Go Back   Steve's Digicams Forums > Digital Cameras (Point and Shoot) > Fujifilm

Thread Tools Search this Thread
Old Apr 28, 2008, 6:54 AM   #11
Join Date: Jan 2004
Posts: 36

Hello FttSniper,

You are producing outstanding results with the S6500fd !

Surely we wereon the right track as Fuji has listened to uswith the newest S100fs. I am now reading theS100fs review, and do get a little bit disapointed by the fact that the lens of the S100fs is so-so with high chromatic aberration (purple fringing) even at moderate zooming levels. They should have kept the zoom a x10, with a quality x1.5attachement to enable a x15 total range.

Maybe such a camera will exist later on, a simplified bridge camera with the2/3 inch sensor inside, this timeproviding a quality x10 zoom, at a lesser price than the S100fs.

This will be needed in order to keep selling bridges, asDSLR bodies are nowbecomingcheap, and mostly with LiveView capabilities. What is the price of a fine x10 zoom for such a cheap DSLR with LiveView ?

How much is the S100fs cheaper ?

Is it worth tradingpicture quality when one realizes that aNikon D60 equipped with a 18-135mm zoom is advertized at 758 EUR ? But no LiveView and only x7.5 zoom. Butless chromatic aberrations (purple fringing),a massive improvement in autofocus speed, and a very significant advantage at iso800 and iso1600. I don't know about the macro possibilities of that Nikon 18-135mm zoom. I doubt they can be as good as the Fuji.

steph_tsf is offline   Reply With Quote
Old Apr 28, 2008, 12:13 PM   #12
Senior Member
Join Date: May 2005
Posts: 1,063

After browsing the review I don't think I would let CA stop me from purchasing this model. The samples in the review as well as some of the other images I have seen posted look pretty nice to me. I can compensate for the CA in Photoshop. I have a 9000 which is also supposed to be bad in this area, but I haven't encountered the problem in enough images to have it be a problem.

I don't like the battery life. That seems to be somewhat limiting. And if this camera is anything like other Fuji cameras I have, I simply don't trust the build quality.

A lot of people really seem to get concerned about CA. But I have found that, in my photos, it's only a problem in a very small number of my shots.
jphess is offline   Reply With Quote
Old May 4, 2008, 7:11 PM   #13
Join Date: Jan 2004
Posts: 36

I own a Fuji F30, a Fuji F31fd and a Fuji S6500fd andcan't complain about build quality ! Especially the S6500fd :itdelivers a taste of quality when used (manual zoomingring).Now if you look through the S6500fd electronic viewfinder, that is another story : it looks cheap.Fuji mustequip their expensive bridge cameras with the best available electronic viewfinders, otherwise theywill feel "cheap" and maybe "ridiculous" when compared to what you get using a DSLR, especially with the DSLR prices now falling.

Comparing the price of the S100fs with the price of the Sanyo HD1000 makes me worrying too. The Sanyo HD1000is a full HD recorder with a x10 quality zoom opening at 1:1.8 ! Image quality in daylight (movies) is amazing ... and this is a native 16:9 format. Put it on your 32 inch or 50 inch LCD, you get a nice image (well, it is a movie) and decent sound.

I think the future is a Fuji S100fs, but equipped with a 2/3inch 16:9 8Mpix tri-CMOS system (like in high-end camcorders), hence each CMOS sensor outputting3840x2160pixels at 8fps. This is quad-HD at the sensor level. So when combining the output of the red, green, blue sensors, you get 8Mpix of "real" pixels at 8fps for JPEG stills and JPEG continuous shooting. You are no more guessing and subsampling the colour using thelow pass filter and theBayer pattern.Using a 2/3 inch 16:9 8Mpix tri-CMOs system,you get8Mpix images that you can upscale to 12Mpix, getting still images having bettersharpness than asingle RGB-Bayer 12MpixCCD, but with a significant advantage in sensitivity, because of the3 sensors.

So, a still camera deliveringexcellent12Mpix JPEGs (say, 5 MByteeach) at max 8 fps so a sustained rate of 40 Mbyte/s. Need a buffer about 128Mbyte, that is all. Flusing the buffer to the SD card in about 8secondswhen using a 20Mbyte/s card. Butnot blocking the machine (multitasking is a reality nowadays),so you can start shooting againas soon as the buffer has room left for a new 5 Mbyte JPEG.

