An interpretation of Phil's review of the S3

[Walter S Matthews]

The most important data presented by Phil in this review and, it seems overlooked by Phil et.al., is the astonishingly good set of photos Phil presented in the gallery part of the review. The color and the depth of color are what I expect form an S3 and represents the main reason why those of us who have bought the camera, actually bought it and continue to enjoy results like those Phil posted.

If he doesn't think those results are significant, that's fine.

But they are significant and they represent what you can get in pictures from the S3.

 

[Gabriele Sartori]

Why is the S3 only "above average" in spite of 2 more stops of
dynamic range and other benefits?

In my book the meaning of above average is very clear. THis is not an average camera, it is better. It is not outstanding for a number of reasons and is not only the price or the slowness, there are the file size, no lock up mirror, the below average SW and many other little things. What he has to write, exceptional just because you own one or you wish to own one?

Phil has been VERY fair with this report. He is the only one showing for the first time the real potential using Photoshop CS. Must give to the man credit. He developed and entirely new system in order to measure DR. He did a great job, not just the classic bla-bla.

--
Regards
Gabriele
California, CA
--------------------
Equipment list in profile

 

[Darrell Wilks]

There goes my forum entertainment.
Darrell

Enough interpretation already.Just tonight, a new poster alleged
that Phil had unfairly tested 20D and S3 images by having shot one
camera in RAW and the other in JPEG,when in FACT he had shot BOTH
cameras in JPEG Fine mode. The newbie poster's allegation of error
in Phil's testing procedure was seconded by a respected S3 user,
and only later was Phil vindicated by someboidy who actually
checked the facts before running down Phil's head-to-head
comparison; so, how about we stop all the "interpretation" of
Phil's tests in which somebody makes 1) a half-witted negative
comment about what Phil did or did not do or 2) states that the
Askey review says 'one thing' when it does, in fact, say NO SUCH
thing. In other words--let's start making doggone sure of what
we've read from Phil,and let's stop making off-based,flat out wrong
or bone-headed comments about what Phil Askey wrote in his review.
Is that too much to ask? Like I said, I found it really,really
annoying to see a newbie poster,and then a well-known poster here
both jump on Phil's methodology--when BOTH of them were flat-out
WRONGLY slamming Phil.
IOW, check your facts, and make sure the 'interpretations' you make
can be backed up with a page URL if need be. I see no reason to be
afraid to speak our minds, but we need to really stop the
half-baked,erroneous claims,and stick to the facts as Phil has
presented them. We've had enough bad interpretations already. If
you make a claim, be ready to cite the Askey review to back it
up,and stop making B.S. claims about what Phil said,and then we'll
have some valuable discussion. But the fanboy stuff is just
ridiculous already.
--
Happy Shooting!
Derrel

 

[Leo Terra]

Ok Phil, then correct the page 25 of Review, because in this page you tell cameras set to solemnity, all settings factory default and factory default in Fuji S3 i Normal JPG, like ISO i 200 and you photographed with ISO 100.

I never said that Phil had compared RAW with JPG.
Read sincerely, he said to have used the configurations standard of
factory, and in this case JPG Fine for to Canon 20D and JPG Normal
for Fuji S3 (take facts of the manual).
If he used JPG fine in all, certainly he didn't use the
configurations standard of factory, what still leaves more
difficult to know how it was made the test and it is also a mistake.
Before criticizing also reads what the " beginners" say, with the
same attention that reads the comments of Phil.

--
Phil Askey
Editor / Owner, dpreview.com

 

[DiDo59245]

Why is the S3 only "above average" in spite of 2 more stops of
dynamic range and other benefits?

In my book the meaning of above average is very clear. THis is not
an average camera, it is better. It is not outstanding for a number
of reasons and is not only the price or the slowness, there are the
file size, no lock up mirror, the below average SW and many other
little things. What he has to write, exceptional just because you
own one or you wish to own one?
Phil has been VERY fair with this report. He is the only one
showing for the first time the real potential using Photoshop CS.
Must give to the man credit. He developed and entirely new system
in order to measure DR. He did a great job, not just the classic
bla-bla.

