Hey Anders! Thanks for the kind words.
Because of how the backstretch element is loaded/displayed, the RICG approach won’t work for that. However, in version 2.3.0, I did adjust that script so that if the user is on a smaller screen, it will load the large version of the image instead of the backstretch, so it is taking less bandwidth. I felt that it was not practical to load even smaller images, since the next smallest default image size is 300 pixels, and most phones are wider than that. Shifting from the backstretch size image to the large image can present a significant file size difference, though, so that was definitely worth implementing.
Since RICG is being rolled into WordPress Core with 4.4, I will be actively looking into implementing it with the large (smaller/not backstretch) featured images. It may be educational for the backstretch images as well, although I’m pretty satisfied with that implementation at this time.
No problem. Thanks for the update Robin! Always a thoughtful answer.
(Do you remember what breakpoint you set for when the shift happens – on a smaller phone/using the larger image instead? )
There is no official breakpoint–it compares the screen width with what the site has set for the large image size. So the default breakpoint would be 1024 pixels, but that will change based on how the site is set up. So if the screen is 1000 pixels wide (as a for instance), you would get the large image instead of the much larger backstretch image.
If you check the source code of the page (using Dev Tools or something), you can see it by loading a page with a backstretch image on a large screen. The source file used will be the backstretch (or original) image size. If you resize the screen, the source file won’t change (this is different than RICG as I understand it), but if you reload the page at the smaller size, you’ll see a different image, the smaller one, loaded as the source. But where that change happens depends on your WordPress settings.
Gotcha! Thanks for the explanation. Very useful indeed.
