Ad
Ad
Ad
Pages: [1] 2 »   Bottom of Page
Print
Author Topic: G10: Michael vs DPREVIEW  (Read 8602 times)
ysengrain
Guest
« on: November 26, 2008, 02:47:17 AM »
ReplyReply

About the new Canon G10: I've just read www.dpreview.com G10 review.
I "feel" very different Michael and dpreview "grasps" particularly concerning image quality.
I saw the pictures Michael showed us comparing Hasselblad and G10 quality, but dpreview seems to be rather far from this.

What is your opinion ?
Logged
Ben Rubinstein
Sr. Member
****
Offline Offline

Posts: 1733


« Reply #1 on: November 26, 2008, 03:44:39 AM »
ReplyReply

Quote from: ysengrain
About the new Canon G10: I've just read www.dpreview.com G10 review.
I "feel" very different Michael and dpreview "grasps" particularly concerning image quality.
I saw the pictures Michael showed us comparing Hasselblad and G10 quality, but dpreview seems to be rather far from this.

What is your opinion ?

The report on this site was that the quality at iso 80-100 was incredible for a camera of that size shown by prints at its native resolution. I don't see DPreview disagreeing. What they do say is that they would prefer less resolution and a camera that isn't limited to it's base iso. I agree entirely. Look at the recent monopod article here. If you can carry a monopod you can carry a small DSLR and no one buying a camera of that size wants it to be a monopod only camera!

Personally I'm waiting to see what m4/3 will come up with in the future.
Logged

httivals
Jr. Member
**
Offline Offline

Posts: 82


« Reply #2 on: November 26, 2008, 06:03:41 AM »
ReplyReply

My experience is that it's an incredible camera; I also found that the LX3 was not superior to the G10 on noise.  I haven't looked at the DP Review article, but my recollection from the past is that a lot of DP Review's analysis is based on comparing JPEGs.  I only shoot raw.  I find that the G10 at iso 400, shot raw, is very good using DPP's limited RAW controls, and excellent using Lightroom's much better sharpening and noise suppression options.

So my experiences are far from those of DP Review.  Thom Hogan's comparison and conclusion (bythom.com) is similar to Reichman's.  

Quote from: ysengrain
About the new Canon G10: I've just read www.dpreview.com G10 review.
I "feel" very different Michael and dpreview "grasps" particularly concerning image quality.
I saw the pictures Michael showed us comparing Hasselblad and G10 quality, but dpreview seems to be rather far from this.

What is your opinion ?
Logged
neil74
Newbie
*
Offline Offline

Posts: 25


« Reply #3 on: November 26, 2008, 07:57:28 AM »
ReplyReply

The one thing that these reviews always seem to miss out is how the camera copes with longer exposures or poor light.  ISO 80 may be good on the G10 but is it still good with 2-3 second exposure in poor light and how would it cope with a 5, 10 second or longer?

I have a G9 which I use as our family holiday camera and I take it everywhere with me but I have not shot a landscape on it yet. I also have a Sigma DP1 which I use for landscapes alongside my 5D, it is very very clean at iso 100 but try a 5-10 second exposure at sunset and it starts to look quite noisy.  The G9 is ok at iso 80 but again start exposing in seconds rather than fractions of seconds and it looks pretty terrible I wonder if the G10 is the same?

Right now I have no inclination to dump the G9 and it also has slightly better macro too.  I hope the next iteration of the G series will be a genuine SLR alternative.
Logged
jmwscot
Jr. Member
**
Offline Offline

Posts: 66



WWW
« Reply #4 on: November 26, 2008, 08:20:03 AM »
ReplyReply

I have a G10 and a 1Ds MKIII and shoot in RAW. The G10 is a great wee camera. It can take shots impossible with the big camera. Macro with subjects 1 inch from the lens, hand held at 1/10 sec because of no mirror bounce, superb bright LCD screen with resolution approaching the iPhone, really sharp lens at both wide and tele, and it has a great fully adjustable flash for casual shots (try taking casual shots with a 1Ds MKIII).  

