Re: resize/zoom ?

On Thu, 3 Dec 2020, Pro Turm wrote:

>> density (which becomes image->x_resolution and image->y_resolution after
> being parsed) is used to set the initial affine scaling values around line
> 3885 of svg.c.
> 1) In case of zoom-in, would this preserve the quality of the image ?
> 2) My problem with the density approach on the whole image is that it would
> blow up at some point (the limit of memory e.g.) and inefficiency (cpu
> time). Suppose I would like to zoom in 20x, would I need to create a whole
> 20x dense image (and why not only part of it)?

That is why you would want to do everything in SVG space.  Wrap SVG in 
SVG.  It would still be wasteful since all of the drawing requests 
would be considered but only the "visible" part would end up being 
rendered (the rest would be clipped).  With any such approach, only so 
much can be done at once so it seems like intelligent pre-processing 
(e.g. only include the parts likely to be included in the view) is 
likely required.

When doing vector rendering, rendering to a different scale does 
produce the normally desired level of refinement.  Things like "blocky 
artifacts" do not appear.

>> I am not aware of a way to easily accomplish this other than to render
> the whole SVG to a raster image and then crop the portion you want from
> it.
> Here arises my problem with the 20x zoom-in.

There are obviously many choices to be made and lots of 
testing/evaluation to do.  Only you know your requirements and the 
possible varations you are facing.  The final result is ultimately 
your own responsibility.  If you are not happy, please let me know 
where I may send your refund. ;-)

If worst comes to worst you might need to turn to commercially 
available solutions costing at least tens of thousands of dollars.

>> Otherwise, you would need to be able to modify the SVG itself so it only
> renders the desired portion.
> I would like to have only one SVG on which to operate. Changing it would
> mean parsing, adding, replacing IO operations. Which is not completely
> magick... Does no GIS have ever used GM/IM yet? GISs have tiles.

I don't have any personal experience with GIS.  I do know that with 
many systems, there are multiple pre-rendered resolutions and the 
tiles are stored in a compressed raster format.  Tiles could be 
individual files, or from multi-resolution tiled TIFF, or from 
proprietary file formats.  Obviously systems like Google Earth are 
doing better than that.

Since you mention GIS, I assume that you have taken a look at GDAL 
(https://gdal.org/) which is a library for raster and vector 
geospatial data formats.

Bob
-- 
Bob Friesenhahn
[email protected], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/
Public Key,     http://www.simplesystems.org/users/bfriesen/public-key.txt