Earthshatteringly Disappointed With 7D

[Jettatore]

Your subjects/time of day are very far from ideal. Most of the content of your images are in shadow, and in one image the sides of objects are being struck by bright highlights while the rest remains covered by moderate shadow.

Sorry, I disagree. There is no excuse for his first picture's noise level... at ISO 200! Geesh! if we are making excuses to justify that kind of performance and that level of "grain" in a daylight shot...its laughable. Either the settings are wrong in one of the menus or the unit is not performing well. I Don't wan't to paint all 7D's as being subpar as owners are apparently happy with their bodies...but suggesting this poster's image is "acceptable" or that his technique could be better is not a fair assessment.

I say return the unit and try a replacement if you want to stick with 7D, or get yourself a 5DII if you can afford it and can't wait for 5D3 to arrive. That level of grain is not acceptable or normal in the pics posted. Cheers!

His images are suffering from additional problems on top of noise, include chromatic aberration, over-all soft focus, additional lens problems with focus on the right side of the image, possible camera shake, lack of highly resolved detail at 100% anywhere in the image, questionable exposure, an non-ideal lighting conditions etc. etc. etc. so it's really hard to tell if his 7D is truly a bad copy or not. I can say I don't have low-ISO noise problems on my 7D under good conditions. I find noise to be the least of the problems with these pictures which makes it difficult to analyze under a microscope and I find looking at noise on it's own a form of taking things completely out of context. Have him post shake free RAW CR2 files with a subject captured in sharp focus and with a variety of in camera options options compared to some free file uploading site with good lighting/exposure and then we can analyze it better. My bet is that I could take his copy of the 7D body with my lenses and make really nice pictures with it that do not suffer from any of the above problems.

 

[jrista]

I've taken one of freidmud's sample images, loaded it into photoshop, and hit the "Print Size" zoom button under the zoom tool. Additionally, I adjusted levels...midtones to 1.3, and applied medium contrast curves, to enhance the image a bit. Both adjustments often have the effect of enhancing visible noise. The screenshot below is of that image in photoshop at print size (for 300ppi on my 103dpi computer screen):

http://img6.imageshack.us/img6/4149/friedmudsampleatprintsi.jpg

I don't see ANY visible noise anywhere it was present in the original 100% crop. I think the worst areas were the shaded parts of the fence posts and the columns of the building in the background. At print size, WITH an added boost to contrast, no noise is visible at all.

I know people like to pixel-peep, and LOVE to complain about the noise levels of the 7D...but I think the results are WAY overblown and overstated. The 7D is a very high resolution camera...possibly as high as it can get without outresolving the current crop of Canon's BEST lenses. The simple fact of the matter is, no one views images at 100%. For on-screen viewing, the final resolution is quite possibly 1/4 or less the native resolution, and all noise is absorbed. For native-resolution prints, all noise is absorbed. For reasonably enlarged prints, noise tends to help keep gradients smooth, and IF it is visible, a bit of noise reduction is usually all thats necessary to fix the problem.

 

[Loswr]

Again, my context is shooting athletes in a race...

Stop right there. The 7D is the camera for you, unless you are considering a 1-series body. No amount of work in post will correct the out-of-focus shots you'll get trying to track runners with the 5DII's AI Servo AF. I would not even consider the 5DII further. In the context of landscape shots, that's different, but for moving subjects, the 5DII is not the body of choice.

...can I do the "a lot of work" once and then copy and paste the develop settings to the other 499 (in Lightroom) or would it differ for each photo and I would have to do "a lot of work" 500 times?

They're pretty stereotyped adjustments, in the sense that the 5DII's images are generally sharper (for the same magnification), have better color, contrast, and saturation, and lower ISO noise. The challenge is increasing the sharpness also increases perceived noise, so tradeoffs need to be made. As I said...close...but not quite there.

 

[awinphoto]

I don't see ANY visible noise anywhere it was present in the original 100% crop. I think the worst areas were the shaded parts of the fence posts and the columns of the building in the background. At print size, WITH an added boost to contrast, no noise is visible at all.

