1/25    When I use my digital camera, there is an obvious difference in photo
        quality depending on the resolution chosen, as would be expected.
        However, changing the compression changes the quality very little,
        but the size in kB a lot.  In fact, a picture taken at medium
        resolution with superfine compression uses 50% more memory, but is
        clearly lower quality compared with a high resolution picture taken
        with normal compression.  So I wonder what compression other sodans
        use on their digital camera.  The imagine quality between normal
        and fine compression seems to be very little, but it's more than 2x
        more kB.
        \_ Just get a big card and take the largest photos your camera can.
           If you throw away detail in the camera, you can never get it back;
           and you won't know what photos you need the detail in until you
           look at the results.  Reducing resolution or increasing compression
           both lose details, in different ways.  512MB flash cards cost $100.
             -tom
           \_ This is a good answer.  --digital camera guy
        \_ my experience shows that the exact opposite is true:
           if space is an issue, I go lower res but keep high quality jpeg.
        \_ I use the finest JPEG quality in case I have to reprocess the
           image later.  I agree that there is diminishing returns as you go to
           higher quality, but you don't want to throw away quality you can
           never get back.  Doing image retouhing or editing on heavily
           compressed photos can be bad because of artifacts and recompression.
            \_ To clarify this: JPEG compression throws away features of the
               image you can't see.  For example, if your picture is dark, you
               won't be able to see the details clearly; JPEG recognizes that
               and throws the details away to save space.  That's fine, until
               you try to lighten the image in Photoshop so you can see it
               better.

               This is the same reason you can get scanners with 48-bit color,
               even though the human eye can't perceive more than 24-bit:
               because the editing you do on the scanned image might bring
               out things that weren't visible in the original.
               \_ more like no output devices can support more than 24-bit...
                  our eyes have incredible dynamic range (high noon to
                  starlight).  [formatted]
                  \_ I seem to recall the human eye has a dynamic range of
                     around 10^5, not all at once, as the following poster
                     noted.
                     \_ the problem is, RGB is linear, but the response of
                        our eyes is not.  Much of that 24 bits is wasted
                        in areas our eyes can't distinguish, and there
                        isn't enough concentration in the areas where our
                        eyes perform well.  -tom
                   \_ That's true: the eye can distinguish a good deal more
                     than 256 shades of grey total, by adapting to different
                     brightness conditions.  However, you can only see 40 to
                     50 shades *at a time*, so that's all you would ever need
                     in a single image.
            \_ Sorry for my ignorance, but what do you mean by reprocessing?
               \_ He probably meant post-processing like via Photoshop.
            \_ Ditto what this guy/gal is saying. I would go one step further
               and take photo's in your camera's raw setting (if it has it).
               From the raw formet you can make high-quality tiff's or any
               level of compression of jpeg you want. You will need to do
               some photoshop stuff to make them look nicer, though.
        \_ So what is the least compression, superfine?
           \_ yes.  Superfine, fine, and then normal.