keithreeder wrote:
For God's sake, how likely is that?
The folk who are throwing images Gabor and Emil's way are likely to the the ones who think they might have a problem in the first place - far more will simply be happily ramping up their shutter actuation count by taking photos and, if they're smart, converting in something that wasn't written by Adobe...
The only thing is it was VERY hard to find such files from 50D or 40D or 5D2 etc. users. I think only about six 5D2 users with such artifacting have ever popped up onto the internet and here we were asking for good samples too and got like 8 bad ones in a week and no good ones (maybe Gabor or Emil got a good one, I've never seen one yet myself).
And as for avoiding Adobe, actually the 5.6Beta seems to show the only conversions entirely free of the artifacting, although it needs to be looked into if it wa accomplished without losing anything else. At first glance it doesn't have the crisp micro-contrast of my 5.5 conversion but it might just be sharpening radius/amount and not smoothing anyway it needs to be looked into.
Jeff Nolten wrote:
DPR mentioned several times that the 7D was better at noise than the 50D, but I was curious how it compared to my two bodies, 5D and 40D. I extracted these numbers by eye from their Gray Luminance noise graphs so I may not be quite accurate (+/- 0.1).
Edit: Sorry if these columns don't line up. How does one match edit box to display
Two things jump out at me. One is how well the 40D fares against its bigger brothers and second is the flatter slope of the 7D's noise curve. Canon have obviously designed the 7D for folks who are going to be mainly working in the higher ISOs. Yes, yes I know this is per pixel and I've been looking at the line pair resolution charts as well, but before I spend the $$ I want to know what the tradeoffs are and guestimate what improvement in my keeper rate there might be.
For those who are going to jump on DPR's not taking pixel density into account, answer this for me: If I average over enough pixels to get the 7D's 100 ISO noise level down to the 40D's, will the 7D image have any crop advantage left over the 40D?...Show more →
1. i believe those are per pixel so the 7D is at a disadvantage there before normalization
2. those may have been from the jpgs and who knows what NR was applied model to model
3. it seems weird they would compromise what could have been fantastic low ISO performance by maybe up to a factor of two for, at best, maybe a 10% gain at high ISO, if any
if they were based on ACR 5.6 maybe ACR just breaks up the mazing somehow into more random looking noise but still produces more than there could have been
UCSB wrote:
Your image looks very nice. I don't see anything wrong with it.
I have noticed that if I use C1 with NR off and sharpening off, I can get much cleaner results. Is that what you are doing here? It seems that NR especially is causing artifacts. I haven't spent enough time seeing what other processing affects the images negatively. Are you doing normal processing on your images (levels, curves, etc.)? Have you shot any ISO 400 images?
Luma NR 3, Chroma NR 37. Sharpening 180/0.8
I usually turn NR off for low ISO.
Levels, curves, contrast, WB etc as default / 0.
At ISO 400 and above, I get more artifacts. It looks more like randomly distributed rectangular grain, not so much arranged in a maze pattern.
For now I have to agree. The 7D has impressive features and I certainly don't want to dismiss them. Whenever I can get enough glass on a subject I use the 5D. I use the crop for pushing the envelope of what I used to get with a "normal" 35 mm camera and the 7D is designed for that just kind of use. I leave next week for a trip to the Peruvian Amazon and will certainly be evaluating my 40D in terms of what the 7D might do for me. Last time I was there I had an Elan II and a 75-300 non-IS so I don't think I'll feel too limited
And thanks for everyone's analysis. There certainly would have been much less information available trying to evaluate upgrading from an Elan II to an Elan 7.
JRIF (Jack Rabbit in flight) : 7D @600mm (300 2.8 IS with 2x tele), this is my bread and butter combo for wildlife. Finally tested it properly today before I leave next week ...the camera didn't disappoint.
Will post more, here's 2 from the session, hand held. As always, exif intact.
skibum5 wrote:
1. i believe those are per pixel so the 7D is at a disadvantage there before normalization
Yes, agreed. I have a G10 as a 1 lb alternative to my 5D for backpacking and it makes stunning landscapes in good light. But it is also true that one tends to use a wildlife camera more towards the pixel peeping end of things.
