5DS-R DR test on DPReview

[rs]

The advantage Sony has at low ISO dwarfs that of Canon at high ISO. At ISO 100, Sony has 1 or 2 stops, if not more, of an advantage. At high ISO, the difference is maybe half a stop, if that.

It's all about relevance. Yes, the high ISO noise difference is not as big as the low ISO DR difference. But which is more useful for most people?

Using random picture selection from flickr, retter DR between ISO 100 and 400 will benefit people the most (or the most people.)

So most pictures on Flickr are shot at < ISO 400 and have the shadows pushed by > 4 stops? I'm sure you'll find the odd photo like that, but certainly not most.

I haven't done the research, but I'd be staggered to find more photos like that than high ISO photos on Flickr.

It's about relevance.

 

[ritholtz]

The advantage Sony has at low ISO dwarfs that of Canon at high ISO. At ISO 100, Sony has 1 or 2 stops, if not more, of an advantage. At high ISO, the difference is maybe half a stop, if that.

It's all about relevance. Yes, the high ISO noise difference is not as big as the low ISO DR difference. But which is more useful for most people?

Using random picture selection from flickr, retter DR between ISO 100 and 400 will benefit people the most (or the most people.)

So most pictures on Flickr are shot at < ISO 400 and have the shadows pushed by > 4 stops? I'm sure you'll find the odd photo like that, but certainly not most.

I haven't done the research, but I'd be staggered to find more photos like that than high ISO photos on Flickr.

It's about relevance.

Hi rs,
Can you please tell me if shadows slider is in terms of stops or levels. Canon has shadow slider range -5 to 5. Does it mean each level is a stop. DXO Optics has shadow slider ranging from -100 to +100. DRreview keeps on talking about pushing shadows by few stops but I am yet to find out how to do that.

Another thing I noticed based on their ISO in-variance test, pushing exposure +5 stops for ISO 100 image is nothing but making it a ISO 3200 image. What is the point of under exposing a lot using high DR camera and turning it into high ISO image. If some one wants to make use of high DR, they need to properly expose image at ISO 100 and make use of sensor DR right?

 

[ritholtz]

Since DPR married DXO, I have been a bit skeptical about the full objectivity of their testing results and reports.

DPR married DXO? Please.

dpreview have been measuring dynamic range and noise since their very first review.

All that they have done now they've just provided a new way for people to compare cameras that didn't exist before. Unfortunately for Canon, it is now easy to see how bad the sensor noise is in their images compared to others. Don't blame dpreview for this nor DxO - unless you are the type to shoot the messenger.

Why do DPR/DXO measure and emphasise DR at ISO 100 and not at 3200?
Simple.
They're looking to measure the best DR that the camera can deliver and this generally occurs at ISO 100.

And yes, more people take photos at ISO 100 - 800 than at ISO 1600 and above.

Why does some one push ISO 100 by +5 stops and turn it into ISO 3200 image. If some one wants to take advantage of high DR at base ISO 100, they need to expose image properly right? How does underexposing is going to help in terms of capturing high DR. Just trying to understand.

 

[jrista]

There is a degree of insincerity in the "DR is everything" crowd. Everyone wants more DR, and they do too. But that's not the whole story. There is a bit of Canon hate in there.

I don't think this is it. It seems pretty clear to me that many of the vociferous DR advocates are upset for basically one reason: they think Canon is intentionally hobbling new cameras with older sensor tech using its market dominance, taking extra profit and planning to up-sell its customers later. The arguments I hear repeatedly are (1)other companies are making "better" sensors in reasonably-priced bodies, so if Canon can do the same, why don't they? (2)If Canon can't make "better" sensors at reasonable prices, why don't they just buy from someone who does?

I understand these arguments, and even think they might be true. I just don't think it's any different from business practices in any other commodity market. Any market leader in their position would do the same, there's nothing illegal about it, and the shareholders like the extra profit. In other words, even if it's true it's my problem. I have the option to buy from another brand if I choose, and that's pretty much the only choice I have. I wish it were different, but I've got more important things to be upset about.

Aye, there are definitely people who believe Canon is intentionally hobbling their equipment. I don't think those people hate Canon, although those who truly believe that are (from their own perspective) justifiably angry, for the very reasons you stated.