I know people like to pixel-peep, and LOVE to complain about the noise levels of the 7D...but I think the results are WAY overblown and overstated. The 7D is a very high resolution camera...possibly as high as it can get without outresolving the current crop of Canon's BEST lenses. The simple fact of the matter is, no one views images at 100%. For on-screen viewing, the final resolution is quite possibly 1/4 or less the native resolution, and all noise is absorbed. For native-resolution prints, all noise is absorbed. For reasonably enlarged prints, noise tends to help keep gradients smooth, and IF it is visible, a bit of noise reduction is usually all thats necessary to fix the problem.

Print and screen pixel peeping are 2 separate things... I think the pixel peeping is a dis-service for the camera as printing is what canon is aiming viewers to be excited about... the final product... not on peoples monitors blown up to 200% looking to find ways to tear apart their products... If you do that with the 5d2 or 3, I got news for you... you'll find noise too...

 

[dtaylor]

I think that "catz" post is the most relevant to my situation. I am often pushing my RAW files to the edge to get more detail / color / contrast out of them.

So am I. Noise is simply not an issue in my prints.

As for "pixel peeping". I don't think that's what I'm doing.

It's exactly what you're doing. But I think there is an issue beyond that and it might be this:

If I try to crop this photo down _at all_ I can see the noise with the full photo being displayed (ie not zooming in) on my 27" iMac. And that's doing a pretty small amount cropping (like only including the whole church). If I go any further, like just the front half of the church... it looks like I took this photo at ISO 800+!

Your description is confusing. I don't know if you are viewing at 50%, 100%, 500%? What are you doing when you say cropping but not zooming in? It sounds like the more you crop, the more noise you see, which implies you are cutting up the photo and viewing those cuts at increasing magnification. But from your description I can't tell what that magnification might be. For all I can tell you're looking at the equivalent of 200" prints.

But even if you don't touch the photo at all and just view at 100% (pixel peeping), you are viewing on a 27" iMac. I love Apple's products, but I *** hate *** their screens for photography work because you cannot tone down the contrast and saturation no matter how hard you try. They look beautiful for anything else, but I think they suck for photography. Even after color calibration they do not give an accurate preview of prints. That's why I have a 3rd party monitor. Out of the box it had the over-the-top-candy-world brightness, contrast, and saturation that is the fad right now, but I could tone it down to match real prints.

So at 100% it's like looking at a 50" or 60" print but with the brightness, contrast, and saturation shoved up. I'm sorry, but you will see noise in a 60" print even from a 5D2, especially if you push those three variables over the top.

How anyone shooting landscapes would think that is acceptable is beyond me. I've cropped ISO 100-200 photos from my XSi _way_ down and been satisfied with the results (and I didn't have to apply any NR at all!).

Here is a screenshot of what I'm seeing on my screen: http://farm8.staticflickr.com/7166/6482020907_8689f39356_o.jpg

Again, I'm confused as to whether you are actually cropping, or just viewing at 100%. That sample looks comparable in size to what I see viewing the image at 100% in PS on my monitor, so let's go with that. You're looking at the equivalent of a roughly 60" print. Do you make 60" prints? If so, you will not be happy with any small format DSLR and couldn't possibly be happy with the XSi which would be mush at 60". You need either the MF 645D, or a pano head to stitch 3 frames from a DSLR.

You made the comment that other people will see the same on their screens. Do you often crop out 1/3rd to 1/4th of your landscape photos for viewing on 27" screens? Or do you frame the scene properly, maybe do a bit of cropping, but otherwise show people complete photos? Because even on an iMac screen a complete photo looks fine. (Except occasionally in Apple's Preview. At some magnifications it just wrecks photos. And it's particular scales to, because I've played with zooming and showing full screen and a difference of just 20% might result in the preview looking fine.)

 

[Michael7]

Everything was going great... until I got home and loaded up those photos in LR3... and saw a ridiculous amount of high frequency noise ALL over the place... even when shooting at ISO 100-200!

I agree. I like to call the 7D, "my little noise pig". Noise levels at low ISO are unnaceptable for a $1500 camera. I found the 50D and 40D files superior in IQ, especially the 40D. My 7D at ISO 800 looks like my 40D at 1600. The 40D files are smooth and clean, giving better fur detail and landscape detail. The 7D files have a bit of mush on top of the noise. I'm using a combination of L lenses and EF-S lenses.

