[Prev][Next][Index][Thread]
re: (meteorobs) NAMN LM Charts
Mark > Those are the charts I thought of first too. I am in a similar position
Mark > without a scanner. I would expect the size of the images to be a possible
Mark > source of problems, but am unsure if there is a way to "compress" them to
Mark > make them load faster.
Sirko> as these are only b/w charts, they can be scanned & stored b/w. This
Sirko> ensures that the files will be quite small, even if the image size is
Sirko> relatively large.
There are large uniform areas so the compression can be significant
(e.g. run-length encoding). Lew and I have done some experiments with
the IMO telescopic charts to make them accessible from the Web. The
saving going from PostScript to GIF was around 98 percent. The stars
aren't crisp circles any more, but the charts were useable. I'm going
to have another go, but at a higher resolution to determine where the
best trade-off between the star images and file size lies. What would
be a reasonable upper limit for the file size (in kB) of a telescopic
chart? There's no point in making the charts available at a size few
can download.
It's amazing how much compression you can achieve. At a recent
astronomical-software conference Rick White from Space Telescope
Science Institute showed images of five people's heads at different
information levels. Even below 0.2 bits per pixel I was able to
recognise three faces. By about 0.3 all the identities were obvious.
By around 1-bit per pixel, the images looked quite reasonable.
Malcolm
References: