The Canon G10: 'I Was Wrong'

ISO 400. Very light processing, and no noise reduction, applied to the original raw file.

By Ken Tanaka

Christopher Hsee et al. may have demonstrated that specifications can be powerful camera selection influences. But it's also possible, albeit less likely, for higher specs to negatively prejudice selection. To some degree that’s what prompted my thumbs-down reaction to Edward Taylor’s excellent October 15 first impressions of the new Canon Powershot G10. I had become somewhat prejudiced by the G10’s higher pixel density on the same 1/1.7" sensor and was not immediately impressed by the G10's list of enhancements over the G9 [which Ken reviewed for us —Ed].

Since that time, however, I've spent much more time actually shooting with the G10 and working with its raw files. Consequently my opinion of the camera has diametrically changed and, in fact, I am now an enthusiastic G10 owner.

I'll not take the space to repeat Edward’s earlier report. I would merely ditto nearly all of his observations. But it's worth noting the key factors that changed my own vote:

Image Quality: Despite Canon cramming more pixels on that tiny CCD (which Canon claims is all-new) the G10's image quality is actually markedly better than the G9's. Raw images are noticeably sharper right out of the camera. Highlights seem less prone to blowing out. Colors are more lively. Noise is well sublimated, more so than the G9. I can only conclude that the new DIGIC 4 processor is largely responsible for these advancements.

Operation: The G10 is certainly no slower than the G9. Its shot-to-shot cycle time (my personal bugaboo) may even be just a bit zippier. Moving the exposure compensation control from a screen-based control to a knurled mechanical dial was actually an excellent design change. I've already begun to wonder how I ever lived without it. 

Size, Build and Feel: Yes, the G10 is slightly bulkier than the G9. But, as Edward noted, the difference is not great enough to force anyone into a different carrying method. The G10's build, like the G9, is very robust and more similar to a Leica M than to a point-and-shoot. Canon's addition of a better right-hand grip (which will put the third-party G grips out of business) makes for a much more secure handling feel. I must also agree with Michael Reichmann’s remark that the G10 feels more like a finished and refined product than its predecessor.

So while Canon's Powershot G10 may not be the ultimate imaginable single-lens camera it's certainly doing its part to close in on that goal. And, contrary to my October remarks, in the final analysis I did find the G10's incremental improvements over the G9 to be worth the investment. The G10 is now my main carry-around camera.

____________________

Ken

ADDENDUM: For those interested in seeing a few more G10 samples, I used the G10 for all of these images except the aerials (Canon 1Ds Mark III) and the Hutchinson Field set (Leica M8). I realize that Web images reveal little about a camera. But I can tell you that the image files rival those from much better cameras.

Comments

[phule]

[[Despite Canon cramming more pixels on that tiny CCD (which Canon claims is all-new) the image quality is actually markedly better than the G9]]

Unless something has changed, Canon purchased this sensor from Sony, just like they purchased all their other CCD sensors for all their other point and shoots.

The same CCD is used in Sony's W300.

[Mike J.]

"Unless something has changed, Canon purchased this sensor from Sony, just like they purchased all their other CCD sensors for all their other point and shoots. The same CCD is used in Sony's W300."

How do you know this? Just something you read online? Unless you got this information from Canon--and not from a rep or from tech support, but from someone in engineering or management--then I'd consider the information suspect. As far as I know, Camera manufacturers don't tell the public where they get every sensor for every camera. Much of what passes for information about the subject online is merely speculation. I wouldn't even be willing to claim that all the sensors for one model of camera are all from the same fabricator, much less to blithely predict where any given sensor came from and what other cameras it's used in.

I'm not saying you're wrong, but I need attribution for statements like this.

Mike J.

[Ken]

Hmmm. Well I certainly don't know if this is true or not. But I do also have a Sony DSC-W300 (which someone wrote-up here earlier this year). It's a very capable shirt-pocket-class p&s. But it features a resolution of 13.6mp versus the G10's 14.7mp, suggesting that they are different CCD sensors.

