Dynamic Range & Camera IQ

[paulgmccabe]

I'm a 550D user who has just yesterday ordered a 7D. I've never even looked through the viewfinder on a FF camera, let alone held one.

I do know my way around my camera, but on these boards I keep hearing people talking about Dynamic Range when comparing cameras or making wish lists of improvements.

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

Cheers.

 

[poias]

This visually explains quite a bit: http://www.flickr.com/photos/dcdead/7091087059/#

 

[V8Beast]

Dynamic range is a very important element in overall image quality, no question. That at said, I find it somewhat surprising that people rarely discuss the impact of tonal range on image quality on your typical online forum. A lot of what I shoot has varying shades of grays and blacks, and in order for the sensor to pick up the subtle shifts in tonality, having a wide tonal range trumps dynamic range by a large margin. Of course, if you're shooting a high-contrast subject, DR is king, but the image can still benefit from a broader tonal range nonetheless.

At the risk of getting shot by the APS-C lynch mob, I'd say that the biggest advantage of a full-frame sensor over a crop - for what I shoot, at least - are the improvements in tonal range.

 

[dtaylor]

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

From Wikipedia: Dynamic range, abbreviated DR or DNR,[1] is the ratio between the largest and smallest possible values of a changeable quantity.... In photography the quantity is luminous intensity (brightness) and the end points are the beginning of pure black and pure white. It's generally measured and expressed in stops of exposure (one stop more being twice as much light; one stop less being half).

For example: you take a picture of a beach sunset with the sun, clouds, ocean, beach, cliff, and a cave all in the scene. As you get closer to the sun details get brighter and brighter until they are pure white. As you get closer to the cave details are darker and darker until they are pure black. The range between the point of brightness where detail transitions to pure white and the point where detail transitions to pure black is the dynamic range.

It impacts you because there are scenes in real life with a brightness range that exceed a sensor's ability to record detail. So you either have to expose to lose some detail to pure black, pure white, or both; or shoot multiple exposures and merge them manually or using HDR software.

I'm going to be blunt: like most things in photography, differences between equipment are exaggerated beyond reason in forums and discussions. It's human nature. If you look for it in a RAW converter, there is a consistent difference in DR between the Canon 7D and the 5D2/5D3 of approximately a stop. The D800 adds approximately 1-1.5 more. These estimates depend greatly on the level of shadow noise you are willing to accept for your purposes, something fans in forums never seem to take under consideration. They will howl and cry about differences seen on a monitor zoomed to 200%, and ignore that the differences are nearly impossible to see when you are purposely looking for them in a 20" print.

The difference is there, and it can be noticed in some situations. But generally speaking you will still get the shot, you might just have a little more noticeable noise / less detail at the extremes. Also, generally speaking, for landscape shots with really impressive DR / noise / detail you will be shooting and blending multiple exposures with any current body.

 

[preppyak]

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

Cheers.

Most explanations are gonna be a little technical, but I think this one does a decent job

http://www.cambridgeincolour.com/tutorials/dynamic-range.htm

As for comparisons, sadly most of what I have seen are video tests, I can't seem to find any straight photo tests. The reality is that it would be extremely difficult to test, as getting an objective test would mean you'd have to use the same lens on each camera, at the same settings at the same time. I dont know a lot of people with multiple 14-24's, a D800 and a 5D Mark III. The moral of the story is generally that the D800 has more dynamic range, it allows you to photograph scenes that have darker shadows and brighter lights than the 5d Mark III. But, that is one aspect of a photo, and depending on what you are shooting, there are many, many other aspects.

 

[jcns]

I'm a 550D user who has just yesterday ordered a 7D. I've never even looked through the viewfinder on a FF camera, let alone held one.

I do know my way around my camera, but on these boards I keep hearing people talking about Dynamic Range when comparing cameras or making wish lists of improvements.

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

Cheers.

did you think 7d is a full frame camera?

 

[K-amps]

I have to think that Canon sensors cannot be all that bad in DR and that coding of RAW data also has a part in it. Sort of reminds me of the Audio compression days when Sony introduced ATRAC compression and while in the Engineers minds, the compression was efficient, the critics hated the sound. They said it lacks something, Sony Engineers countered that they only masked the inaudible parts...

