Wednesday, May 24, 2006

Fuji S3 versus EOS 5D, an on-line comparison

Bernie Ess, former booster and yell king for the FujiFilm FinePix S3 Pro has recently put up a comparison between his new Canon EOS 5D and the FujiFilm FinePix S3 Pro he used to defend so vociferously. Check out his thread on dPreview, and get a few glimpses into the Fanboi culture of Fuji-dom from those who respond to Bernie's simple testing scenarios.
I find it ironic how Bernie spent so much time badmouthing the Nikon D2x (back before the EOS 5D was released),and how he spent so much time berating D2x shooters for their reliance on a "machine gun", as he constantly referred to the D2x. I find it ironic that Bernie used to boast about how having had to slow down his shooting due to the S3's incredibly molasses-like I/O systems helped him improve his photography,and how he felt so,so happy with the S3 Pro and how his experiences with a borrowed D2x showed him that the D2x was an inferior camera to the FujiFilm FinePix S3 Pro. As expected, now that he finally owns a TRUE, 12+ megapixel camera like the EOS 5D, Bernie is now out to show the Fuji world that the S3's output is indeed NOT 12 megapixels. I find it ironic that now that he finally owns a TRUE high-resolution camera, like the EOS 5D, that suddenly his old flame and the love of his life, the S3 Pro, is now relegated to second-tier status.
Even funnier,however, is seeing the way some of the diehard FujiFilm FinePix S3 Pro owners can fail to see the differences between a REAL 12+ MP camera and a Fuji-fied 12 MP camera, even when there is obvious proof right in front of their eyes. One of the BEST comparison tests Bernie did was to shoot the EOS 5D at 6 megapixel resolution,and then he took that Canon 6MP file and up-rezz'd it to 12 MP size, and then he compared the Canon file against the S3 Pro's Fuji-fied "12 megapxel" output. The fact of the matter is, the S3 Pro's images are far,far LESS-detailed than the EOS 5D images are--EVEN when one shoots the 5D at half-quality and then up-rezz's that 6 MP in-camera capture to 12 MP output size. If one is impartial, one can see there is a HIGH degree of artifacting in S3 Pro images. The Phil Askey test comparison shots between the EOS 20D and the S3 Pro show that the unique sensor array FujiFilm uses does NOT, simply does NOT produce clean,artifact-free images to nearly the same degree as more conventional Bayer array d-slr's. I feel bad reading through Bernie's test thread, and seeing the responses that Artichoke writes in defense of the S3 Pro,with his oft-repeated put-downs of Canon's CMOS sensor technology. It's sad to see a man so intensely loyal to his own brand that he has to resort to tearing down Canon's superb CMOS imaging quality in an effort to extoll the virtues of the S3 Pro's imaging quality.
A recent dPreview post explores the Phil Askey comparisons between the FujiFilm FinePix S3 Pro and the Nikon D70 and the EOS 20D. Please check out Phil Askey's S3 Pro review on pages 21,22,and 23. Look at the sample photos,and you can see that the Nikon D70 and S3 are often about neck and neck in terms of detail resolution; look honestly at the EOS 20D resolution sample photos and you can see how heavily artifact-marred the S3 Pro's SuperCCD images are compared to the EOS 20D's 8.2 MP images.
It is finally time to face the facts Nikonistas and simply MUST acknowledge that Canon has a huge lead in sensor technology over the rest of the d-slr field. Nikon,Fuji,Pentax/Samsung ,Konica-Minolta/Sony,and Olympus--ALL of these manufacturers have found it impossible to keep up with Canon in at least a few key areas of d-slr design and production. It does little good to keep hyping the FujiFilm FinePix S3 Pro's resolution abilities as being "comparable to an 8- or 9MP camera" when the Fuji can not even keep pace with a half-quality EOS 5Dcapture shot at "half-quality" or 6MP capture and which has been up-rezz'd in post to a 12 MP file...Canon's EOS 5D full-frame is simply resolving MORE detail, more CLEARLY, and more CLEANLY at 6MP, than the Fuji S3 Pro camera can do at its best-quality setting. The FujiFilm FinePix S3 Pro is,as as Phil Askey called it, a very GOOD but still SIX megapxiel camera.
A few days before Bernie Ess did his Canon versus Fuji test, the FujiFilm SLR Talk forum hosted a thread which explored the differences between 6 and 12 MP,and how the FujiFilm FinePix S3 Pro actually stacks up compared with other cameras. The thread begins here and goes on for a few pages. Please see the rez test charts,and look at pages 21,22,and 23 of the S3 Pro review for some side-by-side comparison images. Also,check out the comments by earthbound_ca in the thread, and see what insights he brings to the discussion.
