Canon images are too soft! Can they be fixed?

Jim B (MSP) said:

CMurdock said:

Jim B (MSP) said:

...

Click to expand...

... However, I have seen similar blurry shots from the G15 in other reviews. ..

Click to expand...

It is comments like this that cause you to lose credibility. The shots were not blurry - however, they may not have been as sharp as you wanted to see. This is not the same. And you said - Canon shots on a beach would be blurry. Give me a break.

But you did accomplish your objective - that was to put down the Canon G16. You didn't say that up front, but your comments, tone, and examples were all in line with that.

If you really wanted to learn how to sharpen a G16 shot - you would have set off down that path. But sharpening never seemed to enter in to a discussion. That is probably because, as you said above - you know all those things.

Click to expand...

I have already stated my purpose in starting this thread: I like the G16 except for the image quality, and I was hoping that someone might know some "secret" (such as a combination of camera settings) that would result in sharper images than I am seeing. The G16 has better color rendition than other cameras that I might buy, and one of the things I would do with it is photograph my colorful products.

As for losing credibility, I don't see how I've done that. Every time I read a review of a Canon camera, the sample images are always softer than I think they could be, and the softness of Canon's images seems to be getting worse as time passes. That's my opinion, whether you like it or not.

clark321 said:

After looking at that comparison I would have to agree with you. The difference is huge. I couldnt say why it is but I doubt it is anything to do with the sensor.

The comparison is actually quite bizarre, DPR comments, "The P7700's version of the photo is much sharper, from one edge of the frame to the other. The G15 is quite a bit softer, especially on the right side. After testing several PowerShot G15s, we are confident that this is just sample variation"

Looking at Imageresource test shots the IQ of the P7700 & G16 look very similar.

Can there really be so much sample variation?

Click to expand...

First of all, I am glad that you are seeing the difference that I am seeing, and that you consider it to be significant.

The question arises: If DPR thinks that the soft Canon image is the result of "sample variation", why did they use it? Furthermore, does a good camera produce images with so much variation?

Before I bought the P7700, I read reviews of both the P7700 and the G15, and I looked at sample images from both cameras. The sample images from the P7700 looked consistently sharper to me than the sample images from the G15. Not only that, I have seen sample images from other Canon cameras -- including their DSLRs -- that looked unacceptably soft to me. It used to be the other way around. It used to be that Canon compact cameras had better overall image quality than Nikon compact cameras, but that seems to have changed.

(In case you didn't read my other posts, I ended up returning the P7700 for reasons that didn't have to do with the image quality.)

You're comparing a $2300 camera with a crappy $200 P&S (s110)? What do you expect? Is this some kind of joke???

--
Tom
Look at the picture, not the pixels
------------
Miss use of the ability to do 100% pixel peeping is the bane of digital photography because it causes people to fret over inconsequential issues.
http://www.flickr.com/photos/63683676@N07/
http://www.flickr.com/photos/25301400@N00/

CMurdock said:

stevo23 said:

If you're trying to get a G16 or S110 to compare favorably to the SD1, I think you will be eternally disappointed. There are fairly large differences all the way around.

And truthfully, while the S110 has carved out a nice place for price and portability, I don't think the image quality compares favorably to many similar models. It is what it is. There is a Canon out there for you I'm sure.

Click to expand...

I've already addressed some of your comments in later posts. I used the S110 because I was considering getting the S120 (the next generation). The S110 has the same innards as the G15, and the S120 should have the same innards as the G16. Honestly, however, I used the S110 crop because it was worse than the G15 crop, and it made my point better.

You should look at my crops comparing the Nikon P7700 to the G15. That's a fairer comparison.

Click to expand...

I have to admit that on these test shots, when I looked at the P7700 and similar compared to the G16 and S110, I was surprised that the Nikons seemed to be more crisp and sharp. I think to some degree, you might look around for other photos and tests of the S110 or S120 to get an idea. The S110 are nice for how small they are and should produce nice photographs.

tbcass said:

You're comparing a $2300 camera with a crappy $200 P&S (s110)? What do you expect? Is this some kind of joke???

Click to expand...

I'm not going to answer this for the 100th time. You need to look down the thread to my post where I made a fairer comparison between the Nikon P7700 and the Canon G15. The point I was trying to make is that I generally like Canon cameras, but I find the image quality to be consistently too soft, and I was wondering if someone knew a way to improve the image quality.