As I look at that D800 pic linked above, in the original shot, it seems underexposed; however the information in the shadows was not "thrown away" and hence retrievable....

Could it be that Canon is throwing away RAW data they think is unusable while users are pulling a lot of it in PP?

 

[dtaylor]

This visually explains quite a bit: http://www.flickr.com/photos/dcdead/7091087059/#

I guarantee you that if you posted that shot but changed the EXIF to say "Canon 7D" people would do nothing but complain about the noise levels ;D

That is better than you could do with a 7D or 5D3, but not by leaps and bounds. I've lifted shadows on a lot of Canon RAW files. For all the howling and complaining about the DR/noise differences between Canon and Nikon sensors (at this time), lifted Canon shadows look about the same in ACR except for banding noise. Banding is what caps how far you can push it. Depending on the shot and target print size you end up only going to 60-80 on the ACR slider, which is where the banding starts to become apparent, where this user reports maxing it out to 100. So the difference is there, but it's not the end of the world.

That's not to say I don't admire the D800 and eagerly await Canon's answer. The 7D and 5D2/5D3 are 24-30" print cameras for critically reviewed landscapes, with the 5D2/5D3 being a bit better at 30". The D800 easily breaks into the >40" range.

 

[molnarcs]

That's exactly my experience working with NEF files from the d7000 - approximately 30pts more on the shadow slider compared to the 7D. That's roughly one stop. For architecture/interiour work (almost always high contrast scenes) this is quite significant. Basically, it's the difference between being able to stay within Lightroom for all the post-processing I do, or having to take a trip Photoshop. This affects about 40-60% of my work.

That said, for JPEG shooters it doesn't matter at all, and it doesn't matter to you if you don't rely on pushing shadows in LR too much. Which reminds me, that it looks like that 2EV difference between Canon and Nikon sensors (as measured by DxO) translates only to about 1 stop in practice when using Lightroom.

This visually explains quite a bit: http://www.flickr.com/photos/dcdead/7091087059/#

I guarantee you that if you posted that shot but changed the EXIF to say "Canon 7D" people would do nothing but complain about the noise levels ;D

That is better than you could do with a 7D or 5D3, but not by leaps and bounds. I've lifted shadows on a lot of Canon RAW files. For all the howling and complaining about the DR/noise differences between Canon and Nikon sensors (at this time), lifted Canon shadows look about the same in ACR except for banding noise. Banding is what caps how far you can push it. Depending on the shot and target print size you end up only going to 60-80 on the ACR slider, which is where the banding starts to become apparent, where this user reports maxing it out to 100. So the difference is there, but it's not the end of the world.

That's not to say I don't admire the D800 and eagerly await Canon's answer. The 7D and 5D2/5D3 are 24-30" print cameras for critically reviewed landscapes, with the 5D2/5D3 being a bit better at 30". The D800 easily breaks into the >40" range.

 

[dtaylor]

That's exactly my experience working with NEF files from the d7000 - approximately 30pts more on the shadow slider compared to the 7D. That's roughly one stop. For architecture/interiour work (almost always high contrast scenes) this is quite significant. Basically, it's the difference between being able to stay within Lightroom for all the post-processing I do, or having to take a trip Photoshop. This affects about 40-60% of my work.

Granted that it can definitely result in more post work and/or a different shooting style if you're always shooting at the edge. I don't want to completely discount the advantage. I just get tired of those who act like Canon should stop making cameras until they improve their sensors

 

[unfocused]

Agree 100 percent with DTaylor's comments.

Adding a very biased personal perspective.

In the film days, this was called the Zone System and it was popularized by Ansel Adams. The basic concept remains the same. Film (now sensors) and the final medium (photographic prints, photo-mechanical reproduction and now, monitors) cannot reproduce the full range of tones that exist in nature.

You have two choices when trying to capture scenes with a broad range of tones: You can "clip" the highlights/shadows by letting them go either completely white (for highlights) or completely black (for shadows). Or, you can try to compress the tonal range into something that is reproducible, preserving detail in both.

Right now, the fad in photography is to try to compress as much of the range as possible, creating "High Dynamic Range" scenes, which is a misnomer, because the range is not actually "higher" instead it is compressed and often looks very artificially so, like the example shown early in this thread. Some people think that looks great. Others think it is artificial looking and excessively dramatic.