Of course, resolution is not everything. Canon EOS 5D owners have to put up with superb low-noise files, excellent High-ISO shooting capabilities,and that awful, smooth,high-detail,high-resolution Canon CMOS sensor capture technology,as well as those nasty full-frame problems like the ability to isolate subjects through creative use of lens depth of field with wide-angles,normals,and telephotos. EOS 5D owners also suffer from the awful fate of having to use Canon's broad lineup of EF lenses with angles of view and DOF characteristics that are exactly like those from the days of 35mm film shooting. I pity the Canon 5D shooters who have to rely on their years' worth of 35mm experience because their D-SLR has done away with the FOV factor. The biggest problem I can see is that the EOS 5D uses EF mount lenses....if only Canon could get the "E" out, things would be very happy here in F-mount land.
Those Poor Canon Guys, they have so little on their plate,and now this...the EOS 5D, with full-frame capture at under $3,500. Without Canon driving the d-slr market's development, the world would be a much,much duller place.It's a pity that there are so many buzz-killers out there who are anxious to prove that,somehow, 1.5x crop sensors are somehow best, and that somehow six million photosites can,somehow, create resolution that's the equivalent of 8- or 9-, or even 12 million photosites. Hey, I've got fifty bucks on me; how about considering that it LOOKS like 85,let's say my fifty dollars are the "equivalent" to one hundred dollars. See how the Fujista analogy falls short,once you REALLY start LOOKING at numbers?
I said,publicly,for a long time that I would not even CONSIDER buying a Fuji S3 Pro until it got to $1699. Well, I actually started to consider the camera once it hit that price, and now that S3 is in the $1299 range in the USA,and is around $1099 US dollars equivalent in the UK, S3 sales are HOT! Finally, finally, finally the camera's price has been lowered to the point where people are buying them in good numbers,willingly,without the need for incentives like a Quantum Q-flash or some promo gimmick. I have been proven right; that $1699 was the beginning of the second wave of S3 adoption,and that at the even-lower prices in the $1299-$1249-$1199 range, are where the S3 belongs "naturally". I honestly think that had Fuji priced their camera lower,months and months ago, that they would have had a massive sales hit on their hands,but the combination of premium-level pricing of the S3 Pro combined with mixed reviews from Thom Hogan and Phil Askey, as well as others, hampered both the S3's reputation,and its lifetime sales potential. Again, let's get this straight: until the last few months, S3's have been languishing on dealer shelves world-wide,with VERY,very slow sales. Now that S3 prices are around $1299 to $1249 from large,reputable USA dealers, Fuji has a modest sales success on its hands,and is winning many happy customers. But there is a huge,huge difference between selling a camera for $2499 and for $1249. I cannot help but feel that,given the S3's slow write/read/display speed problems and very bare-bones buffer and software package that the original retail price was,well, almost unfair to customers. And while Fuji is now seeing brisk S3 sales, I feel like these low,low prices are for the first of the very last units in the channel. What I am seeing on-line is a huge influx of new FujiFilm D-SLR owners who feel prety good about their $1299 camera. And that makes me feel good too. I want people to enjoy their cameras!
There's been this persistent idea that for some reason, I do not "like" the S3 Pro. Which is kind of funny.Who cares if I happen to like it or dislike it? It's just not the camera for me. I'm not willing to archive 25 megabyte a pop raws forward through time. I expect compressed RAW files to be of a storage size commensurate with their era of creation, which is around 10.9 to 11.2 megabytes for a compressed, full-bit 12-megapixel D2x NEF file. Not 25 megabytes for a 6 megapixel image. AND, this is the big thing, the camera is so doggone slow that I cannot use it to leverage my photographic abilities. It's got bottom-tier AF, simplified metering, no metering with older lenses or non-coupled devices, the slowest flash synch in the professional camera class,and extremely LONG delay between the release breaking and the shutter moving, AND it has like 1.8 FPS speed shot-to-shot. The camera has a lot of purely mechanical limitations that really do not HELP me, or any other shooter using it to complete his asignments. Hey, I've shot 50,000 plus frames off the SAME BASIC UNDERBODY in the Fuji S2 Pro. The S3 Pro has a lot of body-based quirks that make the S3 a no-go camera FOR ME. Like the tiny viewfinder image projected by the N80 underbody. The S2 and S3 bodies have very POOR viewfinder images compared to what's on the market these days,and frankly, I cannot SEE thru the N80-class cameras as well as I can the 20D or the pro Nikons I own and use. That anybody would actually CARE that I do not or did not worship their favorite camera is kind of an odd thing. I had one otherwise sane photographer accuse me on dPreview that it was "my bleatings" about the S3 Pro that had "prevented him from buying" himself an S3 Pro for a fairly long time! Wow, such influence! Imagine how I felt when JK blamed ME for preventing HIM from buying an S3 Pro! What, didn't he read the initial Thom Hogan review of the camera? The review that called for FujiFilm's engineers to RE-DESIGN the S3 Pro, and if needed, to re-badge it? Uh.....even people who own and use the S3 have a laundry list of problems it suffers from. So, yeah, I think the S3 Pro is an interesting proposition of a camera. It clearly has a niche of followers who really,truly like what it gives them, but what I think about a camera is meaningless to anybody but me, or somebody who wants to shoot things the way I do. Honestly, I spent my camera money on the MUCH less-expensive EOS 20D back when the S3 was hitting the shelves,and frankly, I can see why so many people shoot Canon digital now. In terms of file size and easy,decent HIGH ISO capabilities and light metering, I think the 20D gives the D2x a real run for its money. If Fuji would have had the S3 Pro at $1699 in February of 2005,and not $2,500 at the same time, maybe I would have had a year and a half to fall in love with the S3 Pro. But I bought an 20D and a 4-lens outfit instead.
