Pse explain "3.1m effective pixels and 6.0m redorded pixels"

I'm considering buying the Fujifilm finepix S5000. One of the advertising
features
is "3.1m effective and 6.0 m recorded pixels:" I see in the review that you
can choose
6m 3m 2m or 1m. Can someone explain?

 

3M, 2M and 1M, the camera will record the photo "as is".
If you take a photo at 6M, the camera enlarges the photo, in the same way
you could do using Paint Shop pro or Photoshop etc.
The photo is actually 3.1 MP, but the camera does the work instead of you
manually enlarging each photo on the PC.
You'd have to ask one of the experts on this group if the quality is the
same as a "true" 6MP, but i would guess that it isn't, as a camera cannot
take photos with more resolution than it's CCD is designed to do. Anything
bigger would have to be interpolated

HTH
maze

 

It's nothing but deceptive marketing if not false advertising, IMHO. They
capture information from 3 million sensors and interpolate it to 6 million.

 

After buying the Fuji F700, I don't believe anything Fuji says anymore.
http://www.photoprojects.net/fujif700.html
Gene

 

3mp is the capacity of the sensor on the camera. It uses software to
interpolate to give you 6mp.

Mark

 

Fuji has their own chipl it uses hexagonal pixels (like a honeycomb)
instead of traditional rectangular pixels (like a checkerboard)

It can take pictures with a pixel counts of
2816 x 2120
2048 x 1536
1600 x 1200
1280 x 960
see http://www.dpreview.com/reviews/fujifilms5000/ (review)
and http://www.fujigroup.net/ (pics and users and user suport)
and http://www.steves-digicams.com/2003_reviews/fuji_s5000_pg5.html
(review)

The top pixel count is equivalent to what a 6mp camera takes.
I have a S602z, and it has a similar chip.
see
http://www.fujifilm.com/JSP/fuji/epartners/bin/3rdGenSuperCCD_2.pdf
for more info than you probably need.

also, see
http://www.dcresource.com/reviews/fuji/finepix_s602-review/index.html
While this is a review of the 602, there may be one of the S5000. More
importantly, this site has full untouched photos of the same place,
same time, at EVERY resolution and EVERY compression the camera can
take. YOU can dl and compare the 6mp, 3mp and even the 17Mb 6mp TIF
file.
http://www.dcresource.com/reviews/fuji/finepix_s602-review/preprod/gallery.html
sample images for above camera / website


Bascially, with the sensors at an angle instead of a strait grid, fuji
has made it possible to 'upsample' a picture with no noticeable image
degradation. I know there are folks here that will disagree with that.
however, if you are a pro or other nit picker, just get out your
$2,000 and buy a dSLR. Opinions are a penny a 100 on the net. the only
one that counts is YOURS.

BTW, the S602 is a step up from the 5000 (7000 replaces the 602). You
shold be ablt to find the 602 for about $400 new online, perhaps less.

chris

 

That's the myth Fuji would like you to believe. The hexagonal arangement may
have advantages but it doesn't increase the number of sensors one iota.
Certainly doesn't double the information collected. If you want to buy into
the myth then do so, but it's a myth with maybe a trace of substance at
most.

 

Its that "Trace of substance" that is important thats
important in the Fuji.

When using the 6mp mode in my Fuji S5000, the photos are
Marginally better than the 3mp mode. Marginally beter is
better.

When I bought my S5000, I read the packaging and display
materials BEFORE counting out the cash. I knew it was a
3mp camera. I bought it because I liked my old Fuji
Finepix, and hoped the S5000 would work as well. It did.

After a week of shooting different subjects in different
modes to see what the camera could do, I decised to see
what the 6mp mode was all about. What it was about was
pictures that were higher in pixel count and looked
"marginally" better on the screen and on paper. I havent
set the camera back to 3mp mode, because marginal
improvements are improvements none the less.

Its like a pig that talks, what it says is far less
important than the fact that it talks at all!

 

In message <gH2Hb.6298$9.net>,

No, only the 6MP will contain everything captured by the camera.

Not at all. The pixels in the camera are laid out so that they fill a
6MP grid, with half the pixels, like a checkerboard pattern. You can
not shift every other row or column, to make a solid grid, because that
would blur the image.

The 6MP output is the only one that captures everything the sensor does.
The 3MP output is the same as the number of sensor pixels, but they are
downsized from the 6MP image, and lose some detail. The 2MP and 1MP
images lose even more detail.

Quality is generally as good or better than a regular 3MP in a full
grid, but not as good as a real 6MP sensor.

--

 

In message <[email protected]_s52>,

Advertising and design issues aside, the camera can't deliver all the
detail it captures at anything less than the 2x-bloated resolution.
--

 

The pixels are octagonal in shape, not hexagonal. They are in a square
grid rotated 45 degrees, not a hexagonal grid. As someone else said,
the arrangement is the same as the set of black squares in a
checkerboard. It's done because rotating the grid like that increases
horizontal and vertical resolution (but decreases diagonal resolution).

If the camera could output into a file format that had
checkerboard-arranged pixels, it would need to write only 3 million
pixels but you'd need special software to read and edit it. Instead,
the camera uses interpolation to fill in the white squares of the
checkerboard with estimated pixels, giving a full square raster that can
be stored in a conventional JPEG file and viewed and edited by standard
software.

If you select 3 MP output, the image has to be downsized, and you lose
all the extra horizontal and vertical resolution.

Dave

 

So then, a 3Mp SuperCCD should be used in the 6mp mode?
There is no winning here...6mp adds info, 3mp tosses it out.

I once read that 35mm film is equivalent to 20mp. So I figure if that
is true, until there is a 20mp camera, all digicams are only
approximating what 35mm can capture.

but then, do you see what i see? How can we possibly know?
We can calibrate our monitors, get 100mp cameras one day, but our eyes
are our eyes, and changing every day due to age.

Is SuperCCD better? Worse? Bayer vs Foveon? For all we know a new
technology will be out next year that renders all our stuff obsolete.

Bottom line? I like my camera. I like the pictures it takes. I can not
decide between 3 and 6 mp. I toss a coin weekly. When i print them at
4x6 or so I cannot tell teh difference. And on my computer screen or
the web, my old 1mp camera looks the same.

Will I make 8x10 prints? Sure, a few for fun maybe. Not nearly enough
that what camera i have will make a darn bit of difference. In the
interim, i will be having a ball!

chris

 

It all depends on what you mean by "should". To get the maximum detail
that the sensor provides, you must select the 6 MP mode.

That 20 MP figure depends a great deal on the lens, film, and scanning
if digital is your final output, or on lens, film, and enlarger lens if
wet print is your final output. I think it's safe to say that most 35
mm shots have nowhere near 20 MP of useful resolution, but that a
knowledgeable photographer with good equipment and favourable shooting
conditions can get there.

The limits of resolution of the human visual system are well known.
There's no point in going beyond that - but it's a high bar indeed.
For example, the resolution limit is about 60 cycles per degree at high
contrast, and realistically it takes 3 pixels per cycle to resolve
detail without artifacts. So 180 pixels per degree is enough resolution
to match reality. But one degree at 12 inches is 0.21 inch, so 180
pixels per degree is about 860 pixels per inch. To fill an 8x10 print
at that pixel density takes 59 megapixels. A larger print at the same
viewing distance takes proportionally more pixels.

In practice, we settle for 200-400 PPI and call that good, even though
it's not as good as reality.

At 4x6, 3 MP ought to do well. It's at 8x10 where you might be able to
see the difference.

Dave

 

I've also read that the limiting factor to digital files capturing that much
info will be the glass. Not to mention it would be overkill even for 11x14,
other than allowing more cropping room.

Of course. 3mp is plenty for 4x6. I even have an 8x10 print from an Epson
2.1 mp cam that looks pretty darn good. Up to 8x10, it's hard to tell the
difference between a file from my D30 and 10D. I just have a little more
cropping room with the 10D.

Mark

 

In message <>,

Most of the most recommended lenses are not bottlenecking the
resolution, especially when stopped down; the sensor is the limit in
most cases. The differences you do see in the very best lenses are more
contrast in the images, especially at the highest frequencies that are
attenuated by an AA filter. In order for an AA filter to cut off at the
nyquist, it must also attenuate frequencies close to it. When images
are taken with a sharp, high-resolution lens, the attenuation still
leaves the high frequencies at levels well enough above the noise floor
that you can boost them back up without artifact with USM.

For example, with that wheel test (the 60 sinusoidal spokes), I get
about 44 pixels as the resolution limit with almost any lens I use. An
exception would be the Canon 75-300 IS at max aperture; that was
somewhere around 50, IIRC. My 300mm f4L IS scored 44, and still scored
44 when I put on either the Tamron SP 2x TC, or the Canon 1.4x TC. With
both, it rises to 52. The main difference between the three combos that
score 44 is in the contrast.
--