Ansel Adams was a big proponent of compression and his Zone System was built entirely on teaching people how to compress the tonal range of film and paper. However, Adams' goal was to try to reproduce a tonal range that looked natural and reflected how the human eye (at least in his view) actually saw the scene. Many of today's HDR images make no attempt to reproduce the scene as a human eye might see it, but instead opt for overly dramatic misinterpretations.

The counterpoint in photography is the image that allows areas to go pure white and pure black while retaining the important detail in the middle tones.

As with most things in life, the decision should be personal and should be based on what your goals are. In my view, moderation is probably the best approach. But, that's a subjective decision.

The appeal of a wider dynamic range in sensors is that the sensor can preserve shadow and highlight detail in the file that can be brought out in the final image. Again, as DTaylor has pointed out, this can often be done in post-processing in Photoshop. I frequently use multiple layers of smart objects, adjusting the highlights and shadows in Camera Raw, masking certain areas and combining them in Photoshop. But then, I'm not a big fan of the HDR look and in most cases prefer the challenge of composing and capturing scenes within the limitations of the camera's existing dynamic range.

 

[jrista]

This visually explains quite a bit: http://www.flickr.com/photos/dcdead/7091087059/#

I guarantee you that if you posted that shot but changed the EXIF to say "Canon 7D" people would do nothing but complain about the noise levels ;D

So true!

That is better than you could do with a 7D or 5D3, but not by leaps and bounds. I've lifted shadows on a lot of Canon RAW files. For all the howling and complaining about the DR/noise differences between Canon and Nikon sensors (at this time), lifted Canon shadows look about the same in ACR except for banding noise. Banding is what caps how far you can push it. Depending on the shot and target print size you end up only going to 60-80 on the ACR slider, which is where the banding starts to become apparent, where this user reports maxing it out to 100. So the difference is there, but it's not the end of the world.

Actually, there is one other aspect of noise in most Canon cameras (the 1D series seems to be largely, though not entirely, excepted from this): Blotchy color noise in the deep deep shadows! I actually am not bothered as much by the banding/hatching noise in Canon cameras as much as I am by that really nasty, blotchy, muddy color noise...which primarily seems to exhibit reddish...when you really reach in deep into shadows with a .CR2 file.

I think the comparison of the D800 to 5D III really exemplified that best: http://www.fredmiranda.com/5DIII-D800/index_controlled-tests.html. This was the review that demonstrated to me the true benefit of the D800's extra couple stops of DR. But if it wasn't for the blotchieness in the 5D III shadows....even if there was some banding...I could still WORK WITH IT! With all that nasty red and blue color-noise blotchieness, though, you have to apply some aggressive extra NR, which eats away at the detail....detail you can see is there...just underneath the noise...probably just as good as the D800....but which will be blurred away by the extra NR needed to clean up that muddy color noise. If it wasn't for all that color noise, pattern noise wouldn't be so bad. There are ways to clean up patter noise....astrophotographers have been cleaning it up with dark frames and bias frames and the like for years. And you can clean it up pretty nicely, too. But color noise....meh. I think that is Canon's true bane in the shadows....blotchy, muddy, red and blue....NASTY!!

 

[dtaylor]

I have a completely different view, with the D800 I can act completely different than what I can do with my Canon cameras, which I have shown a number of times at dpreview
With Canon, you must have the camera on a tripod and take 2-3-4 exposures,

There is no way the D800 sensor can record so much that you need 3 or 4 exposures on Canon to match. That's ridiculous. You might need 2 if you absolutely must have the same noise characteristics in the shadows. But if we're being that picky, at 2 you will get a better final image.

BTW, with AEB I hand hold multiple exposure shots all the time. It doesn't work for action, but I never need it there.

with the D800 I can have the camera by hand and take one shoot and then working with the same raw file and produce one after shadows and the other for highlight, midtone, and then put them together.
A certain difference in freedom of taking photos and with a good results because of 14 stops DR in the d800 compared to 11 stops in 5dmk2 and that not including pattern noise

3 stops? No. Not if you do the same thing with the Canon file (develop 2 versions from the RAW and merge).

The D800 is better, yes, but this is the kind of exaggeration I'm talking about.