I've shot with 2 50D's in the past and hated both of them for the low ISO noise... Compared to my 7D when I sold/unloaded my 50D's the 7D files were superior in every way, especially in low ISO which was surprising to me given it had more MP... I may have been blessed with a really good 7d or cursed with really bad 50Ds... but from my experience, they are not even close. Also food for thought... this format of camera 35mm (7d is smaller than 35mm)... it was only previously designed to output small prints... 8x10's, 5x7's... maybe the occasional 11x14... and that was a full 35mm... now the 7d natively pumps out 11x17's at 300dpi and it's sensor is half the size of a 35mm film... i think you guys are asking for a lot from such as small sensor, pixel peeping without even printing it out...

This test clearly shows the 50D is superior in low ISO noise to the 7D in RAW. I shoot RAW, and agree with these results:

http://www.cameralabs.com/reviews/Canon_EOS_7D/noise_RAW.shtml

The 7D is a little farting noise pig. Oink Oink. Not good for $1500, IMHO.

 

[unfocused]

dtaylor, jrista and awinphoto, I'm giving you all some positive karma. I appreciate that you've taken the concerns of the original poster seriously and provided intelligent responses.

BTW, I really appreciate this explanation:

The blue sensels in most bayer sensors are lower sensitivity than the green sensels. There are also half as many of them as there are green. They require greater amplification to match the luminance level received by green sensels. This is a well-known phenomena with bayor sensors in general.

I've often noticed this issue with shots of a blue cloudless sky. The image looks fantastic at first, but on closer inspection I do see noise in these images. Usually not enough to freak me out, but enough to make me wonder if there is something I could be doing differently. (And, again, it's NOTHING in comparison to what these used to look like on 35mm transparencies.)

I'd love to know more about your individual workflows. Would you consider starting a thread to that effect?

 

[friedmud]

Print and screen pixel peeping are 2 separate things... I think the pixel peeping is a dis-service for the camera as printing is what canon is aiming viewers to be excited about... the final product... not on peoples monitors blown up to 200% looking to find ways to tear apart their products... If you do that with the 5d2 or 3, I got news for you... you'll find noise too...

Read again... I wasn't blowing these "up to 200%". I am looking at the whole photo on my screen. Not zooming at all (ie I can see the edges of the photo). I am _not_ "pixel peeping" (I am _not_ looking at the photos at 1:1 with my screen. Just viewing the photo like anyone would.

Everyone: Print is only one final avenue for my prints. Most often they are enjoyed on high resolution monitors.

I agree that none of the noise present in the shots I've posted would not show up in any reasonably sized print.

THAT IS NOT THE POINT

The point is that I paid $1500 and am receiving _inferior_ quality at ISO 100-200 to the camera I bought 3 years ago for $700. THAT IS NOT GOOD.

No other arguments need to be made.

For my purposes (and for others who are interested in landscape photography) this body does not work.

As for the actual quality of some of the photos I posted: I totally agree guys... I wasn't taking any time at all to actually take attractive photos here. Any CA would have been fixed in lightroom (fixed automatically by just turning on lens adjustments). Blurring, etc would be fixed with tripods, etc. Please stay on target and talk about the noise.

The last photo I posted this morning was shot handheld at 1/1000 in bright daylight. There is no noticeable blurring in the photo beyond a bit of softening at the edges which is normal for this lens.

I can't even believe how far some people go to justify this stuff. There is no way I could keep this camera with this output for landscape photography. The first time I wanted to do a 25% crop (by which I mean crop down to a piece of the photo that is taking up approximately 25% of the original space) of a once in a lifetime photo SHOT AT ISO 100 I would want to blow my brains out.

When shooting at low ISO you should NEVER have to apply NR (or at least not much depending on how much you've cropped) and you should feel free to crank up the sharpness. Anything less is unacceptable.

 

[Cfunkexplosion]

"Earthshattering?"

Talk about First World problems.

It's not an inexpensive camera, I get that. And if you are unhappy with it, you should return it.

I do want to say that I doubt the people looking at your photos are likely to judge them based on what is essentially a forensic examination at high magnification. It seems to me that many people obsess over the minutiae of technical details and forget about the actual composition.