I find that much less likely than simply that Canon hasn't figured out (for whatever reason) how to improve their DR more than they have, and are at the current time simply incapable of doing better. If their most recent sensor patents are any indication, it seems Canon is investing in sensor innovation in a different direction, which is neither surprising nor bad. Different avenues of research are a better way to innovate than if every company pursues the same avenues of research. I just hope the technology doesn't end up sitting in a corner of someone's office somewhere collecting dust.

 

[9VIII]

...Again I think it is naive to think that everyone who wants more DR is just plain and simply a Canon hater.

Jrista, again you and others have somehow come to the conclusion that people here are anti-progress. It couldn't be further from the truth.

I look at the clean shadows and high ISO range with envy, I almost bought an A7Mk2 last month, being held back only by the lack of native Sigma FE mount lenses.
It seems that everyone on either "side" of this debate (even though technically we're all on the same side) thinks people on the other "side" are practising wilful ignorance.
It seems like the whole thing is a giant meta-narrative of people looking at a few strong or sarcastic opinions and labelling an entire group based off of what a few people say.

I have posted more than a few "sarcastic" comments here lately, and then a few weeks ago I took the same type of comment from someone else as "not sarcastic" and began a thorough rebuttal to their comments.
Only to discover the whole thing was said in jest.
If I can misread sarcasm from people who hold entirely the same opinion as me, it follows that a large portion of the joking that has been going on here lately is being taken literally.

No one in their right mind wants "less dynamic range" and I assure you that 99.9% (well, maybe 99.8%) of the people posting here are entirely sane.
I assume that the same applies to other communities, that the majority of the comments implying that Dynamic Range is the defining factor in the quality of photography equipment are an exaggeration and the authors of those comments don't expect anyone to take those comments completely literally.

I will stop this behaviour in my own comments, I will no longer speak in jest about Dynamic Range.
This is my official response on the subject, to be taken entirely seriously:
Canon performs poorly compared to the competition in the area of sensor noise, and delivers less Dynamic Range.
They also have the best lens selection and quality in the industry, and their controls and user interface are second to none.
I wish that Canon would continually improve all aspects of camera performance, including Dynamic Range and Sensor Noise.

 

[Orangutan]

Could you share your sampling and analysis methods?

My sampling was simply look at pictures presented to me when I visit the website and (I think) clicking on "explore". I did this just now, and the top 20 result in:

No ISO: 9
ISO 100 3
ISO 125 1
ISO 200 2
ISO 400 3
ISO 800 2

Now why don't you tell me how my selection process is flawed and biased ...

To do that I would need to know how Flickr chooses photos for the front page, and whether the "explore" feature has an algorithm behind it. I can't say if or how your selection process is flawed and biased. I also can't say that it isn't. For casual opinions your selection methods are fine; however, to draw a conclusion with scientific validity it's not sufficient to say "I have no reason to believe they're biased," you need to be able to say "I have strong reason to believe they're UNbiased." And to make that statement you'd need to know the mechanism used to bring those photos to your attention.

 

[LOALTD]

Not to worry, if the 1Dx Mk II ends up having class-leading DR, the spec will finally matter again.

Base ISO DR of Canon DSLR's already exceeds the DR of all output medium I know of. If the next batch of Canon sensors make Exmor sensors look prehistoric from the DR point of view, I fail to see the advantage for 99% of photographers. While it is a metric which cameras can be measured by, and improvements don't hurt, this aspect of camera performance has been blown out of all proportion.

I have shot tens of thousands of frames with my Mk III, I have hundreds of examples where pushing the shadows in an ISO 100 shot results in noise. The noise is visible in all output mediums. This is ridiculous for a modern camera.


I primarily do landscape/adventure/timelapse photography, so maybe my use case is unique. If I shot sports or people I don't think I would much care about DR.


As it is, I hate having to merge exposures just to get a mild shadow push in a high DR scene. It becomes even more labor intensive if it's a high-DR panoramic (as they usually are)

 

[Sporgon]

I've always said the issues are when you try to expose for the actual light source, and it's not something that I normally do. However yesterday I thought I'd see how far I could go in holding the sun about three hours before sunset, so it was still very bright.