 

[Mt Spokane Photography]

I'm a 550D user who has just yesterday ordered a 7D. I've never even looked through the viewfinder on a FF camera, let alone held one.

I do know my way around my camera, but on these boards I keep hearing people talking about Dynamic Range when comparing cameras or making wish lists of improvements.

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

Cheers.
[/quote


You can take a look here http://forums.dpreview.com/forums/read.asp?forum=1021&message=42598514
There is a huge difference between Canons sensors in 5dmk2 ,5dmk3 and the Sony sensor inside d800
First, Canon has about 12 times higher read out noise than Nikon d800, second, there are no banding or patter noise in the Sony sensors.
This means that you can handle the Nikon camera different and you can expose after the highlights and then adjust the pictures in middle tones and low levels . If you do the same with Canon you have a results as the link to dprevie shows. With Canon you have 2 alternative 1. choose to exposure after the high lights and to keep the high lights and there will be a worse outcome in the shadows 2. if you chose to expose Canon more plentiful the results are earlier high lights clipping.

I'd beware of anyone out to prove a point with deliberately doctored images in a internet forum. The 5D Mark II image is either doctored or from a damaged or defective camera. A 5D mark II does not have poor images like those shown.
However, I will say that Low ISO, the D800 does have a lot of usable DR, but at High ISO, it has less than the 5D MK III. It tends to be more important at very high ISO's for me than at low ISO's, but thats just me.
Here is a image from my D800 which really suprised me with its ability to capture a interior along with a bright sunny outdoors.

 

[dtaylor]

I initially skipped this one because the whole thing was quoted...

You can take a look here http://forums.dpreview.com/forums/read.asp?forum=1021&message=42598514

Sorry, I question that test and presentation. I don't see banding that bad on my crop body.

 

[Mt Spokane Photography]

?

I'd beware of anyone out to prove a point with deliberately doctored images in a internet forum. The 5D Mark II image is either doctored or from a damaged or defective camera. A 5D mark II does not have poor images like those shown.
However, I will say that Low ISO, the D800 does have a lot of usable DR, but at High ISO, it has less than the 5D MK III. It tends to be more important at very high ISO's for me than at low ISO's, but thats just me.
Here is a image from my D800 which really suprised me with its ability to capture a interior along with a bright sunny outdoors.

Do not get rude, I have 2 x 5dmk2 and they show this pattern noise, I have tested cameras since 1978, scanners since 1990 and digital cameras since they come so I know what Im doing.
5dmk2 mk3 has 11 stops DR and then the pattern noise are not including in the measure. Nikon has 14stops
http://www.sensorgen.info/CanonEOS_5D_MkIII.html
http://www.sensorgen.info/CanonEOS_5D_MkII.html
http://www.sensorgen.info/NikonD7000.html
http://www.sensorgen.info/NikonD800.html


Rude? Are those your images? Posting obviously altered images seems to me to be rude.
I have been into photography for quite a few more years than you, started on digital earlier, and have owned or used many digital cameras, including two 5D Mark II's, two 5D Mark III's and a D800.
I've never seen the banding like that shown in the images, except on forums where the images were under exposed and then pulled up several stops. You have to really manipulate the images in post, or have a defective camera. If your 5D Mark II's produce images like those with a proper exposure, and you did not send them back ???
I'd prefer to see information from and intrepreterted by the actual tester.
Sensorgen.org charts are derived DXO information, just a repeat of the same info.
"One source, the one which is used for the data on this site is DxOmark.com."
DXO's own chart illustrates the higher DR for the 5D Mark III at high ISO's and the excellent DR for the D800 at low ISO's. It pretty well matches what I've seen with my cameras.

 

[jrista]

I have a completely different view, with the D800 I can act completely different than what I can do with my Canon cameras, which I have shown a number of times at dpreview
With Canon, you must have the camera on a tripod and take 2-3-4 exposures,

There is no way the D800 sensor can record so much that you need 3 or 4 exposures on Canon to match. That's ridiculous. You might need 2 if you absolutely must have the same noise characteristics in the shadows. But if we're being that picky, at 2 you will get a better final image.

