W3C home > Mailing lists > Public > www-style@w3.org > January 2005

Re: Intrinsic size of jpeg images

From: <leslie.brown@evidian.com>
Date: Fri, 28 Jan 2005 13:27:14 +0100
To: www-style@w3.org
Message-ID: <OF19993631.8B585D0A-ONC1256F97.0041BE92@evcl.evidian.com>

On Fri, 28 Jan 2005, Ian Hickson wrote:
> It's probably because most JPEG images have incorrect resolution data,
> so actually doing this correctly (i.e. taking the real dimensions of the
> image based on the image's theoretical resolution) would break many pages
> on the Web.

I think the word "correctly" is too kind...

To me, most people seem to reason backwards on this issue. Once the picture
is in a jpeg (or whatever) file instead of hanging on someone's wall, the
"real dimensions" of the image are, for example, 600 by 900 pixels.

If a dpi figure is provided, it can be interpreted as indicating desired
or optimal resolution rendering. But the user agent should be able to
whether this is the most appropriate action.

If someone sends me a 600x900 image with "resolution 300dpi" then I hope my
graphics print program is going to print it at 2x3 inches unless I tell it
something different.

On the other hand, I'd expect a web browser to ignore the "300dpi" and
render the image at 600x900 pixels in the absence of any width and height
definitions in CSS or HTML.

To take the idea to its logical extreme, if the "correct" approach is
to treat the dpi resolution as gospel,  then videoprojectors would have
to incorporate a rangefinder so that they could measure the distance
to the screen, calculate the total image size, and tell the PC driving
them to scale the jpeg image accordingly...

Les Brown
Received on Friday, 28 January 2005 14:46:15 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:16 UTC