I shot this as a single exposure. The bottom of the sun itself is clipping just into the top of the frame. There is nothing blown. ( Well there are one or two pixels in the sun itself). So in order to expose for the sun the picture is way underexposed, including the rest of the sky.

The first picture is the straight raw converted to png.

To get all this into perspective the second one is a two stop lift in raw then converted. The third one is how I would present this as a picture in terms of exposure balance. There is no noise etc etc. I post this to show how much a two stop lift is in reality.

Of course jrista will come along and say there's not much DR in this scene anyway

 

[jrista]

...Again I think it is naive to think that everyone who wants more DR is just plain and simply a Canon hater.

Jrista, again you and others have somehow come to the conclusion that people here are anti-progress. It couldn't be further from the truth.

Once again, where do you get that from the statement you quoted? ??? I said that not everyone who wants more DR is a Canon hater...that has nothing to do with claiming people are anti-progress. My comments here are about this communities reaction to anyone brining up DR...it's a vitriolic reaction, and if there is any hate, it is hate for the people who want more DR and talk about it (even if it's in a complaining manner, there should be a different reaction.) That is a matter of people reacting to people...not about progress. At least where I come from, people have the right to freedom of speech. Here on these forums, your free to speak unless its about DR...if you do speak about DR, then everyone else is free to hate on you, belittle you, and be as nasty to you as much as they like. That doesn't strike me as the kind of mature behavior one would expect from a bunch of adults.

There is a truly vitriolic and hateful sentiment here on these forums for anyone who ever brings up DR. There are some good heated debates about the issue on DPR forums, and some good old flame wars between brand fanatics, but I haven't often seen those debates degrade into the kind of vitriol and hate I see here...here on CR, it's taken on a persistent dark and nasty form. I think that reaction prevents any kind of reasonable discussion on the subject...and since it IS one of the areas where Canon technology lags, it's going to be an area that people want to discuss. This community could have a wholly different atmosphere, a brighter one that isn't continually mired by nasty comments and endless DR wars, if everyone could discuss the subject without bringing in the hate.

I prefer to avoid these threads for the most part these days because of the reaction locals have to anyone bringing up DR. No one can discuss this subject rationally, it's all just hate...not hate for the brand...hate for the individuals involved. I would much rather people hate the brand than hate the people...I think hating Canon would be far preferable to the hate people have for each other on these forums just because they have differing opinions about a piece of technology. Why is it that someone like K is upset by the fact that someone else "hates" K's preferred brand? Why does that matter? More so...why is that grounds for K to hate that someone else in turn (not saying he actually does...but there is a lot of that that seems to go on around here)? It's a brand...an unfeeling, inanimate piece of equipment...here today, gone in a year or two, replaced by something better.

This community needs to stop hating on each other for having differing opinions... :-\ (And truly, I think there is significantly less "brand hate" here than some people seem to think...there is undoubtedly "brand frustration" and certainly some strong desires for Canon to produce a better sensor...but I really don't think there is much brand hate, not as much as there is people hate by any means.)

I will stop this behaviour in my own comments, I will no longer speak in jest about Dynamic Range.
This is my official response on the subject, to be taken entirely seriously:
Canon performs poorly compared to the competition in the area of sensor noise, and delivers less Dynamic Range.
They also have the best lens selection and quality in the industry, and their controls and user interface are second to none.
I wish that Canon would continually improve all aspects of camera performance, including Dynamic Range and Sensor Noise.

It's good to hear someone be honest and objective for a change. I totally agree about their lenses. I've warmed up to Sony UI, although I still don't like their control layouts as much.

 

[K]

Not to worry, if the 1Dx Mk II ends up having class-leading DR, the spec will finally matter again.

Base ISO DR of Canon DSLR's already exceeds the DR of all output medium I know of. If the next batch of Canon sensors make Exmor sensors look prehistoric from the DR point of view, I fail to see the advantage for 99% of photographers. While it is a metric which cameras can be measured by, and improvements don't hurt, this aspect of camera performance has been blown out of all proportion.

I have shot tens of thousands of frames with my Mk III, I have hundreds of examples where pushing the shadows in an ISO 100 shot results in noise. The noise is visible in all output mediums. This is ridiculous for a modern camera.


I primarily do landscape/adventure/timelapse photography, so maybe my use case is unique. If I shot sports or people I don't think I would much care about DR.


