Pages: 1 [2]   Go Down

Author Topic: Let's discuss some DR (Dynamic Range) samples  (Read 11199 times)

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4066
    • Photos of Arkansas
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #20 on: November 16, 2014, 07:52:24 am »

Another option would be to have the back mount changed to Mamiya then invest in a used DF or AFDIII. they come on the market quite often.  This would give you AF and a very wide range of lenses.    I don't know how much the Hasselblad AF camera and lenses are.

From what I have seen from your back provides very nice images. You also might want to invest later in a tech camera. You back will allow a lot of shift before color cast comes into play.

On the flip side, Sony is working on a new 35mm camera in the 40 to 50MP range but it seems to be a way off. Bigger issue with that will be lenses that can resolve to the sensor.

Paul





Logged
Paul Caldwell
Little Rock, Arkansas U.S.
www.photosofarkansas.com

Slobodan Blagojevic

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 18090
  • When everyone thinks the same, nobody thinks
    • My website
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #21 on: November 16, 2014, 10:17:52 am »

...
  • Readout noise affects the darkest part of the pictures. How relevant is that?...

As a photographer, it is quite relevant for me. I think the best answer why and best visual example was provided by CptZar in this thread (post #68).

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #22 on: November 16, 2014, 10:49:45 am »

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

I agree, at least with the D810, D800E, D4, a7, a7R, a7S, which are the cameras that I've tested the most. 

In fact, I went looking for pattern read noise (which I define as the component of read noise that doesn't have white spectral response) in D810 images here:

http://blog.kasson.com/?p=7886

I wasn't able to observe it until I got to nosebleed ISOs and average electron counts in single digits.  I was able to measure it, however.

Jim

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #23 on: November 16, 2014, 01:49:25 pm »

Hi Slobodan,

I hear you clear and loud. Personally, I had little issues like that. Why, I don't know:

  • Different subjects and or different light
  • Different shooting habbits
  • Differences in equipment

Personally, I have been quite happy since 2006, when I got my Sony Alpha 100, the Konica-Minolta Dimage 7D was a bit below my acceptance level. The next generation of Sony cameras I had was the Alpha 700, the first one using Sony's new technology with massively parallel conversion. I would say that I shot my best images with that camera.

It was good enough and I had very good shooting opportunities…

Best regards
Erik








As a photographer, it is quite relevant for me. I think the best answer why and best visual example was provided by CptZar in this thread (post #68).


« Last Edit: November 16, 2014, 01:59:03 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #24 on: November 16, 2014, 02:14:30 pm »

Hi,

I have been informed that there is nothing like a mount change. It is more about replacing the back. That may be wrong of course.

One of the weak points of the Hasselblad V package is the lack of a real wide option. My original plan was replace the Hasselblad 555/ELD with a Hartblei HCam with a Canon T&S lens. Right now I don't feel attracted to that option. I bought a Flexbody as an interim and I don't feel tilts work for me. But, the Hartblei may come back.

From where I stand, it seems quite form that Sony will introduce new cameras with 46 and 54 MP in February 2015. I would expect lenses to work decently well, based on my testing with a 3.8 micron APS-C body. On the other hand, I want a couple of really great lenses. The Otus 85/1.4 would fill the bill, but Zeiss will not make it for the Sony FE-mount. Sigma doesn't like the FE-mount , either.

So, the choice boils down to:

  • Live view
  • Good or preferably excellent lenses

With the Hasselblad I feel the lenses are a bit better than expected. Focusing is not easy, though.

Regarding sensor resolution vs. lenses, I made some experiments and I certainly feel that 54 MP has it's benefits: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/78-aliasing-and-supersampling-why-small-pixels-are-good


Best regards
Erik





Another option would be to have the back mount changed to Mamiya then invest in a used DF or AFDIII. they come on the market quite often.  This would give you AF and a very wide range of lenses.    I don't know how much the Hasselblad AF camera and lenses are.

From what I have seen from your back provides very nice images. You also might want to invest later in a tech camera. You back will allow a lot of shift before color cast comes into play.

On the flip side, Sony is working on a new 35mm camera in the 40 to 50MP range but it seems to be a way off. Bigger issue with that will be lenses that can resolve to the sensor.

