News Google targets JPEG with WebP

http://www.bit-tech.net/news/bits/2010/10/01/google-targets-jpeg-with-webp/1

 

If it supports transparency like PNG, I'm all up for this.

 

good in theory and perhaps practice too, but so are dvorak keyboards and i dont see them taking over.

 

That's because Dvorak requires you to physically change something, this is a software change.

 

I think JPEG is too prevalent to be replaced any time soon.

 

I use PNG anyway.
Don't like the lossy nature of JPG

 

PNG's good for drawn pictures but looks terrible with photos, like GIF did in the past.

 

This could be very nice indeed. Isn't JPG a proprietary (albeit ubiquitous) standard?
Anything that helps us get the same quality of content with less bandwidth usage is a win.

 

PNG's can look good too, just that file sizes will be much larger compared to an equivalent JPG.

 

The web is getting faster and faster, HDDs are getting bigger and bigger and images are remaining kinda the same size. Wouldn't this have been more useful 10-15 years ago when we were all fighting with dial up? I just don't see the need for this kind of compression and space saving in this day and age.

Yes it is technically impressive when you can make something a 10th of the size that does the same job but does it really matter when all you're doing is changing the percentage used from 0.01% to 0.001% when cumulative usage is still about 40%?

 

The "web" is not getting faster. It's getting more JavaScripted and Flashy. Don't forget that a lot of "web" consumers now use smartphones and handling traffic on GPRS connection just is not fun. At all. Remember that not every county has ubiquitous WiFi and 4G coverage.

 

Unless they find a way to support legacy browsers without requiring any user interaction (installing plugins, etc), I don't see this catching on anytime soon.

Take a look at how long it took for PNG to be widely used/supported, and IE *still* doesn't support the format properly.

 

That may be true but hard drive space isn't infinite yet so any saving with no image loss is still good news.

 

I think your also only thinking about the consumer side here. I work for a e-commerce company, and we have about 6TB of product images and counting. Server storage and backup is costly, not to mention all the traffic. Serving JPGS accounts for about 70% of our traffic, so any ways to reduce this would be interesting.

It'll will be interesting to see how the format performs vs jpg in the wild and what the browser adoption rate is.

 

Although the web in some areas is indeed getting faster, a large part of the population is still on slow connections. Not to mention the fact bandwidth will still be a consideration (which is why using a smartphone in central london two years ago was a nightmare) even if we have high-speed connections.

Also, as photographic resolutions continue to increase, better image compression will continue to be an important factor.

Not to mention that the better we can compress the things we send down the tubes, the more stuff we will be able to send, regardless of the quality of the infrastructure.

In the western world we are accustomed to "unlimited bandwidth" connections, but there is no such thing as unlimited bandwidth. Every byte sent down the tubes has a (infinitesimally small) cost. Doing more for less is a worthwhile pursuit.

 

Me too, and apparently is planned:
http://blog.chromium.org/2010/09/webp-new-image-format-for-web.html

 

Sounds good. Have to wait and see if it catchs on.

 

Ok maybe web was a poor choice of words. Typical internet speeds are getting faster, more internet is being consumed and web is making up a smaller percentage of that, source. So for the general populace, those who typically browse using laptops or PCs at home/work this change is relatively meaningless. Especially as their tastes are changing and actual web (html, javascript et al) usage is declining as a percentage, this is why I haven't heard of any major breakthroughs in DNS or FTP.

However I do agree that doing more with less is beneficial, though only to a point. We currently live in an age where data and its transmission is in abundance rather then scarcity. So if we all have these pipes entering our homes providing us with access to the internet and the pipes keep growing inline with or faster then our demand for the internet then small optimisations such as reducing image size aren't going to matter as much. Though if we assume that growth in our demand for internet is greater then growth in our ability to supply it then these optimisations matter, but they only matter if our consumption of the types of internet remain constant. Clearly that isn't true, in recent years video sites and internet radio has seen greater growth then traditional web.

Yes smart-phones and other portable devices have created a resurgence in the demand for web but as soon as those devices support the bandwidth to consume current internet regularly we'll see a drop off again in the demand for web.

True this only applies to western countries where the internet is not a scare commodity but then again this is where the money is and unfortunately as history has taught us a great invention isn't worth squat if it isn't worth something to someone else.

TLDR: Yes its good but see this as to why it doesn't matter too much and make your own conclusions.

 

Creating a stills compression technique 40% better than JPEG isn't very hard. The mathematics used in DV (that is, video tape) compression is effectively identical, with a few actually fairly minor changes which mean that 3:1 JPEG is often reckoned to be as good as 5:1 DV. You can work out what that is in percentages, but if you consider that JPEG-2000 is felt to be even better than that, it becomes clear that compression performance isn't really the primary driver for new still image formats on the web.

 

problem is jpeg is like windows xp, its hard to budge even if a better OS is out (7),lol, I remember MS's jpeg XR, had transparency, and also reduces jpeg file sizes, whats better is that it also supported the extended dynamic range yey, jpeg is still here