--
Regards
Gabriele
California, CA
--------------------
Equipment list in profile

--
DiDo
http://www.fotocommunity.de/pc/pc/mypics/2488
http://kulturserver-nrw.de/home/cbruenig/

 

[Robert Smith]

Jeff,

Check this out. I took the same S3 ISO 100 RAW file, converted it in Fuji's HS-V2 converter at the 12mp size, and ultra-sharpened it (USM 500% .4 pixels, TWICE). Did the same with the SAME file, using ACR. The ACR has MUCH more detail, is less digital, more film-like, better tonal gradations, in every way superior except for present noise (which looks amazingly like film grain). These are 200% crops:

Converted in Fuji's HS-V2 converter:



Converted in ACR:



Robert

So you're saying that in Standard mode (you put 'in general') you
can see more dynamic range in an S3 Pro image than other D-SLR's.
I'd say you've not done any objective comparisons. The point we're
making here is that yes the S3 Pro has the extra dynamic range (in
Wide mode) but that it's not using it effectively, a different tone
curve and modification to metering in Wide mode (to expose for
highlights) would have made much more of its extra range. Take
this with the fact that the in-camera JPEG and Hyper-Utility2
appear to be incapable of making the most of the extra R-Pixel data
and it's clear that the camera is unfortunately crippled unless you
shoot RAW all the time and convert in ACR. The camera should be
optimal out of the box, most users will not discover the ACR option
by chance.

My experience is that in Auto DR mode the S3 really does make
excellent choices. I'm getting terrific results in all kinds of
situations that are problematic for any other digital camera I've
used - mixed lighting, backlighting, high contrast, etc. In fact,
I'd say that having an S3 has changed my shooting style, since I
now try things that I avoided in the past because I knew the camera
wasn't up to it.

I have also found that while ACR is better at recovering
information from severely overexposed scenes, there are some
situations in which Hyper Utilities is clearly better. A couple of
nights ago I shot a recital under difficult lighting conditions.
Fuji's raw converter really produced the goods, whereas the ACR
conversion generated ugly colour noise all over the singer's face,
no matter what settings I tried.

Best regards
Werner
"newbee Poster"

This constant slagging of "newbies" (not by Werner or Phil) is
irksome and unneccessary.

 

[Jeff Tsang]

As I said the ACR will have more detail in the shadow area. I like the Fuji pic more when look at it in the zoom out size. That is purely personal. Can you explain what happen to the file size. It seems to decrease more after conversion by ACR taking account of the number of pixels.

Check this out. I took the same S3 ISO 100 RAW file, converted it
in Fuji's HS-V2 converter at the 12mp size, and ultra-sharpened it
(USM 500% .4 pixels, TWICE). Did the same with the SAME file, using
ACR. The ACR has MUCH more detail, is less digital, more film-like,
better tonal gradations, in every way superior except for present
noise (which looks amazingly like film grain). These are 200% crops:

Converted in Fuji's HS-V2 converter:



Converted in ACR:



Robert

So you're saying that in Standard mode (you put 'in general') you
can see more dynamic range in an S3 Pro image than other D-SLR's.
I'd say you've not done any objective comparisons. The point we're
making here is that yes the S3 Pro has the extra dynamic range (in
Wide mode) but that it's not using it effectively, a different tone
curve and modification to metering in Wide mode (to expose for
highlights) would have made much more of its extra range. Take
this with the fact that the in-camera JPEG and Hyper-Utility2
appear to be incapable of making the most of the extra R-Pixel data
and it's clear that the camera is unfortunately crippled unless you
shoot RAW all the time and convert in ACR. The camera should be
optimal out of the box, most users will not discover the ACR option
by chance.

