Thom Hogan's S3 review is up...

http://www.bythom.com/fujis3review.htm

And on his front page he even talks about an UW housing for the D70 (the Fantasea model).

On the S3 front, my immediate response is "meh"

Pretty unflattering. I'd guess that the 1DMk2 matches the S3 in dynamic range based on Thom's review and exceeds it in so many other ways. More expensive but street prices have come down and there are multiple housing options. Not a ringing endorsement for a camera specifically offered for its superior dynamic range.

I find it so odd that a camera designed for extended dynamic range should be targeted at users (JPG shooters) who never intend to use it. Give me RAW only especially with a camera like this.

Mark Thom down as another who thinks Fuji's have bad color. Other than extra DR issue his review pretty much agreed with Reichman's.

maybe it's true that some of the canon's can match the DR of the S3, but for $1500 more. but you have to realize where this camera is fitting for nikon users. i have thousands of dollars invested in nikon glass and as a college student the S3 fits right between what i already have D100 with what i would like, but can't afford D2X. I already own the S3 and i am extremely pleased with it's results. it's slow, that's my only drawback, but i do mostly nature and landscape where 5fps isn't a factor that i need to pay twice the price for the D2X.

it's all about perspective.

I think it's condemning if the S3's DR is matched by a conventional 8MP camera and the street price of the 1DMk2 has dropped. It's a far better camera in every other way. The price difference goes into a very expensive body.

I don't think the point is what's better for F-mount. I think the point is why botther with a dual photosite design if it can't outperform a conventional sensor? Fuji's sensor tech doesn't seem to be superior and it software and ergonomics are apparently substandard. It may be the best available for F-mount for under $4000 but it seems very exposed to me.

I think it's condemning if the S3's DR is matched by a conventional 8MP camera and the street price of the 1DMk2 has dropped. It's a far better camera in every other way. The price difference goes into a very expensive body.

 

I don't think the point is what's better for F-mount. I think the point is why botther with a dual photosite design if it can't outperform a conventional sensor? Fuji's sensor tech doesn't seem to be superior and it software and ergonomics are apparently substandard. It may be the best available for F-mount for under $4000 but it seems very exposed to me.

That isn't quite fair. The sensor on 1DMkII has more than 1.5x larger area. If Fuji can get the same resolution and DR, then their scheme has some merit. Because the CCD has to read out in series, they'll never be able to do 8 frames per second (this is probably the main reason Nikon has gone to CMOS for their pro camera, and Canon understood this earlier and got a technology jump on Nikon) . However, they probably can make the rest of it faster with a better processing and memory.

Heh, I wasn't nearly as un-impressed as you guys. I especially appreciated Thom's snow/stream samples - the way he tweaked the gamma to show the blown highlights in the S3-LowDR shot and then the comparison to the S3-High-DR where the detail was still there and good.

I think it's pretty analagous (sp??) to a sunball shot, where the area around the sunball would be blown with the S2, but the area around the sunball with the S3-High-DR wil have a lot more detail. I think this may be an extremely good sign.

In any case, Thom's review didn't include ANYTHING about shooting the camera in RAW and postprocessing to get the most out of the DR. Hopefully we'll get that soon. From Thom's review:

As I've said elsewhere, it's still early in the RAF experience and I really want to try multiple converters before I make any extensive comments. "

Cheers

James

I'd guess that the 1DMk2 matches the S3 in dynamic range based on Thom's review

I wouldn't have gotten that from Thom's review.

find it so odd that a camera designed for extended dynamic range should be targeted at users (JPG shooters) who never intend to use it. Give me RAW only especially with a camera like this.

I didn't get that from the article either. I think the key assessment was how the DR really works. By taking the higher exposure values and then crunching the rest of the data to fit between 0,0,0 and 255,255,255 the image becomes flat. So the shooter who opts for JPEG to avoid post processing will surely be disappointed.

I think the added dynamic range will improve sunball shots, but the rest of the water will still have that wierd cyan cast. It still has a clumsy histogram implementation, something I couldn't put up with.

I can't see a rush to upgrade, unless more DR is what you need and you don't care that:

1) Still has CAM900 AF

2) Uses d-TTL (what were they thinking)

3) Has the slowest continuos shooting rate in the product range

4) creates massive files that are slow to write

5) Can't easily use histogram to evaluate the exposure

I'll take a pass; at least until I start doing weddings (never), ski photography, or I'm heading to Antartica.

Haha, thanks George, I enjoyed that.

I think a lot of what Craig is talking about is addressed by Thom in this thread on DPreview:

http://forums.dpreview.com/forums/read.asp...essage=12097914

I think the green color cast can easily be corrected if you use the Calibrate feature in ACR the first time you use the camera. Of course, that's no good for JPEGs.

Cheers

James

Herb, you and I must see this differently then. The claim to fame of the Fuji sensor is superior dynamic range, so if it doesn't actually deliver that then I'm unimpressed. It's far easier after all to to deal with a conventional sensor. Regarding the larger size of the 1D2's sensor, I'm assuming that the Canon sensor is not more expensive to manufacture since it's CMOS and it doesn't have the dual photosite design and diagonal layout. I'd rather have the larger CMOS sensor than Fuji one if Fuji can't turn their work into superior performance. The greater cost of the 1D2 is justified by it's superior body BTW, not the size of its sensor and I'd rather have a true 8MP sensor than a faked one.