Paul






« Last Edit: November 16, 2014, 03:27:06 pm 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 #25 on: November 16, 2014, 03:18:53 pm »

From where I stand, it seems quite form that Sony will introduce new cameras with 46 and 54 MP in February 2015. I would expect lenses to work decently well, based on my testing with a 3.8 micron APS-C body

The lenses will work at least as well as they do now: more pixels just samples the image better. There is a caveat though - lenses where the exit pupil is close to the sensor (i.e. many rangefinder lenses or lenses with near symmetrical design) may suffer outside of the center area.

Regarding sensor resolution vs. lenses, I made some experiments and I certainly feel that 54 MP has it's benefits: /index.php/photoarticles/78-aliasing-and-supersampling-why-small-pixels-are-good

Oversampling is what in principle should be strived for to get rid of aliasing and maximize image processing potential. Though, if we assume ideal ("perfect") lens, for proper sampling we need much more pixels than what 54 million would be - a couple of weeks ago I made some fun calculations: https://aberration43mm.wordpress.com/2014/11/03/diffraction-etc/
Logged
Abe R. Ration
amateur photographer, amateur scientist, amaterur camera buff
http://aberration43mm.wordpress.com/

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #26 on: November 17, 2014, 02:03:12 pm »

After having pondered this thread, I can easily say that Dynamic Range is not important, unless it is.  All else equal, I will take more DR than not.

There are many images to be made that easily fit in the DR of almost any modern camera.  And when the important areas fall to one end or the other, blowing the highlights or blocking up the shadows is a good trade off.  However, when important aspects fall at both extremes, then there are issues.  Classic example being sunrise/sunset images or inside/outside images like that posted where important details exist in both areas.  In fact, excluding the garish (or artistic) over saturation of some HDR images, we often think they look unnatural because for so long we were used to blowing out the top end or silhouetting something against the bright sky.  To be able to preserve detail in both violates our traditional expectations, not necessarily the real world or eye's ability to adjust to it.

Having additional 'headroom' though is always a benefit.  It makes pushing and pulling the data around in post processing easier and allows for ETTR when clean shadow areas are desired.  And if we are talking about our ability to composite multiple exposures, that is a 'hack' that is very useful, but also has limitations, especially with dynamic subjects and lighting.

Finally, it is impossible to to judge DR and quality of the RAW data from 2 cameras based on RAW conversion because of the unknown assumptions built into the conversion engines and our lack of confidence that the propriety file was correctly decoded as intended by the manufacturer.  Take the same raw image and decode in multiple engines and while you can get close, very rarely can you get to the point of eliminating visual differences, let alone be able to duplicate the numerical results.  But at the end of the day, a photographer will have a set of tools and set of skills which may be better suited to a particular type of image or images from a particular camera than some other types of images or other cameras.  Of course, this precisely why the user community would prefer a well documented standard RAW file than the camera maker proprietary ones.

Logged

mouse

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 260
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #27 on: November 17, 2014, 03:01:41 pm »

 
Quote
Of course, this precisely why the user community would prefer a well documented standard RAW file than the camera maker proprietary ones.

Amen to that!
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #28 on: November 17, 2014, 03:04:18 pm »

Hi,

I would mainly agree, except a few points.

- Mainly, the raw images can be analysed using RawDigger, that actually shows the image data.

- The other point, which I hoped to demonstrate but failed to a certain extent is that read noise is not very visible in the images as it occurs in the darks, which are often quite suppressed. But, in the images I show I still link shot noise dominates.

- Another point that may be interesting that I have added an HDR image in this mix: http://echophoto.dnsalias.net/ekr/Articles/DRArticle/20140617_lumariver.dng , that image is actually an HDR image generated from raw files by Lumariver HDR, but it is still undemosaiced raw, that can be handled as any other raw file.

Best regards
Erik

After having pondered this thread, I can easily say that Dynamic Range is not important, unless it is.  All else equal, I will take more DR than not.

There are many images to be made that easily fit in the DR of almost any modern camera.  And when the important areas fall to one end or the other, blowing the highlights or blocking up the shadows is a good trade off.  However, when important aspects fall at both extremes, then there are issues.  Classic example being sunrise/sunset images or inside/outside images like that posted where important details exist in both areas.  In fact, excluding the garish (or artistic) over saturation of some HDR images, we often think they look unnatural because for so long we were used to blowing out the top end or silhouetting something against the bright sky.  To be able to preserve detail in both violates our traditional expectations, not necessarily the real world or eye's ability to adjust to it.

Having additional 'headroom' though is always a benefit.  It makes pushing and pulling the data around in post processing easier and allows for ETTR when clean shadow areas are desired.  And if we are talking about our ability to composite multiple exposures, that is a 'hack' that is very useful, but also has limitations, especially with dynamic subjects and lighting.

Finally, it is impossible to to judge DR and quality of the RAW data from 2 cameras based on RAW conversion because of the unknown assumptions built into the conversion engines and our lack of confidence that the propriety file was correctly decoded as intended by the manufacturer.  Take the same raw image and decode in multiple engines and while you can get close, very rarely can you get to the point of eliminating visual differences, let alone be able to duplicate the numerical results.  But at the end of the day, a photographer will have a set of tools and set of skills which may be better suited to a particular type of image or images from a particular camera than some other types of images or other cameras.  Of course, this precisely why the user community would prefer a well documented standard RAW file than the camera maker proprietary ones.


« Last Edit: November 17, 2014, 03:09:02 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #29 on: November 22, 2014, 12:47:49 pm »

Here is an example of needing more DR. The camera was already set for snow shots at I'm guessing +1.5ev. I had < a second to see, frame, capture, as the couple of birds came to be fed. There was no time for multi-frame capture of the same scene.



To be able to capture the low sun on the water, unblown, while capturing the snow near white, I would probably need several more stops beyond what cameras currently provide.

BTW, that was the Sony A350 (CCD) which is basically stuck at ISO100.
« Last Edit: November 22, 2014, 12:58:47 pm by Fine_Art »
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #30 on: November 22, 2014, 12:58:31 pm »

Snow Ghosts!  Making me long to go back to Whitefish, Montana and ski Big Mountain!

With respect to the image, that is another example why I vote more DR and lower shadow noise.  As an ameatuer, I don't get to make images as frequently as I would like.  And so, I make more mistakes, especially when I encounter situations I don't shoot often.  If a camera has 2 stops more DR and can save me from myself, I consider myself grateful!


Here is an example of needing more DR. The camera was already set for snow shots at I'm guessing +1.5ev. I had < a second to see, frame, capture, as the couple of birds came to be fed. There was no time for multi-frame capture of the same scene.



To be able to capture the low sun on the water, unblown, while capturing the snow near white, I would probably need several more stops beyond what cameras currently provide.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #31 on: November 26, 2014, 03:20:48 am »

Hi,

I shot an image a few days ago which may illustrate some good points. The original DNG image is here (it contains the ARW image that can be extracted by Adobe DNG Converter): http://echophoto.dnsalias.net/ekr/Articles/DRArticle/CanoeStadium/20141109-_DSC6262.dng

This image has a DR of about 10 EV (-7EV to +3EV), see the raw histogram below


The image has a pretty wide range of tones, so a nice photographic interpretation involves some tone mapping, I am not there, but say something like this:


The rendition below is intended to show the recoverable range using extreme manipulation in LR:


The area around the sun has some channel clipping, but LR can till reconstruct the edge of the sun disc:


An actual pixel crop of a small part, still decent detail. I probably don't want detail in the dark jacket.


Looking at a small crop of the jacket I see this histogram:

The clipped area seems to have something like SNR between four and eight. The bell shape indicates it is shot noise and it is little affected by readout noise.

Best regards
Erik
« Last Edit: November 26, 2014, 03:24:38 am by ErikKaffehr »
Logged
Erik Kaffehr
 

Isaac

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3123
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #32 on: November 26, 2014, 10:31:50 am »

To be able to capture the low sun on the water, unblown, while capturing the snow near white, I would probably need several more stops beyond what cameras currently provide.

Just curious -- Have you tried processing for the reflections, processing for the snow separately, and then merging with exposure fusion or blending?
Logged

Fine_Art

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1172
Re: Let's discuss some DR (Dynamic Range) samples
« Reply #33 on: November 26, 2014, 12:15:13 pm »

Just curious -- Have you tried processing for the reflections, processing for the snow separately, and then merging with exposure fusion or blending?

I am pretty sure I took separate exposures after that shot for the water. I never got around to putting it all together. I used to find masking unworkable in CS3. It would kind of do it, then you would have to pick at it pixel by pixel. Now I have Topaz Remask so I should look at it again.
Logged
Pages: 1 [2]   Go Up