Google drops another turd in the punchbowl

You may have heard that Google went and invented a new still-image format, because the zillion we already have apparently aren't good enough. It's a disaster and Mozilla has rejected it, but they're putting it in Chrome anyway.

Oh well, despite that, I'm sure it will be every bit as successful as VP8, Orkut, Wave and Buzz were. (And Ogg, though we can't pin that one on them.)

Jeff Muizelaar:

WebP also comes across as half-baked. Currently, it only supports a subset of the features that JPEG has. It lacks support for any color representation other than 4:2:0 YCrCb. JPEG supports 4:4:4 as well as other color representations like CMYK. WebP also seems to lack support for EXIF data and ICC color profiles, both of which have be come quite important for photography. Further, it has yet to include any features missing from JPEG like alpha channel support. [...]

Every image format that becomes "part of the Web platform" exacts a cost for all time: all clients have to support that format forever, and there's also a cost for authors having to choose which format is best for them. [...]

Where does that leave us? WebP gives a subset of JPEG's functionality with more modern compression techniques and no additional IP risk to those already shipping WebM. I'm really not sure it's worth adding a new image format for that. Even if WebP was a clear winner in compression, large image hosts don't seem to care that much about image size. Flickr compresses their images at libjpeg quality of 96 and Facebook at 85: both quite a bit higher than the recommended 75 for "very good quality". Neither of them optimize the huffman tables, which gives a lossless 4--7% improvement in size. Further, switching to progressive JPEG gives an even larger improvement of 8--20%.

Tags: , ,

47 Responses:

  1. Sean B says:

    As suggested at your link, it's probably not really any better than JPG for compression ratios people care about in the first place. So a total waste of time in every direction.

  2. Google: 15,000 engineers and not a single manager capable of saying "no."

    • Elusis says:

      Well, when you're allowed to use 10% of your time on any project you like, inevitably there is a certain amount of, um, "uncurated results."

    • moocow says:

      fucking engineers. They need adult supervision, amirite?

  3. gfish says:

    Actually, Orkut is quite popular -- just not in the US.

    This image format thing is really dumb, though.

  4. Anonymous says:

    WebP certainly seems unnecessary; making images marginally smaller doesn't seem worth introducing a new image format when we already have several good image formats with no restrictions on their use. About the only interesting feature it has: transparency, which other lossy formats like JPEG don't seem to have.

    WebM (VP8) on the other hand, seems incredibly useful since we don't have any other video formats with the same features and no legal restrictions.

    • Anonymous2 says:

      Even with WebM, we don't have any other video formats with the same features and no legal restrictions.

    • Yuhong Bao says:

      Yea, jwz comparing webp with webm and wave is IMO ridicolous.

    • Edouard says:

      > WebM (VP8) on the other hand, seems incredibly useful

      That's what they told me about Ogg Theora, right up until the point where when they told me it was no good and WebM was great.

      It's like someone telling you the world is going to end on a particular day, and then after it doesn't, telling you it's going on another day. At some point you start to lose your confidence in them.

      • Anonymous says:

        Ogg Theora solved the "free format" problem, but didn't actually keep up with any modern video compression techniques. People used it because it worked passably, but it didn't really have anything going for it except for the lack of legal restrictions on it.

        WebM, on the other hand, actually uses modern video compression techniques, and it credibly competes with other video formats. Plus, it has sufficient momentum behind it to get widespread adoption in browsers and other tools.

        • Adolf Osborne says:

          Define "momentum."

          The only time I've ever been presented with WebM content is when I've gone looking for it, and I'm not exactly new in these parts.

    • phuzz says:

      Making one image marginally smaller doesn't make much difference to most of us, but if Google could reduce the size of all the images they store by just 1%...
      Although I think this is more just because they can, rather than any long term goal.

  5. tylernol says:

    VP8 is a turd, and it is ready to be ripped apart by patent infringement lawsuits. It is just that no-one really uses it, so H.264 bullies like Alcatel/Lucent dont really care about it...yet.

    • Also Sprach Thusnelda says:

      Stop with the silly FUD. Google (YouTube), Mozilla, and Opera all support it, and all of them have a LOT to lose if it was really ripe for patent infringement lawsuits. Not to mention that Skype uses it, and Flash is adding support for it.

      The only people who seem to be raising that FUD are the ones with a vested interest in H.264. I wonder if that just _might_ have something to do with the Justice Department launching an investigation into MPEG-LA? Hmm!

      • tylernol says:

        read up:

        in summary, VP8 is basically a dumbed-down version of h.264, and just because Google says it is patent free does not make it so.

        usually patent trolls wait until a product has been widely adopted and it would be painful to switch away. Hence Alcatel going after LG, Apple, etc for h.264 patent infringements recently. If Alcatel or someone else went after Vp8 right now and it got painful people would just switch back to h.264.
        And on top of all this is that VP8 makes absolutely no sense for mobile devices -- there is no hw support and hence performance and power consumption is poor.

    • Joshuag says:

      VP8 is indeed a turd. Try creating an implementation based on their "documentation". The whole thing is a joke. Pure political posturing. Ain't worth a damn except to the people holding the On2 sourcecode, i.e. Google.

    • You really think Google transcoding all of YouTube counts as *nobody*? I think all the patent talk is just FUD unless someone has data to suggest otherwise.

      • tylernol says:

        since youtube == google then yes, nobody else has committed. Even google has not committed wholesale.

  6. CJ says:

    JPEG-2000 abuses WebP and leaves it sobbing in the corner feeling ashamed for being such a low-quality format. But since the current images formats are Good Enough, JPEG-2000 adoption is also pretty much zero.

    Also 4:2:0? Doesn't anybody remember that the ratios in X:X:X are the number of samples per NTSC subcarrier? What the hell does that have to do with image formats? I demand that someone care about this esoterica!

    • Mike Hoye says:

      Caring about historical esoterica would require actually understanding the reasons things are designed and built the way they are. If you're just going to be rewriting it from scratch regardless, why waste time on something like that?

      • LionsPhil says:

        I have this wonderful little fantasy that one day jwz's CADT Model article will cease to be frequently relevant.

  7. Nicetryguy says:

    .ogg is a godsend for indie app / game developers, even if it makes for a lacking web standard

    MP3 has all kinds of licenses attached

  8. Brian says:

    Ogg is a container format, and I was unaware that it was anything but successful at its purpose.

  9. James says:

    I can imagine a meeting where a Google manager asks "Well, we paid $150 mil for this video format, which is kinda a lot - d'ya think we can get anything else out of it?" And so they shove it into an image format. What could go wrong?

  10. I think my IQ dropped 50 points after reading this crap.

  11. James says:

    It's Google, this is how they operate - make a load of crap, throw it at the nearest wall (the Internet) and wait for stuff to stick. If something sticks (Chrome, Android, Google Porn Search*) run with it. If it slides off, kill it based on how slippy it is.

    At Google failure is an option and seems to be encouraged. It'd be nice if they failed quietly in private though, it might stop everyone else from releasing half-baked crap on us.

    * Run the image search with safe-search off...

    • grue says:

      * Run the image search with safe-search off…

      They really need to extend the options two more slots. "Moderately unsafe" and "Completeley unsafe" would be very popular search options.

  12. Edouard says:

    The article links to two Google studies on the relative quality of WebP vs JPEG images, and then doesn't show you the resulting images.

    That's like going to a wine tasting, and getting a "tannin parts per million" breakdown instead of ACTUALLY TASTING THE WINE.


    • Art Delano says:

      To be fair to Google on that count, all you'd see is a broken image icon next to a nice picture of a mountain range or something.

      • Landa says:

        What you do in that case is turn both images you want to compare into some kind of lossless format and let the reader view these.

      • badc0ffee says:

        They could decode the images, and then re-encode the images as PNGs for illustrative purposes.

        • Breton says:

          Or you write a webP decoder in javascript. Come on, do it Mr. "The web is the new OS!!", put your money where your mouth is.

  13. Anthony says:

    Oh, look - they've re-invented the Johnson-Grace *.art format.

    Uh-oh. Does that mean that Opera is the new AOL?

  14. Jay Paroline says:

    But guys! Webp is 38.9% faster! That is a very specific number, so I can only assume that it is correct! Write your congresspersons!