Shrinking and wrapping the copyrights to own imagery

The new web site is "responsive" or so they say. Large images are sliced and diced into smaller versions for display on a small device window, saving load time. The smarty-pants part is it figuring out your screen across the board.

Checking what was viewable and what the image tags looked like, I grabbed a bunch  / with a tip of the sombrero to online:

https://stackoverflow.com/questions/4602153/how-do-i-use-wget-to-download-all-images-into-a-single-folder-from-a-url

$ ls -l *rocky*

spot lighters   77360 Jun 20  2025 6855cb[*]_rocky horror_9235210_orig-p-500.jpg

spot lighters  211763 Jun 20  2025 6855cb[*]_rocky horror_9235210_orig.jpg

I used the A list "ico, SVG, svg, jpeg, jpg, bmp, gif, png, heic" but no SVG came through. And PDF failed in a robotic way.

The p-500 means the file got downsized by the automat to fit in a 500-pixel holey window.

$ exiftool *rocky* | egrep "^File Size|^Image Size"
File Size                       : 76 KiB
Image Size                      : 500x752

File Size                       : 207 KiB
Image Size                      : 532x800

Cut about 2/3 in size for only 500 pixels versus 532.

One has the tags I embossed, at least on the largest chunk.

  31698 Dec 21 01:00 69473f12[*]0c_D48a-p-500.jpg

  68603 Dec 21 01:00 69473f12[*]0c_D48a-p-800.jpg

  81388 Dec 21 01:00 69473f12[*]0c_D48a.jpg

PDF tag? sure, why not.

$ strings *D48* | grep -i spath

  <pdf:Author>Jim Spath</pdf:Author>

$ strings *D48*500* | grep -i spath

$ strings *D48*800* | grep -i spath

On the other site, images over 2000 pixels wide get de-tagged.

Spotlighters-Purchase-Tickets-Icon_Ticket-2.jpg

Your image was automatically resized to 2000x1085 pixels. 

Spotlighters-Logo.jpg

Your image was automatically resized to 2000x882 pixels. 

Once that happens, the tell-tale is a different processor. I'm guessing GD beneath, maybe Python middleware?

$ jhead  Spotlighters_Purchase_Tickets_Icon.jpg

File name    : Spotlighters_Purchase_Tickets_Icon.jpg

File size    : 126559 bytes

File date    : 2026:01:31 22:13:11

Resolution   : 2000 x 1085

JPEG Quality : 75

Comment      : CREATOR: gd-jpeg v1.0 (using IJG JPEG v62), default quality

Fixer upper: 

$ time exiftool   -author="Art Ist"   -copyright="CC BY-NC-SA 4.0 Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International This work is licensed under CC BY-NC-SA 4.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/4.0/" -subject="Art"    -title="For Spotlighters"  -datetime="2026-01-31 00:00:00"  Logo-2000x882.jpg

    1 image files updated

What else?

 Oh, yah. Pictures or it didn't happen.

Still looking for where I snipped off the orientation.

Note to Self, Mainly, on Shrinking Map Colors

 I am helping a theatre redo their website, from a post-Y2K relic to one that promises responsiveness (their phrase). They make a big deal about putting warning flags on images over 1MB, or a certain pixel size (like 800x1340). On the "directions and parking" section the old site had text directions, probably copied from Mapquest in 2002 or so. I kept the narrative and added static maps, since I didn't want to try a "live" google map image, figuring most folks would put the address in the own GPS/map app.

  The OSMAND app works great, so I figured I'd do screenshots and show the suggested routes, whether they lined up with the old text or not. They looked good, full screen PNG files with nice colors. Only they ended up way too big for a quick page load.

-rw-rw-rw-  1 user  users  1586661 Dec 14 22:19 map_from_westminster.png

When I saved them as JPEG, they didn't shrink much, just got a little fuzzier. Smaller size images didn't look great either. So I turned to the old standby library of graphics fixer-uppers, NetPBM [ netpbm.sourceforge.net/doc/ ]

The classic set of 216 colors (AKA "Web Safe") made sense as a target palette, as long as the altered image looked decent. I recall doing color reduction in the past and making some pictures that didn't. The docs said create a palette file, like so:

pamseq 3 5 >websafe.pam

Now, I had a 3 color band file, 5 values deep, just like Netscape likes. Applying that to a large image took almost no time:

pnmremap -map=websafe.pam map_from_westminster.pnm >websafe_west.ppm

Note this skips the PNG to PNM and PPM to PNG steps, which I added in a shell script that rotated through the 8 compass points of map/directions I wanted (N, NE, E, etc.)

#!/bin/sh

for img in \

 img1 \

 img2 \

do

  pngtopnm $img.png >$img.pnm

  pnmremap -map=websafe.pam $img.pnm | pnmtopng > ${img}_c216.png

done

Even on a slow drive, this is quick.

$ time pngtopnm map_from_westminster.png | pnmremap -map=websafe.pam - | pnmtopng >west_c26.png

pnmremap: 216 colors found in colormap

real    0m0.246s

user    0m0.260s

sys     0m0.017s

$ ls -l west_c26.png

-rw-r--r--  1 user  users  282243 Jan 12 00:01 west_c26.png

From 1.5 MB to 276 KB: good shrink!

$ perl $EXIFTOOL *west* | egrep "File Size|Image Size|Color Type"

File Size                       : 1549 KiB

Color Type                      : RGB

Image Size                      : 800x1340

File Size                       : 276 KiB

Color Type                      : Palette

Image Size                      : 800x1340

Afterward, I reapplied metadata such as Author and Copyright to the PNG file with the awesome exiftool.

Yet to do: add an overlay showing the OpenStreetMap sources used.