Expensive resizing to multiple output files

[yosh]

Hello, I'm looking for less expensive way to resize one image to multiple output files.

Let's say I have one input JPEG image (i.e. 3500 x 2500 px, around 3MB). Is there a way to read this file once and resize it to multiple output formats, without reading the source file again? Currently I'm doing something like this:

Code: Select all

my $img = Image::Magick->new;
foreach my $outputfile(@outfiles) {
  $img->Read($inputfile);
  $img->Scale(geometry => "${scale}x${scale}");
  $img->Write($outputfile);
  @$img = ();
}
undef $img;

The problem is with Read() which takes around 0.5 second (which is fast, but with 10 output formats reading alone takes 5 seconds).

If I put Read() outside loop then I do operations on the same picture. It's fast, but each scaling is using already scaled image. It causes problems when 1 format is very small and the next one is larger. I believe there has to be a way to always reach for source image without using the expensive Read() every time.

Thanks in advance for any hints or suggestions.

Yosh

[yosh]

Ok, actually just after I wrote this I've found the Clone() method and changed the scaling to:

Code: Select all

my $img = Image::Magick->new;
$img->Read($inputfile);
foreach my $outputfile(@outfiles) {
  my $temp = $img->Clone();
  $temp->Scale(geometry => "${scale}x${scale}");
  $temp->Write($outputfile);
  undef $temp;
}
undef $img;

I'll leave this here in case it helps someone. Perhaps there is a better way?

[snibgo]

You found the answer before I had time to post it. It is the obvious solution.

======

The answer is to read once, before the loop. Inside the loop, clone the image and operate on that.

For cloning in perl, see http://www.imagemagick.org/script/perl- ... cellaneous

[100find]

Several years ago, I met the same problem, and my solution is reading the image file into Mem and use Blob model to load the image from mem several times .
The Clone method is very smart.