Overall Image Quality
Up close, average, reduced down in size the images do look rather good. But then you'd not expect this little compact digital camera to perform like a "prosumer model". Let down a fair amount by noise from Fujifilm's "SuperCCD" and a slightly odd black-point. Colours were pretty good, not S1 Pro level good by any stretch but good enough for happy snapping (better than some of the competition). The camera really does seem to be bundle of compromises.
It's probably also worth noting that shooting at 2400 x 1800 NORMAL quality JPEG allocates only 1.5 bits per pixel, this in itself is pretty heavy JPEG compression and although JPEG artifacts are not visible it may well be amplifying the CCD noise. Shooting at 2400 x 1800 FINE produces rather large 1.6 MB images...
There was some noticeable barrel distortion from the 40i's lens. Measured at about 0.7% barrel distortion. Noticeable but not too bad, you'd only come across it in certain circumstances.
As seems to be a common trait with Fujifilm's SuperCCD we've got visible noise in shadow / dark areas of the image. The second sample here shows it most clearly, camera metered the image well attempting to balance between the dark side of the image and the bright sunlight, unfortunately that put any detail on the right in a shadow which has suffered from noise (only amplified by Fujifilm's processing techniques it would seem).
Shots here are of the PIMA/ISO 12233 standard resolution test chart (more available for comparison in our comparison database). Studio light, cameras set to auto, Exposure compensation +0.7 EV for all cameras. How to read the charts: All values are 1/100 th lines/picture height/width. So the "10" value equates to 1000.
Before you start reading these charts I had to explain that I had a complete nightmare getting 40i shot. Because it has a fixed lens at pretty wide angle I had to get really close to the res chart, this had the side effect of blurring the right hand corner (not sure why it's just the right, but there you go) and visible barrel distortion. I took the res shot about 30 times with slightly different settings, exp. compensation etc. and this was the best of the lot...
|Fujifilm Finepix 40i - 2.4 million effective pixels*|
|Fujifilm Finepix 4700Z - 2.4 million effective pixels*|
|Canon PowerShot S100 (Digital IXUS / Digital ELPH) - 1.92 million effective pixels*|
|Nikon Coolpix 990 - 3.12 million effective pixels*|
* Effective pixels being the true output of the CCD (eg. 1600 x 1200 - 1.92 million) rather than the quoted pixel count (which includes masked-out pixels used for black current calculation.
Because of the problems shooting the resolution chart I've decided not to publish resolution measurements taken from the chart, I'll let you decide that visually from the chart originals above. Sufficient to say it's almost identical to the 4700Z (which was much easier to measure because of its zoom) and somewhere less than a 3 megapixel digital camera. Not a bad performance but not stellar and certainly not the 4.3 megapixels claimed by the SuperCCD.
|Fujifilm Finepix 40i||Fujifilm Finepix 4700Z|
|Canon Powershot S100 (IXUS / ELPH)||Nikon Coolpix 990|
We're only measuring colour here. RGB values below were taken from a VGA reduced image (to average colours and eliminate JPEG artifacts) using the Eyedropper tool in Photoshop with a 5 by 5 dropper average. (Red, Green, Blue)
(IXUS / ELPH)
Relatively good job, though both the Fujifilm samples showed a blue cast under daylight (all the colour patches are shot under daylight) and reds came out a bit orange.