2. those may have been from the jpgs and who knows what NR was applied model to model
Yes I used the jpg graphs since the raw conversion is still iffy. I hope DPR uses similar in-camera settings,
3. it seems weird they would compromise what could have been fantastic low ISO performance by maybe up to a factor of two for, at best, maybe a 10% gain at high ISO, if any
Indeed. I already carry two bodies while traveling, so now I'm supposed to carry a low iso crop and a high iso crop? A One series would then be lighter and cheaper.
if they were based on ACR 5.6 maybe ACR just breaks up the mazing somehow into more random looking noise but still produces more than there could have been
I look forward to seeing what the 3rd party converters eventually do since this will be critical to using the 7D.
alundeb wrote:
Luma NR 3, Chroma NR 37. Sharpening 180/0.8
I usually turn NR off for low ISO.
Levels, curves, contrast, WB etc as default / 0.
At ISO 400 and above, I get more artifacts. It looks more like randomly distributed rectangular grain, not so much arranged in a maze pattern.
Yes, that is what I'm seeing also. Thanks for the reply.
keithreeder wrote:
Some pretty obvious conclusions to be drawn here, I think...
(And to be absolutely clear - I've done nothing to these files to "improve" them in any way...)
The obvious conclusion is that all the other RAW converters have capitulated and given up resolution to deal with the column gain fluctuations. The LR3ß has higher detail despite the artifacting. If the column gains were better balanced, one would have the freedom from artifacts of the others, as well as the detail and sharpness of the LR3ß.
kirry007 wrote:
JRIF (Jack Rabbit in flight) : 7D @600mm (300 2.8 IS with 2x tele), this is my bread and butter combo for wildlife. Finally tested it properly today before I leave next week ...the camera didn't disappoint.
Will post more, here's 2 from the session, hand held. As always, exif intact.
Nice captures. I thought about getting the 2X tele. Would love a 500L but I need to win a lottery first. I heard using the 2X will not autofocus but your lens is a 2.8. My 300 is an F4. I guess going over 5.6 is the issue?
I switched to this method --- I am getting sharper and better controlled noise in images
- process image in DPP with sharpening at 0, all noise reduction at 0
- output from DPP at 16 bit per pixel TIF
- open 16 bit per pixel TIF in Adobe Photoshop CS 4 64bit
- arbitrary rotate if required
- apply noise reduction using 64bit Neat Image v6.1 (auto profile, auto fine tune)
- apply smarp sharpen at 121% (with extra precision option selected)
- apply auto contrast, or auto color if required
- convert to 8 bits
- save as maximum quality JPG
The skies look good in the images, and the sharpness is better.
This works great at least until Adobe releases a non-beta Adobe Camera Raw converter update to try.
You can see recent updates here. Go to the bottom of NM birds
skibum5 wrote:
The only thing is it was VERY hard to find such files from 50D or 40D or 5D2 etc. users. I think only about six 5D2 users with such artifacting have ever popped up onto the internet and here we were asking for good samples too and got like 8 bad ones in a week and no good ones (maybe Gabor or Emil got a good one, I've never seen one yet myself).
My 7D is currently at the service center for inconsistent AF problems. If the technician does not give up and swap out my camera, I can send you RAW files from my camera for you to examine. Gabor has done it and found no issues as far as unequal channel response, low ISO banding or mazing is concerned. This is not an inherent 7D problem. It's just another perennial Canon QC issue.
thw2 wrote:
My 7D is currently at the service center for inconsistent AF problems. If the technician does not give up and swap out my camera, I can send you RAW files from my camera for you to examine. Gabor has done it and found no issues as far as unequal channel response, low ISO banding or mazing is concerned. This is not an inherent 7D problem. It's just another perennial Canon QC issue.
if you get the same copy back, i'd love to see a raw sample or two of a certain type
