Backstory here, gleaned from eavesdropping on Mozilla employees and Gecko engineers: Google and Facebook are leaning on Mozilla to implement WebP in Firefox, while Mozilla's been resisting due to (in their eyes) the lack of measurable improvement over JPEG. Furthermore, as joosters alludes to elsewhere in this thread, most JPEG libraries set the default compression level way too conservatively, and most users of the libraries don't bother to ever deviate from that default. Merely increasing the compression level from the default can yield a 10% decrease in image size without degrading quality whatsoever as measured by the standard metrics mentioned in the OP.
But there's the rub: the standard metrics kinda suck! Nobody's yet come up with a really solid algorithm to compare the quality of different image formats and compression ratios. So all this research we're doing may very well be useless and misguided. :P
(I'd love it if someone who actually knows what they're talking about can pitch in. Corrections welcomed!)
I personally keep expecting to see a schism develop between Mozilla and Google, considering they compete on two fronts: web browsing and mobile OS. Hopefully, nothing more than just healthy competition will happen.
Backstory here, gleaned from eavesdropping on Mozilla employees and Gecko engineers: Google and Facebook are leaning on Mozilla to implement WebP in Firefox, while Mozilla's been resisting due to (in their eyes) the lack of measurable improvement over JPEG. Furthermore, as joosters alludes to elsewhere in this thread, most JPEG libraries set the default compression level way too conservatively, and most users of the libraries don't bother to ever deviate from that default. Merely increasing the compression level from the default can yield a 10% decrease in image size without degrading quality whatsoever as measured by the standard metrics mentioned in the OP.
But there's the rub: the standard metrics kinda suck! Nobody's yet come up with a really solid algorithm to compare the quality of different image formats and compression ratios. So all this research we're doing may very well be useless and misguided. :P
(I'd love it if someone who actually knows what they're talking about can pitch in. Corrections welcomed!)