My experience is that in Auto DR mode the S3 really does make
excellent choices. I'm getting terrific results in all kinds of
situations that are problematic for any other digital camera I've
used - mixed lighting, backlighting, high contrast, etc. In fact,
I'd say that having an S3 has changed my shooting style, since I
now try things that I avoided in the past because I knew the camera
wasn't up to it.

I have also found that while ACR is better at recovering
information from severely overexposed scenes, there are some
situations in which Hyper Utilities is clearly better. A couple of
nights ago I shot a recital under difficult lighting conditions.
Fuji's raw converter really produced the goods, whereas the ACR
conversion generated ugly colour noise all over the singer's face,
no matter what settings I tried.

Best regards
Werner
"newbee Poster"

This constant slagging of "newbies" (not by Werner or Phil) is
irksome and unneccessary.

 

[Emerson]

Hi Bernie,

I agree with you do a point, but I like the fact that he gives an opinion. When it comes to the nuts and bolts of cameras there is no one I trust more than Phil Askey. I don't think we should be killing the messenger, all Fuji had to do was be reading this board for the past two years to find out what most of the users wanted in the next camera.

Hopefully Phil's opinion will make for a better S4. I bought the Fuji S2 because of my love for the final image for skin tones in portraits and weddings, I also have Canons which I like better for other reasons. The Fuji S3 addressed none of the problems I have with the S2, I did and still do love the image of the S2, so tweaking the image for more DR was not high on my list.

Emerson

In this case it would be enough to list the "Plus" and "minus"
points without a final rating. Even a "recommended" is a subjevtive
rating, it's "recommended by Phil Askey".
I for one would "highly recommend" the S3 to anyone with a passion
for photography. I did not like the 20d as much. So doing reviews
without a final judgement would be better.

Bernie

 

[Robert Smith]

.... but that doesn't mean it is incapable of reconstructing a 12mp image from that data, and resolve more real detail at that setting. I'm not an expert in how ACR does this, but Thomas Knoll (Photoshop designer) has weighed in here and explained that ACR does indeed create the 12mp interpolated file first, even when rendering the 6mp image (in S2 and S3 files), and also uses the R pixels, in the case of the S3 files. That's what I've read here, and I believe it, based on these results.

Robert

Check this out. I took the same S3 ISO 100 RAW file, converted it
in Fuji's HS-V2 converter at the 12mp size, and ultra-sharpened it
(USM 500% .4 pixels, TWICE). Did the same with the SAME file, using
ACR. The ACR has MUCH more detail, is less digital, more film-like,
better tonal gradations, in every way superior except for present
noise (which looks amazingly like film grain). These are 200% crops:

Converted in Fuji's HS-V2 converter:



Converted in ACR:



Robert

So you're saying that in Standard mode (you put 'in general') you
can see more dynamic range in an S3 Pro image than other D-SLR's.
I'd say you've not done any objective comparisons. The point we're
making here is that yes the S3 Pro has the extra dynamic range (in
Wide mode) but that it's not using it effectively, a different tone
curve and modification to metering in Wide mode (to expose for
highlights) would have made much more of its extra range. Take
this with the fact that the in-camera JPEG and Hyper-Utility2
appear to be incapable of making the most of the extra R-Pixel data
and it's clear that the camera is unfortunately crippled unless you
shoot RAW all the time and convert in ACR. The camera should be
optimal out of the box, most users will not discover the ACR option
by chance.

My experience is that in Auto DR mode the S3 really does make
excellent choices. I'm getting terrific results in all kinds of
situations that are problematic for any other digital camera I've
used - mixed lighting, backlighting, high contrast, etc. In fact,
I'd say that having an S3 has changed my shooting style, since I
now try things that I avoided in the past because I knew the camera
wasn't up to it.