Regarding the 1D2 matching the DR of the S3, that is a guess and not justified by Thom's article alone. Thom claims that the S3 offers a 1 stop DR advantage and probably 2 stops putting the implied DR at 9-10 stops. I'm confident that the 1D2's DR is at least 9 stops based on other camera tests unrelated to Thom's review. It's at least a stop better than the 20D which is accepted to be at least 8 stops itself.

Regarding the irony of JPG output, that's strictly an observation. I can't see why shooters interested in a 6MP camera that offers "superior DR" would choose to work exclusively in a format only capable of 7 stops when every camera on the market can already do that.

At this point for $2500 I want a camera that does something more than fake it's >6MP output and the operational ease of at least a D100. Companies have had a while now to figure out basic things like instant on and proper histograms and image reviews. The operational difficulties of the S3 as described by both Thom and MR shouldn't be accepted in the dSLR market at any price point.

Herb, you and I must see this differently then. The claim to fame of the Fuji sensor is superior dynamic range, so if it doesn't actually deliver that then I'm unimpressed. It's far easier after all to to deal with a conventional sensor. Regarding the larger size of the 1D2's sensor, I'm assuming that the Canon sensor is not more expensive to manufacture since it's CMOS and it doesn't have the dual photosite design and diagonal layout. I'd rather have the larger CMOS sensor than Fuji one if Fuji can't turn their work into superior performance. The greater cost of the 1D2 is justified by it's superior body BTW, not the size of its sensor and I'd rather have a true 8MP sensor than a faked one.

I see your point. I'm more interested in the merits of their design approach, and comparisons only make sense when scaled to the same sensor size. It's alway possible to trade off dynamic range for resolution (more pixel) or cost (smaller chip). If they managed to do as well as a sensor that's 1.5 to 2x bigger, then they have something there. The cost of a silicon chip goes up as a high power of the area for chips this size: the yeild drops rappidly with area, and packing of large rectangles on a round wafer becomes less efficient when they get big. I would bet that's more significant than the difference between CCD and CMOS processing, and the differences in the circuit design is probably negligible.

As a quick check: the D70 now goes for around $800 and the 300D for around $650 retail. The manufacture cost can't be more than $200 to $300 for either sensor.

From Fuji's point of view, they probably do not consider the 1DII a significant competitor. It may be a good business decision to make the best DR camera body for Nikon lenses at the lowest cost.

Regarding the irony of JPG output, that's strictly an observation. I can't see why shooters interested in a 6MP camera that offers "superior DR" would choose to work exclusively in a format only capable of 7 stops when every camera on the market can already do that.

I think the gist of Thom's JPEG comment is that ones who do not want to deal with raw files prefer this camera because it's JPEG output is the best. This seems reasonable to me. It's a nonlinear map from the 12-14 bits raw to the 8 bits jpeg. A larger underlying dynamic range raw output would give jpegs that have more details in the shadows and less prone to blown highlights.

It's clear that the reviewers believe the S3 is an excellent JPG camera. My point is that JPG is the worst format to judge high dynamic range with so it should theoretically handicap the S3 more than the competition. Thus the irony. I find it interesting that the S3 owners at dpreview have latched onto the 14 bit converter feature as if that makes the S3 inherently better (and everyone else inherently lamer).

The Fuji design may offer advantages at the same sensor size but my point is the larger sensor of the 1D2 may offer the same or better performance and that it's cost is not necessarily higher. CMOS yields at larger sizes are better than CCD yields are as far as I know and the ADC circuits can be integrated.. The Fuji design has led to the awful performance characteristics as well though that could be overcome. If you feel married to the DX sensor size (which you might with F-mount) the issue is moot but remember that DX was chosen in the first place for cost considerations. I don't accept now that the goal is best performance in a DX size.

Ignoring the 1D2, the Kodak SLR/n uses roughly the same pixel pitch as the S3. At base ISO I'd be curious to see a direct comparison of dynamic range between the two. I think dynamic range comparisons between the S3, SLR/n and 1D2 would be very enlightening not because the bodies are of similar price but because Fuji has hyped its sensor design and I don't think its dynamic range is nearly the improvement they claim it is. As much as people have maligned the Kodak's usability it appears the S3 is much worse.

if you look around, eveyone who has the camera is in love with it and people who don't own it just complain about various things. take that as you may.

I don't know what you're suggesting but that can be said for just about any camera. In the end, forums are frequently short on objective information as owners of equipment are more interested in validation than real dialog. One of the good points of wetpixel is that it's relatively free of this type of thing.

If you have evidence the S3 criticism is nothing more than envy please speak up. Thom's review camera was purchased with his own money and I wouldn't say he's in love with it. Once upon a time Thom used the S2.

you're totally right craig. but as an owner it just seems that the s3 is getting more criticism then most other cameras get. it's not a fast action camera, as it wasn't designed to be. it's really good at what it does, which is the best image quality for the money. some day soon i'll post some shots of my own. i just have to have the time to.