Canon S5 IS needs badly better IQ (*)

Thanks. I hope I'm still allowed to post S3 pics since I 'upgraded' to the S5 - LOL! ... Matt

Jun2 said:

Very nice picture.

Jun

Click to expand...

Matt,

Well said. That's what I was getting at in my little exchange above with the OP -- there's just nothing to be gained from going on and on and on and on about tiny little details found in 100% crops of test review images. There's an utter lack of perspective that you gain if you just go out and use one of these cameras.
--
Tom Hoots
My PBase galleries:
http://www.pbase.com/thoots

Like I said in another posting this forum has become a 'Pixel Peeping Extravaganza' lately - I can however relate to the OP because I've posted comparisons myself and have since gotten away from it - because it brings out the real 'freaks' (not just me - LOL!) ... Matt

Tom Hoots said:

Matt,

Well said. That's what I was getting at in my little exchange above
with the OP -- there's just nothing to be gained from going on and on
and on and on about tiny little details found in 100% crops of test
review images. There's an utter lack of perspective that you gain if
you just go out and use one of these cameras.

Click to expand...

Agreed - 'hands on' is the best approach.

Tom Hoots said:

--
Tom Hoots
My PBase galleries:
http://www.pbase.com/thoots

Click to expand...

I think beyond pixel peeping..for the vast majority of people the S3, S5, Panny FZ 8, Panny TZ 3, and others are capable of producing 8 x 10 printed images at ISO 100 that any of use would be hard pressed to identify one source from the other. The colors might be different..then agai printing makes such a difference in that as well. Especially true since we look at photos to enjoy them and do not use two cameras under identical conditions and then compare results side by side.

Having said that, the advantage of 28 mm vs. 36 mm is a real world one that has obvious implications with scenics, family pictures, weddings etc. Now factor in the price that Canon is charging for the S5 and compare it to the MSRP (not even the lower selling price) and the Panny FZ 18 and the Fuji S8000sd (IF their IQ is equal to..slightly better than..slightly worse than the S5) and they become compelling and perhaps significantly better alternatives for many potential owners. Even the Panny TZ 3 with its small size, lower price (although with far fewer customizable controls) is a very very attractive camera for many. The FZ 18 even offers RAW...something Canon should with the G7 and S5. We live in intersting times and have wonderful choices. I do think that Canon needs to move along faster and have more competative pricing. This in no way says that Canon doesn't make wonderful cameras. The A series is a very very good buy for wonderfully capable cameras. I don't think that the S5 price is justified at all and I think that Canon's sales will be hurt by the very strong competition. Time will tell. Tests will tell. IF the FZ18/S8000sd test poorly..We'll see. I do feel that Canon has fallen asleep at the wheel when it comes to SuperZooms. Small changes. Little true progress... The competition is aggressive and bold.

Through the introductions of several models, I've refrained from purchasing an S-series cam because I can't tolerate the image quality (tonality etc) of the images they output. Referring to the photos within this thread as examples, I think we all "know" that stripes on a tiger should be black (not some shade of gray). Also, I think we all "know" that the glossy surface of a squirrel's eyes shouldn't look like they are coated with speckles of (noise) sawdust. I don't care what ISO was used; it's a daytime shot and that amount of image noise is just unacceptable (intolerable) to me. Without post-processing (who has time to mess with post-processing EVERY image?) most of the S2/S3/S5 images have looked too "soft", to my eye -- as though Canon tuned the processor to output flattering, softened "portrait" images. For well-lit scenes, the most immediately noticeable and most undesirable quality of images from the S-series cams, across the board, is their hazy(?) washed-out(?) appearance.

For an S-series cam, the tiger photo posted above reflects surprisingly good output. Most of the tiger, and zebra etc photos from Sx cams displayed on pBase contain blown highlights.

Do digital cameras still use a separate sensor to meter light like in film cameras?

I don't understand the link. To me, it is just saying that there are calibration errors in cameras. This has always been true and variations can be observed within different samples in a single line of cameras.

All it says is that the lighting conditions is that it simulates daylight, not that all shots are taken with an EV of X on a separate calibrated light meter. I'm not sure you can assume that the lighting is exactly the same in both tests. (I believe the posted shots are not side by side tests, but were taken at different times.)

I'm new to the digital world and am merely trying to understand it. Please do read into my posts that I criticizing the original poster or the dpreview test procedures.

Photofreak7 said:

• I also think

Canon knows the targeted market will do well below ISO200 - that's
why the S5's 'AUTO ISO' setting maxes out at ISO200 (High ISO setting

Click to expand...

I have to keep wondering where and when you guys use your cameras? Such excuses as 'Canon knows that their customers will do fine below ISO200' are incomprehensible and in fact my personal observations tell really a different story. I could as well throw in that Canon should know that their customers seldom print their photos in large sizes...

