Mozilla Rejects WebP Image Format, Google Adds It
icebraining writes with a link to Ars Technica's look at the recent rejection of WebP by Mozilla Developer Joe Drew."Building mainstream support for a new media format is challenging, especially when the advantages are ambiguous. WebM was attractive to some browser vendors because its royalty-free license arguably solved a real-world problem. According to critics, the advantages of WebP are illusory and don't offer sufficient advantages over JPEG to justify adoption of the new format. (...) 'As the WebP image format exists currently, I won't accept a patch for it. If and when that changes, I'll happily re-evaluate my decision!' wrote Mozilla developer Joe Drew in a Bugzilla comment.'" However, as the article explains, Google sees enough value in WebP to add it as a supported image format for Picasa.
Why do we need yet another image format?
New file format's can't cure something that user education requires.
That mere fact that I am reading this article indicates that WebP has enough momentum to potentially be useful. The fact that other browser(s) are adding support is even more relevant. So the real question I believe is what wouldn't they add it? It's not costing anything, and (apparently) it's already been developed. So what's the issue?!
Is there any actual downside to including it? Bloat, perhaps, but doesn't Firefox already support obscure/archaic formats like APNG, PPM and XBM? It might be wasted effort, so I can understand not making it an active task, but refusing to add any patch adding it seems... dumb.
The world is confusing enough w/o having multiple formats to deal with. Imagine if, instead of DVD, we would have had another Betamax vs. VHS war. (Call it DVD vs. BetaDVD.) Nothing good comes out of these things, at least not for consumers.
And I don't see any benefit from a JPEG v. Webp war either. GIF, JPEG, and PNG works just fine for us casual web surfers.
I also found this part of the article informative:
Muizelaar's complaints about Google's WebP testing methodology are familiar because they echo some of the concerns that were raised early on by other WebP critics like x264 developer Jason Garret-Glaser. The gist of it is that Google [1] used peak signal-to-noise ratio (PSNR) as its basis for quality comparisonsâ"a technical benchmark that experts say fails to account for how images are actually perceived. Another problem is that Google [2] recompressed existing JPEG images rather than starting with uncompressed source files..... WebP's lack of basic feature parity with JPEG in areas like metadata handling and ICC color profiles is identified by Muizelaar as another major problem with Google's format..... [Muizelaar says] the time that Google is putting into WebP would be better spent by improving JPEG encoders or contributing to existing next-generation image format efforts.
"I disapprove of what you say, but I will defend to the death your right to say it." - historian Evelyn Beatrice Hall
post photos much? png kinda sucks for textures.
It is somewhat interesting to see an image format brought to the table without something basic like support for EXIF storage of some kind, or some feature(however crudely hacked on) that makes it clearly superior to JPEG(like an Alpha channel).
I can understand that somebody the size of Google probably gets real worked up about how to shove more images through slightly less bandwidth; but that actually seems like kind of a niche concern: For icon/branding/graphic design purposes, much of the heavy lifting is done by lossless(for clean, non-crunchy look); but small because of limited color palettes, broad areas of flat color, etc. images. That's mostly GIF and PNG, with some Flash and SVG.
For everyone from people who barely care to people who care how it will look as an 8*10 or a desktop background, you have JPEGs of various sizes and compression levels. On the low end, people will put up with some seriously grain-tastic shit, so long as it loads fast. Anybody who is too good for JPEG entirely is probably either slamming around some fancy print-ready flavor of TIFF, or storing whatever flavor of RAW their preferred camera back spits out.
I'm just not seeing the under-served niche here.
Moz's way of getting back for http://en.wikipedia.org/wiki/APNG
I won't accept a patch for it. If and when that changes, I'll happily re-evaluate my decision!"
Quality community driven, bottom up open source software at work!
Picasa? I would think the stronger indicator of support would be Chrome, but then again, Google's schizophrenic position on codec support ("We're rejecting H.264 video in the name of openness! Now enjoy the bundled Adobe Flash plugin and MP3/AAC playback.") makes them difficult to gauge.
Google sees enough value in WebP to add it as a supported image format for Picasa.
No kidding? Google sees value in a format they themselves developed? Next you're going to try to tell me Microsoft sees a lot of value in OOXML.
WebM has a clear advantage when the alternative is not letting your users view video on pages that serve WebM. Other than that, evaluating the advantages of one video format versus another is up to the video producer not the video consumer.
Look how long it took for IE to get transparent PNGs, and then there was the GIF patents issue. Imagine if Firefox was to adopt every hobbiest image format. How about the Goatse Image Format or the Tubgirl Integreated File Format, or even Bluewaffle Media Picture.
Well, of course. From Wikipedia:
PNG is lossless, better for photos then JPEG.
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. This cost is no less for WebP than any other format because progressive decoding requires using a separate library instead of reusing the existing WebM decoder. This gives additional security risk but also eliminates much of the benefit of having bitstream compatibility with WebM. It makes me wonder, why not just change the bitstream so that it's more suitable for a still image codec?
WebP, by Jeff Muizelaar.
There aint no pancake so thin it doesn't have two sides.
How is png worse for textures if it's lossless?
Grammar nazis are to this community what excrements are to gold.
If webp supported alpha transparency it would be useful. png is a lossless format and therefore much bulkier. A png is normally 5 times bigger than jpg image. But jpg doesn't support transparency
How about baseing such as decision on considering what users want / need / might find useful, rather than some developers opinion of whether the technology has merit. Failing all that, because it gives users and web content creators an open source alternative choice?
After logging in slashdot still does not take you back to the page you were on. It's been that way for 20 years.
PNG is lossless
More specifically it's a lossless representation of a single layer RGB image.
better for photos then JPEG.
For display of photos on the web the huge filesize advantages of JPEG outweigh the minor reduction in quality.
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
PNG files can be 10 times bigger than JPG's, and I doubt you always need lossless compression, especially when JPG actually does a good job.
The entire premise of the article is that Firefox, a web browser, didn't add support for a new fringe picture format, something that isn't really the purpose of the program but they're falling behind because Picasa, a program exclusively for showing pictures did? I should think an image program would be the first to add a new image format.
Am I the only one who thinks the author is an idiot?
I'll meet you at the intersection of "Should be" and "Reality"
"However, as the article explains, Google sees enough value in WebP to add it as a supported image format for Picasa."
s/Google/Microsoft/
s/Picasa/Windows/
Because that makes the compression ratio of PNG go to shit.
File size.
Is this spam I smell.... on Slashdot????
amen
Google will create and maintain it and it can be any kind of file at all--image, document, movie, slide deck, virtual machine HDD image, whatever. There will be a few bytes at the beginning of the file to tell Chrome how to deal with it. It will integrate nicely with all of Google's services. Everyone else can either support it or not.
Dear Slashdot: next time you want to mess with the site, add a rich-text editor for comments.
There is no native support for MNG in Opera, Chrome or Safari. Konqueror is the only one I know of that still has MNG support.
That google makes their own browser? At this point, they can add support for any feature or format that they want and then they can start using that feature or format on youtube or piccasa or one of their other sites and basically strong arm other browsers to support that feature because google controls some of the most popular sites on the internet. I could see how this might turn into an issue especially if chrome takes a sufficiently large chunk of the browser market.
"In America, first you get the sugar, then you get the power, then you get the women..." -H. Simpson
I can see how this would be a problem in real-time... for pre-rendered, lossless would be preferred in most situations.
Grammar nazis are to this community what excrements are to gold.
You download a 4000x4000 full-color PNG over a low-end DSL line. Not everyone has a 100 Mbit connection and some of us like to see the latest crazy-resolution NASA images without having to spend ten minutes downloading the thing. Also, a quality 100 JPEG image will have virtually imperceptible artifacts and still be a good bit smaller than PNG.
Also look at audio: For most applications, lossy formats like MP3 and Vorbis are perfectly fine. Even video games with bombastic sound use Vorbis rather than Flac to store the music. There is a trade-off between perfection and file size.
USE HOT GRITS WITH STATUE OF NATALIE PORTMAN (NAKED AND PETRIFIED)
Opera has already added support for WebP. Not just support, they are using it in their Turbo feature to recompress JPEG images because it produces smaller files and they think it looks better. See for example this release from April: http://my.opera.com/portalnews/blog/2011/04/12/opera-11-10-barracuda-released-to-the-wild
(In another post they said they thought it made people look younger.)
1) We know that JPEG 2000 (part 1) is most likely truly freely licensable. It was designed this way, and has been around for many years and used by deep pocketed companies for digital cinema, this I suspect any submarine patents would have surfaces by now. I can't say the same for WebP, WebM, or even H.264.
2) JPEG 2000 can have whatever color components you want. If you want a component to be an alpha channel, that is great, do it.
3) JPEG 2000 was developed by and international standards organization, so you know a lot of eyes saw the specification during development to ensure it is a well defined standard.
4) JPEG 2000 has a lossless option.
This article is about WebP, not WebM. Firefox does very much support WebM, just as do Chrome, Opera, Safari and IE (these last two browser require the WebM codecs to be installed, all the other just work). And YouTube is serving WebM video (among other formats).
There's a hidden treasure in Python 3.x: __prepare__()
If you're going to add a new image format, add jpeg-2000.
better lossy compression
good lossless compression
alpha channel.
Web P was weird from the start. Somebody said "hey, we can just rip that part out and call it a still image format". It's some google programmers hello world that crossed some PR guys desk...
Still, it's further proof of my motto... The earliest codes were based on decades of amazing research and development, and frankly, did everything right. They were designed for transparent reproduction at high bit rates, not low quality junk, so there's room for alternatives there. But across a wide range of scenarios, the early lossy formats get impressively close to the maximum possibly compression, per perceptual entropy, and do so at a tiny fraction the resource requirements of the shiny new trendy junk that is often technically inferiori.
In short, MP3 will be around forever, layer2 at 192k is frankly perfect, Mpeg 2 (dvd, hdtv, etc) will be the high quality video standard effectively forever, jpeg will always be awesome, and perceptual encoding simply doesn't evolve at internet speeds, no mater how much we'd like it to, or how much you spend marketing it.
A can we please get a better name for it?
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
The real problem with WebP is that the name just doesn't trip off the tongue. Try as you might, you'll never get it down to less then 2.5 syllables - it'll always be "Web-uh-P", or it'll get mangled to "Weppy". WebM, JPEG, GIF, P(i)NG - none have this problem.
If Mozilla.org still get 85% of their revenues from Google, this looks like a good gamble .... NOT.
"I love my job, but I hate talking to people like you" (Freddie Mercury)
But is'nt that just a matter of time? Network throughput will most like increase much faster than size of jpegs(due to higher resolution displays), so give it 10 years and it doesn't really matter much whether you use jpeg or png speedwise.
But this ain't it. What we need:
* Support for alpha channel
* Support for arbitrary image metadata
* Support for deep color (more than 8 bit; preferably, arbitrary int with an arbitrary number of bits, or float32). This would be a killer. Right now, cameras do either no processing/deep color (RAW), or full processing/uselessly quantized color (JPEG). For the vast majority of users, processed/full color would be strictly better. This is also true when editing images.
What would be nice:
* Support for animation
* Support for 3d (Fuji has a nice, simple extension to JPEG to do this)
* Support for lossless
* Support for 8 bit, greyscale, LAB, monochrome, etc.
* Optimized for lossless editing (in the way that JPEG supports rotations and certain crops)
* Support for setting white point, black point, and gamma
Google sees enough value in WebP to add it as a supported image format for Picasa.
given that google owns / develops both it'd pretty obvious to "push" it in to picassa
you should also note how half of the browsers that support webp are developed by google
Couldn't WebP be useful for fast extraction of stills (frames) from a webm movie without reencoding?
Google are basically doing the same thing here with WebM, WebP, NaCl, SPDY etc. It may well be these products (or some of them) are open source, it may well be that some have technical merits and require serious consideration. But the way that Google is rapidly reinventing the web in its own image and unilaterally declaring support for formats of its own making is not good for anyone at all. Chrome will be the reference standard for browsers and all the rest are going to look inferior by comparison, forever chasing a moving target and risking second class status. There needs to be a process that should be followed before formats are adopted and it's not just the case of printing a spec, or releasing source code. These things need to be subjected to scrutiny and all major browsers need to be stakeholders. If they're not then Google will surely attempt to hijack the web as did Microsoft before it.
Network throughput will most like increase much faster...
Not everyone lives outside the US!
I remember way back in the early 90's there was this program that came out with it's own image format. The pictures looked great and were about 10KB in size. I remember being blown away by it back then. There wasn't that many pictures with it and nobody else used the format.
Anybody else remember it?
We already have PNG. Why the need for another format?
Sure, keep adding more formats. But good luck trying to unseat the established formats.
WebP has a reference implementation whose license is compatible with free software licenses. JPEG XR, on the other hand, has patents licensed under terms that do "not provide the freedom that the GPL requires" according to the Software Freedom Law Center. And the copyright license of its reference implementation, the HD Photo Device Porting Kit, is not compatible with the GPL or any other copyleft license. Therefore, a JPEG XR decoder and GPL code cannot be combined to form one larger program. See Wikipedia:JPEG XR#Licensing and Wikipedia:Microsoft Open Specification Promise#Scope limitation.
Actually, JPEG doesn't lack ICC, or EXIF.
I know that. I've seen my camera's Exif data show up in several JPEG images that I have uploaded to Wikimedia Commons. But unlike some languages, which provide one third-person pronoun for a sentence's topic and one for other nouns, English provides only one ("it") for nouns of inanimate gender. Please allow me to rephrase the last sentence of my post:
But WebP lacks an alpha channel, and unlike JPEG, WebP lacks Exif and ICC.
I want a new image format.
I want alpha, I want CMYK or whatever colourspace. I want exif or whatever metadata.
But more than anything I want it to support both lossy and lossless algorithms so we can finally see an end to people using jpg for everything, including hard contrast logos.
It would just take a checkbox on the save dialog with some wording to encourage lossless where appropriate.
This signature intentionally left blank
FTW
I'm by no means an image expert, so read this as a question and not a suggestion: why isn't this implemented as a compression format for TIFFs? My understanding is that a TIFF is basically a bunch of metadata wrapped around a chunk of image data. I mean, look at the output of tiffinfo sometimes. I have a hard time believing WebP could require metadata than TIFF already supports, but you could add new private fields if it does. Given that you already have a (to my eyes) perfectly usable container, it seems like a waste not to use it.
Dewey, what part of this looks like authorities should be involved?
pretty much all his argument could have been used on why NOT to use Mozilla when it was new.
Something that's being overlooked here is size. WebP is about half the size of an equivalent Jpg.
Of course, ti really won't go too fer until the add the alpha ability.
After which I expect it to make some quick jumps on the way to becoming the standard.
For no other reason then every cloud based storage is going to want users to use it.
I have 40+Gigs of Jpgs. We aren't even very active with the picture. Some people know seem to be constantly taking picture, so I'm sure there have 100's of Gigs worth of Jpgs. The value add of cutting the image size in half is huge. Not just for storage, but for transferring from smart phones.
It also occurs to me that cutting the image size in half would be desirable form telco that support smart phones on their systems.
The Kruger Dunning explains most post on
I think that's pretty crappy of Mozilla. They're just mad they didn't come up with it first.
Actually, lack of ICC color profile is a good thing as it's happily ignored by Internet Explorer (ICC v2 and V4), Firefox (ICC V4, V2 ignored by default) and Opera (V2 and V4).
eh? Several popular browsers don't render color properly so an image format shouldn't allow for rendering it properly? There's "the future" you know...
My God, it's Full of Source!
OUTSIDE_IP=$(dig +short my.ip @outsideip.net)
So the cosine transform was invented before both of our lifetimes put together. But that doesn't mean the grid of 8x8-sample cosine transforms was invented then, or that transforming followed by quantization was invented then, or that the zigzag sample ordering was invented then, or that the specific kind of entropy coding used in the JPEG bitstream was invented then. Combining two methods from the prior art in a non-obvious way can still produce something novel and worthy of a patent.
How is this trolling, it is actually true! DNG is Adobe's version of RAW. When making adjustments, you don't modify the image, you modify the metadata.
You could say it is even more lossless than PNG.