Same with the rocks and stones. We see the same thing in these examples too

I'm not telling a person with your experience what to like or not... but at some point the image speaks for itself, no?
Logged

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #11 on: November 16, 2014, 01:23:20 am »

Erik,


I processed both files in C1P V7 and believe me when I say this, there is absolutely no comparison between the images. The P45+ image has better color separation, cleaner shadows, overall integrity and so on. Attached are two screenshots of areas where I think these are demonstrated well.

If you sell that back, you will be making a very big mistake. Instead of trying to equalize everything, please invest more time to improve your processing skills with respect to the Phase files. There's a lot more that can be done with them if you put in a bit more effort.
Logged
my portfolio: www.sandeepmurali.com

synn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1235
    • My fine art portfolio
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #12 on: November 16, 2014, 01:24:21 am »

Second screenshot:
Logged
my portfolio: www.sandeepmurali.com

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #13 on: November 16, 2014, 02:57:39 am »

Hi Synn,

Thanks for feedback, much appreciated.

I have actually warmed up quite a bit to my Hasselblad stuff in the last six months.

Dissatisfaction is not really the cause I am considering to sell of my Hasselblad equipment, it is more for practical reasons, something like:
  • Manual focusing is not easy, I lack magnified live view. This is one area experience helps a bit.
  • Having dual kit. I shoot from ultra-wide to telephoto, tripod + backpack is 18 kg, not getting younger
  • I see 50 MP around the corner, and some very good lenses (?).
  • If I get 50 MP gear, how much use will the Hasselblad see?
  • Right now, I don't feel I will go with technical cameras, that has been part of my original plan.

Thinks like that. Comments like yours and Paul's actually matter, definitively.

I have made some arrangements to buy another lens for the Hasselblad, if I find no buyer for the stuff.

I am sort of like this, I have considered going MFD several times, looking at both the Mamyia ZD and the Pentax 645D, but both cameras were late to arrive. I also considered ditching Sony for Nikon (part because of the lenses and partly because they were slow with live view).

Best regards
Erik

Erik,


I processed both files in C1P V7 and believe me when I say this, there is absolutely no comparison between the images. The P45+ image has better color separation, cleaner shadows, overall integrity and so on. Attached are two screenshots of areas where I think these are demonstrated well.

If you sell that back, you will be making a very big mistake. Instead of trying to equalize everything, please invest more time to improve your processing skills with respect to the Phase files. There's a lot more that can be done with them if you put in a bit more effort.
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #14 on: November 16, 2014, 03:21:27 am »

Hi,

Yes the reason I started this thread was DR, which I feel is overrated in many cases. I felt that these two images were usable to talk DR.

The reason I choose these is that there is a large contrast range, including detail on the outside, illumination is trough small windows and I don't think they cause a lot of lens flare.

A couple of issues I considered were:

  • Some people talk about photographic DR, something like SNR = 10.
  • I would suggest that photographic DR is essentially a function of exposure.
  • Readout noise affects the darkest part of the pictures. How relevant is that?

So I felt it was a good reason to post some real world samples. Much of the discussion used to be about DxO-mark, which I guess is quite adequate, but I don't know how relevant the data is.

Best regards
Erik


Erik,

I understand that you are coming to this in a technical and graphic way, so please don't take my impressions the wrong way...

This post, taken with your post on a previous thread about tech aspects you showed both cameras and concluded from the graphs that it could be "lens flare" causing what you see in the graph. I think I understand the graph but I'm not the Engineer you are...obviously

However I can see...

Whatever you see in the graphs, the +45 clearly takes MUCH better, MUCH richer files and colors than the Sony. It's not even close to my eye and I'll bet some others here too. The other post showed crystals that appeared to be sprinkled ON the photo with the +45 and a pale comparison in the Sony. Same with the rocks and stones. We see the same thing in these examples too

I'm not telling a person with your experience what to like or not... but at some point the image speaks for itself, no?
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #15 on: November 16, 2014, 04:44:01 am »

Hi Paul,

Thanks for your time and recommendations.


Here is an 8s exposure with the SLT 99 from the same scene, it is amore valid comparison with the P45+ that was exposed 7.5 s. In this case the windows are blown out beyond rescue, too. The two exposures may be a pretty close match (to be coming from a set of real world images).

http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC4756.ARW

I may call your attention to this file, that was generated using Lumariver HDR from another set of P45+ images. This is exposure fusion done on the raw image. I would recommend developing it in Lightroom as DNG support in Capture One may be weak.

http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/20140617_lumariver.dng

Best regards
Erik




Hello Eric:

I worked with both files, and to me, the P45+ is the better file overall.  The only problem with the P45+ is as you noted, the light coming in the windows is too harsh to really capture much of information outside, whereas with the Sony, you can pull down the windows and still hold some image.

However.

1.  I thought that C1 worked much better on the overall image of the p45+, especially shadow recovery, both overall and on the roof.  The Sony file I added an adjustment layer to pull up the roof, I did not need this on the P45+
2.  I did a keystone adjustment to the Sony just for looks, I left the P45+ as is was.
3.  P45+ image had a good bit of CA around the windows, but that cleaned up in C1, Sony image did not
4.  Sony image loaded too red to me especially on the bricks, but no knowing what the interior really looked like, I sided with the P45+ as it gave a much better tone to the bricks
5.  P45+ easily wins on details, I did add a bit of structure to the P45+ as the image seemed just a bit dulled down, I saw it was 7 seconds.  When you look at the seating in the background, wood details stand out much better in the P45+ file

Both files are clean, and I assume both were at base iso, I saw the P45+ was at 50.  It would be very interesting to try the same test, at a shorter shutter speed with higher iso, and see if the same holds true.  In this test, you have proven to me that the CCD can do excellent work, when given enough light and time.  

I have attached a side by side of both images in C1, I used Capture 1 vr8 as it's much better than 7 in so many areas.  

Edit:  I could not have gotten such a clean transition with my P45+, the roof area would have looked much worse and the seats.  After looking that this image the first one you posted from the dock scene, I am quite sure that Phase One made some changes to the P45+, in the later models, and or Kodak was able to change something in the chip fab (which I doubt).  I have been of the belief that most camera companies do this, (under cover improvements), and looking at your P45+ images really convinces me of that.  Keep that back as it's a winner for sure.

Sincerely
Paul


Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #16 on: November 16, 2014, 04:47:02 am »

Hi Synn,

Thanks for the recommendations.

This image is exposed pretty close to the P45+ image, with windows blown out past recovery:

http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC4756.ARW

Best regards
Erik
Logged
Erik Kaffehr
 

Abe R. Ration

  • Newbie
  • *
  • Offline Offline
  • Posts: 24
    • Abe R. Ration's lens and camera blog
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #17 on: November 16, 2014, 05:35:55 am »

A couple of issues I considered were:

  • Some people talk about photographic DR, something like SNR = 10.
  • I would suggest that photographic DR is essentially a function of exposure.
  • Readout noise affects the darkest part of the pictures. How relevant is that?
As I see it, photographic DR can be defined in two ways:
  • Relative to output image
  • Relative to the data

In the first case one needs to consider the output image size and distance of observation. The latter doesn't needs to consider that (though one has to understand the relevancy of those). Regardless, a photographic DR requires a minimum SNR.

One thing we need to notice is that the SNR needs to be normalized some way. If we do not do that, then we're not talking about image SNR, but pixel SNR.

An interesting question is if your suggestion of PDR being essentially a function of exposure holds water. I assume your arbitrary SNR=10 floor is for pixel SNR. As N(total)=sqrt(N(shot)^2 + N(read)^2), we can easily see that for small read noises on most modern sensors read noise is indeed irrelevant for PDR. Without any read noise a signal of 100 photons (or electrons) would create SNR=10. With most sensors the read noise at base ISO is about 3 or 4 electrons, so to have SNR=10 we'd need about 109 to 114 photons. Unforunately there are also sensors - notably from Canon - which has much higher base ISO read noise. The new 7DII has ISO 100 read noise in the ballpark of 13 electrons. Thus to have SNR=10 you'd need to collect about 190 photons!

Thus clearly with such SNR=10 we can not say that PDR is essentially a function of exposure, but read noise needs to be considered.

(Also I think it is better to consider the storage capacity of pixel (FWC), instead of exposure itself and the latter would still have to consider the former and also QE, and the number of pixels.)

I am pretty sure that for PDR pattern noises is quite irrelevant.

About normalizing: the most straightforward way of normlizing is just to consider the whole image sensor - as we can relatively easily measure the relevant metrics of a pixel, we can also easily calculate the what that means to the whole sensor:


S(total) = S(pixel)*n where S()=number of photons, and n=number of pixels.
N(total) = sqrt(n*N(shot)^2 + n*N(read)^2), where N()=noise, read=read noise, shot=shot noise=sqrt(photons)

Of coure such normlization needs a bit larger SNR minimum for PDR :)

Alternatively we could use for example the normalization DxOMark uses.

Note:
None of the above considers colours influence to noise. Creating colour out of the data increases noise and this is not constant for all cameras. It is likely that for conventional cameras this is not much of a topic (weaker colour separation, like of Canon's should increase noise a bit for colour images), but if one considers Sigma's sensors it can be very significant. Calculating the relevancy is beyond my limited skills though, I suspect.
« Last Edit: November 16, 2014, 05:40:05 am by Abe R. Ration »
Logged
Abe R. Ration
amateur photographer, amateur scientist, amaterur camera buff
http://aberration43mm.wordpress.com/

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #18 on: November 16, 2014, 06:25:25 am »

Hi Abe,

You are quite right! Must have been temporary (hopefully) loss of brain function on my part.

I agree with everything you say, actually. Just one thing about FWC, to make utility of FWC we need to expose to the right.
 
Best regards
Erik
Quote
An interesting question is if your suggestion of PDR being essentially a function of exposure holds water. I assume your arbitrary SNR=10 floor is for pixel SNR. As N(total)=sqrt(N(shot)^2 + N(read)^2), we can easily see that for small read noises on most modern sensors read noise is indeed irrelevant for PDR. Without any read noise a signal of 100 photons (or electrons) would create SNR=10. With most sensors the read noise at base ISO is about 3 or 4 electrons, so to have SNR=10 we'd need about 109 to 114 photons. Unforunately there are also sensors - notably from Canon - which has much higher base ISO read noise. The new 7DII has ISO 100 read noise in the ballpark of 13 electrons. Thus to have SNR=10 you'd need to collect about 190 photons!
« Last Edit: November 16, 2014, 06:36:33 am by ErikKaffehr »
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples updated:
« Reply #19 on: November 16, 2014, 06:43:20 am »

Hi,

Here are the present set of samples on my site:

A 99 SLT 2.5s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC4758.ARWWindows not clipped
A 99 ALT 8s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/_DSC475.ARWWindows clipped
P45+ 7.5s exposurehttp://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/CF045286.IIQWindows clipped
P45+ Lumariver HDR
multiple exposures
fused
http://echophoto.dnsalias.net/ekr/Articles/DRArticle/NativeRaws/20140617_lumariver.dngHDR, Windows not clipped
« Last Edit: November 16, 2014, 06:52:47 am by ErikKaffehr »
Logged
Erik Kaffehr
 
Pages: [1] 2   Go Up