70D and Dxomark....

[hutjeflut]

I know everybody has invested in lenses and accessories, so jumping ship is not practical, but here is what that 3 stop of extra DR can mean (5D3 on top, D800 on the bottom):

what does this have to do with DR it seems a iso comparison to me nothing more nothing less?

maby im missing something but you cant tell me these 2 are shot on the same iso setting and shutter speed.

 

[jrista]

I know everybody has invested in lenses and accessories, so jumping ship is not practical, but here is what that 3 stop of extra DR can mean (5D3 on top, D800 on the bottom):

what does this have to do with DR it seems a iso comparison to me nothing more nothing less?

maby im missing something but you cant tell me these 2 are shot on the same iso setting and shutter speed.

You should read the original review:

http://www.fredmiranda.com/5DIII-D800/index_controlled-tests.html

The noise exists because of a fairly significant shadow lift. The beams in the samples originally showed up as entirely black, and were lifted enough to bring out color and detail (at least a few stops). The rest of the photos were "correctly exposed", so the tests were not an ISO comparison...they were quite explicitly DR comparisons. Dynamic range is ultimately limited, in the shadows, by noise. Read noise, to be specific. Sony's Exmor sensor simply has less at low ISO (in exact terms, the 5D III has 33.1e-, where as the D800 has a mere 2.7e-).

 

[Aglet]

These are all largely pointless debates...Nikon vs. Canon color depth bits, minute 1-2% differences in SNR, etc.

I agree with pretty much everything you've stated but the SNR differences are much greater than 1-2%. They may be that at the highlites end but SNR difference is actually huge where it matters in post, on the dark end of the curve.
there's a good 2-4 stops difference between Canon and …everybody else at 0db SNR.

 

[jrista]

These are all largely pointless debates...Nikon vs. Canon color depth bits, minute 1-2% differences in SNR, etc.

I agree with pretty much everything you've stated but the SNR differences are much greater than 1-2%. They may be that at the highlites end but SNR difference is actually huge where it matters in post, on the dark end of the curve.
there's a good 2-4 stops difference between Canon and …everybody else at 0db SNR.

Sorry, I wasn't clear. I was specifically referring to higher ISO settings, and the SNR 18% measure at DXO. At higher ISO settings, the differences are miniscule, if they exist at all. Canon maintains an edge thanks to a WEAKER CFA, but in the end color quality is so phenomenally good on all modern DSLRs those differences don't matter much. For high ISO users, other factors matter much more...AF system, metering, frame rate, etc.

At low ISO, improved SNR does indeed lead to more stops of dynamic range, no one disputes that.

 

[jrista]

Just to demonstrate a point. The Canon image in the samples above looks quite terrible. While you can probably never extract the same amount of shadow detail as the D800 with a 5D III, it need not look quite that bad. Here is a copy that was run through both Topaz DeNoise 5, with vertical debanding, 100% chroma noise removal, and minor luma noise removal, as well as maximum dynamic range recovery. It was then run through Nik Dfine 2 in order to remove more luma noise. DeNoise 5 does an excellent job removing chroma and banding noise, while Dfine 2 does a better job at reducing luma noise without smudging detail.

Here is the result:

Clear loss of detail on the upper beam relative to the D800. Still some noise on the door frame and window relative to the D800. Some artifacts on the no smoking sign (although more likely due to the use of a JPEG as my source...denoising from an original RAW or high quality TIFF would produce better results.) Despite those, the results are FAR more acceptable than the original 5D III RAW without any noise reduction. Dynamic range has improved considerably with some effective NR. Not the same 13.2 stops as the D800 has, but definitely more than the 10.97 stops that DXOMark measured.

So, yes, the 5D III, like most other Canon sensors, has some pretty crappy shadow noise. Doesn't mean the buck stops there, however. ;D

 

[Pi]

Here is a copy that was run through both Topaz DeNoise 5, with vertical debanding, 100% chroma noise removal, and minor luma noise removal, as well as maximum dynamic range recovery. It was then run through Nik Dfine 2 in order to remove more luma noise. DeNoise 5 does an excellent job removing chroma and banding noise, while Dfine 2 does a better job at reducing luma noise without smudging detail.

Can you denoise now the crop from the Toronto's Yonge-Dundas Square shot I posted earlier without losing detail? If I like it, I may send you the whole image, and pay you for that.

 

[jrista]

Here is a copy that was run through both Topaz DeNoise 5, with vertical debanding, 100% chroma noise removal, and minor luma noise removal, as well as maximum dynamic range recovery. It was then run through Nik Dfine 2 in order to remove more luma noise. DeNoise 5 does an excellent job removing chroma and banding noise, while Dfine 2 does a better job at reducing luma noise without smudging detail.

Can you denoise now the crop from the Toronto's Yonge-Dundas Square shot I posted earlier without losing detail? If I like it, I may send you the whole image, and pay you for that.

Hmm, I think I missed that one...I came into this thread in the middle. I can read from the beginning if its in this thread, and I'd be happy to give it a whirl.

 

[Pi]

Here is a copy that was run through both Topaz DeNoise 5, with vertical debanding, 100% chroma noise removal, and minor luma noise removal, as well as maximum dynamic range recovery. It was then run through Nik Dfine 2 in order to remove more luma noise. DeNoise 5 does an excellent job removing chroma and banding noise, while Dfine 2 does a better job at reducing luma noise without smudging detail.

Can you denoise now the crop from the Toronto's Yonge-Dundas Square shot I posted earlier without losing detail? If I like it, I may send you the whole image, and pay you for that.

