Were they acting responsibly when they implemented Theora into Firefox even though it has inferior quality too? Are we now saddled with a new video format spread to an entire ecosystem and websites required to support it? Are users pissed off that they "Save As" Theora files but can't use them in other apps?
WebP isn't just about efficiency over JPEG, the biggest benefit comes from replacing PNG and GIF, which people use for alpha transparency and animation respectively. These are practical, immediate problems that today's web developers face, can they wait a few more years until Xiph approved codec does better?
The issue of widespread usage of alpha transparency and animated images do apply however. Rather than hold up WebP over a future imagined codec that doesn't exist pending research, a never-going-to-implement-anyway HEVC, or a doesnt-have-all-the-features JPEG, why not work with the WebP folks to fix any deficiencies?
It's highly reminiscent of fight over the ACA in Congress. "Repeal it". Rather than work with making the bill better or fixing the problems, just filibuster.
> Rather than hold up WebP over a future imagined codec that doesn't exist pending research, a never-going-to-implement-anyway HEVC, or a doesnt-have-all-the-features JPEG, why not work with the WebP folks to fix any deficiencies?
The bitstream format of VP8 (which WebP is a derivative of) is frozen.
> It's highly reminiscent of fight over the ACA in Congress. "Repeal it". Rather than work with making the bill better or fixing the problems, just filibuster.
Neither of those mean Mozilla can't work with Google together on a next-gen spec. Simply saying the current status quo is ok (jpeg + gif/png) is a disservice. Pushing JPEG XR isn't a solution either, it doesn't support animation for example.
What I'm saying is, rather than say "no", be constructive. I see Mozilla saying "No", but I don't see them actually proposing anything that solves the other problems WebP is trying to solve. No studies have been presented on lossless, transparency, or animated encoding efficiency for example.
WebP offers huge immediate savings on PNG/GIF use cases today, and this is especially important on mobile. Developers need these options now, not 2 years from now.
> I see Mozilla saying "No", but I don't see them actually proposing anything that solves the other problems WebP is trying to solve. No studies have been presented on lossless, transparency, or animated encoding efficiency for example.
Mozilla supports (and developed) APNG, which has existed for a lot longer than WebP. Google has thus far not implemented APNG (bug here, open since 2008 [1]). You might want to be careful about casually throwing around words like "NIH" when it comes to animated image formats…
The fact is that both browsers are perfectly free to not implement standards that they think are harmful, without having to come up with some better solution right away. Whenever we implement something on the Web, if it gets traction we are stuck with it forever. So we should move carefully.
Don't you think we should be making sure standards are sizable improvements over the status quo instead of just accepting every marginal improvement? If WebP is indeed worse than JPEG, then authors have to choose between better quality (JPEG) and alpha channels/transparency (WebP). That's not a very good decision to force them to make.
Is WebP really "worse" than JPEG, or just about the same, or not a substantial improvement? The real question is, what if you don't compare WebP to JPEG, but instead, compare it to APNG, then how does the calculus come out?
If I prepare two animated/transparent images of 100 frames today, what is the filesize of APNG vs WebP?
To me that's the real question. My social network streams are filled full of animated GIFs now and they are huge. The G+ mobile client uses WebP for these and the savings are substantial.
> Is WebP really "worse" than JPEG, or just about the same, or not a substantial improvement?
I don't know. Those are subjective terms. You can draw your own conclusions from the study…
> The real question is, what if you don't compare WebP to JPEG, but instead, compare it to APNG, then how does the calculus come out?
I don't know off the top of my head. I agree that this should be studied.
What I'm quibbling with is the notion that we should just rush to adopt large proposals like WebP if they're a marginal improvement in any area. Like I said, on the Web we are stuck with mistakes forever if they gain traction. Maybe the story would have been different if WebP had been just the WebP-lossless-animated portion. (I don't call the shots, except in Servo, but I would personally have no issues with adopting WebP-lossless-animated.) But WebP is tied together with the lossy compression that seems to have issues relative to JPEG. "Two steps forward, one step back" is less attractive than "two steps forward".
Google says WebP versus PNG is 28%, so I would expect the results for WebP-animated versus APNG to be similar. IMHO, a 28% improvement is not at the level of "so good we must rush to standardize this now, warts and all".
Cutting bandwidth usage on mobile clients by 28% seems like a pretty good win to me. If the gains were 5%, I agree it would be iffy, but ~30% is nothing to shake a stick at.
This whole issue seems sort of like SPDY. HTTP-NG went on for a more than a decade and produced nothing. Google rolled out SPDY, and suddenly we have an HTTP 2.0 spec being discussed. Sometimes the perfect is the enemy of the good, and vendors rolling out implementations forces the community to act and put resources on an issue.
Even if WebP doesn't end up as the ultimate new replacement for JPEG/PNG, the fact that Google, Facebook, and others are rolling it out will likely force some movement to get serious about finding a replacement. Ideally, one doesn't roll out an implementation until there is a consensus spec agreement, but the real world usually doesn't match up the the ideal.
Look at how asm.js's rollout forced V8 to adapt. Even if V8 isn't adopting asm.js, Mozilla shipping it forced the V8 team to spend more time optimizing those paths.
I don't particularly care if WebP is the final image standard or not. What I don't want to see is Mozilla sitting on their hands waiting for a solution to fall into their lap, and putting all their resourcesin Daala.
I expect standards working groups to actually have working members, actively engaged in putting forth positive proposals, not people who just sit around and raise objections. That was one of the problems ARB had for years. If that doesn't happen, then you end up getting steamed rolled by proprietary proposals (e.g. Cg, CUDA, et al)
WebP isn't just about efficiency over JPEG, the biggest benefit comes from replacing PNG and GIF, which people use for alpha transparency and animation respectively. These are practical, immediate problems that today's web developers face, can they wait a few more years until Xiph approved codec does better?