April 18, 2014, 11:17:21 AM

Author Topic: 5D III Dynamic Range  (Read 25314 times)

CarlTN

  • 1D X
  • *******
  • Posts: 2159
    • View Profile
Re: 5D III Dynamic Range
« Reply #30 on: February 09, 2013, 05:02:16 AM »
Bottom line, unless you are viewing the final result on some kind of very exotic, purpose-built, lighted device (whether back or front-lit)...then a picture with a lot of stops of DR...once it is printed on any kind of paper, will necessarily look "painterly".  As in, you are reduced to seeing very few stops of DR with a printed image.  However, even a printed image can in some ways exceed the information your computer monitor is able to display.  It just looks less bright, less vivid...than a lighted display device.  It gets dynamically compressed.

Under normal room lighting, my prints usually look a lot darker, duller, and less colorful, than they did on the computer screen (let alone if the room lighting isn't of neutral color temperature).  Hold the print up close to a good light, and it looks a lot closer to what was on the screen.  View it outdoors in the sunlight (preferably slightly diffuse), and it can also look close.

It's important when editing (and then printing) pictures, to allow for how they will usually be lit, wherever they wind up being hung, or displayed.  If the final destination is only electronic on the web, then you have one less variable...but of course other considerations. 

But you're never going to see 14 stops of DR, no matter what it is viewed on.  Your eye saw more than that as you took the shot, so just try to remeber that as you ponder the picture, and enjoy it.

canon rumors FORUM

Re: 5D III Dynamic Range
« Reply #30 on: February 09, 2013, 05:02:16 AM »

mrmarks

  • Rebel T5i
  • ****
  • Posts: 162
    • View Profile
Re: 5D III Dynamic Range
« Reply #31 on: February 09, 2013, 05:14:04 AM »
Bottom line, unless you are viewing the final result on some kind of very exotic, purpose-built, lighted device (whether back or front-lit)...then a picture with a lot of stops of DR...once it is printed on any kind of paper, will necessarily look "painterly".  As in, you are reduced to seeing very few stops of DR with a printed image.  However, even a printed image can in some ways exceed the information your computer monitor is able to display.  It just looks less bright, less vivid...than a lighted display device.  It gets dynamically compressed.

Under normal room lighting, my prints usually look a lot darker, duller, and less colorful, than they did on the computer screen (let alone if the room lighting isn't of neutral color temperature).  Hold the print up close to a good light, and it looks a lot closer to what was on the screen.  View it outdoors in the sunlight (preferably slightly diffuse), and it can also look close.

It's important when editing (and then printing) pictures, to allow for how they will usually be lit, wherever they wind up being hung, or displayed.  If the final destination is only electronic on the web, then you have one less variable...but of course other considerations. 

But you're never going to see 14 stops of DR, no matter what it is viewed on.  Your eye saw more than that as you took the shot, so just try to remember that as you ponder the picture, and enjoy it.

So what would be a typical DR range of a LCD display and a print?

Txema

  • PowerShot G16
  • **
  • Posts: 22
    • View Profile
Re: 5D III Dynamic Range
« Reply #32 on: February 09, 2013, 05:44:13 AM »
In my experience Hdr with only one shot is a very destructive process if you start up with just one shot on a limited dr camera as my 5D mark III.  So, for me,  HDR only works well on static subjects. I wish my 5D III would ha 14 stops DR.

Anyone knows the real difference in DR on the 5DIII between ISO 100 normal and 200 with hilight tone priority?

CarlTN

  • 1D X
  • *******
  • Posts: 2159
    • View Profile
Re: 5D III Dynamic Range
« Reply #33 on: February 09, 2013, 05:46:16 AM »
So what would be a typical DR range of a LCD display and a print?

I have no idea exactly, supposedly the photo-editing approved monitors are better than the cheap ones.  Supposedly prints can display 10 stops (I think I read that somewhere).  But that is 10 stops that is effectively compressed by the net result of the reflectivity of the print and whatever constitutes its emulsion, ink, or pigment...in direct light. 

And people are editing 32 bit HDR files on monitors that can't even display a 16 bit file.  And when that HDR file gets printed, it looks like a painting.  Paintings, however, look nice.  They just can't be mistaken for seeing the original event with your eye.  This is especially true, when someone produces an image that looks worse than a cheesy cartoon, drawn with spray cans.

I would dare say, that the main reason a (good) HDR image looks nice, when printed, is precisely because it has been artistically altered and arranged, to compress the overall DR, and highlight the parts that make the color and/or the composition look good.  I'm sure I'm not stating anything new by saying that.

This is why I don't understand the obsession with dynamic range at the low ISO end.  Sure, it gives you more lattitude.  However, for those of us who actually prefer to shoot things that aren't lit by a strobe, or the midday sun...the dynamic range recorded by the camera as the sensitivity is boosted, is what really affects image quality.  In which case, Nikon and Canon aren't that far apart.  They will both get better as time goes on.  Maybe Canon will surpass overall, maybe they won't.  Either way, it's better just to pick a system and go out and take pictures, and enjoy.  Otherwise, you will never catch up to the world class shooters.

David Hull

  • Rebel T5i
  • ****
  • Posts: 169
    • View Profile
Re: 5D III Dynamic Range
« Reply #34 on: February 09, 2013, 10:55:36 AM »
Many scenes have less than 11 stops of actual DR in the scene, in which case there's no 'deficiency'.  Other scenes have more than 14 stops of DR, in which case all current dSLRs are 'deficient' and multiple bracketed exposures and HDR are the only way to capture the full scene DR, assuming you need to do so.
While not false, I think this description is misleading.

A given scene usually have dark parts and bright parts. When we do image processing we might want to "dodge & burn", apply curves and levels or whatever to make the scene look "pleasing", to direct the viewers attention to something, to overcome the limitations of printers etc.

If your camera has 2 stops more of DR, then it is an indication that (depending on how you choose to expose your scene), you will have 2 stops more of highlight information, 2 stops more of shadow information, or some combination of those two. If a scene has "15 stops of DR" you will be able to make more of it with a 14-stop camera than an 11-stop camera.

I think that all of the above is relevant and fair information about camera vs scene DR. Now there are many views on the importance of DR > 10 stops. For some reason it is deemed highly important to some Nikon users, while some Canon users find it utterly unimportant. I expect that situation to flip if ever Canon release a 15-stop DR DSLR.

-h

What really matters is what sort of photography you can do with it and so far we haven't been treated to any earth shattering breakthroughs in image making as a result of these two stops of additional DR.  If it was as significant as some would have us believe, then we would easily be able to walk into any gallery and pick out the Nikon images from the Canon ones – but we cannot.  We continue to be treated to the same me tired, contrived examples that are completely devoid of any artistic value whatsoever and I doubt that THIS will change even if Canon were to come out with a 15 stop camera.

mrmarks

  • Rebel T5i
  • ****
  • Posts: 162
    • View Profile
Re: 5D III Dynamic Range
« Reply #35 on: February 09, 2013, 11:40:31 AM »
The human eye is estimated to be able to see detail across a range of 12 stops, or so, but that is estimated as without the iris adjusting, so a fair comparison to a single image. Our inbuilt aperture works so well, fast and automatically, that we can actually see detail in scenes far wider than that.

Cameras, in a single shot in RAW, can record detail across around 12 stops, some a little better, some a little worse. Now if you shoot jpeg you can only record 8 stops. When you shoot RAW 12 bit you get a potential 12 stops. True 14 bit has a ceiling of 14 stops. No camera that records in 14 bits can possibly record detail across more than 14 stops. Until camera manufacturers start releasing true 16 bit RAW files we won't see any genuine increases in DR.

Monitors display around 8 stops, very close to jpegs hence the standard. Monitor range is very dependent on ambient light, if your monitor is in a bright room it loses DR, if it is in a dark room it gets some of that lose back, but never more than the 8 or so stops.

Prints are down in the 6 stop range, this is entirely dependent on paper reflectivity and the amount of light thrown onto the print. Think about it, in a dark room a print has zero stops of discernible data!

Now this is just a very brief generalisation, each minutia could be argued about ad nausea. I am not interested.

Further to say 32 bit prints must look painterly is not true, it all depends where the ranges of interesting detail fall within the scene. For instance, if you shoot an interior but are interested in holding  detail outside in the sunlight, the detail outside might be 5 stops higher, but there will be very little info between the two, you can then compress the area between the inside and outside without introducing painterly qualities. Very few scenes with huge dynamic range have smooth histograms, most have several peaks where the interesting stuff is, all you are doing is bringing those interesting bits back within displayable ranges.

This is an example of that, to retain the clouds and sky I have lowered the top end, but there are very few tones between the room and the outside that were compressed, so no "un-natural" look.

Thanks for the great info!

mrmarks

  • Rebel T5i
  • ****
  • Posts: 162
    • View Profile
Re: 5D III Dynamic Range
« Reply #36 on: February 09, 2013, 11:44:28 AM »
What really matters is what sort of photography you can do with it and so far we haven't been treated to any earth shattering breakthroughs in image making as a result of these two stops of additional DR.  If it was as significant as some would have us believe, then we would easily be able to walk into any gallery and pick out the Nikon images from the Canon ones – but we cannot.  We continue to be treated to the same me tired, contrived examples that are completely devoid of any artistic value whatsoever and I doubt that THIS will change even if Canon were to come out with a 15 stop camera.