I have also found that while ACR is better at recovering
information from severely overexposed scenes, there are some
situations in which Hyper Utilities is clearly better. A couple of
nights ago I shot a recital under difficult lighting conditions.
Fuji's raw converter really produced the goods, whereas the ACR
conversion generated ugly colour noise all over the singer's face,
no matter what settings I tried.

Best regards
Werner
"newbee Poster"

This constant slagging of "newbies" (not by Werner or Phil) is
irksome and unneccessary.

 

[Jeff Tsang]

The following is the result of search of DP review.

"This is a misquote of me. What I said is that Adobe also does a doubling of the pixel count during the demosasicing step (to fill in the holes in the 45 degree rotated matrix), and then an optional downsampling back to the original pixel count. This is the same sizing logic that Fuji is using. However, Fuji has not provided Adobe their actual algorithms.

Thomas Knoll. "

The point is that Fuji has not provide Adobe their actual algorithms. So Adobe is guessing and that explain why conversion by HU2 and ACR is different. Anyway thanks for your information. I tend judge a photo by its overall look rather than the minor detail. That is not to say minor details are not important. Just a personal preference.

Robert

Check this out. I took the same S3 ISO 100 RAW file, converted it
in Fuji's HS-V2 converter at the 12mp size, and ultra-sharpened it
(USM 500% .4 pixels, TWICE). Did the same with the SAME file, using
ACR. The ACR has MUCH more detail, is less digital, more film-like,
better tonal gradations, in every way superior except for present
noise (which looks amazingly like film grain). These are 200% crops:

Converted in Fuji's HS-V2 converter:



Converted in ACR:



Robert

So you're saying that in Standard mode (you put 'in general') you
can see more dynamic range in an S3 Pro image than other D-SLR's.
I'd say you've not done any objective comparisons. The point we're
making here is that yes the S3 Pro has the extra dynamic range (in
Wide mode) but that it's not using it effectively, a different tone
curve and modification to metering in Wide mode (to expose for
highlights) would have made much more of its extra range. Take
this with the fact that the in-camera JPEG and Hyper-Utility2
appear to be incapable of making the most of the extra R-Pixel data
and it's clear that the camera is unfortunately crippled unless you
shoot RAW all the time and convert in ACR. The camera should be
optimal out of the box, most users will not discover the ACR option
by chance.

My experience is that in Auto DR mode the S3 really does make
excellent choices. I'm getting terrific results in all kinds of
situations that are problematic for any other digital camera I've
used - mixed lighting, backlighting, high contrast, etc. In fact,
I'd say that having an S3 has changed my shooting style, since I
now try things that I avoided in the past because I knew the camera
wasn't up to it.

I have also found that while ACR is better at recovering
information from severely overexposed scenes, there are some
situations in which Hyper Utilities is clearly better. A couple of
nights ago I shot a recital under difficult lighting conditions.
Fuji's raw converter really produced the goods, whereas the ACR
conversion generated ugly colour noise all over the singer's face,
no matter what settings I tried.

Best regards
Werner
"newbee Poster"

This constant slagging of "newbies" (not by Werner or Phil) is
irksome and unneccessary.

 

[Robert Smith]

The following is the result of search of DP review.

"This is a misquote of me. What I said is that Adobe also does a
doubling of the pixel count during the demosasicing step (to fill
in the holes in the 45 degree rotated matrix), and then an optional
downsampling back to the original pixel count. This is the same
sizing logic that Fuji is using. However, Fuji has not provided
Adobe their actual algorithms.

Thomas Knoll. "

The point is that Fuji has not provide Adobe their actual
algorithms. So Adobe is guessing and that explain why conversion by
HU2 and ACR is different. Anyway thanks for your information. I
tend judge a photo by its overall look rather than the minor
detail. That is not to say minor details are not important. Just a
personal preference.

