I think Mozilla is right to tread carefully and not commit to yet another format for a minor improvement. In my opinion our web clients are bloated enough as i

I think Mozilla is right to tread carefully and not commit to yet another format for a minor improvement. In my opinion our web clients are bloated enough as is, and they're only getting bigger. Once you add support for a format and it gains mainstream traction on the web, it'll be really hard to remove, and we'll be stuck with software that has to support so many things for reasons that do not make sense in the near future. As for the study, I have to admit I did not read it. But the first thing to note is that computed metrics (SSIM, PSNR) are not reliable approximations of perceived quality. Optimizing for either will produce a codec that pleases the computer, not the eye. This is why the tuning of image and video codecs for psychovisual effects is such a laborious, manual process -- you have to have a human viewer to assess the quality. In the long term I think it would be good if a new still-image format was derived from the intra portion of some reasonably advanced video codec, as they all beat JPEG handily. This is essentially what WebP is, but if I recall correctly, Google didn't really give VP8 the love it needed when they acquired On2. Instead they rushed with a "release early, release often" mentality, presumably hoping the community would jump the boat first and do the grunt work for them later (this was during html5 video wars when some thought Theora sucks and others thought h.264 is too encumbered -- VP8 was Google's resolution). The details are shady but the curious should search for the debate around early VP8 (I think the x264 folk did some extensive notes). Needless to say I think, VP8 (and WebP) wasn't and isn't the game changer Google would've hoped. To avoid patents, a reasonable choice would IMHO be to look at Xiph.Org's upcoming video codec Daala. Curiously Mozilla has the right man in house since Monty moved from Red Hat to Mozilla like, almost like yesterday... Daala however isn't quite ready yet as the Xiph guys are busy researching out new ideas and unicorns. But the results so far are very interesting to say the least; anyone interested should check out Monty's demo pages.[1] As others have said, JPEG is "good enough" for most of us, and even JPEG is used rather inefficiently, often with too high quality. So there's no compelling reason to push for a new format here-and-now. Give it some time and commit once there's something significant I say. In the future we'll thank ourselves for not having deployed all that junk over minor increments. [1] http://people.xiph.org/~xiphmont/demo/ http://people.xiph.org/~xiphmont/demo/