Home · Register · Software · Software · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Sony Forum | Join Upload & Sell

1       2       3       4      
5
       end
  

Sony A7R III Dynamic range at 14.7 Evs
  
 
Viramati
Offline
• • •
Upload & Sell: Off
p.5 #1 · p.5 #1 · Sony A7R III Dynamic range at 14.7 Evs


Matt Grum wrote:
Ignoring thermal noise from long exposures, then it is always better to use a longer exposure and lower ISO, and this applies to all cameras.

NR kicks in at ISO 32,000 on the A7RIII, so I would stay below that. In fact if I were shooting at ISO 25600 I would have a good long think about what I'm doing and if there was a better way.

Related point, if you can stand the dark viewfinder or lack of exposure preview, then you should never need to go beyond ISO 640, as there's no gain [pun unintended] compared to fixing
...Show more
Maybe but the venue was one where I have shot before with the A7rII and the lighting was exactly the same as was the iso and unless my eyes a tricking me I am seeing a slight difference. Nothing to really get excited about but in processing the images I just found the seem a 'little' cleaner. No very scientific I know



Dec 05, 2017 at 01:39 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #2 · p.5 #2 · Sony A7R III Dynamic range at 14.7 Evs


Matt Grum wrote:
It's worth noting that the chart shows the reduction in photon noise with stacking, but doesn't take into account the read noise. Read noise variance scales linearly with the number of images stacked.

That's why you only get 1 stop increase in DR when stacking 4 images (which is 2 extra stops of light). If you're interested in DR then bracketing will provide vastly superior results to stacking (e.g. you can easily get an extra 9 stops with 3 exposures). Bracketing also reduces photon noise in the darker parts of the image (which is the only place it matters, photon
...Show more

Sounds right. In practice though, averaging 8-15 shots gives me shadows that are so clean, I can push 2 or even 3 stops while still keeping color and detail integrity. Bracketing is very useful when shooting very high contrast scenes that exceed the limits of what's possible with averaging but I get more natural results from the latter.



Dec 05, 2017 at 04:07 PM
bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #3 · p.5 #3 · Sony A7R III Dynamic range at 14.7 Evs


Matt Grum wrote:
It's worth noting that the chart shows the reduction in photon noise with stacking, but doesn't take into account the read noise. Read noise variance scales linearly with the number of images stacked.

That's why you only get 1 stop increase in DR when stacking 4 images (which is 2 extra stops of light). If you're interested in DR then bracketing will provide vastly superior results to stacking (e.g. you can easily get an extra 9 stops with 3 exposures). Bracketing also reduces photon noise in the darker parts of the image (which is the only place it matters, photon
...Show more

This is exactly right any why Pixel Shift for example, only improves PDR 1.15 stops and not 2.

BTW, "9 stops with 3 exposures" is clearly a typo



Dec 05, 2017 at 07:20 PM
Matt Grum
Offline
• • •
Upload & Sell: Off
p.5 #4 · p.5 #4 · Sony A7R III Dynamic range at 14.7 Evs


bclaff_too wrote:
.BTW, "9 stops with 3 exposures" is clearly a typo


Yeah, I was thinking -3, 0, +3 which would get you 6 more stops. However with bracketing you can basically get any increase you want, and can very quickly outstrip the dynamic range the lens is capable of delivering.



Dec 05, 2017 at 11:26 PM
bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #5 · p.5 #5 · Sony A7R III Dynamic range at 14.7 Evs


You may find my post Noise Reduction in Stacked Images interesting. Charts the real effect of stacking 1 to 16 images.


Dec 06, 2017 at 03:09 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #6 · p.5 #6 · Sony A7R III Dynamic range at 14.7 Evs


bclaff_too wrote:
You may find my post Noise Reduction in Stacked Images interesting. Charts the real effect of stacking 1 to 16 images.


Very interesting info Bill!
So FPN is the culprit and stacking too many images could actually reduce IQ. In my own real world images, I've noticed that when stacking more than 64 images, there's visible posterization in some areas. I attribute this to the lower noise which is no longer dithering transition areas.

What would you say is the optimum number of stacked frames to reduce black and illuminated read noise while keeping FPN under control? By optimum, I mean, the lowest number of images for noticeable S/N ratio improvement.



Dec 06, 2017 at 04:59 PM
bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #7 · p.5 #7 · Sony A7R III Dynamic range at 14.7 Evs


Fred Miranda wrote:
Very interesting info Bill!
So FPN is the culprit and stacking too many images could actually reduce IQ. In my own real world images, I've noticed that when stacking more than 64 images, there's visible posterization in some areas. I attribute this to the lower noise which is no longer dithering transition areas.

What would you say is the optimum number of stacked frames to reduce black and illuminated read noise while keeping FPN under control? By optimum, I mean, the lowest number of images for noticeable S/N ratio improvement.


Given the findings I'm not surprised you see "posterization" in some areas when stacking large numbers of images.

It's hard to know what's "optimum". I did add a chart to the post I linked that shows stops of improvement rather than noise.



Dec 06, 2017 at 06:18 PM
Stoffer
Online
• • •
Upload & Sell: Off
p.5 #8 · p.5 #8 · Sony A7R III Dynamic range at 14.7 Evs


bclaff_too wrote:
You may find my post Noise Reduction in Stacked Images interesting. Charts the real effect of stacking 1 to 16 images.