Fair enough. But have you tried to make large prints from S3 files?? At 18x27 inches (maybe even smaller), the jaggie artifacts from HS-V2 converted files are really obvious, and very annoying. These jaggies are absent in ACR-converted S3 files, and in the prints from these files. I've made the same comparisons on a number of images, viewed on screen and in print, and there is missing detail in the HS-V2 versions. The Fuji files sometimes appear to be "sharper" because of the acute horizontal and vertical lines they produce (look at my samples to see these), but this same processing is what makes the moire and jaggies appear. Be fair: which one looks more like a real photograph, more like film?? Which one prints better, especially large? As you say, the results are what is important, not the pixel peeping. But when the pixel peeping shows dramatic differences, these appear in real prints as well.

Here's a crop from a landscape image I was working on, and had really awful jaggies in print from the HyperUtility version. The ACR version was converted at its "max rez" (6144x4096 pixels), and I up-rezzed the HyperUtility version to the same size and perceptually sharpened both to the same degree:

HS-V2:



ACR:



Here's a sample (same pic of the eye above) of a "hotly" exposed portrait in bright sun (not very artistic, I admit), and the skin tone transitions were much better in ACR. This is my daughter, and she has no green or yellow skin on her nose!:

HS-V2:



ACR:



Add to this the now infamous "pink" highlights when we reduce an overexposed S3 image in HyperUtility. (If you haven't seen this, I can post a copy. It's really awful, and it's NOT in the file, because ACR doesn't do this.)

You've either got to be putting your head in the sand, or are for some reason loyal to Fuji, to not see this. Believe me, I WANT to support Fuji in this argument, but their software stinks (and it's SLOW as He!! on a Mac). If they improve it, I'll be the FIRST to support them, and I'll proclaim it loud and clear.

Robert

.... but that doesn't mean it is incapable of reconstructing a 12mp
image from that data, and resolve more real detail at that setting.
I'm not an expert in how ACR does this, but Thomas Knoll (Photoshop
designer) has weighed in here and explained that ACR does indeed
create the 12mp interpolated file first, even when rendering the
6mp image (in S2 and S3 files), and also uses the R pixels, in the
case of the S3 files. That's what I've read here, and I believe it,
based on these results.

Robert

Converted in Fuji's HS-V2 converter:



Converted in ACR:



Robert

Totally agree. We should not open raw file with C1 or ACR. The file
size will decrease by half compare with that open by HU. The color
is lighter and more grains. This will effectively gives more detail
in the shadow at a price of course. The price is too heavy for me.
The color of the photo will be less rich, less smooth and less look
like a photo. I suspect the DR element has been removed by opening
with ACR or C1.

 

[TedS85150]

Why is the S3 only "above average" in spite of 2 more stops of
dynamic range and other benefits? Apparently because it is too
slow, and too expensive. These criticisms are valid and important
to many photographers, but it seems eccentric to give a single
overall rating to a camera (or any other product) that has many
facets. As an obvous example, a Leica M-camera has a terrible
viewfinder that shows only about 85% of what the lens sees, can't
focus closer than about 30", can't accomodate telephoto lenses, and
is very expensive. But many great photographers have used this
camera series because of its well known virtues in certain
situations. If Phil's approach had been used to rate the Leica M3
when it came out, it would have earned a below-average rating.
DPReview's evaluations are important for the information they
convey. The overall ratinga are much less useful, and should be
abandoned.
--
Bruce Douglas

--
Ted S

 

[Mark127996]

I own the camera and even though I picked at a few minor details like the significance of DR these reviews he does are really quite excellent.

I am not sure Phil has an allegiance to one brand or another but say that Fuji was not his brand of choice, don't you think the time and effort he put in merits some praise here?

I have had my S3 since Dec 29 now and even I lerned a few things from the review.

I would save even his final rating is generous given the typical DSLR buyer. I have been thinking about this and if you are a techie like myself, you expect to see features and advances in those areas.

