F11 or e900 Tragic Choice. I just made it ..

[Kim Letkeman]

On the E900, is it possible to set the ISO to 800, underexpose by 1
stop, and correct the brightness in postprocessing? How would the
results compare with the F11 at ISO 1600?

That works fine if you don't lose any shadow detail. But forget about it approaching the F10 and F11 in quality. That's just not gonna happen from that sensor. And less light always brings out a lot more noise.

I ask because in one of the other dpreview forums--I think the one
for Fuji SLRs--there's a long thread extolling the practice of
setting the S3 (I think) to ISO 1600, underexposing by 1/2 stop,
and fixing it in Photoshop.

So they get the equivalent of ISO 2400 ... interesting. But the S3 has a bigger sensor, no?

 

[Unda Covalava]

I would like to see these tragic results if possible. Is the F11 that much better?

 

[Unda Covalava]

That works fine if you don't lose any shadow detail. But forget
about it approaching the F10 and F11 in quality. That's just not
gonna happen from that sensor.

How do you know this? I was not able to find reviews, much less head-to-head comparisons anywhere. Did Fuji really get it right with the F10/11, then blow it with subsequent chipsets?

 

[Unda Covalava]

That's a comparison I'd like to see. How does the image look downsized to match an f10/f11 shot. Does the noise become less apparent? Is there then a more detailed, less noisy shot?

 

[Kim Letkeman]

How do you know this? I was not able to find reviews, much less
head-to-head comparisons anywhere. Did Fuji really get it right
with the F10/11, then blow it with subsequent chipsets?

Actually, they didn't blow it ... rather, they were handicapped by two things:

1) E900/S9000z -- 9mp on a sensor about the same size as the 6mp sensor in the F10 is pretty much going to guarantee more noise. Fuji knew that but they needed to compete in the high mp bridge category.

2) S5200 -- that 5mp sensor is not today's chip ... so it just cannot quite compete with either of the others. Still, Fuji has very good noise processing and has been able to offer 1600 ISO for the first time on a chip like this.

In order to demonstrate why I have come to these conclusions, I've put together a little compendium of noise graphs from http://www.digitalcamerainfo.com ... this juxtaposes a number of cameras to see how their noise performance compares. The graph first, then my interpretation of the results. Note: when readig the graph, note that the Y axis is not always scaled identically, so you need to look at each point individually to see how the cameras stack up.



My interpretation of the cameras in order:

1) D50 -- the winner by far. The best noise measurements at every ISO. Wow.

2) Rebel XT -- An excellent sensor. Interestingly, it blooms a bit at 800, letting the F10 slip under it, but regains the lead again at 1600. Note that the F10 has higher in camera noise reduction, so I think the visual quality of the Rebel XT's noise is better at the same measurement level.

3) F10 -- huh? What is a tiny sensor like that doing here? Well, that's what the fuss is all about, isn't it?

4) Rebel -- just squeaks in ahead of the S9000z. But definitely behind the F10 by a more than a hair.

5) S9000z -- well done! At 800, it's still very low noise (although much higher than the F10.)

6) S5200 and E300 tie -- The SLR starts with lower noise, is a hair better at 800, but worse at 1600. Wow! Again, it is possible that the visual quality of the noise is better on the Oly. But they are too close to call by the measurements.

7) F828 -- not great ...starts pretty low at ISO 64 (or 50, can't tell on the graph) then rises rapidly. A hair less noisy at 800 than the S5200 at 1600.

8) FZ30 -- starts really poorly ... finally competitive with the F828 at 400 ISO. Low or even moderate noise is just not this camera's "thing".

Summary
======

So there you have it ... the buzz is real ... the results are amazing ... and if you want more features and still want the ability to shoot pictures inside theatres, museums, parties, and so on but do not want a dSLR for whatever reason ... well, the answer is still Fuji at this time.

 

[Unda Covalava]

Wow, you went to a lot of trouble for these graphs. I see what you're saying about what happened after the F10 came out.

I don't quite trust the source referenced, they really have an odd-ball way of presenting, or making up data. For example, it's not clear how they determine % of actual pixels used - seems they just throw numbers out there.

Also, based on personal experience, the FZ30, which looks horrible on those graphs, is certainly better than the 828.

And finally, the graph does not tell the whole story, because an E- image is going to outdo a smaller sensor camera every time, which you allude to in the quality of the noise, but also in other parameters such as preserved detail, color accuracy, dynamic range, etc.

 

[Kim Letkeman]

Wow, you went to a lot of trouble for these graphs. I see what
you're saying about what happened after the F10 came out.

I don't quite trust the source referenced, they really have an
odd-ball way of presenting, or making up data. For example, it's
not clear how they determine % of actual pixels used - seems they
just throw numbers out there.