You hit the nail on the head.

canon rumors FORUM

Re: 5D III Dynamic Range
« Reply #36 on: February 09, 2013, 11:44:28 AM »

well_dunno

  • 7D
  • *****
  • Posts: 356
    • View Profile
Re: 5D III Dynamic Range
« Reply #37 on: February 09, 2013, 12:12:56 PM »
Is Don Quixote over? Which is the one we are watching now?

*hands out some more popcorn*

DanielW

  • EOS M2
  • ****
  • Posts: 245
    • View Profile
Re: 5D III Dynamic Range
« Reply #38 on: February 09, 2013, 01:09:07 PM »
A little while ago a friend asked me what she needed to know about DSLRs in order to take "good" pictures.

I created an extensive list. I never mentioned dynamic range. Shame on me.

Not really. A good friend you are...
Like a number of people have said on this forum (and I am sure many have said so on other forums elsewhere too)  that it is the finger that half clicks and then completes the click is all that matters...

They have not? Huh!



Aglet

  • 1D Mark IV
  • ******
  • Posts: 814
    • View Profile
    • A2B Art
Re: 5D III Dynamic Range
« Reply #39 on: February 09, 2013, 01:21:43 PM »
.. For some reason it is deemed highly important to some Nikon users, while some Canon users find it utterly unimportant. I expect that situation to flip if ever Canon release a 15-stop DR DSLR.

+1
but I aint waitin' for Canon to produce such.  Even 11 or 12 stops of CLEAN DR provides a workable imaging tool.

The aspect that's constantly under-rated by most who crow about the adequacy of Canon in comparison to the sensor superiority of other mfrs is the damn pattern noise.
I'd have kept my 5d2 and 7d with the DR they had if only the base iso noise didn't look like a plaid overlay or picket fence, respectively.

..We continue to be treated to the same me tired, contrived examples that are completely devoid of any artistic value ..

Use your imagination.
I'ts nice to know you have gear that CAN do whatever you want in a challenging situation or to allow recovery from an under exposure error than to have gear which has technical limitations that would require more labor and effort to overcome.
If you're gonna pound that DR nail you could beat on it with that model 5d3 wrench or you could smack it down in one clean blow with a d800 hammer.

Aglet

  • 1D Mark IV
  • ******
  • Posts: 814
    • View Profile
    • A2B Art
Re: 5D III Dynamic Range
« Reply #40 on: February 09, 2013, 01:23:53 PM »
To me, the primary benefit of the technology is that if you have a bad exposure for whatever reason, you can lift the shadows a bit and fix it.  You clearly have more leeway in this regard with the later Nikon gear than the Canon stuff and that is mainly due to the pattern noise, not the random noise based DR difference (that DxO measures).  This is a huge benefit for the Nikon gear; however, in fairness to Canon, I have never really had a problem with their equipment for any real world work I have done.

OK, you  DO get it. :)

ishdakuteb

  • EOS M2
  • ****
  • Posts: 324
    • View Profile
Re: 5D III Dynamic Range
« Reply #41 on: February 09, 2013, 02:47:43 PM »
he will not being able to simply because all of this images are crappy, from color to composition through number of years being photographer unless he goes out there and steal one.  you can tell by taking a look at his images via provided link in previous post.

sigh... i thought that i am saying i am getting tired of talking to these type of people, who have been talking about the same topic for a year (over a year when including other forums like dpreview, photographyblog, etc...), nothing else...

proof:  http://forums.dpreview.com/forums/post/50383414

umm.... "IARNA International anti-banding and read out noise Association" never heard of it... might be there are two member one is president and another one is vice president LOL...
« Last Edit: February 09, 2013, 03:16:52 PM by ishdakuteb »

Radiating

  • EOS M2
  • ****
  • Posts: 323
    • View Profile
Re: 5D III Dynamic Range
« Reply #42 on: February 09, 2013, 03:22:16 PM »
Folks,
Greetings,
As we all know the dynamic range of 5D MK III is not better than its counterpart Nikon D800. I was wondering if this deficiency could be addressed by using single shot HDR for batch processing. Has anyone tried HDR batch process to improve dynamic range, what would be good software for this purpose or any other comments. Thanks in advance.
Raj
:)

Dynamic range is the result of the highlight saturation ceiling and the noise floor. You can't use software to improve it because you can't create data that doesn't exist in the first place.

Canon cameras have a significant noise floor in the shadows, due to the way the sensor data is read that adds noise to it. Sony/Nikon's method of reading the sensor is inherently less noisy so they have more dynamic range.

Canon improved the 5D3 dynamic range by about a stop from the previous generation, and you don't really need more DR unless you're doing architecture or landscape, in which case you can use multi shot.

canon rumors FORUM

Re: 5D III Dynamic Range
« Reply #42 on: February 09, 2013, 03:22:16 PM »

RMC33

  • 7D
  • *****
  • Posts: 424
    • View Profile
Re: 5D III Dynamic Range
« Reply #43 on: February 09, 2013, 03:27:42 PM »
Jee another thread that was going good.. with a great video offering up some wonderful techniques... that has now sunk to the bickering of other threads.

Aglet

  • 1D Mark IV
  • ******
  • Posts: 814
    • View Profile
    • A2B Art
Re: 5D III Dynamic Range
« Reply #44 on: February 09, 2013, 06:49:33 PM »
PBD: I called Mikael an irrelevance, which by my understanding of the word is not an insult, because his comments in these threads are so often irrelevant.

IMO, you are directly disrespecting the man with that statement.
You seem to have an adequate grasp of english to be able to understand that you did not refer to his IDEAS with that comment, which you've repeated more than once.

If you were to do so, you'd have said, "Hey, your opinion or comment is irrelevant."
FWIW, irrelevance is not even a real word, but you're putting it into a context that comes across as disrespectful.
I also find your badgering of people for their raw files a put-off as well.
What do you hope to accomplish from someone else's raw files?
If you want samples of D800 or 5d2 files, you can find them, shot under very controlled conditions, on various sites.  Imaging-resourse is a good example.
Download them and play with them.  See how they compare for yourself.
I see little need to duplicate their efforts to appease such requests.

PBD: Like his eye comment..
you did not need to respond to that, he didn't contradict what you said, he merely added to it

PBD: Can you show us some images where the DR of the Canon just can't work but the slightly higher DR of the Nikon saved the day and made a worthwhile image? It appears not.

yes.
I can show you where, if I had shot Nikon, it would have provided an image that _I_ would have liked much better.  My client didn't care.  I'm a much fussier customer than my clients.

PBD: However, what we are constantly asking for is examples of where the Canon equipment lets you down in real world shooting,

Right, and when I actually started a thread on this exact topic a couple days ago I got 2 pages of crap from various fan-boys on this site who weren't patient enough for some examples to be posted by me or anyone else.  Very few others even braved the flames to actually try to participate.  I feel bad for them because they were interested in the topic but were likely turned off by the resulting brawl.  I certainly was.
Why should I, and some others, waste our time trying to share our experience and knowledge when the response is ridicule or disrespect?  Many of the people on this forum have something worthwhile to contribute, even if they're newbies.  However, I've often seen too many responses in a condescending attitude from certain participants; that does not create a healthy atmosphere for participation.
And it must the crankiness that comes with old age but I'm gonna call it when i see it.

PBD: where maximising the lower capabilities of the Canon are not enough, where worthwhile real world images could only be shot with a Nikon. In truth I have never been shown one, not one single real world image, only lots of artificially set up and badly exposed "tests".

What I've generally stated is that the superior sensor performance of the other mfr's cameras allow a lot more creative freedom in post and simplified shooting in the field because we don't have to try wring the best from a mediocre product.
I and others have also stated that it's certainly a benefit to be able to recover from an underexposed shot, or to be able to compress the contrast in a shot with lots of dynamic range, so that the shadows aren't wrecked by pattern noise.  There are enough real-world situations where this is an issue.
I've made prints from my 5d2 and 7d that people think are great!  Sure.  I wonder what they'd have thought about the same images if I could have lifted the dark levels a little like I wanted to, so a hint of the detail hiding there could be seen - preferably without the stripes.

I don't know who said it but taking your dark print outside to view in sunlight to see the shadow details is not often an appropriate option!
This is when shadow range compression (lifting) is beneficial and when patterned noise cameras, like most contemporary Canons, are not up to the task in the more extreme cases.

Artificially contrived tests are what you use when you do research.  It's called a controlled environment to minimize variables.  It's daft to suggest such tests do not have merit, people need to understand how they work and how the results will apply to real world shooting.  If someone can't understand that, it's gonna be pretty hard to teach them much of anything.

PBD: I shoot a lot of very high dynamic range images, a Nikon would not serve me better.
That's true, but it is irrelevant! ;)
HDR can even be done with an 8-bit compact camera.

canon rumors FORUM

Re: 5D III Dynamic Range
« Reply #44 on: February 09, 2013, 06:49:33 PM »