BTW, with AEB I hand hold multiple exposure shots all the time. It doesn't work for action, but I never need it there.

with the D800 I can have the camera by hand and take one shoot and then working with the same raw file and produce one after shadows and the other for highlight, midtone, and then put them together.
A certain difference in freedom of taking photos and with a good results because of 14 stops DR in the d800 compared to 11 stops in 5dmk2 and that not including pattern noise

3 stops? No. Not if you do the same thing with the Canon file (develop 2 versions from the RAW and merge).

The D800 is better, yes, but this is the kind of exaggeration I'm talking about.

read it again
I did not say it is the same thing, I wrote that the result of one exposure and then developed in 2 ways give a good results and i do not need a tripod, freedom

There is something so seriously wrong with that 5D II image. There is red banding and FPN throughout the entire image, even the highlights. There is NO WAY that image was properly exposed in the first place. I've seen that kind of banding in sample black-frame 5D II images, but only after pushing exposure by about 8 stops, or by opening the image in PS/ACR and using the levels tool to drop the white point to within a fraction of the black point. The 5D II's maximum saturation is 64600e- and its read noise at ISO 100 is 28e-. That is a ratio of 2307:1! No friggin way your going to see that kind of banding with a minor curve bump like you've demonstrated. That exposure would have to be underexposed by many stops to exhibit like that. I cry fowl!!

BTW, banding noise IS READ NOISE. Read noise is a bit too specific, as that generally refers to the noise introduced by the ADC. I tend to refer to electronic noise, which comes in a variety of forms, but is all very low relative to maximum signal (even when there is 28 electrons worth!) Its a bunch of bullhonkey when you mentioned before that banding is not taken into account when determining DR...absolutely it is, its simply that DR is computed as the AVERAGE of read noise to maximum saturation. Since it is the average of electronic noise, that that would mean your computing DR as the ratio between what would roughly be 14e- and 64600e-, a ratio of 4614:1 (or 11.something stops), so some electronic noise...such as banding...will show through in very deep shadows.

 

[ishdakuteb]

Can somebody please point towards a good explanation of DR and also comparisons between similar cameras (e.g. D800 and 5D M2/3) so I can understand it better and how it impacts me? Also any explanation of camera IQ would be helpful.

The difference is there, and it can be noticed in some situations. But generally speaking you will still get the shot, you might just have a little more noticeable noise / less detail at the extremes. Also, generally speaking, for landscape shots with really impressive DR / noise / detail you will be shooting and blending multiple exposures with any current body.

I have a completely different view, with the D800 I can act completely different than what I can do with my Canon cameras, which I have shown a number of times at dpreview
With Canon, you must have the camera on a tripod and take 2-3-4 exposures, with the D800 I can have the camera by hand and take one shoot and then working with the same raw file and produce one after shadows and the other for highlight, midtone, and then put them together.
A certain difference in freedom of taking photos and with a good results because of 14 stops DR in the d800 compared to 11 stops in 5dmk2 and that not including pattern noise

i am not a pro. photographer or whatever. i am a software/system guy. however, i am willing to match your d800 pix that you have shown with just one shot (do not need multiple shots) using my canon 7d (do not need to use my 5d mark III), except the sharpness (you should know why).

i am the one who just recently start to learn photography (little bit more than 6 months starting from a t3, move up to 7d and buy another 30d for learning purpose, and recently again bought another 5d mark III for low light support. i can prove it that i am not a pro but i can match your image with just one shot)

both canon and nikon have their own pros and cons. they are both good. i really hate when coming to talk about technical data and apply that to favor one over another. if i take pictures with two different cameras and you can recognize which one is nikon, which one is canon then i will BUY YOUR THOUGH.

note: "Here is a image from my D800 which really suprised me with its ability to capture a interior along with a bright sunny outdoors". image will show all details when taking with right exposure and white balance (thought that you should know that).

-- au revoir

 

[dtaylor]

read it again

I don't need to. Several people have now pointed out that they do not see banding that bad on 5D's, and I don't see banding that bad even on a 7D. Don't blame the camera for mistakes you might have made in your test, processing, or presentation.

 

[dtaylor]

Do not get rude,

He wasn't.

I have 2 x 5dmk2 and they show this pattern noise, I have tested cameras since 1978, scanners since 1990 and digital cameras since they come so I know what Im doing.

And that means nothing if your results cannot be replicated.