As predicted, more Microsoft news from MIX10, and this is some big stuff: Internet Explorer 9. As we all know, Microsoft really let Internet Explorer rot away, allowing competitors to make much better browsers with better standards compliance and performance. With IE9, Microsoft is aiming to not just close that gap – but to overtake the competition. Update: Ars has an in-depth look at the platform preview.
There’s really a lot to tell about Internet Explorer 9, so I don’t really know where to start. First of all, you can download a test version, called the IE9 Platform Preview. This preview will be updated every eight weeks, and customer feedback will be taken into account – presumably much like the Windows 7 development phase.
So, what’s new in IE9? Fist, there’s a new JavaScript engine called “Chakra”. It is a much, much faster engine than IE8 has, and it currently performs somewhere in the middle compared to other major browsers. The interesting thing about it is that IE9 will run Chakra (and thus, JavaScript code) on a separate thread, on a separate core (if available).
That’s not the only thing that IE9 will hand off to hardware – text, graphics, and SVG rendering will be hardware accelerated as well, providing, according to Microsoft at least, better performance than other browsers.
A hot iron is obviously standards compliance, and it appears that Microsoft has finally learnt its lesson with this one. The post on the IEBlog announcing the IE9 MIX10 details is full of holding hands and flowers and free hugs, and the current Platform Preview build scores 55/100 on the ACID3 test. Microsoft is asking for feedback on this stuff.
“The main technologies to call out here broadly are HTML5, CSS3, DOM, and SVG,” Microsoft writes, “The IE9 test drive site has more specifics and samples. At this time, we’re looking for developer feedback on our implementation of HTML5’s parsing rules, Selection APIs, XHTML support, and inline SVG. Within CSS3, we’re looking for developer feedback on IE9’s support for Selectors, Namespaces, Colors, Values, Backgrounds and Borders, and Fonts. Within DOM, we’re looking for developer feedback on IE9’s support for Core, Events, Style, and Range.”
At the Platform Preview site, there’s a whole bunch of browsers tests to perform, and, of course, you can download the latest IE9 build – new territory for Microsoft. This is what competition can do, and it’s good. Especially the hardware acceleration sounds incredibly interesting.
Quote from the MSDN blog post:
Interesting. We soon have support for the video tag in every major browser, but we also find ourselves in the middle of another format war… *sigh*
Edited 2010-03-16 17:42 UTC
I don’t think there will be a format war – only a war over how h264 is to be implemented; will Firefox hook into gstreamer and then allow the end user to install the ffmpeg gstreamer plugin manually through a European (mirror site in a country where software patents aren’t upheld) website. Although we can all jump around wishing for Theora the cold hard reality is that h264 gives better results at lower compression rates which then corresponds to lower bandwidth costs. Oh, and no Limuer (or how ever you spell his name), I don’t give a crap if you link to 12 websites by unknown people claiming that their website with 12 visitors a year have ‘great scalability’ – the facts are self evident – either accept them or just move along.
The question is not whether H.264 is technically better or not, it’s the licensing issues. Besides paying MPEG-LA for the encoder and decoder, you’ll have to pay them for each video as well.
Except MPEG-LA has said they’re not going to do that…
Which doesn’t say they won’t change their mind again sometime in the future.
Someone has already corrected you on the costing but the question again is this; what is the cost of the bandwidth, how much bandwidth does it use up (or more correctly, save) and the cost of the CODEC versus the cost of the bandwidth used by Theora; then on top of that the quality has to be taken into account – if the video is crap quality that it won’t matter whether you can get the bandwidth is equal between Theora and h264, if the video is of lower quality then it won’t bring the views thus won’t get the traffic thus the advertisers won’t back your website.
There are a number of factors that need to be taken into account and not just the immediate cost of the CODEC alone.
Edited 2010-03-17 00:34 UTC
This is a bit confused. Either codec can be made to deliver video of better or worse quality. It is actaully the quality:bandwidth ratio that is a single parameter. At this time, for the same bandwidth, h264 delivers slightly better subjective quality than Theora. The difference is barely perceptible, most people can’t tell the difference in blind tests.
Experimental builds of Theora right now are apparently achieveing appreciably better quality than h264 at the same bandwidth. By the time IE9 is out, this currently-experimental build of Theora is likely to be released.
Edited 2010-03-17 02:06 UTC
1) How is it confusing if you have two videos being compressed; both of them set at 1mbps on the encoder; both of them use the same bandwidth but one will have a higher quality than the other.
2) Experimental builds – thats very nice but guess what? the majority of the videos encoded and sitting on servers is either in an old On2 Technologies format or in H264; imagine the millions of videos on YouTube – is it really productive to convert all the videos on there at the risk of quality loss (transcoding)? the processing power used up? then there I’m sure many more other considerations that need to be taken into account. It isn’t a matter of open and shut.
I wasn’t confused, it was your original text that was confusing. We agree, what one needs to do is either: (1) to set the same bandwith and compare quality, or (2)to set the same quality and compare bandwidth used. The latter is impractical compared to the former, so we use the first method.
A video website worth its salt would of course encode from uncompressed video, or at least transcode from higher-resolution compressed video if the uncompressed video data is unavailable.
As for quality and processing power required … Dailymotion have done it:
http://openvideo.dailymotion.com
http://www.fsf.org/blogs/community/dailymotion-theora/?searchterm=d…
If Dailymotion can do it, where’s the risk?
As far as the non-free nature of h264 is concerned, and hence its utter unsuitability for use as a codec for the public access web, there is no question at all. This IS a matter of open and shut, if you will pardon my intentional pun.
Edited 2010-03-17 05:45 UTC
Ok, which goes back to my original point regarding the choice of h264 over theora – there has to be a reason why YouTube chooses h264 over Theora besides just wanting to do it because ‘thats what everyone else is doing’. In a perfect world though I’d prefer to see Google open up VP8 and make it into a royalty free, patent free format for the web.
Yes, but as I said previous; there has to be a reason why h264 is preferred over VP3/Theora besides just Youtube or what evre company following the crowd. You’ve provided a set of case scenarios but it hasn’t addressed the reason why the powers that be still go with h264 over VP3/Theora.
Edited 2010-03-18 02:49 UTC
There is. There is a list of patents for h264 which apparently is 47 pages long. Just the list. So, every company on the following list has a strong vested interest in getting everyone else to use h264.
http://www.mpegla.com/main/programs/AVC/Pages/Licensors.aspx
That is a longish list of powerful interests that are being quite successful in generating a fair amount of anti-Theora FUD right now. They have a lot of people believing that Theora is rubbish, and risky, and that h264 is the only way to go. The companies on that list just happen to also be IT companies who supply codec software, video acceleration graphics hardware, operating systems and web access client devices. Hmmmm, nad they want you to use h264, so you will pay them money, and pay and pay and pay again for the same research they did years ago.
Self-interest of a powerful group of IT companies is the reason why h264 is pushed so hard … by commercial interests.
Meanwhile, in spite of solid eveidence that Theora performance is overtaking h264 (see for yourself, don’t take my word for it):
http://www.osnews.com/permalink?413958
it takes the likes of me dozens of posts and considerable effort rebutting gullible people who just won’t hear any suggestion that Theora is the way to go.
Edited 2010-03-18 05:15 UTC
Agreed – much as I might wish otherwise, Mozilla aren’t going to win that battle.
While Microsoft stays out of the fight, <video> is irrelevant, since the major browser can’t use it anyway. And if they do join the fight, whatever they do will probably be definitive, and it surely won’t be Vorbis. The best to hope for is that MS won’t pick a single format, and will just leave it up to their platform video APIs (DirectShow?). Much as Mozilla will eventually be forced to do with GStreamer or other platform-native options…
Vorbis is an audio code, the open video codec from Xiph.org is called Theora.
