hi guys! just want to get some feedbacks especially to all 7D owners out there, i was browsing the net and reading some reviews about the 7D and there are good few blogs and reviews that are saying that the image quality of the 7D is SOFT and lot of NOISE even on low ISO. any thoughts? i know there are a lot of people very exicited with the 7D and its features but does it really lives up to our expectation?
For starters, the 7D places higher demands on the sharpness of your lens than pretty much any camera out there. This is mitigated somewhat by the crop factor when using full-frame lenses, since you will mostly be using the center of the lens.
Given a sharp lens, even a 50mm 1.8 stopped down to f/5.0, I am amazed at the detail that comes through at any ISO less than 3200. Yes, if you pixel peep, you will see some noise starting about 400 ISO, but even then, the graininess is not that objectionable.
At 3200 ISO, which the Auto-ISO visits frequently indoors with no flash, the images have more noise and reduced dynamic range, but they are still usable. You just lose the luxury of cropping the heck out of it that you have at lower ISOs.
Shoot, depending on the image, I've got some 6400 ISO shots I like. Print these up to 5x7 and you will never be able to tell.
As Bryan's review says, and I've seen it on mine too, Canon seems to have changed the scale on the sharpening level. The default of 0 or 1 isn't always enough. I usually use sharpness=2.
In capable hands, it's a very capable camera. No, as much as the haters would have liked, Canon did not botch this one. I hold the opinion that there are too many people on the internets who are more into photography for the gear and the gear's specs, since they can drool/agonize over the specs while they are bored at work. More of these folk need to be out taking pictures. They will find the specs are secondary to the art.
Just sayin' [A]
200 ISO example:
[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.35.62/IMG_5F00_1526R_5F00_lzn.jpg[/img]
3200 ISO example:
[img]/cfs-file.ashx/__key/CommunityServer.Components.UserFiles/00.00.00.35.62/IMG_5F00_1598.JPG[/img]
Originally Posted by msmiley
Agreed. I have the 7D and I love it. But, with any camera, most of the time the problem is not the camera, but the idiot behind it [:P]
brendan
thanks guys! just want to get some thoughts from this issue. i don't have a 7D though, just curious with the early issues regarding the 7D which i think is a very good camera and very early to have issues since its release date. i'd like to have one though much better if the 5D mark II.
Your ISO 3200 example is amazing. I'm using a 24" monitor and I can see very little noise. I am a proud owner of a 7D and have several images shot at ISO 2000 that show very little noise. Noise is more visible at ISO 3200 but very usable photos.
Most of the blogs and reviews are wrong and misleading: they fail to account for differences in scale (spatial frequency) between the 7D and other cameras. That was the same mistake that lead many reviewers to think the 50D was actually worse than the 40D, when in fact it is the same or better.
That said, there are indeed some issues with the 7D. The "softness" is caused by an interaction between a certain camera flaw (mismatched spectral QE between the two green channels) and the raw conversion. There may have been a manufacturing blunder that caused the 7D to have imbalanced greens. With a normal raw conversion, that results in mazing artifacts -- which causes the perception of noise at low ISO, even in the midtones and highlights. That's the artifact that early raw converters had (the ACR beta, for example). But later raw converters fixed the problem by averaging the two green channels together; the downside being a loss of contrast/resolution.
Another problem is line noise (FPN in this case) at just 8-9 stops of dynamic range at low ISO. This may be the "NOISE even on low ISO" that you read about. Fortunately it is very easy to remove during raw conversion with a simple bias frame subtraction, which boosts the usable dynamic range by several stops. Unfortunately, none of the consumer raw converters support that feature. Only astrophotography converters like IRIS have it, and they are suboptimal for other reasons (lackluster demosaic and color matrix).
So the 7D does have some flaws that will hopefully be fixed in the next camera with this sensor (e.g. 60D or 550D), but overall, I'd still say that the 7D is excellent.
Daniel, can these flaws beaddressedby Canon via firmware? By the way, I am very happy with what I get with my 7D.
Hard to say. For the second problem (line noise at low ISO), the solution is easy to implement in firmware, but I'm not sure if the hardware is fast enough to do it without slowing down the frame rate.
For the green channel issue (softness), there are several possibilities. One is that it's not a flaw at all, and Canon did this on purpose. Maybe there is some advantage that I don't know about, and they felt it was worth losing resolution to gain whatever the benefit is.
Another possibility is that the problem may be simpler than what I described. If it is just a simple gain imbalance on most units -- not an actual difference in spectral QE -- then that could be corrected in firmware. It would require the user shoot some "calibration" shots, then calculate the gain imbalance factor, and include that factor with the metadata of the raw file.
If instead it really is the more complex problem (QE imbalance), then it cannot be corrected for more than one color temperature at a time. For example, they could correct it for daylight, but then tungsten would still be flawed.
Both of these are sufficiently complex fixes that I find it unlikely for Canon to go through the work. It's much simpler to just tell photographers to increase sharpening.
Posting Permissions
Reply With Quote