I made a little comparison between the 40MP Pentax 645D and the G15. comparison page

I sharpened a bit and downsized the 645 sample image. The result is quite Foveon-sharp. I think the G15 also produced a sharp image with its small sensor and resolution. View them at 100%:


Pentax, resized from 40 megapixels to match Canon G15


Canon G15 ACR RAW sample, sharpened

bizi clop said:

I made a little comparison between the 40MP Pentax 645D and the G15. comparison page

I sharpened a bit and downsized the 645 sample image. The result is quite Foveon-sharp. I think the G15 also produced a sharp image with its small sensor and resolution. View them at 100%:

Click to expand...

Great test, they are the same but with one exception, the G15 is showing more definition of the red fluffy balls, whereas the pantax has saturated the colour and lost the details.

Brian

I recommend DXO for a top of the notch imge correction. This requires however that your cam is supported by DXO. For obtaining the best results your cam should be able shooting pictures in raw format I owe a Canon S100 that is showig some image softness, but thanks to DXO I get afte image correction pictures without almost any visible blur, that means being sharp to 1 pixel.

Most people these days will look at photos on monitors. Most monitors today are at 1920 X 1080 or can display 2MB (2,073,600) pixels. Divide that by your monitor size and you are lucky if you are getting 100 ppi. Magazines typically print at 300, while fine art magazines print at 720. So unless you are taking 2MB photos, there is some serious downsizing that is occurring when you view your photos full size on your monitor. That is why there is no shortage of folks here on DPReview who will tell you that if you want to really examine the quality of your photo, you print it, large - to take advantage of significantly higher pixel densities in print and there are those who prefer the qualities of reflected light. Pixel peeping might be fun, but when you view the entire image (not a massively enlarged small portion of an image) on the monitor - what differences - in resolution and other properties - can you really discern? I don't know the answer but I think it is an interesting question.

edispics said:

Most people these days will look at photos on monitors. Most monitors today are at 1920 X 1080 or can display 2MB (2,073,600) pixels.

Click to expand...

While today its true, in a few years we will look them on televisions/monitors/tablets/glasses/etc having UHD/4K resolutions.

If you like Canon and sharpness is your main focus (pun intended) I'd recommend grabbing an EOS-M with 22mm/f2.0 (35mm equivalent) lens, before they're all gone - which will be shortly. There's been a fire sale for the last couple of months, but Amazon still has some Warehouse Deals and third party sellers offering the camera/lens.

The 18-55 zoom is also quite good, but the 22 is faster and slightly sharper, and would probably be better suited to you stated needs.

It's an APS sensor camera, with a great touchscreen. There are a few quirks, but if the DP2 is your dream camera, from what I've read you'll have to get used to quirks.

As previously mentioned in this thread, you can adjust sharpness, saturation, contrast, etc. in various Canon cameras, like the S95, which I have. With adjustments (including sharpness) I've always gotten very pleasing pictures for my purposes of everyday use. (I shoot JPEG with it and do minor pp). I've seen similarly nice photos from a G16.

However, when I want the best image quality, a small sensor compact isn't what I take with me.

If you like the S120/G15-16 form factors and menus, I'd suggest getting the S110 for $219 and seeing what you can do to set it up to your liking. From what you've said in this thread, I doubt it will please you, but at least you won't be out much money.

CMurdock said:

tbcass said:

You're comparing a $2300 camera with a crappy $200 P&S (s110)? What do you expect? Is this some kind of joke???

Click to expand...

I'm not going to answer this for the 100th time. You need to look down the thread to my post where I made a fairer comparison between the Nikon P7700 and the Canon G15. The point I was trying to make is that I generally like Canon cameras, but I find the image quality to be consistently too soft, and I was wondering if someone knew a way to improve the image quality.

Click to expand...

OK. In the menu turn up sharpening and you will get a sharper photo. The default in camera jpg sharpening varies from camera to camera model. It is a big mistake to judge cameras based solely on web samples.