But why not support Theora (and Vorbis for that matter) out of the box? What is stopping them? What would it hurt?
Default support for Theora in all HTML5-capable browsers is entirely in line with the original design intent of the web. h264 is not.
More code to manage, audit (for IP purposes), maintain for a very limited scope of customers that can’t be justified when compared to other things that are of more importance. If the Theora boosters want it to be adopted then they should provide a DirectShow CODEC that allows encoding and decoding. Nothing has stopped the Theora boosters except for apologists like you who expect everyone else to do the heavy lifting so that you can feel warm and fuzzy about your technology of choice.
If the burden of auditing code for license compliance is a concern, then Theora is by far the better choice to support. Infinitely so.
There are no parties asking for license fees for Theora.
Despite your ranting, it is the browser and OS vendors who must provide the support, not the designers of the codec. This is just as true for h264 as it is for Theora.
When the CODEC is a niche the it is up to the programmers of the CODEC to provide the CODEC as to move it from a niche to a viable alternative to then the mainstream software vendors take it from a alternative to a provided piece of technology. There is more work involved than just jumping from development of CODEC to getting it in side Windows simply because “its a good idea, and I think you should do it”.
Sigh!
You just don’t get it, do you?
Xiph.org do not write IE, or Safari, or Firefox, or Windows, or OSX, or Linux. It is up to the parties that do write that software to implement Theora within that software.
This is the way it is for h264. Why should it be different for Theora?
Edited 2010-03-17 05:29 UTC
Agreed; even if they hired 20 of the worlds greatest programmers you’re eventually going to push Theora so hard that you’ll eventually hit the patent wall where you can go no further because you’ve exhausted all existing technology you can exploit and the only thing you can do is start cherry picking off patented algorithms one at a time thus adding cost to the mix.
Well for Microsoft it doesn’t matter; they already provide VC-1 and h264 playback for Silverlight as well as via Windows Media Player – so for Microsoft they’re caring neither here nor there knowing that OEM’s if they wish to ship a version of Linux with h264 supported out of the box with Firefox will end up licensing it thus driving up the cost to the point that the marginal difference between Windows and Linux are so small that it pushes Linux out of the equation as so far as it being the low cost alternative.
In the case of Firefox on Windows it can hook into directshow, on Mac OS X hooking into quicktime but in the case of Linux/*NIX, it is going to be very difficult but ultimately it won’t impact on Linux/*NIX, it’ll become an issue for vendors like Red hat and Novell as their customers will start demanding that they start providing these features as part of their distribution – after all, as a customer they pay for a product and expect that as part of paying for it that they receive support for these emerging technologies. So for the non-profit community based organisation it’ll become a non-issue but for commercial Linux vendors it’ll raise questions whether they’ll provide it as their customers demand it (the customers who pay for the support contract).
Edited 2010-03-17 00:21 UTC
Thusnelda and Ptalarbvorm give the lie to this assertion. Between them, these two successive optimisations of the patented VP3 codec methods have vastly improved the performance of the Theora encoder. This is illustrated by the fact that earlier versions of the Theora decoder can still play Theora videos encoded by Thusnelda and Ptalarbvorm.
http://en.wikipedia.org/wiki/Theora
The exploitation of existing VP3 technology has not yet been exhausted, and there has been no need to start cherry picking off any (other) patented algorithms.
But there are also features of h264 that haven’t been implemented; so you’re up against a CODEC that is already performing very well and has yet to be fully optimised with all the features being exploited when it comes to encoding. You may talk about VP3 not optimised to the max, but there is considerably more features of h264 whose features haven’t been taken advantage of as based on the h264 encoders wikipedia entry.
AFAIK, many of the features of h264 that aren’t being used in the context of video for the web aren’t actually suitable for video for the web.
Furthermore, you cannot in the same breath claim that Theora is not an option for the video codec for the web because of inferior bandwidth performance, and then when it turns out to achieve better bandwidth performance than the way that h264 is currently being used, then turn around and say h264 is potentially better.
