I have seen it before in a few images when shooting and processing large raw ..i noticed noise in my low iso images.. sometimes when opening up the shadows using exposure controls in lightroom and Dpp (less with DPP) i noticed this pattern that u are talking about.
the last few weeks ive been using Sraw1 and the nosie at low iso is gone.. i dont even notice it anymore at low iso's ..and i have more than enough detail in my images ..the skys are nosieless and the graduation of color is superb.. and the file size is less to boot!
I know this sounds crazy but.. there seems to be more room for expoure adjustment with a sraw1 file .. I dont know whats happening here with sraw1 but i love it.
Nchakos wrote:
I have seen it before in a few images when shooting and processing large raw ..i noticed noise in my low iso images.. sometimes when opening up the shadows using exposure controls in lightroom and Dpp (less with DPP) i noticed this pattern that u are talking about.
the last few weeks ive been using Sraw1 and the nosie at low iso is gone.. i dont even notice it anymore at low iso's ..and i have more than enough detail in my images ..the skys are nosieless and the graduation of color is superb.. and the file size is less to boot!
I know this sounds crazy but.. there seems to be more room for expoure adjustment with a sraw1 file .. I dont know whats happening here with sraw1 but i love it. ...Show more →
ISO 100 I see absolutely 0 noise. As soon as I get to ISO 200 I immediately see chroma noise in the shadows.
I've seen this in a 20D, 5D 1DsII and 5DII. DPP deals with it better than other RAW processors and I've never found Canon shadows as 'pushable' as for instance Nikon shadows.
Yeah, well, you're not going to see any change in that for awhile. I'm sure there's a good technical reason for it (noise in the amplifier circuitry or something) but there's no DSLR out there that doesn't do this right now. This issue is just a matter of people pushing their equipment beyond the current capabilities, nothing to do with the 5DII per se.
Another hardware flaw is that if you overexpose your image by 5 stops you can't pull it back down, you just get this pesky solid white / grey blob
Learn to work with it. Or get an electrical engineering degree and go work for Canon
Yes, but patterned noise is not desirable and means it's in the hardware. It should be random.
digilloyd.com found exactly this pattern noise when comparing to the D3x and he used dpp.
here a few samples i put in the other thread where the pattern noise doesn't seem so bad:
well it is dark now but i quickly found this ISO200 with HTP on the RAW values went over very wide range, I used like 62 highlight recovery in ACR anyway here are the (oh, also i took these on a FREAKING cold day, like I think it was about 12-14F at the time):
still not seeing that nasty pattern noise you two guys seem to be easily getting.
maybe this is not the best sample, but still seems like has to be doing better
although maybe it is not quite an extreme enough scene, since I didn't really have a huge area of truly deep shadow, the black necks are pretty skinny. I do have an ISO6400 shot under very feeble indoor lighting and there the shadows do have lots of horizontal banding (still not really with the vertical though). to really be sure need to do eact same scene setup with two different bodies.
also don't mind the blur, those geese in the 100% crop are not within the DOF
and then 100% crop with 50% (!) shadow/highlight shadow recovery tool applied:
noisy yes, but this is pulling near blackness and the banding is not there vertically that i can see and barely there horizontally (that said ISO6400 under truly dim conditions can show quite a lot of horizontal banding): http://skibum4.smugmug.com/photos/487551821_cND38-X3.jpg
FYI, the first image was pushed as was mentioned, and was done so to dramatize what was in the shadows, namely non random noise patterns (which I think are atypical of a well functioning camera).
However, the second photo is not pushed really at all, and comes from the overall shot shown below, which is an unadulterated, low quality jpeg conversion of the RAW file (which BTW I sent to Canon). So it is not a function of pushing. I pushed the other photo because I happened to make the shadows too deep. I'm guessing I could have pushed less had I exposed more carefully in that experiment, and still obtained the cross hatch banding.
My camera should be at the NJ Canon repair facility this afternoon I'm guessing.
I think the 5DII is stunning in real-world shooting, but then I like my shadows BLACK.
One thing I've noticed is that ACR seems to dislike HTP and creates some macro blocking in the shadows that is not evident in DPP. Because of this, I always disable HTP. (With RAW capture this feature is of questionable value anyway.) However, those of you pushing the shadows might try testing the interaction of this feature and your chosen RAW converter.
OMG I thought I was going crazy, the noise in low contrast pushed ISO 100 shots is crazy, thanks Braniac for raising this, I thought I was the only one
walter23 wrote:
Yeah, well, you're not going to see any change in that for awhile. I'm sure there's a good technical reason for it (noise in the amplifier circuitry or something) but there's no DSLR out there that doesn't do this right now.
Neither the 1Ds3 nor the 1D3 nor the 5D does this.
This issue is just a matter of people pushing their equipment beyond the current capabilities, nothing to do with the 5DII per se.
Canon stated that this camera would have as good image quality as the 1Ds3. The 1Ds3 was capable of being pushed to this limit. The 5D2 is not.
>Another hardware flaw is that if you overexpose your image by 5 stops you can't pull it back down, you just get this pesky solid white / grey blob
Nobody is proposing 5 stop overexposure. How is this relevant?
brainiac wrote:
Neither the 1Ds3 nor the 1D3 nor the 5D does this.
+1 on the 1Ds3
I just pushed some 1Ds3 images by 5 stops. Lots of noise, but all random. No cross pattern (or any other pattern, for that matter) to the noise at all.
Yes - and NR software is pretty good at dealing with random noise whereas it's crap at dealing with larger scale structure. Likewise, if you just downrez to 10 megapixels random noise averages away nicely whereas structured noise doesn't.
Skibum's noise is nice and random, whereas erichard's and mine isn't.
One more point. At high isos my camera is fine. It's only when I try to eke out maximum dynamic range at low iso and low contrast that I run into noise pattern problems.
brainiac wrote:
Canon stated that this camera would have as good image quality as the 1Ds3. The 1Ds3 was capable of being pushed to this limit. The 5D2 is not.
Maybe in their view as good does not necessary mean identical performance across board. Better high ISO and worse lower ISO cancels each other out to become as good.
Beside, it is totally understandable that they would not let a presumer have better over IQ than their flagship model.