The G9 which I replaced was pretty good but I always wished I had taken the shot, especially tele shots, with the big camera. Not so with the G10. There is a touch of noise at 80 and 100ASA on blue skies which you can see at 100% on a monitor and 400ASA shows definite noise on screen especially after sharpening. However if you are into making prints up to 11 x 14 the quality is superb and noise is not a problem. 800ASA is fine for stuff on the web up to 800px wide and 1600ASA is unquestionably coarse. There's stuff it's no use at such as sport, action or anything requiring a modest or fast frame rate (it's slow to write to file and it can seem like an awful long wait until you can fire the shutter again)

So stop peering at reviews and pontificating about whose review is most accurate. If you want to make great prints and upload superb images to the web go and buy one - it will help the economy. But if you are only going to do endless pixel peeping on your monitor at 100% you really should be looking at medium format !!
Logged
jani
Sr. Member
****
Offline Offline

Posts: 1604



WWW
« Reply #5 on: November 26, 2008, 08:35:55 AM »
ReplyReply

Quote from: neil74
The one thing that these reviews always seem to miss out is how the camera copes with longer exposures or poor light.  ISO 80 may be good on the G10 but is it still good with 2-3 second exposure in poor light and how would it cope with a 5, 10 second or longer?
I don't think that's what these compact cameras are made for, and I think that most of them positively suck at this, compared to most dSLRs.

Quote
I also have a Sigma DP1 which I use for landscapes alongside my 5D, it is very very clean at iso 100 but try a 5-10 second exposure at sunset and it starts to look quite noisy.  The G9 is ok at iso 80 but again start exposing in seconds rather than fractions of seconds and it looks pretty terrible I wonder if the G10 is the same?
It probably is.

However, you could try taking interval shots (continuous shooting) with shorter exposures and merging them later on, like astrophotographers have been doing for years.

I don't know how well suited the DP1 or the G9/G10 are for this kind of use, though.
Logged

Jan
BJL
Sr. Member
****
Offline Offline

Posts: 5121


« Reply #6 on: November 26, 2008, 11:12:11 AM »
ReplyReply

The difference could be just a matter of reviewer priorities; that is why I seek reviewers whose priorities are close to mine, like Michael's attention to ergonomics and hands-on evaluation.

DPReview seems obsessed with lab measures of per pixel performance like S/N ratio while underestimating the advantages of higher pixel counts, as shown by its near crusade on " pixel density". Its reviews persist in ignoring the fact that files from a sensor of higher pixel count and somewhat lower per pixel S/N ratio can be processed (by downsampling or such) to a lower resolution/lower noise/higher DR option, potentially allowing it to match the resolution and noise levels of a sensor with fewer pixels.  The 50D vs 40D comparison is a good example of this: full res. files from the 50D have better resolution but worse noise that the 40D at equal ISO, but with suitable downsampling, high ISO 50D images can end up slightly better the 40D images for both resolution and noise levels.

More generally, DPReview seems to put a very high weight on IQ in low light (high ISO) situations, while being rather dismissive of the desire for getting highly detailed images at lower shutter speeds, or having plenty of cropping latitude to stretch telephoto reach and such. As a consequence, the two most recent Canon reviews at DPReview are some of the most unfavorable Canon product reviews there for a long time: the 50D and G10.

Ironically, along with frequently claiming that a camera under review would be better with a sensor of lower pixel count, DPReview then often criticizes cameras with "too high" pixel counts for then using NR processing selectively at higher ISO speeds to give the option of lower noise in exchange for lower resolution. What is strange about this to me is that
1. the "bigger, fewer pixels" option limits one to lower resolution all the time, even in good light situations where noise is not a problem,
while
2. the "more, smaller pixels" option with extra NR only as needed at higher ISO speeds can offer a lower resolution and lower noise combination when most needed (in low light situations) while offering higher resolution at other times (good light situations).

Logged
Ben Rubinstein
Sr. Member
****
Offline Offline

Posts: 1733


« Reply #7 on: November 26, 2008, 12:38:00 PM »
ReplyReply

Quote from: BJL
The difference could be just a matter of reviewer priorities; that is why I seek reviewers whose priorities are close to mine, like Michael's attention to ergonomics and hands-on evaluation.

DPReview seems obsessed with lab measures of per pixel performance like S/N ratio while underestimating the advantages of higher pixel counts, as shown by its near crusade on " pixel density". Its reviews persist in ignoring the fact that files from a sensor of higher pixel count and somewhat lower per pixel S/N ratio can be processed (by downsampling or such) to a lower resolution/lower noise/higher DR option, potentially allowing it to match the resolution and noise levels of a sensor with fewer pixels.  The 50D vs 40D comparison is a good example of this: full res. files from the 50D have better resolution but worse noise that the 40D at equal ISO, but with suitable downsampling, high ISO 50D images can end up slightly better the 40D images for both resolution and noise levels.

