On 23/03/2012 3:14 PM, Dave "Crash" Dummy wrote:
Wolf K wrote:
On 23/03/2012 1:43 PM, Dave "Crash" Dummy wrote:
Peter Jason wrote:
I have Win7 SP1. I take pictures of about 4 MB with a good
quality camera. These are excellent, but I lose definition when
these are reduced to about 500KB with software such as Photoshop.
When I take small pictures with the camera the quality/definition
is very good. I don't want to take two shots of everything just
to be able to get small pictures for emailing. Why cant large
quality pictures be reduced to smaller ones - with the same
definition?

Convert the large JPG image to a lossless format, like PNG, then
reduce that image.

Won't work if by "reduce the image" you mean "change pixel
dimensions."

Depends on what you mean by "won't work." Any reduction is going to lose
detail because there are fewer pixels, but I have found that I get
better quality when I scale down a lossless format than I do when I
scale down a lossy one.

That's interesting, since AFAIK when the JPG is displayed or converted,
it's first reconstituted as a bitmap. Thus the quality of the resulting
PNG image would depend on the quality of the source JPG. AFAIK, the
scaling algorithm works on the actual JPG/PNG/etc data. If, as you say,
PNG reduced is better quality than JPG reduced, then there must be a
subtle differences in the algorithms. Or else the data structures differ
in what they conserve. Dash it all, now I'll have to do some research to
find out what's going on!

Anyhow, a JPG isn't necessarily lossy, it depends on the compression
ratio. Cameras generally use the lossless ("high quality") compression
ratio. Anyhow, all our cameras do so.

Many simple image-viewers/processors use a lower quality ratio by
default. You can find out if your does this by doing Save As on the
original image with small name change, then comparing the file sizes.
The programs I use offer setting the default compression ratio, so apart
from the effects of the processing itself, there is no change/loss in
quality.

HTH
Wolf K.