--
Tom
Look at the picture, not the pixels
------------
Miss use of the ability to do 100% pixel peeping is the bane of digital photography because it causes people to fret over inconsequential issues.
http://www.flickr.com/photos/63683676@N07/
http://www.flickr.com/photos/25301400@N00/

edispics said:

Most people these days will look at photos on monitors. Most monitors today are at 1920 X 1080 or can display 2MB (2,073,600) pixels. Divide that by your monitor size and you are lucky if you are getting 100 ppi. Magazines typically print at 300, while fine art magazines print at 720. So unless you are taking 2MB photos, there is some serious downsizing that is occurring when you view your photos full size on your monitor. That is why there is no shortage of folks here on DPReview who will tell you that if you want to really examine the quality of your photo, you print it, large - to take advantage of significantly higher pixel densities in print and there are those who prefer the qualities of reflected light. Pixel peeping might be fun, but when you view the entire image (not a massively enlarged small portion of an image) on the monitor - what differences - in resolution and other properties - can you really discern? I don't know the answer but I think it is an interesting question.

Click to expand...

I think the time when we judge image quality by the printed result has passed. My guess is that 99.9% of all digital photos are never printed. A typical home user might print only the occasional photo that is so nice that s/he wants to frame it. Many more photos than that will end up on the web. The cameras that have the best image quality will look sharp on-screen. Actually, computer screens are good because they reveal softness in images that you wouldn't notice if the image was just printed.

For me, pixel peeping isn't just "fun". It is how I view most of my photos. I will reduce the photo to the width of the screen and appreciate it that way, and then I will enlarge it to 100% and appreciate the details. I have been doing this since I purchased my first digital camera around 2001. I actually get more pleasure from "pixel peeping" than I get from viewing the full-size image. That Sigma image that I posted of the red tree looks glorious when you pixel peep, but it loses its magic when viewed at screen width. Even in an 8.5 x 11 print, I think it wouldn't look as good. To really appreciate that photo printed, you would have to print it at poster size.

wexie said:

If you like Canon and sharpness is your main focus (pun intended) I'd recommend grabbing an EOS-M with 22mm/f2.0 (35mm equivalent) lens, before they're all gone - which will be shortly. There's been a fire sale for the last couple of months, but Amazon still has some Warehouse Deals and third party sellers offering the camera/lens.

The 18-55 zoom is also quite good, but the 22 is faster and slightly sharper, and would probably be better suited to you stated needs.

It's an APS sensor camera, with a great touchscreen. There are a few quirks, but if the DP2 is your dream camera, from what I've read you'll have to get used to quirks.

As previously mentioned in this thread, you can adjust sharpness, saturation, contrast, etc. in various Canon cameras, like the S95, which I have. With adjustments (including sharpness) I've always gotten very pleasing pictures for my purposes of everyday use. (I shoot JPEG with it and do minor pp). I've seen similarly nice photos from a G16.

However, when I want the best image quality, a small sensor compact isn't what I take with me.

If you like the S120/G15-16 form factors and menus, I'd suggest getting the S110 for $219 and seeing what you can do to set it up to your liking. From what you've said in this thread, I doubt it will please you, but at least you won't be out much money.

Click to expand...

Wexie, you have given me some very good advice. Thank you.

I want the Canon only for product photography (because it has reasonably good color rendition). Since product photographs on the web do not require the best camera, why shouldn't I get an old model that is now discounted? Even if the images are slightly soft, that won't be apparent because the images will be shrunk for the web. The problem is that my 8-year-old Canon compact doesn't have the best color rendition, and I can tear my hair out trying to get the images to look right in post-processing. But last year's Canon model should certainly have better color rendition than my 8-year-old model.

Actually, I was hoping to use my product-photography camera for street shooting until I can afford the Sigma camera that I want. But I guess I will have to give up on that.

Thank you!

CMurdock said:

edispics said:

Most people these days will look at photos on monitors. Most monitors today are at 1920 X 1080 or can display 2MB (2,073,600) pixels. Divide that by your monitor size and you are lucky if you are getting 100 ppi. Magazines typically print at 300, while fine art magazines print at 720. So unless you are taking 2MB photos, there is some serious downsizing that is occurring when you view your photos full size on your monitor. That is why there is no shortage of folks here on DPReview who will tell you that if you want to really examine the quality of your photo, you print it, large - to take advantage of significantly higher pixel densities in print and there are those who prefer the qualities of reflected light. Pixel peeping might be fun, but when you view the entire image (not a massively enlarged small portion of an image) on the monitor - what differences - in resolution and other properties - can you really discern? I don't know the answer but I think it is an interesting question.