Now, let us design the video department.Mode"1080p30fps"would read the sensor at 30 fps, butin a special readout mode binning 2x2pixels together. Having no RGB-Bayer filter, you can do that. So you trade resolution in favour of sensitivity, as each captured pixel has now 4 times more surface.You get full HD 1920x1080 images. Non-interlaced. At 30 fps. So thelong awaited 1080p format. And you have 3 sensors. So youcan expect the quality to be real HD and the sensitivityto be in the order of 1 lux. Mode "720p60fps" would read the sensor using a 3x3 binning scheme (getting 1280x720 pixels : do the maths yourself from the 3840x2160 pixels), but this time at60Hz. So a native 720p format at 60fps for action and sports, with an even better sensitivity as each captured pixel surface is now 9 instead of 4.

Knowing this, I feel we are wastingboth time and money using all the currently marketed gadgets.All what is currently available is ill-designed stuff, compared to what just got described here.

The real answer to photography is a 2/3inch 16:9 8Mpix tri-CMOS system outputting3840x2160pixels ! And this is also the real answer to HD video, using a 2X2 binning and 3x3binning in the readout !

All the rest isfutile.

What is not going to be futile, for DSLR makers, is when machines built like this, considered as "video bridges", will provide a DSLR mode : back to the pleasure of using aTTL viewfinder ! It would be disabled for video, of course, but it would appear asan (expensive) option for still capture.

As a conclusion, we may say that past 10 years of digital consumer Imaging Technique (1998-2008) only has revealedthe emerged part of theiceberg.The next ten years will see a concentration, withall quality devices sharing thesame acquisition modules (2/3inch 16:9 8Mpix tri-CMOS system outputting3840x2160pixels) and being NATIVELY used both for stills and video.The current single-sensor RGB-Bayer systems will be considered as low-cost toys, because of "faking" pixels, and they will be dumped by all serious brands.Progresses in automated manufacturing associated to very high volume production will enable tri-CMOS devices at a decent cost.Consumer-grade Imaging Technique is going to surprise us ... a lot !

What can be expected too, is that someconsumer league will force the equipment manufacturers to quote the real pixel counts in their product announcements and specifications.An example : in aFull HD LCD screen outputting 1920x1080 "true" pixels,each pixel is consisting on a RGB triplet, so it is valid to say that the image has a resolution of2 Megapixels in colour. But in adigital camera using a1920x1080 pixel CCD sensorcovered bya lowpass filter and a mosaic RGB-Bayer filter, is it honest to advertize it as "2 Megapixel" ?Not at all ! If one is calling the digital camera a "2 Megapixel" device, then one should call the Full HD screen a "6 Megapixel" device !Or,preferably, if one is calling the Full HD screen a "2 Megapixel" device, the RGB-Bayerdigicam using a single 2 Megapixel sensorsould be called a 675K pixels device !

This approach is necessary forboosting tri-CMOS sales. Abudget camera using software antishake (instead of optical antishake) would use atiny 1/6inch 16:9 2Mpix tri-CMOS system outputting1920x1080pixels. The output format would be1980x1080 pixelswith anti-shake being off,and 1280x720pixelswith anti-shake being on. The advantage of the 1920x1080 pixels setting would also be a "wide angle"capture with an equivalent focal of 28mm insteadof43mm in the 1280x720 pixels setting.Thiswould beadvertizedas "6 Megapixel". And this would reuse the vast army of low-cost big zooms one can find in current entry-level camcorders, as they are all equipped with tiny 1/6 inch sensors today. And yes, such entry-level camcorderswould deliver decent upscaled JPEG stills of 6 Megapixels, with busrt shooting up to 8fps. With 1080p30pfs and 720p60fps video. All this for 299 dollars !That is the evolutionwe are going to see betwen 2010 and 2020.

Then, horizon 2020,we'll get the first 2MegapixelParametric Maser sensors, each individual pixel relying onMaser-amplification and Parametric-amplification (nano machining, transverse electric field, bias excitation, non-linear coatings,quantic pumps, embeddedper pixel Peltier element),hence shutter speeds always faster than 1/250th of a second even at 1 lux ! One million ASA equivalent, as standard for video.With such a high sensitivity, no need to have three sensors. One can use only one sensor,sequentially detuningthe wavelenght sensitivity peak, so for each aquired pixel, there may be sequentially a lot more sampling points than just red, green, and blue.

Such sensors are genuine light traps,appearing as darker than totalblack, so no diffraction nor optical aberrations.
steph_tsf is offline   Reply With Quote

Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off

All times are GMT -5. The time now is 3:55 PM.