[stevierose]

Hey Ken-

What program are you using for RAW conversion on the G10? Last I checked ACR hadn't yet included it.

Steve

[David]

Tie me to the mast. Let me slip past these sirens without succumbing.

[Edward Taylor]

Ken -

I am glad you are enjoying the G10 and I really enjoyed looking at your photos.

In my G10 review, one point that got lost in my editing is this - The G10 can produce images that rival a DSLR, but will produce them less frequently than a DSLR will. Again, the right conditions are more important with a small sensor camera. A DSLR is more forgiving and more consistent. The G10 is a great carry around camera though, and you have gotten some great results with it.

Thanks for taking the time to share your thoughts.

Ed

[Ken Tanaka]

@ stevierose: Good question. I have been using Canon's own Digital Photo Pro to process the G10's raw image files until Adobe's Camera Raw is updated (reportedly within a week or two).

@ Edward: Indeed, the tiny sensor camera can sometimes have narrow tolerances. The G10's new exposure compensation dial has encouraged me to explore it further and make better use of the facility. Thus far I'm finding that setting it to -1/3 for bright-ish conditions and +1/3 for dim light seems to produce a more tolerant raw image file. Give it a try.

[WeeDram]

Ken: One problem I have with small digicams (and it is often a BIG problem) is camera shake @ slow speeds. With a good film RF I can get good negs at 1/8 or even 1/4s. Given that a digicam's big bugaboo is arguably ugly digi-noise at ISOs over 200, slow speed "handholdability" seems pretty important, irrespective of of the processor in this Canon.

[Bron Janulis]

Ken,
Very nice "Roger Brown" painting.

Bron, who is keeping his G9, in spite of it's inferiority, due to finance.

Bron

[Nightening]

I guess 14,700,000 cupcakes aren't too many after all.

[John Roberts]

"Yes, the G10 is slightly bulkier than the G9. But, as Edward noted, the difference is not great enough to force anyone into a different carrying method."

I'm glad Ken re-emphasized this point. I keep seeing the G10 being knocked for being too bulky to be pocketable. Was this ever intended to be a "pocketable" camera? Canon makes several other cameras for those in the market for one that rides unnoticed in one's pocket. I think Canon had photographers with other priorities in mind when they designed the G10.

[jay moynihan]

2 questions:

Any comments or observations on its JPEG quality, from base ISO up through ISO 400?

I understand that the viewfinder is 80%. is it centered on the image or is it shifted to a side?

Thanks

[Shaun]

Regarding the optical viewfinder, while not having done extensive testing, I have found the viewfinder on the G10 centers itself on what the sensor captures, a step in the right direction vs the G7 and G9 that I had. I find the viewfinder usable in the G10, where I wouldn't have bothered on my G9 or G7.

[Ken]

@ Jay: Good question. I don't use JPG very often, and had never used it yet with the G10. So I gave it a quick whirl. Like the G9 (and other Canons), the G10 offers so many in-camera JPG processing adjustments that you can almost get anything you want. But for purposes of briefly exploring the relationship between ISO and noise in JPGs I prepared a series of six "standard" setting images cycling through the camera's ISO settings, from 80 to 1600. I maintained the scene, the aperture, and the focal length throughout. I think inspection of the smoothest area of the spoon, as well as the spoon's shadow, will give you some idea of the camera's ISO/noise relationship.

http://www.pbase.com/tanakak/g10isosamples

Caution: These are full-size Large-Fine JPG files each approximately 3 Mb in size, although PBase will let you view smaller sizes.

@ Bron: Your comment reminded me that I really should have given credit to the artist whose work I photographed above. The sculpture is titled "King Lear" by J. Seward Johnson. It's currently installed in Chicago's Pioneer Court, located on Michigan Avenue across the street from the Wrigley Building and in front of the Equitable Building (in background).

@ Nightening: Yes, I guess the G10's eyes really are not too big for its tummy.