As it is, I hate having to merge exposures just to get a mild shadow push in a high DR scene. It becomes even more labor intensive if it's a high-DR panoramic (as they usually are)

You need to buy a Nikon DSLR that has a Sony sensor.

 

[msm]

...Again I think it is naive to think that everyone who wants more DR is just plain and simply a Canon hater.

Jrista, again you and others have somehow come to the conclusion that people here are anti-progress. It couldn't be further from the truth.

You should be careful of speaking on behalf of others. And I think it just reeks of jealousy and I find it quite pathetic. When some people probably spend several hours every day for the last 3 years defending Canon at every opportunity, bashing cameras of competing brands and the people who mentions Canons deficiencies I have a really hard time believing they are actually indifferent :

 

[Loswr]

The third one is how I would present this as a picture in terms of exposure balance.

Your demonstration is ineffective. Why? Because there are still shadows in your image, both in the foreground and at the distant treeline. The point of having a 6-stop lifting capability is so you can totally obliterate shadows from your images.

Not to worry, technology will eventually solve this problem for Canon users. No, I'm not suggesting they'll change their sensors, but rather we will soon have space flight and planoforming capability such that we will colonize planets with several suns, thus eliminating shadows effectively even for those poor fools still using Canon cameras.

 

[Orangutan]

The third one is how I would present this as a picture in terms of exposure balance.

Your demonstration is ineffective. Why? Because there are still shadows in your image, both in the foreground and at the distant treeline. The point of having a 6-stop lifting capability is so you can totally obliterate shadows from your images.

Not to worry, technology will eventually solve this problem for Canon users. No, I'm not suggesting they'll change their sensors, but rather we will soon have space flight and planoforming capability such that we will colonize planets with several suns, thus eliminating shadows effectively even for those poor fools still using Canon cameras.

What's funny over a pint isn't always funny over the interwebs.

 

[Sporgon]

The third one is how I would present this as a picture in terms of exposure balance.

Your demonstration is ineffective. Why? Because there are still shadows in your image, both in the foreground and at the distant treeline. The point of having a 6-stop lifting capability is so you can totally obliterate shadows from your images.

Not to worry, technology will eventually solve this problem for Canon users. No, I'm not suggesting they'll change their sensors, but rather we will soon have space flight and planoforming capability such that we will colonize planets with several suns, thus eliminating shadows effectively even for those poor fools still using Canon cameras.

What's funny over a pint isn't always funny over the interwebs.

I think two suns is a great idea as long as they both rose at the same time, otherwise we'd get weak sunsets and sunrises.

Anyway, why don't cartoons have shadows ?

(Of course I know the answer to that one, but hopefully with advanced computerization we will get cartoons with shadows, and then lifting shadows by six stops will look out of fashion - as if it was ever in fashion).

 

[rs]

Hi rs,
Can you please tell me if shadows slider is in terms of stops or levels. Canon has shadow slider range -5 to 5. Does it mean each level is a stop. DXO Optics has shadow slider ranging from -100 to +100. DRreview keeps on talking about pushing shadows by few stops but I am yet to find out how to do that.

Another thing I noticed based on their ISO in-variance test, pushing exposure +5 stops for ISO 100 image is nothing but making it a ISO 3200 image. What is the point of under exposing a lot using high DR camera and turning it into high ISO image. If some one wants to make use of high DR, they need to properly expose image at ISO 100 and make use of sensor DR right?

DPP is certainly not measuring stops, and DXO can't be at +/- 100 stops. Nothing appears to be documented about a direct correlation between sliders and stops, and I can't see how it could be - it is tuned to a range of tones, mainly the shadow areas, but typically tailing off in the deep shadows, and also tailing off in the mid range.

I have had a bit of a play with the shadows in DPP using the before/after comparison and the shadows alerts (threshold of 32/255). In my limited testing pushing the shadows by +5 requires the exposure to be pushed down by 1 stop to get the alerts area to match.

You're right; there is little point in exposing at ISO 100 when 3200 would do exactly what you want.

 

[3kramd5]

Should be able to correlate it with some experimentation. Take a series of photos, adjusting exposure in camera by 1 stop at a time. Then go into software and adjust the middle exposure up and down until the luminance matches the others. Ideally it's linear (e.g. if +25 is one stop, +50 is two stops), but I suspect it may not be.

