Canon 5Ds Dynamic range--any accurate information not just speculation??

[Ray]

DPReview has added a page of tests and discussion of the 5Ds with regard to dynamic range:
http://www.dpreview.com/previews/canon-eos-5ds-sr/7

In summary, a slight improvement but not up to the level of cameras equipped with Sony sensors.  They attribute the difference to off chip A/D converters on the Canon cameras vs integrated on chip A/D coverters on Sony sensored cameras.

Thanks for the link. Those dpreview comparisons really highlight the deficiencies of the 5DSR dynamic range compared with Nikon cameras.

The impression I get is, if you're shooting a contrasty scene with a DR (or Subject Brightness Range) of 6 stops or greater, and you want to raise the deepest shadows, then the 5DSR will show significantly more noise in those shadows than the cropped-format Nikon D7200.

As I understand, the 5DSR pixel is of similar quality to the 7d Mk II pixel. There might be a slight improvement, maybe 0.1 or 0.2 EV better, but let's get things into perspective. Any improvement less than 0.5 EV is not worth getting excited about.

The Nikon D7200 pixel is slightly smaller than the Canon 7D2 pixel, yet it's DR is 2.68 EV better at ISO 100. At ISO 800, the DR of the D7200 pixel is still around 1 EV better than the 7D2 pixel.

The images below tell the story.

[jjj]

The real question then becomes, "Is that a problem?"

Yes, if parts of the image are incorrectly exposed.

With film I used fill-flash if feasible, with digital I use a bit of shadow recovery, but if the exposure makes the shadows dark, then they probably should be dark. If the exposure blows out the background, maybe it should blow out the background (otherwise one loses the sense of the actual lighting conditions).

But if you do not want black/murky shadows or blown highlights….
Using fill flash changes the look of a picture and if it's a look you do not want, it's not practical.

[Guillermo Luijk]

Looking at Bart/Jack's DR figures for the 5DS, it seems the final DxO's values will not be far from the 7DII at pixel level. Applying to those figures a SNR correction for 8Mpix ("Print" criteria in DxO), I have obtained the following simulation showing where the 5DS's DR will probably fall:



The 12,44EV figure comes from 7DII's 11,11EV, statistically normalised from 50Mpx to 8Mpx:

SNR linear improvement: [(8688*5792)/(3464/2309)]^0,5=2,508 -> pushing 7DII "Screen" curve by 1,33EV: 11,11+1,33=12,44EV.

Regards

[Bernard ODonovan]

DR tested in this 5DS-R vs D810 vs 5D Mk III vid:

https://www.youtube.com/watch?v=Gq3ISUHsfsQ

Colour accuracy mentioned in this one:

https://www.youtube.com/watch?v=sQ7NC9nSESg

Meanwhile Sony US has a fire sale:

http://www.sonyalpharumors.com/sony-us-lists-crazy-low-prices-a7r-for-199-only/

 

[Ray]

Regardless of issues of dynamic range, one advantage for owners of Canon gear, who might wonder if they should upgrade to a 50mp camera, or wonder if they really need 50mp, is the effect those extra pixels have on upgrading all of one's Canon lenses.

The resolution difference between the 36mp of the Nikon D810 and the 50mp of the 5DSR is marginal, as shown in the youtube video above. However, the resolution difference between the 22mp of the 5D3 and the 50mp of the 5DSR must be very significant.

In terms of sharpness and detail, a 5DSR with a single good quality zoom should be equivalent to a whole basket load of high quality primes for use with a 5D3. 

[Guillermo Luijk]

Looking at Bart/Jack's DR figures for the 5DS, it seems the final DxO's values will not be far from the 7DII at pixel level. Applying to those figures a SNR correction for 8Mpix ("Print" criteria in DxO), I have obtained the following simulation showing where the 5DS's DR will probably fall:



The 12,44EV figure comes from 7DII's 11,11EV, statistically normalised from 50Mpx to 8Mpx:

SNR linear improvement: [(8688*5792)/(3464/2309)]^0,5=2,508 -> pushing 7DII "Screen" curve by 1,33EV: 11,11+1,33=12,44EV.

OK OK, now you can congratulate the Oracle 



Regards

[Ray]

OK OK, now you can congratulate the Oracle 



Regards

Brilliant! The tested DR values are so close to those you calculated, they are the same for all practical purposes.

However, I have to admit I'm a little disappointed at the huge DR difference between these two cameras. 2 & 1/3rd stops at base ISO is huge. I was hoping for at least a slight improvement, but I guess Canon just doesn't have the manufacturing capability.

[caribsurf47]

So the Sensor in the two new cameras does not have the DR of the Nikon D810 or the Sony Alpha 7R, so where does that leave Canon users with collections of L Series glass and/or Zeiss ZE lenses? Do we sell up and start again (at what cost)? If these Camera Bodies are pitched at Professional or Semi-Professional photographers for studio work using studio lighting, what is the actual consequence of having improved potential detail retrieval but essentially similar DR to the Canon EOS 1DX? How crucial will this be to studio output made under artifical lighting?

[Petrus]

So the Sensor in the two new cameras does not have the DR of the Nikon D810 or the Sony Alpha 7R, so where does that leave Canon users with collections of L Series glass and/or Zeiss ZE lenses? Do we sell up and start again (at what cost)? If these Camera Bodies are pitched at Professional or Semi-Professional photographers for studio work using studio lighting, what is the actual consequence of having improved potential detail retrieval but essentially similar DR to the Canon EOS 1DX? How crucial will this be to studio output made under artifical lighting?

If professionals have been using Canon gear, why would the results now be suddenly worse? If there was no reason to switch to Nikon before, now is even less with the new Canon giving similar or theoretically better resolution. DR has stayed the same, but if t has been good enough until now, why not also tomorrow?

[Guillermo Luijk]

So the Sensor in the two new cameras does not have the DR of the Nikon D810 or the Sony Alpha 7R, so where does that leave Canon users with collections of L Series glass and/or Zeiss ZE lenses? Do we sell up and start again (at what cost)? If these Camera Bodies are pitched at Professional or Semi-Professional photographers for studio work using studio lighting, what is the actual consequence of having improved potential detail retrieval but essentially similar DR to the Canon EOS 1DX? How crucial will this be to studio output made under artifical lighting?

Extra dynamic range is only an advantage for certain types of photography, basically when the RAW has to be heavily processed to lift the shadows.

Canon vs Sony war in brief:

• Base ISO applications with controlled light (studio) and JPEG shooters: no winner.
• Base ISO applications (typ. using a tripod) where uncontrolled high contrast scenes can be involved (landscape, architecture, interioism): Sony wins by 2 stops
• High ISO applications with low light (sports, photojournalism, night photography): no winner (at high ISO Canon performance equals Sony's)

Regards

[caribsurf47]

Agreed. It is the point I was hoping to get out of this Canon VS Nikon VS Sony debate. It does depend what you are going to use the camera for.

[gazwas]

Shoot at ISO400 which is a pretty standard ISO these days and there is basically ½ stop benefit to using a Nikon/Sony over the Canon which is not going to make any practical difference to the images. I'm hoping the new BSI sensor in the A7RII addresses this and gives the Sony sensor a much shallower drop in DR as ISO increases as the new sensor makes little difference to the high ISO noise performance looking at examples I've seen.

[Bart_van_der_Wolf]

Extra dynamic range is only an advantage for certain types of photography, basically when the RAW has to be heavily processed to lift the shadows.

Exactly. Single shot high dynamic range scenes, that also have shadows that need to be lifted a lot in post-processing will benefit.

And although the shadows of the recent Sony sensors are less noisy due to lower read-noise, they are still shadows that received very few photons (and thus are affected by a lot more shot noise). There is simply no alternative for real detail and color from (lots of) photons, so (if possible) exposure bracketing will still produce better quality than a single high DR capture.

Cheers,
Bart

[Ray]

How crucial will this be to studio output made under artifical lighting?

I wouldn't think it would be much concern at all in any environment where one has the luxury of being able to arrange artificial lighting, or where the circumstances are suitable for bracketing exposure to merge to HDR.

The increased resolution afforded by 50mp is a big upgrade for those who already use Canon equipment. If  I didn't use Nikon equipment, I'd be thrilled to bits at the prospect of upgrading to 50mp with no loss of DR at the pixel level, compared with the 5D3, at least from base ISO to ISO 400.