Fuji put out a great camera that advances Image Quality, but it isn't going to be a camera after many hearts. Even me, who defended the thing to a fault realizes that we have to respect peoples rights to choose a camera for whatever reasons they want to.

We also keep saying that if Fuji Put the Sensor in a D2h body, or if only Nikon would put a Fuji Sensor in its pro body.

I think the rating is pretty generous if you think about what could have been, yes, even today not even having to wait for the future.

I also think though that if we feel he made some errors that we should have civil discussion about it.

Phil, I still think you did a great job and remained very objective in your analysis. Thanks for all the hard work and for this wonderful free site most of us enjoy.

--
Thanks!
Mark
Fuji DReam Machine shooter
My new website!
http://www.eyesalive.com

 

[dean dugger]

Amen!

deano

I own the camera and even though I picked at a few minor details
like the significance of DR these reviews he does are really quite
excellent.

I am not sure Phil has an allegiance to one brand or another but
say that Fuji was not his brand of choice, don't you think the time
and effort he put in merits some praise here?

I have had my S3 since Dec 29 now and even I lerned a few things
from the review.

I would save even his final rating is generous given the typical
DSLR buyer. I have been thinking about this and if you are a
techie like myself, you expect to see features and advances in
those areas.

Fuji put out a great camera that advances Image Quality, but it
isn't going to be a camera after many hearts. Even me, who
defended the thing to a fault realizes that we have to respect
peoples rights to choose a camera for whatever reasons they want to.

We also keep saying that if Fuji Put the Sensor in a D2h body, or
if only Nikon would put a Fuji Sensor in its pro body.

I think the rating is pretty generous if you think about what could
have been, yes, even today not even having to wait for the future.

I also think though that if we feel he made some errors that we
should have civil discussion about it.

Phil, I still think you did a great job and remained very objective
in your analysis. Thanks for all the hard work and for this
wonderful free site most of us enjoy.

--
Thanks!
Mark
Fuji DReam Machine shooter
My new website!
http://www.eyesalive.com

 

[EdVT]

Ed wrote

I think All of you on this forum have been way to hard on Phil and his review, Take it for what it was meant to be "A REVIEW" Not a "Thesis" for wich he had a year and unlimited funds and equipment. proffesional expierience, preference and personal opinions, are alowed in a REVIEW, A lot of you contradict your selfs in the next sentence you write, and make some pretty silly statement. read and think before you put your opinions to paper. I take my hat of for guys like Phil, I learn from him an combine that knowledge with what I learn from other reviewers and yes from ALL of you here on this forum. then I make up my own mind on what I think will serve my purpose best. That how I decided to purchase my S2, and that's what this forum and these reviews are meant for. DISCUSS do'nt BASH.
Thanks Phil for your time and effort spend on these endeavors! Ed

But as everything is a commercial subject he preferred to do a
comparison of S3 with entrance cameras, making a review distorted.
I think 2 points of more latitude it is a considerable value,
mainly for fashion photographers and architecture. Any one that
sees the page 19 of Review will agree with me.

yes, Leo
Phil got it wrong ...he didn't even mention what is perhaps the
greatest advantage of the S3 ...14 bit AD conversion
this along with the additional DR makes for lifelike, smooth color
transitions that benefits virtually every shot taken with the S3
--
pbase & dpreview supporter
Fuji SLRT forum member since 5/2001
http://www.pbase.com/artichoke

 

[Robert Smith]

Hi Mark,

I applaud your fairness (your restraint?) in praising Phil's review, or at least defending it. I have searched deeply my own sense of fairness, and I have come to a somewhat different conclusion. If you'd like to know more on what I think, please read the linked post:

http://forums.dpreview.com/forums/read.asp?forum=1020&message=12719487

The bottom line for me is that I think Phil was fair and honest in his own mind, but his job of reviewing, IMO, was at least as Average as the rating he gives the camera, and I have some specific reasons for saying so.