I've found my 7D to be a good camera. It's a durable tool that shoots quickly and has perfectly acceptable ISO performance for the price. The failings of my photography have to do with the skills, or lack thereof, of the person behind the camera. For me, noise at 100% has never been much of factor in the creation of compelling images.

 

[7enderbender]

Maybe I'm a little thick again but could some one show me where there are issues with "noise" in those samples? I see issues, especially with CA but all of that seems more a lens issue. I just don't see it.

Seems like a typical shot by the way where digital cameras still struggle due to the unforgivingly low dynamic range. I'd play with the exposure a little first and see what that does. Also: any chance some of the little software gizmos like "highlight priority" or something like that are turned on?

 

[awinphoto]

Everything was going great... until I got home and loaded up those photos in LR3... and saw a ridiculous amount of high frequency noise ALL over the place... even when shooting at ISO 100-200!

I agree. I like to call the 7D, "my little noise pig". Noise levels at low ISO are unnaceptable for a $1500 camera. I found the 50D and 40D files superior in IQ, especially the 40D. My 7D at ISO 800 looks like my 40D at 1600. The 40D files are smooth and clean, giving better fur detail and landscape detail. The 7D files have a bit of mush on top of the noise. I'm using a combination of L lenses and EF-S lenses.

I've shot with 2 50D's in the past and hated both of them for the low ISO noise... Compared to my 7D when I sold/unloaded my 50D's the 7D files were superior in every way, especially in low ISO which was surprising to me given it had more MP... I may have been blessed with a really good 7d or cursed with really bad 50Ds... but from my experience, they are not even close. Also food for thought... this format of camera 35mm (7d is smaller than 35mm)... it was only previously designed to output small prints... 8x10's, 5x7's... maybe the occasional 11x14... and that was a full 35mm... now the 7d natively pumps out 11x17's at 300dpi and it's sensor is half the size of a 35mm film... i think you guys are asking for a lot from such as small sensor, pixel peeping without even printing it out...

This test clearly shows the 50D is superior in low ISO noise to the 7D in RAW. I shoot RAW, and agree with these results:

http://www.cameralabs.com/reviews/Canon_EOS_7D/noise_RAW.shtml

The 7D is a little farting noise pig. Oink Oink. Not good for $1500, IMHO.

To each their own... as I said it very well could be I got a good sample of the 7d and it has worked well for me, my clients whom are household names whom you would know, and gotten rave reviews review website after review website... I know your website you are referring to... i've commented in the blog section below commenting it was odd that the 7d with an inferior lens they used to test the camera was sharper than the 5d2 and lens they used for that test... If you or derek or anyone else dont like the camera, return it, sell it, do whatever with it... of all the client whom I shoot for, none of them has ever complained about noise, and I have used this camera to pump out tradeshow booth murals for that stretch over 15 feet wide (albeit printed at like 72dpi or so) No one is telling you that you should love the camera or even keep the camera... It has worked wonders for people like me and others and has been a steady workhorse for my photography and will continue to assist my 5d2 until the day comes where I see no need for the camera...

For what it's worth, i've had coldwell banker display my photos screen res on large HDTV monitors throughout the vancouver olympics... I did not hear any grumblings about noise in the photos displayed... They were all shot with the 7d.

 

[jrista]

dtaylor, jrista and awinphoto, I'm giving you all some positive karma. I appreciate that you've taken the concerns of the original poster seriously and provided intelligent responses.

BTW, I really appreciate this explanation:

The blue sensels in most bayer sensors are lower sensitivity than the green sensels. There are also half as many of them as there are green. They require greater amplification to match the luminance level received by green sensels. This is a well-known phenomena with bayor sensors in general.

I've often noticed this issue with shots of a blue cloudless sky. The image looks fantastic at first, but on closer inspection I do see noise in these images. Usually not enough to freak me out, but enough to make me wonder if there is something I could be doing differently. (And, again, it's NOTHING in comparison to what these used to look like on 35mm transparencies.)

I'd love to know more about your individual workflows. Would you consider starting a thread to that effect?

Thanks for the vote of confidence.