However, I wouldn't be nearly as thrilled to upgrade from a 36mp D800E to a 50mp Canon with such a significant drop in DR. One of the advantages of the ISO-invariant nature of the latest Nikon cameras, such as the D810 and D7200, is that one doesn't have to be so careful and obsessive about 'correct' or ETTR exposure. If the lighting conditions and desired shutter speed require a raising of ISO, one can simply underexpose at a lower ISO with reduced risk of inadvertently blowing highlights, and with no disadvantage of greater noise compared with raising ISO. It makes for quicker and easier shooting in manual mode.

[Abdulrahman Aljabri]

Very interesting discussion, I will add that I found very little advantage using the a7r over my 5d mark3 for most of my hotel photography. Hence, most of it involves hdr bracketing. Now in the instances when I am limited to 1 picture like natural light people photography I think there might be a little difference, but I have not tested it enough to make any conclusion. I am thinking about making a food photoshoot for that purpose just to see how much I can process shadows with the a7r over the 5diii. 

[BernardLanguillier]

Very interesting discussion, I will add that I found very little advantage using the a7r over my 5d mark3 for most of my hotel photography. Hence, most of it involves hdr bracketing. Now in the instances when I am limited to 1 picture like natural light people photography I think there might be a little difference, but I have not tested it enough to make any conclusion. I am thinking about making a food photoshoot for that purpose just to see how much I can process shadows with the a7r over the 5diii.  

How do you convert you raw files if I may ask?

I have found that C1 Pro does a much better job than LR for large DR images, probably as a result of their digital back heritage.

Cheers,
Bernard

[Ray]

I have found that C1 Pro does a much better job than LR for large DR images, probably as a result of their digital back heritage.

Can you show us some comparison shots, Bernard, of the same shadows in the same image, processed with the two different converters?

How would you quantify the differences? For example, would you say that the shadows in the C1 Pro converted RAWs look as good as the shadows in the LR converted RAWs which have had 1/3rd of a stop greater exposure, or 1/2 a stop greater exposure, or more?

Are there perhaps certain disadvantages in using C1 Pro, compared with LR? Are there also perhaps certain disadvantages in using only LR as opposed to the full Photoshop program with options of further selective noise reduction after conversion in Camera Raw?

Hope I haven't opened a can of worms. 

[Abdulrahman Aljabri]

How do you convert you raw files if I may ask?

I have found that C1 Pro does a much better job than LR for large DR images, probably as a result of their digital back heritage.

Cheers,
Bernard

I merge the raw pictures with the Photoshop CS6 Merge to HDR tool in 32bit mode and save as tiff. Then I open it in Adobe Camera Raw and do the exposure, contrast and color adjustments. I guess I can open the same tiff file in Capture One. Correct me if I am wrong, but Capture One Pro 8 does not have an HDR merge feature, right? If you use Capture One you have to relay on another program to do the HDR processing, right?   

[Guillermo Luijk]

Very interesting discussion, I will add that I found very little advantage using the a7r over my 5d mark3 for most of my hotel photography. Hence, most of it involves hdr bracketing.

If your regular workflow involves HDR bracketing, there is no advantage for you in a high perfomance sensor regarding dynamic range, because you'll be always using RAW data with a good exposure, i.e. a high SNR. Two shots from my 10-year old APS Canon 350D manage to capture the same DR as a single shot from the Nikon D810. If I shoot three times vs one on the Nikon, the little Canon easily outperforms the FF body.

The advantages of having a high dynamic range sensor become practical in single shot applications where the RAW file is to be strongly processed.

Regards

[Jack Hogan]

• Base ISO applications with controlled light (studio) and JPEG shooters: no winner. ...

...

The advantages of having a high dynamic range sensor become practical in single shot applications where the RAW file is to be strongly processed.

Generally agreed.  However, what used to be considered 'strong' processing with DR-limited cameras in the past is becoming more 'normal' in these days of 12+ stops of usable tones - and many current cameras can produce OOC images that intelligently squeeze more of the captured DR into them automatically and pleasingly, based on pre-analyzing the scene.  For instance Nikon has Auto ADL and Highlight-weighted metering for just such a purpose.  I am sure other manufacturers have their own version of it.

Jack