More generally, DPReview seems to put a very high weight on IQ in low light (high ISO) situations, while being rather dismissive of the desire for getting highly detailed images at lower shutter speeds, or having plenty of cropping latitude to stretch telephoto reach and such. As a consequence, the two most recent Canon reviews at DPReview are some of the most unfavorable Canon product reviews there for a long time: the 50D and G10.

Ironically, along with frequently claiming that a camera under review would be better with a sensor of lower pixel count, DPReview then often criticizes cameras with "too high" pixel counts for then using NR processing selectively at higher ISO speeds to give the option of lower noise in exchange for lower resolution. What is strange about this to me is that
1. the "bigger, fewer pixels" option limits one to lower resolution all the time, even in good light situations where noise is not a problem,
while
2. the "more, smaller pixels" option with extra NR only as needed at higher ISO speeds can offer a lower resolution and lower noise combination when most needed (in low light situations) while offering higher resolution at other times (good light situations).

I agree with DPReview. The G10 is not a landscape camera, it's a take anywhere camera and as such fails badly if you can only expect good results at very low iso's (not that much else is any better). MR loves to say 'horses for courses' and the G10 is has missed the mark for the purpose of a p&s camera, what it should have been. I hear a lot about downsizing to match noise but have never seen the truth of it in the real world myself, certainly not managed it myself. Dpreviews point, that they could have made a 10 megapixel camera far more suited to the purpose of a p&s camera is personally more to the point than landscape photographers saying that it makes a nice niche product 'by mistake'. They could have made a great p&s and instead they made a low iso camera in a body format foreign to the need for low iso. You don't use a p&s for landscape or in studio - you use it as a p&s so who gives a damn if it makes a good carry around landscape camera at a push for smaller print sizes.
« Last Edit: November 26, 2008, 12:43:34 PM by pom » Logged

Dale_Cotton2
Jr. Member
**
Offline Offline

Posts: 94


WWW
« Reply #8 on: November 26, 2008, 12:47:50 PM »
ReplyReply

Ben/Pom writes:
Quote
I hear a lot about downsizing to match noise but have never seen the truth of it in the real world myself,
You've got company: DPReview Blog
Logged
barryfitzgerald
Sr. Member
****
Offline Offline

Posts: 566


« Reply #9 on: November 26, 2008, 01:04:11 PM »
ReplyReply

I have to agree with Dp. Down sampling has long been touted as a quick fix for noise. I have tried it myself, and I did not find it effective, or not very significant.

I don't see anything wrong with the review on dp, it still got a good one, and you can see the pros and cons of the camera, and make your own subjective view on it. Course, the G series has always had a nice quality feel in the hand, that can not be conveyed in a review, handling is important..very to some.

I would have to say Canon has been "defeated with megapixels", and I am pleased to see dp punish pixel stuffing just for the sake of it. I have said this for some time, but image quality is a LOT more than megapixels alone, many areas to it, and some areas such as dynamic range, deserve far more attention than a few lousy extra pixels. About time some folks stopped peering into res charts, and 200% crops, and more into the big picture of digital image quality. I hope other sites stop beating the tiresome megapixel drum. Fact is, we are well into double digit megapixels, you could moan a few years back that your 2mp camera didnt have enough resolution. I highly doubt anyone can complain nowadays. Let's move on to bigger fish...
Logged
Tony Beach
Sr. Member
****
Offline Offline

Posts: 452


WWW
« Reply #10 on: November 26, 2008, 03:53:27 PM »
ReplyReply

Quote from: ysengrain
I "feel" very different Michael and dpreview "grasps" particularly concerning image quality.
From MR's review:

"As good as they can be, (see my G10  vs Medium Format article) these aren't really a replacement for a DSLR let alone a medium format camera when it comes to image quality. They're good, but not that good.

The real issue is seen when noise can be easily detected even at base ISO."


From DPR's review:

"...at base ISO and in high contrast conditions, the G10 can produce detail and resolution that is astounding (better than some DSLRs). If your main use is in good lighting conditions outdoors, and you are prepared to take care with correct exposure, this camera is almost as good as it gets in compacts."