Smooth gradients of any prime color can be problematic from a noise perspective, simply due to the nature of noise and how luminance is generated from a bayer sensor (i.e. if you shoot an even red gradient, you have to amplify red sensels more to get a proper luminance, as green and blue sensels will be capturing minimal light...that can produce a very noisy red image even at ISO 100). Even at ISO 400, I often ran into noise issues in moderately even green scenes with my 450D. The more even and smooth a prime-colored thing, the more likely the *random* nature of noise will exhibit. The random effects of noise tend to be most visible in even tone things of the prime colors red, green, and blue; when SNR is low (i.e. black and dark shadows); and in under-exposed high-key tones (light gray and white). Smooth gradients of any of those will usually exhibit more noise than non-prime colors, not just blue. I think the problems friedmud is having are due to the fact that the scene was a bit under-exposed (I think some judicious ETTR would have probably helped a LOT), and the fact that there were large, unbroken expanses of even blue sky.

Regarding my workflow, I'm sure I could provide answers...I'm not 100% certain what your asking though. If you start another thread, I'll take a look and offer what insight I can.

 

[Michael7]

Congratulations, that's quite impressive. But that still doesn't mean the 7D is superior to the 50D in low ISO noise. It's not. And it's even worse compared to the 40D.

The amount of noise in wildlife shots (fur, darker areas at the wildlife standard ISO 400) and landscapes (low ISO is where most people shoot here) is inexcusable for a $1500 camera. It can never be called an "upgrade" if the result is more noise.

 

[jrista]

This test clearly shows the 50D is superior in low ISO noise to the 7D in RAW. I shoot RAW, and agree with these results:

http://www.cameralabs.com/reviews/Canon_EOS_7D/noise_RAW.shtml

The 7D is a little farting noise pig. Oink Oink. Not good for $1500, IMHO.

I find that test VERY suspect. They bumped up sharpness for the 7D by 5! They also reduced the sharpness for the 5D by 2!! The 7D shots in that comparison are FAR sharper than either the 50D or the 5D, and the sharpening is guaranteed to enhance the levels of noise. The 7D is known to be a little soft, and usually most objective reviewers bump up sharpness in DPP by 2, at most 3, to even things out. At 3, I've found the 7D to be sharper than the 5D, so I think a fair comparison would have been to keep the 5D sharpness at default, bump the 7D to 2, and THEN compare noise. As it stands now, the linked comparison has way over sharpened the 7D, which feels very much like the author is biasing against the 7D. Its clear as day how much softer the 5D shots are...if you visually balance out SHARPNESS between all the samples, then compare noise, I think the 7D would fair a hell of a lot better, like it should.

 

[jrista]

Read again... I wasn't blowing these "up to 200%". I am looking at the whole photo on my screen. Not zooming at all (ie I can see the edges of the photo). I am _not_ "pixel peeping" (I am _not_ looking at the photos at 1:1 with my screen. Just viewing the photo like anyone would.

That might actually be the issue. It looks like your using Lightroom, and one of the things that annoys the hell out of me about Lightroom is that when viewing images "fit" to the screen, it does a very nasty, quick, and DIRTY scaling. Its not a quality downscale, its cheap nearest-neighbor downscaling. I think the problem your having is not that the image is truly as noisy as it appears on your screen...it might be that the larger size of the 7D images requires a greater amount of downscaling to "fit", more so than the 40D. Cheap nn filtering tends to get worse the more you down scale, resulting in more and more noise, not less.

I would suggest downscaling with bicubic smoother in photoshop to roughly the same size, and see how things look. You seem to be pretty thoroughly poisoned against the 7D already, so maybe its too late to change your mind...but don't trust Lightroom's standard "fit" previewing as a measure of IQ...its simply not.

 

[Cfunkexplosion]

Here are some landscape photos taken with this "unacceptable" camera.

http://500px.com/alwaysbj182

That I can't do this with my 7D, or that you can't do this with yours, is not the fault of the camera.

 

[dtaylor]

Read again... I wasn't blowing these "up to 200%". I am looking at the whole photo on my screen. Not zooming at all (ie I can see the edges of the photo). I am _not_ "pixel peeping" (I am _not_ looking at the photos at 1:1 with my screen. Just viewing the photo like anyone would.