 

[9VIII]

...Again I think it is naive to think that everyone who wants more DR is just plain and simply a Canon hater.

Jrista, again you and others have somehow come to the conclusion that people here are anti-progress. It couldn't be further from the truth.

You should be careful of speaking on behalf of others. And I think it just reeks of jealousy and I find it quite pathetic. When some people probably spend several hours every day for the last 3 years defending Canon at every opportunity, bashing cameras of competing brands and the people who mentions Canons deficiencies I have a really hard time believing they are actually indifferent :

What reeks of jealousy? The people posting about the same deficiencies we've known about for the last three years on every possible thread?
There's an antagonist for every defender. We could play the Chicken and Egg game but that's pointless now.
It really is just a game, the pictures and technology are just a front for getting in heated arguments. You will find these people in every corner of every society.
I will not participate in the games anymore.

 

[ritholtz]

Hi rs,
Can you please tell me if shadows slider is in terms of stops or levels. Canon has shadow slider range -5 to 5. Does it mean each level is a stop. DXO Optics has shadow slider ranging from -100 to +100. DRreview keeps on talking about pushing shadows by few stops but I am yet to find out how to do that.

Another thing I noticed based on their ISO in-variance test, pushing exposure +5 stops for ISO 100 image is nothing but making it a ISO 3200 image. What is the point of under exposing a lot using high DR camera and turning it into high ISO image. If some one wants to make use of high DR, they need to properly expose image at ISO 100 and make use of sensor DR right?

DPP is certainly not measuring stops, and DXO can't be at +/- 100 stops. Nothing appears to be documented about a direct correlation between sliders and stops, and I can't see how it could be - it is tuned to a range of tones, mainly the shadow areas, but typically tailing off in the deep shadows, and also tailing off in the mid range.

I have had a bit of a play with the shadows in DPP using the before/after comparison and the shadows alerts (threshold of 32/255). In my limited testing pushing the shadows by +5 requires the exposure to be pushed down by 1 stop to get the alerts area to match.

You're right; there is little point in exposing at ISO 100 when 3200 would do exactly what you want.

Thanks RS. Does it mean pushing all 5 levels of shadow slider up in DPP is equivalent to 1 stop. Canon added another slider (Auto Light optimizer) which seems to do good adjustment instead of playing around selective sliders. DXO optics also has similar slider and much broader selective sliders (-100 to 100). But pushing these sliders (specifically shadow slider) seems to add some kind of orange caste. Any ideas to fix color issues arises while pushing these sliders.

 

[mb66energy]

The third one is how I would present this as a picture in terms of exposure balance.

Your demonstration is ineffective. Why? Because there are still shadows in your image, both in the foreground and at the distant treeline. The point of having a 6-stop lifting capability is so you can totally obliterate shadows from your images.

Not to worry, technology will eventually solve this problem for Canon users. No, I'm not suggesting they'll change their sensors, but rather we will soon have space flight and planoforming capability such that we will colonize planets with several suns, thus eliminating shadows effectively even for those poor fools still using Canon cameras.

Theses statements confirm my opininion that canonrumor debates are sometimes far away from real concerns about photography and photographic tools - on the other hand: It helps me to stop reading forum entries and to try to do some photography ... so thanks for that.

 

[Sunnystate]

+100

I was always saying that certain, most ferocious Canon advocates here in the end cause more damage than immediate benefits.
Not only on Canon rumors anyway. Unfortunately in the large scope of things and in the long run the consequences can be dire.

The third one is how I would present this as a picture in terms of exposure balance.

Your demonstration is ineffective. Why? Because there are still shadows in your image, both in the foreground and at the distant treeline. The point of having a 6-stop lifting capability is so you can totally obliterate shadows from your images.

Not to worry, technology will eventually solve this problem for Canon users. No, I'm not suggesting they'll change their sensors, but rather we will soon have space flight and planoforming capability such that we will colonize planets with several suns, thus eliminating shadows effectively even for those poor fools still using Canon cameras.

Theses statements confirm my opininion that canonrumor debates are sometimes far away from real concerns about photography and photographic tools - on the other hand: It helps me to stop reading forum entries and to try to do some photography ... so thanks for that.