There is some concern over their test photos ... all are not perfect. But they run the photos through IMAGEN ... and that doesn't throw numbers around. They have recently changed their reporting methods to match the way dpreview reports (lines of resolution instead of measured pixel count), so it is sometimes hard to get a perfect resolution comparison. But for noise they simply plot the numbers they get, also from IMAGEN.

Also, based on personal experience, the FZ30, which looks horrible
on those graphs, is certainly better than the 828.

Read the review here and on digitalcamerainfo ... both cameras suck where noise is concerned. Neither camera is noise free even at its lowest ISO.

They are fighting for last place ... who really cares who wins?

And finally, the graph does not tell the whole story, because an E-
image is going to outdo a smaller sensor camera every time, which
you allude to in the quality of the noise, but also in other
parameters such as preserved detail, color accuracy, dynamic range,
etc.

Preserved detail is simply a matter of what IMATEST sees. And there I think you need to go back and read the reviews again ... the F10 recorded an astounding 90% of advertised pixels at about 5.5mp ... while the E300 recorded a dismal 4.55mp ... they had this to say:

"After recording numerous exposures of the test chart, we concluded the EVOLT E-300, using its 14-45mm kit lens, only utilized 4.55 of its 8 advertised megapixels for imaging. This is an extremely low score that is inferior to any DSLR we have tested to date and much more in line with substandard point-and-shoot models. The test shots were conducted at various aperture openings throughout the lens's focal range. We reported the highest score we could attain with the camera, which was shot at f/11 at 45mm (35mm equivalency)."

EVOLT against F10 ... a fairer fight than it should be. F10 has better retained resolution and much better noise numbers.

 

[Unda Covalava]

Read the review here and on digitalcamerainfo ... both cameras suck
where noise is concerned. Neither camera is noise free even at its
lowest ISO.
They are fighting for last place ... who really cares who wins?

LOL- that's funny. I have the FZ30 and find it to be a fine photographic instrument, capable of capturing amazing detail with the finest gradations of color and tone. And this comes after 5 years of experience with over a dozen digital cameras. Noise on a graph simply doesn't tell the whole story. Might as well buy a camera simply on megapixel count!

EVOLT against F10 ... a fairer fight than it should be. F10 has
better retained resolution and much better noise numbers.

So according to this review site, the F10 takes better pictures than the E-300. Not sure about that. In any case, I'm interested in the F10/F11 for its performance and size, and hopefully Fuji will release a more pocketable version of the F11 at PMA.

 

[Poochdp]

hopefully Fuji
will release a more pocketable version of the F11 at PMA.

I think you will have to hold your breath longer then that!

 

[Lou Chen]

The Fuji rep at CES said that the F10/11 will be replaced by the F30 in June. Don't know about the form factor, but major differences will be image stab, higher screen pixel count, and ISO range out to 3200.

 

[Unda Covalava]

That sounds too good to be true. Could you imagine the possibilities of a useable iso that high with image stabilization? And the higher screen pixel count would be nice, as the F10 seems like it's a little on the low-res side. Reps usually don't disclose this type of info, but if accurate, sounds like a very impressive camera. Not sure if I'm going to wait that long, though.

The Fuji rep at CES said that the F10/11 will be replaced by the
F30 in June. Don't know about the form factor, but major
differences will be image stab, higher screen pixel count, and ISO
range out to 3200.

 

[Kim Letkeman]

LOL- that's funny. I have the FZ30 and find it to be a fine
photographic instrument, capable of capturing amazing detail with
the finest gradations of color and tone. And this comes after 5
years of experience with over a dozen digital cameras. Noise on a
graph simply doesn't tell the whole story. Might as well buy a
camera simply on megapixel count!

I'm sure it is a fine instrument. There are certainly enough people convinced of that to make it one of the leading sellers. That does not mean though, that it does not suck compared to others where noise is concerned. Most people never print big and never do a lot of photoshopping ... for those who just print small or look on screen, I'm sure it is enough.

EVOLT against F10 ... a fairer fight than it should be. F10 has
better retained resolution and much better noise numbers.

So according to this review site, the F10 takes better pictures
than the E-300. Not sure about that. In any case, I'm interested
in the F10/F11 for its performance and size, and hopefully Fuji
will release a more pocketable version of the F11 at PMA.

I didn't say it took better pictures overall. But it takes somewhat sharper pictures when the E300 is mated with the el cheapo kit lense. and it takes better pictures at 1600 ISO in low light. You decide. If these things don't matter, then the SLR would easily win otherwise, wouldn't it?