Click to expand...

I think the time when we judge image quality by the printed result has passed. My guess is that 99.9% of all digital photos are never printed. A typical home user might print only the occasional photo that is so nice that s/he wants to frame it. Many more photos than that will end up on the web. The cameras that have the best image quality will look sharp on-screen. Actually, computer screens are good because they reveal softness in images that you wouldn't notice if the image was just printed.

For me, pixel peeping isn't just "fun". It is how I view most of my photos. I will reduce the photo to the width of the screen and appreciate it that way, and then I will enlarge it to 100% and appreciate the details. I have been doing this since I purchased my first digital camera around 2001. I actually get more pleasure from "pixel peeping" than I get from viewing the full-size image. That Sigma image that I posted of the red tree looks glorious when you pixel peep, but it loses its magic when viewed at screen width. Even in an 8.5 x 11 print, I think it wouldn't look as good. To really appreciate that photo printed, you would have to print it at poster size.

Click to expand...

It sounds to me like you better start saving for that Sigma camera. That is the output you would really enjoy. Or perhaps you should be the Pentax 645. Now those cameras will let you pixel peep and really enjoy what you have. No Canon, or any other brand of small sensor P&S camera is going to do it for you.

Curious if you feel these are soft at 100% The second has a fairly light unsharp mask applied to it. They are jpg output from my Powershot SX10IS, it doesn't do raw unless you bust out CHDK. There are lots of sharpening methods besides an unsharp mask; deconvolve, inverse diffusion, octave, and shock-filters to name a few. Really though, comparing Bayer and Fovean sensors, it makes more sense to compare a 33% or so Bayer with 100% Fovean. After all, what you might expect to see at 100% has a lot to do with "pixel" density. Maybe you still do feel this is too soft, and that's fine, just thought I would put this out there.

Not sure why my second image isn't showing in the post, you can see it by clicking the second view original size link.

Nevermind it's showing now, guess it just took a little bit.

ttbek said:

Curious if you feel these are soft at 100% The second has a fairly light unsharp mask applied to it. They are jpg output from my Powershot SX10IS, it doesn't do raw unless you bust out CHDK. There are lots of sharpening methods besides an unsharp mask; deconvolve, inverse diffusion, octave, and shock-filters to name a few. Really though, comparing Bayer and Fovean sensors, it makes more sense to compare a 33% or so Bayer with 100% Fovean. After all, what you might expect to see at 100% has a lot to do with "pixel" density. Maybe you still do feel this is too soft, and that's fine, just thought I would put this out there.

Click to expand...

Were you talking to me? People are posting in so many areas of this thread, I've gotten lost.

Most people seem to think that the best way to evaluate the sharpness of a camera is to look at macro images or portrait images, but I disagree. I think the best way is to look at the details in the distance of a landscape photo.

Yeah, I replied to the thread, which you started, so speaking mostly to you, though I don't mind if anyone else would like to chime in. This isn't macro or portrait. Landscape is an extremely poor way to evaluate sharpness in actuality because there are a lot of environmental factors that can make shots appear blurry and hazy. So, umm, is what you're saying that you can't evaluate the sharpness of this shot? Personally I feel that most shots that don't turn out sharp from a recent camera (and this shot isn't even from a recent model) are due to either user error or poor conditions, particularly poor light. Now that it's Winter I don't think I can get something quite this sharp with this camera, the aperture just isn't quick enough (without tripod or flash that is). Honestly, you might really be up for troll of the year, at every turn you either avoid the issue (sharp or not?, you also brushed off Lee Jay's examles: http://www.dpreview.com/forums/post/52642381), change your tale (your first comparison was too ridiculous so you have revised it but are still cherry picking shots that seem unusually bad), or switch to another metric that makes no sense (seriously... landscape to judge camera sharpness). It's possible that you just have a poor understanding of photography, but you seem to know how to say things in just that kind of way that makes people want to correct you. Furthermore, I mentioned various methods of sharpening, which you seem to have indicated was the goal of this thread when you replied to Lee Jay. The alternative being that you wanted some magic sauce camera settings. The magic sauce is the same for any camera, it's about getting the right combination of iso, shutter speed, and aperture (yes there are more complicated things, but for most stuff). Anyway, I don't actually care what camera you get, just stop spouting misinformation.