But you've also said you're cropping, and the mix of terminology and lack of details leaves people confused as to what the final magnification is.

The screenshot you posted that I took a look at showed roughly the same view I see at 100% (1:1, i.e. pixel peeping) on my screen. The noise looks the same as well, and while it's there, it's certainly not Earth shattering. I know that noise looks worse on a typical Apple screen with typical settings. But, again, this is 100%, roughly equivalent to a 50-60" print. No small format DSLR is noise free at that size.

Everyone: Print is only one final avenue for my prints. Most often they are enjoyed on high resolution monitors.

Viewing the entire original jpeg on my screen is 50% magnification. I see no noise at that scale.

The point is that I paid $1500 and am receiving _inferior_ quality at ISO 100-200 to the camera I bought 3 years ago for $700. THAT IS NOT GOOD.

No you're not. You're not viewing them at the same scale, and you're viewing images that, due to other issues, are not exhibiting the fine detail the 7D is capable of.

For my purposes (and for others who are interested in landscape photography) this body does not work.

I've got a couple 22" portfolios that say differently. And I know a landscape photographer who shoots DSLRs, MF, and 4x5. He has said that through 24" it's extremely hard for him to tell the difference between his 7D shots and his 4x5 shots, and that he will grab his 7D before grabbing his FF (12 MP) for landscape work.

I can't even believe how far some people go to justify this stuff.

Nobody is justifying it. We're simply pointing out that people do not expect 60" views out of small format cameras. Not any currently shipping small format camera. Not the 5D2 or D3x, and I wouldn't expect it out of the 1Dx either. If you actually make 60" prints, you need a larger format or stitching.

There is no way I could keep this camera with this output for landscape photography. The first time I wanted to do a 25% crop (by which I mean crop down to a piece of the photo that is taking up approximately 25% of the original space) of a once in a lifetime photo SHOT AT ISO 100 I would want to blow my brains out.

Do I understand you to mean that you would take a 1296x864 crop out of the full file? That's a 1.1 MP image. You can't expect to view or print a 1.1 MP image at 27". 5x7 would be about the limit if there's any fine detail.

Do you mean a 4.5 MP image? 8x10 or maybe 11x17, depending on subject and shot quality.

Do you really shoot, crop out 25% of your images, and view/print at 27"? How could you be satisfied with any small format DSLR doing this? Why not just frame the photo the way you want it to appear?

When shooting at low ISO you should NEVER have to apply NR (or at least not much depending on how much you've cropped) and you should feel free to crank up the sharpness. Anything less is unacceptable.

Whether or not I apply NR at low ISO depends on the subject, exposure, and how hard I've pushed the RAW file. As for sharpening, well, I'm careful about it, but definitely fall on the sharp side. Nothing I print goes without both LCE and some sharpening.

 

[awinphoto]

Congratulations, that's quite impressive. But that still doesn't mean the 7D is superior to the 50D in low ISO noise. It's not. And it's even worse compared to the 40D.

The amount of noise in wildlife shots (fur, darker areas at the wildlife standard ISO 400) and landscapes (low ISO is where most people shoot here) is inexcusable for a $1500 camera. It can never be called an "upgrade" if the result is more noise.

From my personal experience working with both cameras, this is what my personal opinions are of the two cameras... I found the 50d awful and hated it so much i sold it to my father-in-law the day the 7d was announced/released... If you find the 50d better, use it... i'm not saying you shouldn't.. you could sell your 7d if you haven't and buy 2 50d's if you wish... All I can give you is my own personal findings... A company I shoot for still have a 50D and comparing shots I've taken with that camera vs shots I've taken with my 7d or 5d2... i'd take my cameras any day of the week... but that is me... I'm not debating or telling you what you should do... I'm just stating my POV from my experience with my equipment... that's all.

 

[awinphoto]

Here are some landscape photos taken with this "unacceptable" camera.

http://500px.com/alwaysbj182

That I can't do this with my 7D, or that you can't do this with yours, is not the fault of the camera.

Great shots... +1

 

[jrista]

Here are some landscape photos taken with this "unacceptable" camera.

http://500px.com/alwaysbj182

That I can't do this with my 7D, or that you can't do this with yours, is not the fault of the camera.

Great shots... +1

Ditto! +1