Testing the limits of Canon 6D RAW files.

The sensor in the Canon 6D is one of the best Canon-made sensors to be put into a DSLR camera for both low and high ISO levels. I have conducted some tests to see where the limits might be for this latest generation Canon sensor.

The following test shows all of the native ISO levels compared. I'm happy using the 6D up to ISO 12800 for landscape photography. The noise actually doesn't look bad as far as noise goes.

557 x 1886

EMBED PREVENTED, IMAGE TOO LARGE: https://farm8.staticfl​ickr.com …84002355_917834​74b8_o.jpg Click here to see our image rules.

Canon_Lover wrote in post #16707429
(...) I always bracket my exposures, so 3 stops is plenty for my work. I think I would need a sensor with over 15 stops of dynamic range to avoid bracketing in many cases, so even the current Sony sensors would still benefit from bracketing.

Bingo!

I've found this thread particularly interesting, to say the least. I am on the exact same page as you are.

That's about the magic #: close to 15 stops, for handling most challenges (not all, but pretty much almost all that practically matter).

In an attempt to evaluate the benefit of a substantially lower read-out noise at lower ISOs (translation: more "pushable" shadows), I've been experimenting with the good-old 1D3 which, together with the 1Ds3, 1Dx and maybe 40D, shows pretty well controlled banding, and high-quality files.

By applying techniques and basic concepts from Astro-Photography (all in Lightroom + Photoshop), here's what 1D3 files would look if read-out noise would be decreased by approx. a factor of SQRT(9) = 3 (pushed in Photoshop with Exposure=+5, not coming from bracketed frames):

1920 x 1200

EMBED PREVENTED, IMAGE TOO LARGE: http://www.pbase.com …ge/154572671/or​iginal.jpg Click here to see our image rules.

1920 x 1200

EMBED PREVENTED, IMAGE TOO LARGE: http://www.pbase.com …ge/154572672/or​iginal.jpg Click here to see our image rules.

When I run the latest version of DPP against 8.3 ACR, and push all NR sliders down, and set a custom wb temperature in both tools, the resulting JPGs are worse with ACR. I am a bit leary of the claims that ACR does a better job, my personal opinion is that folks are either running old versions of DPP, or aren't matching apples to apples.

I could have done something wrong, but in regards to noise, with both tools set to 0/0 and a custom 4500K WB, the ACR result is worse, if I did nothing wrong.

TeamSpeed wrote in post #16707671
When I run the latest version of DPP against 8.3 ACR, and push all NR sliders down, and set a custom wb temperature in both tools, the resulting JPGs are worse with ACR. I am a bit leary of the claims that ACR does a better job, my personal opinion is that folks are either running old versions of DPP, or aren't matching apples to apples.

I could have done something wrong, but in regards to noise, with both tools set to 0/0 and a custom 4500K WB, the ACR result is worse, if I did nothing wrong.

If you turn off ACR chroma noise reduction, it is pretty bad. If you leave to default it works really well.

For my tests, I went with the best results I could get with DPP and ACR for both sharpness, contrast, and noise reduction.

DPP files make me have to fight more to keep a good looking photo with good contrast and details. It really wants to go flat with awful colors above 2 stops.

ACR files only need a little dose of NR to make them work their best when pushed over 2 stops.

Canon_Lover wrote in post #16707748
If you turn off ACR chroma noise reduction, it is pretty bad. If you leave to default it works really well.

For my tests, I went with the best results I could get with DPP and ACR for both sharpness, contrast, and noise reduction.

DPP files make me have to fight more to keep a good looking photo with good contrast and details. It really wants to go flat with awful colors above 2 stops.

ACR files only need a little dose of NR to make them work their best when pushed over 2 stops.

DPP NR is a little lacking, but the rest of the results are good. This is why I use Noiseware instead. I take DPP NR down below in-camera values when processing images, along with sharpness down as well. I then deal with noise outside of DPP, and then resharpen.

For example here is a high ISO from a 7D. The bottom is OOC, the middle is a DPP output where I take the settings down, and the top is the final processed version.

I haven't yet found a workflow that works for me using the Adobe tools. Anyways carry on....

Canon_Lover, I have likewise been impressed by the ability to recover shadows from the 6D files.
Appreciate your efforts to illustrate its capabilities.

TeamSpeed wrote in post #16707756
DPP NR is a little lacking, but the rest of the results are good. This is why I use Noiseware instead. I take DPP NR down below in-camera values when processing images, along with sharpness down as well. I then deal with noise outside of DPP, and then resharpen.

For example here is a high ISO from a 7D. The bottom is OOC, the middle is a DPP output where I take the settings down, and the top is the final processed version.

I haven't yet found a workflow that works for me using the Adobe tools. Anyways carry on....

I don't use NR or Sharpening with DPP. Even with the NR off in DPP, it still kills details in the shadows when pushing more than 2 stops. The software isn't even designed to go over 2 stops. I will post the default results when I get a chance. DPP looks horrid without lots of intervention past a 2 stop push.

I go by how an image looks overall after processing. While ACR might not win every category, I think the results just look better for the work I do and cameras I use.

Canon_Lover wrote in post #16707849
...The software isn't even designed to go over 2 stops.

How do you accomplish more than a 2 stop push in DPP?

I have never had a need to go past 2 stops anyways, so I will continue to use DPP, the workflow is easier to bulk update in DPP and mass create JPGs, then run bulk CS3 actions on those files.