Hmm, I think I missed that one...I came into this thread in the middle. I can read from the beginning if its in this thread, and I'd be happy to give it a whirl.

Please don't! It was a joke (the image is real though).

 

[jrista]

where is the banding?
5D3
ISO 4000
extreme cropped

You shot at ISO 4000. Thanks to Canon's use of a Bias Offset, by that ISO, the black point has moved above the level at which banding noise exists. The only noise remainder is slight quantization noise during ADC...and photon shot noise.

 

[jrista]

Re: Black-Cap event

70D in stock at my local store.
Goin' in for some lens cap shots.

Case in point - this is how Aglet evaluates a new camera: 'pictures' with the lens cap on. He'll then boost the black images 4-5 stops, and tell us all how crappy the 'pictures' are, and how Canon still hasn't made their sensors cameras any better.

Yawn.

raw files are pushed 4 stops

Do you look for bio-luminescent markers with the room lights on?

(BTW, are you a real scientist?... cuz you sure spend a LOT more time on this forum during the work day than I'd like to see any employee of mine doing)

Lens cap shots? Pushed 4 stops?? I've never heard of a worse way to evaluate a camera. That's not even photography. It shows something, but has no practical relevance.

It has practical relevance for certain kinds of photography. The most obvious being landscape photography. Another, maybe not so obvious, would be interior design photography (where you often encounter bright windows resulting in scenes that easily top 13 stops of DR.) It is not practical for EVERY kind of photography, and once you get past ISO 400, on Canon cameras, black point is usually above the banding noise floor and physics is the primary driver of noise, so it doesn't matter.

 

[jrista]

Here is a copy that was run through both Topaz DeNoise 5, with vertical debanding, 100% chroma noise removal, and minor luma noise removal, as well as maximum dynamic range recovery. It was then run through Nik Dfine 2 in order to remove more luma noise. DeNoise 5 does an excellent job removing chroma and banding noise, while Dfine 2 does a better job at reducing luma noise without smudging detail.

Can you denoise now the crop from the Toronto's Yonge-Dundas Square shot I posted earlier without losing detail? If I like it, I may send you the whole image, and pay you for that.

Hmm, I think I missed that one...I came into this thread in the middle. I can read from the beginning if its in this thread, and I'd be happy to give it a whirl.

Please don't! It was a joke (the image is real though).

LOL

Well, if you ever pick up DeNoise 5, it should solve the problem nicely. It is very configurable. The debanding can work in horiz, vert, or both, and each are independently adjustable. You have fine control over how denoising is applied to different color channels, the ability to recover black level (DR), etc. If you've been having trouble cleaning up shadow noise, it's pretty much the tool that will do the job. I use either Nik Dfine 3 or some other NR tool for luma noise reduction (photon shot noise), as Topaz DeNoise 5 past a certain "strength" level starts to heavily blur detail.

 

[FunkyCamera]

I can't wait for Nikon to go out of business. Their sensors are worse in every way that matters, their bodies are low quality and the controls make no sense and their pathetic excuses for lenses perform worse than coke bottles.

I have to try hard not to laugh out loud every time I see a moron who spent thousands on a "pro" Nikon body. Pathetic. The 70D wipes the floor with everything Nikon can put out at any price level.

 

[jrista]

I can't wait for Nikon to go out of business. Their sensors are worse in every way that matters, their bodies are low quality and the controls make no sense and their pathetic excuses for lenses perform worse than coke bottles.

I have to try hard not to laugh out loud every time I see a moron who spent thousands on a "pro" Nikon body. Pathetic. The 70D wipes the floor with everything Nikon can put out at any price level.

Well, I think we can boil this down to a single word:

T R O L L

 

[Aglet]

I can't wait for Nikon to go out of business. Their sensors are worse in every way that matters, their bodies are low quality and the controls make no sense and their pathetic excuses for lenses perform worse than coke bottles.

I have to try hard not to laugh out loud every time I see a moron who spent thousands on a "pro" Nikon body. Pathetic. The 70D wipes the floor with everything Nikon can put out at any price level.

Well, I think we can boil this down to a single word:

T R O L L

That troll is a lovely single-shot HDR exposure capturing him in action and clearly showing us his dingy candle-lit domain while retaining excellent fidelity in the outdoor scene visible beyond his entrance. No obvious banding or read noise problems so must be an image created with Exmor-based tools.

 

[rpt]

I can't wait for Nikon to go out of business. Their sensors are worse in every way that matters, their bodies are low quality and the controls make no sense and their pathetic excuses for lenses perform worse than coke bottles.

I have to try hard not to laugh out loud every time I see a moron who spent thousands on a "pro" Nikon body. Pathetic. The 70D wipes the floor with everything Nikon can put out at any price level.

Well, I think we can boil this down to a single word:

T R O L L

That troll is a lovely single-shot HDR exposure capturing him in action and clearly showing us his dingy candle-lit domain while retaining excellent fidelity in the outdoor scene visible beyond his entrance. No obvious banding or read noise problems so must be an image created with Exmor-based tools.

May be by Ex-evenmor based device

 

[Pi]

T R O L L

That troll is a lovely single-shot HDR exposure capturing him in action and clearly showing us his dingy candle-lit domain while retaining excellent fidelity in the outdoor scene visible beyond his entrance. No obvious banding or read noise problems so must be an image created with Exmor-based tools.

Yeah, flare like this can only come from a Nikon lens...

 

[tnargs]

Well, I think we can boil this down to a single word:

or counter-troll..... maybe he just read enough trolling...