For example, anyone of you ever seen those thousands of flashing bright points in the middle of audience in sports and cultural spectacles? Guess what, they are people photographing the back head of their fellow spectators. Since nowadays for security reasons in many events large cameras are not allowed for audience, a compact lowlight capable S-series style super-zoom camera would be ideal. Furthermore, I don’t even want to bring large, heavy and expensive DSLR equipment to every crowded event.

There are also normal lighting situations where small aperture and high shutter speed are required to get the shot. Increasing the sensitivity is the only way to cope with those situations.

November night in Helsinki at Finnair Stadium. It is dark but the stadium is well lighted, luckily with daylight HID-lamps (daylight white balance gives the best results noise wise with high ISO).



S2 IS making the best it can. I used constantly ISO200 (new 400) and ISO400 (new 800) settings in order to get shutter speeds short enough. Here is one of my ISO400 (new 800) shots. BTW, the lawn is artificial and always looks suspective in photos, with high NR even more so.

S2IS, 72mm (full zoom), ISO400 (new 800), 1/100, f4.5, -1/3EV, low sharpening. PP includes chroma and luminance NR, downsizing and sharpening.



S2IS, 72mm (full zoom), ISO400 (new 800), 1/100, f4.5, -1/3EV, low sharpening. 100% crop of the original, PP includes now only chroma NR.



That shot would have been considerably lower quality with S5 as a result of both: its noisier sensor and in consequence aggressive noise reduction.

Photofreak7 said:

While I agree that the NR is to strong with Digic III
(wonder what Digic II on the S5 would have been like?)

Click to expand...

The 8MP 1/2.5 sensor used in S5 is so noisy that it needs Digic III's aggressive noise reduction. In case you didn't notice the S5 image is noisier AND has more noise reduction applied AND holds generally less detail. With the Digic II S5 would be excessively noisy even at base ISO.

-Virvatulet

I'm afraid that perhaps we're at the end of significant improvements in IQ based on current sensor technology. I doubt if lens resolution can dramatically improve so the only things that might change are sensor technology and processor technology. But there's only so much you can do there, and I think that's probably mostly going to be about bringing sensor noise down at higher ISOs.

So yeah, I don't know. I personally think the IQ's been good enough from these little cameras to satisfy most of their users for the last few generations, so I don't really think IQ's a problem. Is it ever going to be as good as a DSLR, I doubt it, but it's good enough for what it is.

I personally think we need some point and shoots with a more conservative zoom range and big beautiful DSLR sensors in them. My little old Oly Stylus Epic had a fixed focal length 35/2.8 lens and a big 35mm piece of film in it. In the right conditions it took indistinguishable photos from my SLRs; however, I can't really honestly say that about any of the point and shoots around today. Maybe someday.

Curious5 said:

Do digital cameras still use a separate sensor to meter light like in
film cameras?

Click to expand...

Only in DSLRs. P&S cameras use main sensor for AF, AE and image capture.

Curious5 said:

To me, it is just saying that there are
calibration errors in cameras.

Click to expand...

Not actually errors, more of intentional design parameters.

Curious5 said:

This has always been true and
variations can be observed within different samples in a single line
of cameras.

Click to expand...

Due to their inherent characteristics and structure digital cameras have very little if any real variation in AE.

Curious5 said:

calibrated light meter. I'm not sure you can assume that the
lighting is exactly the same in both tests. (I believe the posted
shots are not side by side tests, but were taken at different times.)

Click to expand...

The setup is the same and artificially lighted with high-end equipment. The idea is that these images are comparable. One can reasonably expect same conditions.

-Virvatulet

The mp race has become ridiculous.....people aren't really buying an 8mp camera anymore, yes it has 8mp but the resolution isn't there and the image quality isn't there. I do tend to think that older lower mp cams have better quality than the destructive high mp images now. And those who say it doesn't show up in print are in denial because even my 20d noise at moderate iso shows up in print.
--
http://www.pbase.com/shhe

Steven Noyes said:

At print sizes most people do (8X10 and below), there will be no
visible difference (but then why the extra MP?).

Click to expand...

Exactly. But there will be visible difference when noisy pixel crammed sensor is used at higher sensitivities. For what the extra megapixels are good for if they make the image worse?

S2IS, 72mm (full zoom), ISO400, 1/100, f4.5, -1/3EV, low sharpening. 100% crop of the original, PP includes only chroma NR.



That shot would have been considerably lower quality with S5 as a result of its noisier sensor and in consequence aggressive noise reduction.

Steven Noyes said:

But for a good light, ultra-light solution, these hi-MP super zoom
cameras do very well.

Click to expand...

Actually, there is a contradiction in that too. These noisy pixel crammed sensors need a lot of light but due to lack of good dynamic range handling capability they can not cope well with it. Paradoxically, both problems i.e. lack of sensitivity and lack of DR are the result of megapixel race.

-Virvatulet

Lets face it 9 times outta 10 these current P&S cameras look horrible when you do a 100% crop on a pic. In the original post I fail to see why you would use a high ISO on a still life type image. As well as a high ISO outdoors on still images.

Other than trying to capture moving shots I always keep my ISO as low as it'll go.

--



Jeff
http://www.members.shaw.ca/onepunch/cars.htm

hitrack said:

Lets face it 9 times outta 10 these current P&S cameras look horrible
when you do a 100% crop on a pic.

Click to expand...

It really doesn't have to be that way.

hitrack said:

In the original post I fail to see
why you would use a high ISO on a still life type image. As well as a
high ISO outdoors on still images.

Click to expand...

We are comparing two cameras’ performance. This is the only way to do it. Higher sensitivities have many applications, more explanation in this message:

http://forums.dpreview.com/forums/read.asp?forum=1010&message=24177838

-Virvatulet

You explained it well and held your up your end of the arguments.

While I thought this would be a "dogfight", it turned out to be quite informative.
--
Bill

Tom Hoots said:

when "in the real world," once you resize that picture to show on a
forum like this or to send via e-mail, or if you print the picture,
the details you're making such a ruckus about virtually won't even be
visible in the output.

Click to expand...

If that is so then why you seem to defend the megapixel race? There is a raving contradiction between what you write here and what you do. I suggest that you would take the comparison test image from the OP and try to resize it, you will be surprised how much you'll have to downsize it until the quality differences start to diminish. Try it "in the real world", as you sought for.

Tom Hoots said:

noise reduction post-processing. Does it truly distract you from
otherwise appreciating the image?

Click to expand...

No, as I have already clearly expressed.

The rest is already written here:

http://forums.dpreview.com/forums/read.asp?forum=1010&message=24177838

-Virvatulet

This subject should be brought up as often as possible. The goal of these cameras should be to provide a tool that is easy to use and produces great images. Amazingly, technology is being used to actually harm the latter! And as we all should know by now, marketing is entirely to blame.

I've always known that people in these forums (and most forums, for that matter) have a tendency to verbally attack and vehemently defend new iterations of their favorite technology, no matter what. Yet I'm still astonished at the lack of support shown to the blatant truth demonstrated in this thread. If it's not questioning the test (which, by the way, is only one of dozens of similar examples that can be made across many camera brands) it's questioning the importance of the test! Unbelievable.

How can people defend the image quality of these new cameras when there are so many examples available of how it is slowly degrading? This has been explained time and again, by the people who run this site if nothing else!!!

Right before our eyes, point and shoot cameras are being shoved into a world where they can only take usable photos at ISO100 or below. This is not just the Megapixel Myth anymore! No, instead of just producing a few new cameras with unnecessary resolution, the Megapixel Myth is actually starting to kill the image quality! Not only that, it's making the cameras perform slower, and making our image editors perform slower as well.

Fellow photographers, please don't let this "resolution" ignorance propagate any more. High megapixels should be fooling casual buyers in Wal-Mart, not people with access to educated community discussion and reviews!

Virvatulet said:

If that is so then why you seem to defend the megapixel race? There is a raving contradiction between what you write here and what you do.

Click to expand...

Well, it would help whole bunches if you could comprehend what I'm actually writing about. Oh, please, I'm not defending the megapixel race -- I'm amazed at how you could reach such a conclusion. I'm simply making the point that the megapixel race is an actual fact of life, and so long as millions of people keep on buying mega-megapixel cameras, a veritable handful of folks whining about diminishing image quality here on dpreview.com isn't going to change anything at all.

I'm just "virtually rolling my eyes" at the thousands of posts in which folks like you are just pounding this dead issue into horseburgers.

Pixel density will go up. Image quality will go down. Millions of cameras will be sold, anyway. If you truly want better image quality, the industry has DSLR bodies and lenses it would love to sell to you.

You can leave it there and get on with the rest of your life, or you can keep on spending your time arguing over teensy little differences between 100% crops from cameras that haven't been designed with "ultimate picture quality" as a very high priority.

THAT is what I'm saying.
--
Tom Hoots
My PBase galleries:
http://www.pbase.com/thoots

Tom Hoots said:

simply making the point that the megapixel race is an actual fact of
life, and so long as millions of people keep on buying mega-megapixel

Click to expand...

And you are one of them, defending here your purchase.

Only one question: Did you try it "in the real world", that is took the comparison image of the original post and try to downsize it until the quality differences start to diminish?

When you do, you will understand that the quality drop is much more than "teensy little differences between 100% crops".

-Virvatulet

I didn't remember to mention that you can start testing with the DPR’s "Image control" panel which is above every image (needs JavaScript enabled).
Just click "Zoom Out".

-Virvatulet

J1000 said:

This subject should be brought up as often as possible. The goal of
these cameras should be to provide a tool that is easy to use and
produces great images.

Click to expand...

Obviously not everybody like to think the same way. I really wonder why.

Cheers,

-Virvatulet