Robert

I own the camera and even though I picked at a few minor details
like the significance of DR these reviews he does are really quite
excellent.

I am not sure Phil has an allegiance to one brand or another but
say that Fuji was not his brand of choice, don't you think the time
and effort he put in merits some praise here?

I have had my S3 since Dec 29 now and even I lerned a few things
from the review.

I would save even his final rating is generous given the typical
DSLR buyer. I have been thinking about this and if you are a
techie like myself, you expect to see features and advances in
those areas.

Fuji put out a great camera that advances Image Quality, but it
isn't going to be a camera after many hearts. Even me, who
defended the thing to a fault realizes that we have to respect
peoples rights to choose a camera for whatever reasons they want to.

We also keep saying that if Fuji Put the Sensor in a D2h body, or
if only Nikon would put a Fuji Sensor in its pro body.

I think the rating is pretty generous if you think about what could
have been, yes, even today not even having to wait for the future.

I also think though that if we feel he made some errors that we
should have civil discussion about it.

Phil, I still think you did a great job and remained very objective
in your analysis. Thanks for all the hard work and for this
wonderful free site most of us enjoy.

--
Thanks!
Mark
Fuji DReam Machine shooter
My new website!
http://www.eyesalive.com

 

[JohnB23144]

That's always bugged me on the S2, and it's an example of a left-over from the original film nature of the body.

For that matter, I've been wondering why some new shooting modes couldn't be introduced on digital cameras - along with shutter speed and aperture, let ISO be a parameter that can auto-adjust based on conditions.
JohnB

 

[Radu Grozescu]

For that matter, I've been wondering why some new shooting modes
couldn't be introduced on digital cameras - along with shutter
speed and aperture, let ISO be a parameter that can auto-adjust
based on conditions.

Actually, some Nikon bodies have Auto ISO, and maybe other cameras I do not know about.

--
Radu Grozescu

http://www.RaduGrozescu.com
Corporate & Editorial Photography

 

[Darrell Wilks]

The only drawback to ACR I have found is some halo-ing around specular highlights. I could not find settings to remove them in ACR. I will grant that these are visible at the near pixel zoom level, but not so much when viewing the full image. The ACR conversion does seem more pleasing overall.
Darrell

Fair enough. But have you tried to make large prints from S3
files?? At 18x27 inches (maybe even smaller), the jaggie artifacts
from HS-V2 converted files are really obvious, and very annoying.
These jaggies are absent in ACR-converted S3 files, and in the
prints from these files. I've made the same comparisons on a number
of images, viewed on screen and in print, and there is missing
detail in the HS-V2 versions. The Fuji files sometimes appear to be
"sharper" because of the acute horizontal and vertical lines they
produce (look at my samples to see these), but this same processing
is what makes the moire and jaggies appear. Be fair: which one
looks more like a real photograph, more like film?? Which one
prints better, especially large? As you say, the results are what
is important, not the pixel peeping. But when the pixel peeping
shows dramatic differences, these appear in real prints as well.

 

[Lost Horizon]

My guess is Phil "held his nose" and gave the camera an "above average". If you really factored in the price and what you get wrt other cameras, I bet "below average" is what he really wanted to assign.

Why is the S3 only "above average" in spite of 2 more stops of
dynamic range and other benefits? Apparently because it is too
slow, and too expensive. These criticisms are valid and important
to many photographers, but it seems eccentric to give a single
overall rating to a camera (or any other product) that has many
facets. As an obvous example, a Leica M-camera has a terrible
viewfinder that shows only about 85% of what the lens sees, can't
focus closer than about 30", can't accomodate telephoto lenses, and
is very expensive. But many great photographers have used this
camera series because of its well known virtues in certain
situations. If Phil's approach had been used to rate the Leica M3
when it came out, it would have earned a below-average rating.
DPReview's evaluations are important for the information they
convey. The overall ratinga are much less useful, and should be
abandoned.
--
Bruce Douglas