The G10 is not the final word on what we can expect from a small P&S camera.  It's still too big for what I want (I want it to fit in my pocket, and I don't mean my jacket's pocket), it's a little expensive, and the sensor is just too small.  Personally, I don't want a "do anything" P&S camera that is the proverbial jack of all trades (at best) when that's what my DSLR is for; what I want is a discrete, fixed (slightly wide angle) focal length, high quality P&S camera that does candids really well.  I don't think I'm alone in what I want.

DPR in part concludes in their review that:

"The G10 leaves that lingering question. Just how good could this camera have been..."
« Last Edit: November 26, 2008, 06:40:19 PM by Tony Beach » Logged
eronald
Sr. Member
****
Offline Offline

Posts: 3871



WWW
« Reply #11 on: November 26, 2008, 06:24:07 PM »
ReplyReply

I bought a G10 after reading Michael's review. I also own a Mamiya-Phase P45+. On balance I'd say it was an accurate review.

Edmund

Logged

Edmund Ronald, Ph.D. 
Pete Ferling
Guest
« Reply #12 on: December 01, 2008, 12:09:54 AM »
ReplyReply

Michael does mention comparable on prints up to super B, and that he doesn't plan to sell his MF camera anytime in the near future.  I bought a G5 when it was new for $800, and thought it to be worth every penny for a 5MP camera that captures RAW and challenged 35mm.  Why should the G10 be any different in regards to MF?  I still shoot 35mm and on my Mam 645 because I like the look those cameras give me.  There's a whole lot more than to compare a PS with system camera, (but I can't fit a M645 in my pocket either   My G5 is looking old and xmas is coming...
Logged
reburns
Full Member
***
Offline Offline

Posts: 142


WWW
« Reply #13 on: December 01, 2008, 03:10:08 PM »
ReplyReply

I sided with this:  http://www.imaging-resource.com/IMCOMP/COMPS01.HTM, plus took a few jpgs in the store.   In these test photos I saw that the G10 didn't resolve as well, and more importantly didn't handle highlights well.  I figured that between the higher sensitivity sensor and faster lens, the LX3 easily has a stop+ advantage.  For my use I decided that the wider angle would be better, that the small optical viewfinder would be inadequate anyway.  I weighed on the downside that the LX3 suffers from more lens distortion.  It really was a tossup between the two, and I decided to patronize the maker who wasn't pushing pixel count, even if downsampling produced similar results.  Now that Adobe DNG converter is available you could buy both, carefully test and return one in pristine condition.  Maybe it's like getting the daily world news, you need several differing sources and then come to your own conclusions.

Logged
eronald
Sr. Member
****
Offline Offline

Posts: 3871



WWW
« Reply #14 on: December 01, 2008, 04:07:48 PM »
ReplyReply

[quote name='reburns' date='Dec 1 2008, 10:10 PM' post='240919']

I have printed 44" wide several times from my G10. It can take it, but needs careful handling in shooting and post. ISO 100 is just about the highest you really want to push it, the noise pattern is ugly if you can see it.

Used the right way, in the hands of someone who knows what he's doing, like Michael,  the G10 can deliver astonishing results.  So can any other camera above $500.

Edmund

Logged

Edmund Ronald, Ph.D. 
Czornyj
Sr. Member
****
Offline Offline

Posts: 1371



WWW
« Reply #15 on: December 01, 2008, 04:40:59 PM »
ReplyReply

Quote from: eronald
Used the right way, in the hands of someone who knows what he's doing, like Michael,  the G10 can deliver astonishing results.  So can any other camera above $500.

I agree - apperture should be kept open, and ETTR also helps to keep noise under control. As one should expect, problems with such a smallish sensor are diffraction and noise in the shadows and midtones.
http://members.chello.pl/m.kaluza/g10tst.jpg

When you keep them under control, you can achieve very good results (D2X with 24-70 on the top vs G10 on the bottom):
http://members.chello.pl/m.kaluza/zagadka.jpg
« Last Edit: December 01, 2008, 04:44:06 PM by Czornyj » Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8853


« Reply #16 on: December 01, 2008, 06:02:37 PM »
ReplyReply

Quote from: pom
I hear a lot about downsizing to match noise but have never seen the truth of it in the real world myself, certainly not managed it myself.

Pom,
I don't understand why you've never seen the truth of this in the real world. I see it myself clearly all the time. It seems usual practice when examining noise in an image to do it at 100% on the computer monitor, when pixel peeping. My current 19" CRT monitor is set at a resolution of 1280x1024. A bit of basic maths indicates that's approximately 92 ppi. A G10 or 50D image at 100% on this monitor represents a print size of 34" x51". If I were to examine image noise at 100% on my 50" HD plasma, which has a resolution of approximately 45 ppi, I'd be effectively looking at a crop of a 6ftx9ft print.