This past week, I shot EACH and every one of my D-SLR cameras, to get an idea of what they each can offer me. I have to say, the one thing that jumped out at me was the overall degree of artifacting from the S2 Pro images, the easy-breezy light metering and exposure of the 20D, and the tempermental and unforgiving nature of the D2x's imager,as well as the good color ACR in CS-2 now offers for the original D1's NEF files. While I liked the S2's color palettte, I was shocked to see how badly the imager performed on architectural subjects I was shooting,as well as on strong natural diagonal lines on macro subjects. And while I shot the D2x the most, I found time to shoot the 20D as well, and was muchly pleased at the high amount of detail both of those cameras (D2x and 20D) could resolve. I also had got in a bit of practice at autofocusing using several vastly different AF systems.
I think if you take a look,honestly, at the current state of affairs in digtal SLR-dom, the upper-level cameras from Canon have some advantages. Some real advantages,not just some imagined advantages. I'm a longtime Nikon fan,and am pretty committed to their system of lenses, but I must say that I am concerned that Nikon has made a series of really big mistakes and miscalculations over the past four years, which have in many cases, really hurt the company,and its loyal users. Product design and production flaws in the 70-200 VR lens, the D70, the D2h's dead meter problems, the D200 banding issues, and so on are some of the big mistakes Nikon has made in just the last couple of years. Nothing is perfect, but Nikon as a company has had some problems that I as a customer have found very annoying. But on-line there are some very interesting disagreements over what people are seeing with their own eyes. I think it's clear though: Bernie's S3 and EOS 5D comparison photos show what a real 12+ megapixel image looks like,and what an in-name-only 12 megapixel image looks like. He is taking some heat for using the wrong RAW converter,and shooting the S3 in some wrong-headed manner, but honestly, it's pretty clear; Six MP cannot compete with 12+ megapixels,resolution- and detail-wise.
Part II in this saga continued the day after I wrote the above. Bernie put up another post using different raw converters...frankly, the amount of difference between four good raw converter software applications is,well,miniscule at the limits of sensor resolution. Bernie took an urban landscape and examined small sections of some of the MORE-distant landscape, that part where every little bgit of added detail is nixe, and guess what? ACR and Fuji's own HU give about the SAME amount of detail resolved near the upper limits of the sensor's capabilities. I'm NOT buying that Fuji's HU raw software really betters ACR by enough to move the S3 out of the 6 million photosite category and even into the Canon 8.2 million photosite territory. Face it--all Pentax,Minolta,and most Nikon D-SLR's have a 6.1 MP or smaller imager in them, and the same with the S2 and S3 Fuji cameras. Canon has several 8 and 8.2 MP cameras, as well as higher MP models. Nikon has the D200 and the D2x. Canon goes 12.8 MP and up to 16.7 MP, with 1.6x, 1.3x,and 1.0x or Full Frame sensor sizes. The idea that Fuji's SuperCCD in either the S2 or S3 models resolves as much as a real 12 to 13 MP camera's sensor is just not consistent with the facts; in fact, the 8.2 MP 20D can resolve more REAL-world detail than an S3 can, if you really look at the images. The S2 and S3 do well on horizontal and vertical test chart lines; the real problem with the honeycomb sensor arrays is the problem with strong aliasing on diagonals,as well as just scrambled eggs-like detail on fine,high-frequency details in natural world scenes. Of course, the Fuji cameras have always been about COLOR and FEELING, and about the pictures, not about pixel-peeping. There is a certain subset of photographers who love the Fuji cameras,and who are willing to overlook a lot of flaws in order to get at the strength of the Fuji cameras. I see no problem with that. Fuji pictures often have a "look" about them. I see what Fuji-lovers see,picture-wise. But the details is where the Fuji cameras, S2 and S3, come up short. Who cares,though, if the pictures are good. Right?

No comments: