Hi,
I used Lanczos-3 filter for resampling in a program called Thumbnailer. After reading your post I resampled the image in Photoshop Elements using bicubic and there result turned out to be much better. I'm not going to use Thumbnailer for enlarging pictures anymore.
I have to agree with you. The differences are not that distinct between the native 6 mp and 3 to 6 mp interpolated image. At least I would not care for the difference and waste storage space on the native 6 mp. However, for large prints the native 6 mp is probably preferable.
I used Lanczos-3 filter for resampling in a program called Thumbnailer. After reading your post I resampled the image in Photoshop Elements using bicubic and there result turned out to be much better. I'm not going to use Thumbnailer for enlarging pictures anymore.
I have to agree with you. The differences are not that distinct between the native 6 mp and 3 to 6 mp interpolated image. At least I would not care for the difference and waste storage space on the native 6 mp. However, for large prints the native 6 mp is probably preferable.
--Thomas http://www.pbase.com/thla/finepix2800zThomas,
Thank you for the info. Just out of interest, I also downloaded the
same images from the same site. I resampled the 3 mp file to 6mp in
PS6 using Bicubic. The differences between the resampled 3 mp and
the native 6 mp are not as distinct as the images you posted (How
did you do yours?). I am more incline to say that they are very
similar. (Sorry, I do not know how to embed images stored on my
hard drive. Cut & Paste didn't work.) When both files are resampled
to 12 mp for further comparison, then I can tell that the native 6
mp file is better. However, I cannot say that the difference is
enough to justify the claim by Fuji. One other consideration is
that the resampling was done from a compressed JPG file. The
differences could be even less if I had an uncompressed file for
resampling. I would think this is the case, because the native 6 mp
file was "resampled" in the camera using original data. Further
resampling is unfair to the 3 mp file.
In any case, I still think the Super CCD is a very interesting
design. The 3 mp image is in the top 10% of the 3 mp prosumer
digicam league and the 6 mp file is the resampling result of a very
clever hardware/firmware combination. I look forward to the 602.
Shii
You can find some comparison shots between 3 and 6 mp on:
http://myalbum.ne.jp/cgi-bin/a_menu?id=fa782151 (very slow site!!).
The shots are beta samples from the F601 with the new super CCD. I
know that the optics are better on the 6900, but the difference
between 3 and 6 mp is evident. Below are two cropped samples, that
I also posted last week, of a 3 mp picture interpolated to 6 mp
compared with a native 6 mp.--I sneered at the Fuji marketing hype before but now, after
following numerous discussions on this site, I am looking at the
Super CCD with a lot of interest. Not that I am accepting the
argument that one could generate 6 mp worth of information from a 3
mp sensor, but I am interested in the argument that the Super CCD
can get more out of a 3 mp sensor than the conventional CCD can. Of
course, I understand that the hexagonal layout allows bigger sensor
area and hence better dynamic range and lower noise, etc. I am
looking at other possible benefits of a hexagonal layout. At the
risk of offending some Fuji fans here, I wonder if I could look at
the hexagonal layout, inconjunction with the proprietary software,
as a clever way of doing image resampling. In other words, this is
a hardware/software combination to image resampling rather than
just a software approach such as Bicubic and GF. I further
understand that the 6900 can generate both 3 mp and 6 mp files.
Could somebody post, or point me to a link, 3 mp and 6 mp files of
the same scene so I can "play" with them?
Thanks,
Shii
Thomas
http://www.pbase.com/thla/finepix2800z