[Juan P. Valdés]

I am been strange with your affirmations about the G10, have one and all the tests in Raw with DPP show a previous approach that fills with paste the image. Regards.
Valdés

[Huw Morgan]

I was curious about the quality of the G10 sensor and thought I'd look it up in the DxO database (see yesterday's MJ blog entry) and, from a quantitative point of view, the raw image quality of the G10 is marginally better than the G9 (DxO score of 37.8 vs 34.7). The raw image score of the G10 is significantly inferior to Canon DSLR's. For example, a 50D scores 62.9 and a 1000D scores 61.6. The key difference is in the low light ISO performance where the larger pixels of the DSLR's really kick butt.

Real-world comparisons seem to indicate that colour depth and dynamic range are comparable to DSLR's and this is borne out by the DxO numbers.

[Dave Fultz]

"Any comments or observations on its JPEG quality, from base ISO up through ISO 400?"

As mentioned above it takes the right conditions but when you have them, JPEGs at 80, 100, and 200 are excellent; 400 requires a little noise reduction work.

"I understand that the viewfinder is 80%. is it centered on the image or is it shifted to a side?"

78% and centered. Pretty much useless for anything more than seeing if you're pointing the right direction.

Dave

[Ken]

@ Jay: Sorry, I just noticed that I'd not answered your second question, regarding the viewfinder. (I assume you're referring to the optical viewfinder?) The viewfinder image is, indeed, centered with the lens coverage. Note, also, that at the widest focal length you will see the lens barrel at the bottom of the viewfinder.

[Dogman]

I'm a photographic muddler. I've never been very technically-minded but I've cultivated a willingness to try anything with a Boy Scout's preparedness. Thus I've been able to successfully muddle through my picture-taking for over three decades. After packing as many as a half dozen cameras and etc. for various trips, I concluded recently that this has to be re-thought. Edward's review and now Ken's, along with several others on various websites, have convinced me. I ordered a G10 on Saturday and it's supposed to be delivered tomorrow. Maybe I can successfully leave the Leicas behind as I learned to leave the MF Pentaxes after I bought a DSLR. I'm sure the G10 will prove to be more than satisfactory for my muddlings.

[jekkherg]

I bought a G10 even before Ed posted his review and was actually surprised after seeing the negative comments from those specification-influenced readers. The obvious doesn't really add up to the projected sum but the big variable X here is the unproven effect of the DIGIC 4 in the equation. Im glad you changed your point of view Ken.

[Ken]

Addendum:
In case anyone is still watching this subject, Digital Photography Review has just published its full review of the Canon PowerShot G10 (http://www.dpreview.com/reviews/canong10/) . DPreview has been the gold standard in digital camera reviews for many years. They have a well-earned reputation for being thorough and consistent in their evaluations.

But since their sale (last year?) and since the necessary expansion of their review staff their conclusions have sometimes given me a smile. In the case of the G10 review I do not strongly disagree with any the reviewer's (Don Wan) points. In fact, just the contrary. Some sound very much like Edward's original remarks and closely reflect my reactions, too.

In the final "Conclusion" section, which is the section that everyone familiar with DPreview reviews reads first, Don has rated the camera "Recommended", one notch down from the highest "Highly Recommended" rating which, for example, the G9 received last year. This, despite having given the G10 ratings ranging from 8.0 to 9.5 in each of the six specific review categories. He apparently feels that the Panasonic LX3's slightly lower price is a better value...despite its lower categorical ratings.

OK Don. If you say so,

[peter]

dpreview has not been so consistent with their ratings. they sometimes contradict their own marks. at the end of the conclusion, G10 averaged 8.65 at the ratings and was given RECOMMENDED level. LX3 and Canon G9 averaged 8.5 and 8.43 respectively, conversely giving them HIGHLY RECOMMENDED level.

[Rene]

Hi! Ken great pic on the top of this page!

Got a G10 recently but still trying to re-learn manual settings and basically taking pics by trial-and-error presently.

Would it be possible for me to ask for your e-mail address please?

I am really intrigued by the picture at the top of this page. Were you using a tripod? Could you please send me your settings to my e-mail address so I may also take of note of your e-mail address?

Thank you VERY MUCH!

Rene from the Cebu City, Philippines 6000