At such magnification, any noise becomes very apparent. Small differences in noise in images from different cameras can clearly be seen. But who makes 6ftx9ft prints? The presence of noise is only of practical significance when the image is viewed as a whole, either as a print of a specific size or on a monitor display at a specific size.
Logged
Pete Ferling
Guest
« Reply #17 on: December 01, 2008, 11:11:37 PM »
ReplyReply

I have to agree with Edmund, as there's more to it than cramming pixels x into sensor size y.  Even with a 40D and a 1Ds (mark I), I have to apply proper technique to match what I get with MF, (i.e. if I'm too lazy to use a tripod, I shouldn't be surprised with with crappy detail on a 24" print).  I have printed and verified both formats with large prints on an Epson 9800, and some have left me scratching my head as to why I should continue with film, and then others tell me to I should have taken the 645 with me instead.  

The resolution gap is fading, and what's left is DR and a particular look.  Currently I take my digital with me, and then pick and chose which shots I would like to have as film as an option.  However, I have more digitals than film making it to the selects stage.  All this positive talk about the G10 doesn't surprise me, (and I can buy it as gift for my wife...
Logged
Ben Rubinstein
Sr. Member
****
Offline Offline

Posts: 1733


« Reply #18 on: December 02, 2008, 03:52:49 AM »
ReplyReply

Quote from: Ray
Pom,
I don't understand why you've never seen the truth of this in the real world. I see it myself clearly all the time. It seems usual practice when examining noise in an image to do it at 100% on the computer monitor, when pixel peeping. My current 19" CRT monitor is set at a resolution of 1280x1024. A bit of basic maths indicates that's approximately 92 ppi. A G10 or 50D image at 100% on this monitor represents a print size of 34" x51". If I were to examine image noise at 100% on my 50" HD plasma, which has a resolution of approximately 45 ppi, I'd be effectively looking at a crop of a 6ftx9ft print.

At such magnification, any noise becomes very apparent. Small differences in noise in images from different cameras can clearly be seen. But who makes 6ftx9ft prints? The presence of noise is only of practical significance when the image is viewed as a whole, either as a print of a specific size or on a monitor display at a specific size.

Again a tired old argument, the 100% on screen represents billboard size, etc, etc. In my experience sharpness/noise seen at 50% on screen is exactly how it looks on print when viewed from proper distances (another outmoded and innaccurate concept) and close up to the print 100% is 100%. The eye can resolve much more than 96DPI, 300-360DPI is the usual amount given. Is it strange then that the screen has to show such an enlargement until you can actually see the detail, at lower magnifications the fine detail is lost to the low resolution of the display whereas the detail is readily apparent on a print made at a much higher resolution which the eye can readily comprehend?

The argument is that downrezzing files reduces grain. I've never been able to make that work. Even if it did, the concept of someone buying a 24 megapixel camera only to downrez the files to match the noise from a lesser megapixel camera is preposterous.

Logged

Ray
Sr. Member
****
Offline Offline

Posts: 8853


« Reply #19 on: December 02, 2008, 09:23:48 AM »
ReplyReply

Quote
Is it strange then that the screen has to show such an enlargement until you can actually see the detail....

No it's not strange. The enlargement relates to the size of your monitor. 100% is pixel for pixel. The larger your monitor, the larger the image becomes (at 100%), even though the screen resolution might be the same and therefore the ppi (or dpi) lower.

Quote
The argument is that downrezzing files reduces grain. I've never been able to make that work. Even if it did, the concept of someone buying a 24 megapixel camera only to downrez the files to match the noise from a lesser megapixel camera is preposterous.

Pom, have you been drinking too much? You should go steady on that wine   .

At a reasonable ppi of 240 for printing purposes, a 24mp image will get you a 17"x26" print without interpolation; a 50D a 13"x20" print, and a 40D an 11"x16" print. Are you seriously claiming that it is preposterous that anyone might sometimes want to make prints smaller than 17"x26" from a 24mp image, or that it is preposterous that anyone would want to make a 17"x26" print from a 10mp 40D file?

I've made a few 24"x36' prints from 6mp images taken with my Canon D60, which look just fine. However, they were taken at base ISO.





Logged
Pages: [1] 2 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad