Visual Adjustments SVG Image Enlargement Issue
Opened this issue · 4 comments
I just noticed that when I switched from using the circle.svg to the Flower-Frame.svg since that SVG image is so big, the test to try to increase the size fails because it is getting resized automatically down to size of the screen.
We need to limit the size to 200 / 200 in order to allow the OS/Reading system to increase its size in order to satisfy the test.
It looks like the test as implemented in v 2.0 book is not working in practice.
Changing the font size does not alter the size of the new image. In the previous version of the test, the gray circle would usually change size.
The SVG image is included with:
<img src="../images/Flower-Frame.svg" alt="illustration: a black and white detailed symmetrical flower" width="150" height="150"/>
and it is rendered in most cases to the maximum width of the reading view.
However, this rises the question as to whether this test is conceptually sound. Should changing the text size also alter the size of an image?
The accessibility provision valued by users is that they can zoom in, perhaps to the extent that it is larger that the screen size, and they can pan around the magnified image. This is possible in some apps (Apple Books, Thorium Reader) but would be considered an Advanced feature.
I wonder if the image if enlarged in proportion to the text might lead to unexpected behavior. A large image to begin with might become too large for the viewport resulting for the need to pan around to see the whole image.
I think that we should be testing for the enlargement of a SVG, and as an advanced feature of enlarging other image types.
Would we want 2 images one PNG like we had and the new SVG image, but with improved CSS size that Lars is investigating?
In an advanced book we can be a little creative, with SVG, PNG with transparent background, etc. But we do seem to be gravitating towards removing the image test from the Fundamental Visual Adjustments title, until we reach a clear position on the test and the functionality is supported in multiple reading systems.