 

[Unda Covalava]

Most people never print big and never do a lot
of photoshopping

I guess I'm unlike most, as I usually print big ( 11"x14" on up) and always do a lot of photoshopping.

... for those who just print small or look on
screen, I'm sure it is enough.

For these purposes, it's decidedly overkill.

 

[normankoren]

"After recording numerous exposures of the test chart, we concluded the EVOLT E-300, using its 14-45mm kit lens, only utilized 4.55 of its 8 advertised megapixels for imaging. This is an extremely low score that is inferior to any DSLR we have tested to date and much more in line with substandard point-and-shoot models..."

I would caution against interprting the "mpxls ideal" in Imatest quite so literally. I added it to Imatest at the request of friends, but I've found it creates considerable misunderstanding. It is based on the idea that an "ideal" camera would have MTF = 1 up to the Nyquist frequency and 0 thereafter. So MTF50 would equal the Nyquist frequency. Of course no real camera can have such a response. But a large values of "mpxls ideal" can be achieved with heavy sharpening (with a small radius;
--
Norman Koren

 

[Kim Letkeman]

Ok ... then what conclusion can one draw from their data, ignoring their strongly negative interpretation of these results? That is ... are the data meaningless for judging relative sharpness between cameras?

And just curious regarding an issue that was brought up in another thread ... in a few instances, they (digitcalcamerainfo.com) appear to have included more of the test pattern in the shot than they should have. Does this skew the results signifcantly? Or can imatest account for that?

 

[Unda Covalava]

BTW-
really enjoyed your writing on film vs digital, back around 2003-'04.

 

[Unda Covalava]

Most people never print big and never do a lot
of photoshopping ... for those who just print small or look on
screen, I'm sure it is enough.

"The Panasonic FZ30's images produced unusually sharp 13x19 inch prints, and would likely hold together well enough even at considerably larger print sizes, or with a fair bit of cropping."

From the review at imaging-resource:
http://www.imaging-resource.com/PRODS/FZ30/FZ30A13.HTM

 

[tbcass]

Really? Noticeably better?

I spent a few minutes looking at the images on Steve's Digicams,
which is the only place I know of that has photographs of the same
subject taken under similar lighting.

I have to say that, at 300%, the E900 images look no better than
the F10 images. In fact, the E900 and S9000 have more edge
artifacts and noise in shadows. But not much more, and the 50%
pixel count advantage no doubt makes up for that when both are
printed or downsized for display.

I agree, there isn't much dfference between the two until you magnify above the F10's native resolution where the s9000 becomes superior. I did notice one thing in the iso 1600 images. I believe the alleged superiority of the F10 at high iso's is due to higher noise reduction, not some superiority in the sensor. While a little noiser the s9000 images were much sharper. I believe the lower noise in the F10 are achieved through softering the image because the noise is there, just softened.

 

[tom sugnet]

'The Fujifilm FinePix S9000 produced very crisp-looking 13x19 inch prints on our i9900 printer. We've observed in the past that Fuji's SuperCCD sensor technology really shines in printed output, more so than when viewed on-screen, and the S9000 once again proves that out. Even 13x19 inch prints stood up very well to close inspection, appearing sharper than we'd normally expect 9-megapixel prints to look when printed at that size.

As noted above, the S9000 does fine at lower ISO settings, but its image quality degrades fairly rapidly above ISO 400. ISO 800 shots will probably be usable as 8x10 inch prints for most folks interested in displaying them on a wall or table, where they won't be scrutinized too closely'

(imaging-resource.com)

Most people never print big and never do a lot
of photoshopping ... for those who just print small or look on
screen, I'm sure it is enough.

"The Panasonic FZ30's images produced unusually sharp 13x19 inch
prints, and would likely hold together well enough even at
considerably larger print sizes, or with a fair bit of cropping."

From the review at imaging-resource:
http://www.imaging-resource.com/PRODS/FZ30/FZ30A13.HTM

--
Fujifilm Finepix S5100 + Fujifilm Finepix F10

 

[Kim Letkeman]

Most people never print big and never do a lot
of photoshopping ... for those who just print small or look on
screen, I'm sure it is enough.

"The Panasonic FZ30's images produced unusually sharp 13x19 inch
prints, and would likely hold together well enough even at
considerably larger print sizes, or with a fair bit of cropping."

From the review at imaging-resource:
http://www.imaging-resource.com/PRODS/FZ30/FZ30A13.HTM

I'm sure the FZ30 prints great at 13x19 at ISO 80. Go above that and you better be running Neat Image. But since I like Neat Image, who am I to criticize the Panny for its little problem? I shoot almost exclusively at 1600 ISO with the Fuji, so I am the perfect candidate to buy that thing anyway