Fixed Pattern Noise is worse on a7rII than the new a7rIII, right? Donít know if that changes things a bit in regard to Fredís findings with the a7rII.



Dec 06, 2017 at 06:22 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #9 · p.5 #9 · Sony A7R III Dynamic range at 14.7 Evs


bclaff_too wrote:
Given the findings I'm not surprised you see "posterization" in some areas when stacking large numbers of images.

It's hard to know what's "optimum". I did add a chart to the post I linked that shows stops of improvement rather than noise.


This chart is very helpful Bill. I would think that the "illuminated" numbers are more relevant to real world images. Correct? If that's the case, a mere 4-8 images could be all that we need for improved SNR without any side-effects.








Dec 06, 2017 at 07:15 PM
 

Search in Used Dept. 

        


bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #10 · p.5 #10 · Sony A7R III Dynamic range at 14.7 Evs


Stoffer wrote:
Fixed Pattern Noise is worse on a7rII than the new a7rIII, right?


Yes, definitely.
Stoffer wrote:
Donít know if that changes things a bit in regard to Fredís findings with the a7rII.


Unclear but FPN will be more visible as you increase the number of stacked images.
So the more FPN, the more likely you'll see something.



Dec 06, 2017 at 07:23 PM
bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #11 · p.5 #11 · Sony A7R III Dynamic range at 14.7 Evs


Fred Miranda wrote:
This chart is very helpful Bill. I would think that the "illuminated" numbers are more relevant to real world images. Correct? If that's the case, a mere 4-8 images could be all that we need for improved SNR without any side-effects.


My thinking is the opposite.
Normally we have plenty of SNR in the brighter areas; that wouldn't be my motivation to stack.
I'd stack to get cleaner shadows (better low signal SNR)
So stacking to improve "black" would be my motivation but I'm getting some highlight improvement as a bonus.
(With that caveat that too much improvement could expose FPN.)



Dec 06, 2017 at 07:27 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #12 · p.5 #12 · Sony A7R III Dynamic range at 14.7 Evs


bclaff_too wrote:
My thinking is the opposite.
Normally we have plenty of SNR in the brighter areas; that wouldn't be my motivation to stack.
I'd stack to get cleaner shadows (better low signal SNR)
So stacking to improve "black" would be my motivation but I'm getting some highlight improvement as a bonus.
(With that caveat that too much improvement could expose FPN.)


Makes sense, that's what I meant, I got it backwards. From your chart, after 16 shots, 'black' stops continue improving or start to flatten out?



Dec 06, 2017 at 07:31 PM
bclaff_too
Offline
• •
Upload & Sell: Off
p.5 #13 · p.5 #13 · Sony A7R III Dynamic range at 14.7 Evs


Fred Miranda wrote:
Makes sense, that's what I meant, I got it backwards. From your chart, after 16 shots, 'black' stops continue improving or start to flatten out?


The exact number of images depends on your personal tolerance for "diminishing returns".
After some testing here I chose 16 images as the number for my FPN testing; but these aren't real world images.



Dec 06, 2017 at 07:39 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #14 · p.5 #14 · Sony A7R III Dynamic range at 14.7 Evs


Great info as always Bill.
Best,
Fred



Dec 06, 2017 at 07:47 PM
snapsy
Offline
• • • • •
Upload & Sell: On
p.5 #15 · p.5 #15 · Sony A7R III Dynamic range at 14.7 Evs


If one is expending the effort to use a stacking workflow to reduce noise then it doesn't seem like it would represent much more effort to run the images through a third-party debanding plug-in to remove FPN banding. I haven't used them in years - it would be interesting to see how well they work on Sony's flavor of FPN. I'm sure they'll eliminate the visible/perceptual FPN - the question is whether they can do so without introducing any artifacts.


Dec 06, 2017 at 07:54 PM
mttran
Offline
• • • • •
Upload & Sell: Off
p.5 #16 · p.5 #16 · Sony A7R III Dynamic range at 14.7 Evs


Are we able to produce the final pixel shift in raw/tiff? if we can then we can simulate some HDR from the raw/tiff file as well.


Dec 06, 2017 at 09:27 PM
Matt Grum
Offline
• • •
Upload & Sell: Off
p.5 #17 · p.5 #17 · Sony A7R III Dynamic range at 14.7 Evs


mttran wrote:
Are we able to produce the final pixel shift in raw/tiff? if we can then we can simulate some HDR from the raw/tiff file as well.


Yes Imaging Edge can export the pixel shift results as a TIFF in 16-bit.

When I get some time I'll write a tool to convert pixel shift files to linear DNG for import into your RAW processor of choice (and automatically handle movement).



Dec 06, 2017 at 10:47 PM
Fred Miranda
Offline
Admin
Upload & Sell: On
p.5 #18 · p.5 #18 · Sony A7R III Dynamic range at 14.7 Evs


Matt Grum wrote:
Yes Imaging Edge can export the pixel shift results as a TIFF in 16-bit.

When I get some time I'll write a tool to convert pixel shift files to linear DNG for import into your RAW processor of choice (and automatically handle movement).


I hope you get some time soon!!
We definitely need that!



Dec 06, 2017 at 10:51 PM
1       2       3       4      
5
       end






FM Forums | Sony Forum | Join Upload & Sell

1       2       3       4      
5
       end
    
 

You are not logged in. Login or Register

Username     Reset password