Lossless and Transparency Encoding in WebP
blog.chromium.org
blog.chromium.org
Man, PNG was such a nice binary format, though. sniff
2) "If Google can encourage Firefox to start supporting it". Yeah, like they did with WebM, and then completely backstabbing Firefox by never actually removing H.264 support from Chrome, despite claiming to going to do it for many months. If the Firefox developers don't have total amnesia, I think they're going to need a lot of encouragement.
Chrome and Firefox in desktop and Opera and Android in mobile have enough combined share to shift the market on this if they want to. I'd hope and assume that Mozilla's decision making would be a bit more related to their mission to support an open web than to petty recriminations for percieved slights.
Yes, meaning dropping support for H.264. Firefox did this, Chrome did not (I'm not sure what Opera does?). What's the conclusion? Google doesn't particularly care to shift the market in favor of their own format, or thinks it will burn its other fingers too much in doing so. Meanwhile, H.264 is used everywhere, including for Firefox users, who get to enjoy it over the now-deprecated Flash.
Do you see a problem looming here for "supporting the open web"?
I'd hope and assume that Mozilla's decision making would be a bit more related to their mission to support an open web than to petty recriminations for percieved slights.
Does WebP advance an open web? Let's see, is there some actual standard describing it, or do you have to blindly lump in a big chunk of Google code to support it? (That's a real question)
(Opera, who originally proposed the video tag, with a suggestion to use Theora as the only viable video format at the time with licensing compatible with the web, didn't drop H.264 because they never supported it. I mentioned them in this context because they already support and use WebP in their bandwidth-saving page-transcoding features which are particularly useful in mobile devices on 3G networks. This solves the chicken and egg problem and gives browsers, like e.g. Amazon's Silk, a reason to support WebP even if no-one uses it yet.)
edit: and to answer your last question, WebP was up until recently a subset of WebM, which can be decoded by at least two other independent implementations I know of (FFMPEG and one in Java). Standards settings bodies have been corrupted by politics and are no match for multiple, independently produced, open source codebases.
Does anyone know a good way to provide webp with a fallback, short of sniffing the UA? I've tried using mod_negotiation (http://httpd.apache.org/docs/2.0/mod/mod_negotiation.html), but Chrome doesn't sent the webp mime type, so I can't get it to work...
Anyone know of a solution?
http://stackoverflow.com/questions/5573096/detecting-webp-su...
I've also heard from people working with web browsers that the Accept-header wouldn't work well for the real web anyway. Sigh. :(
It goes against the web design principle of making your site usable to all users regardless of what they are browsing with to use an image format that is not widely supported. This principle is backed by users who will simply leave your site in disgust if they are asked to jump through hoops to use it. Very few designers will be willing to maintain two versions of all their images just because one of the formats is a bit smaller.
Microsoft has it's own range of newish web image formats that never caught on. History suggests IE is unlikely to support anything "not made here" unless it is forced to. The competition to Microsoft's image formats, JPEG2000, never caught on either. Firefox has an infamous open bug about adding it in dating back to 2000: https://bugzilla.mozilla.org/show_bug.cgi?id=36351 - over 11 years! Despite what it looks like in the last comment, my Firefox 8 still does not have support.
The last image format that managed to make it was PNG, and that was almost over a decade ago, although alpha support in PNG was added comparatively recently to IE. Last time I checked MNG (PNG's answer to animated GIFs) still wasn't widely supported.
Sorry to be such a downer, lack of support for JPEG2000 was a let down for me. As I see it Firefox, Opera and Safari will all need to support this first, before web designers in general will even consider using it. Only once a significant number of web designers are clamouring for it will IE slowly add support for it. Only then will it see wide use.
Tool support is vital too, most importantly Adobe PhotoShop. No web designer I know of is going to muck about in the commandline with imagemagick.
The bug was WONTFIX-ed more than 2 year ago, so it will probably never support it.
As I see it Firefox, Opera and Safari will all need to support this first, before web designers in general will even consider using it.
...and this was the reason for the WONTFIX: https://bugzilla.mozilla.org/show_bug.cgi?id=36351#c98
I sense some chicken and egg problems here. I'm guessing neither side though the improvements compared to complexity of adding support are worth it. I'm not sure WebP changes the equation.
mng is currently used for simple animations in UIs (for example, Qt has built-in support for it). On the web, I guess there just isn't that much of a need for an animated format (other than fully-fledged movies).
jpeg2000 is used by some 3d tools and games as texture/asset format.
WEBP already has the advantage that it has been implemented by one of the larger browsers, and that it supports everything. Lossy, lossless, transparency, animation. And smaller files as a bonus. It could be the image format to end all image formats :-) I can see a lot of uses outside the browser as well, even if it fails as browser format. I also think google is going to push it as the defacto image format for Android.
Support in design tools isn't that important. One can always convert image formats as part of deployment. In many cases this is needed anyway (for example for CSS spriting / inlining).
This new format produces smaller files, and is therefore faster on the internet.
You can choose a benefit that you derive from that - less bandwidth cost per user, more users per bandwidth cost, faster page load times for users.
Faster page load times for users has measurable impact on user behavior - people like faster pages better.
If you don't speed your page up, and your competitor does, all else being equal, customers will prefer your competitor.
There are costs of supporting multiple formats, but if the benefits outweigh the costs, people will put in the effort. It's not a chicken and egg problem at all.
> Very few designers will be willing to maintain two versions of all their images just because one of the formats is a bit smaller.
I disagree. I think you're seeing Mobile versions of sites, already.
Also, in stating that PNG was the last successful new format, I think you're forgetting SVG, CSS, and HTML5 Canvas, all of which can be used to produce images on browsers.
> No web designer I know of is going to muck about in the commandline with imagemagick.
Not even if there's a measurable financial impact? I think you're overstating their future unwillingness. The burden, of course, is evidence.
... and of course, it's been done: http://jim.studt.net/jpeg-alpha/ :) It's even been on Hacker News, though not to much attention: http://news.ycombinator.com/item?id=1772760
Although PNG8+a compression isn't as good as lossy WebP, it is availble now, and it is compatible with all browsers.
Comparison: http://pngmini.com/vs-webp/
Source code: https://github.com/pornel/improved-pngquant
* flashes the DarkShikari signal *
I wonder if just adding LZMA or PPMd compression to PNG (instead of deflate) would have the same size improvements, with the benefit of keeping PNG's very nice binary format.