Either bandwidth performance is important, or it is not. Theora has all but caught up with and is now apparently passing h264 the achieved quality:bandwidth ratio for use with video over the web, and it makes no difference if h264 has some unexploited potential improvements or not when Theora IS realising actual improvements.
Edited 2010-03-17 05:28 UTC
For those interested, here is a directory of files for comparison of Theora (Thusnelda) versus Theora (Ptalarbvorm):
http://people.xiph.org/~greg/video/ptalarbvorm/
No audio yet. This was about a month ago: the improvement from Thusnelda (current) to Ptalarbvorm (experimental) is quite noticeable and worthwhile.
Here is a directory where Theora (Thusnelda), Theora (Ptalarbvorm) and Youtube h264 are compared (4th Feb):
http://people.xiph.org/~greg/video/ytcompare/
In particular, here are three screenshots of the same frame:
http://people.xiph.org/~greg/video/ytcompare/bbb_theora_486kbit_366…
http://people.xiph.org/~greg/video/ytcompare/bbb_theora-ptalarbvorm…
http://people.xiph.org/~greg/video/ytcompare/bbb_youtube_h264_499kb…
The three pictures are of almost indistinguishable quality.
I think YouTube h264 was 499kbit, Theora (Thusnelda) was 486kbit, and Theora (Ptalarbvorm) was 376kbit.
So, Theora development (Ptalarbvorm) is apparently an appreciable improvement in quality:bandwidth ratio over the h264 standard that YouTube is using currently.
So I ask, once again, since the Theora client software (the decoder, or player if you like) doesn’t require changes in order to get these improvements, why not include a Theora player in one’s OS and browser?
Why shouldn’t Vista, Windows 7 and IE9 support HTML5/Theora out of the box? Why not? What is the downside?
The license terms to implement Theora are roughly these: “Anyone may implement, anytime, anywhere, in any context, for any platform. No royalties apply. Fill your boots”.
These are more or less the same license terms as HTML5, CSS3, ECMAScript, DOM and SVG, which IE9 apparently IS going to implement.
Edited 2010-03-17 10:40 UTC
“A hot iron is obviously standards compliance, and it appears that Microsoft has finally learnt its lesson with this one.”
This is just the cycle of “embrace, extend, exterminate” that is typical of MS. I’m not saying this is either good or bad, as I believe it has led to both.
what are you basing that on? this will be the third major release they have done that actually moves it to standards compliance, and we still haven’t seen anything remotely like “extend” or “exterminate”
Well, there use to be a ton of IE specific tags that were not part of any specification. They tried to do the same thing with the MS Java VM. And are currently doing it with XML in Office.
I think the non standard tags helped kill off Netscape’s popularity, though they had their own non standard tags as they were trying to do the same thing to IE. It back fired with the MS JVM partly because Sun owned Java. XML in Office appears to be working. This is, of course, simply my opinion.
Also, I am not arguing moving more towards a standard base. You can support a standard base while still adding in extra stuff that doesn’t break the standard. It can be part of what helps one to differentiate yourself from the competition.
Edited 2010-03-16 18:52 UTC
So, your first two examples were good examples of the whole “embrace, extend, extinguish” thing, but we’re talking about 15 years ago now, and they got in a lot of trouble for both. Your third example is wrong, they are not trying to extend xml in any non standard way, peoples problems were binary bits in the xml that weren’t documented (although that is still valid xml)
When it comes to web stuff, the problem wasn’t html, it was javascript and css. IE7 fixed most of the worst css problems, IE8 was fully css 2.1 compliant (so propriatary extensions are prefixed by a -ie-, just like the other vendors do)IE9 is actually adopting emerging standards, which is something we haven’t seen the IE team do before.
As a web developer, the only thing better then this is if IE suddenly disappeared overnight one day and I didn’t have to worry about it at all. We haven’t seen any evidence of extending standards into propriatary places yet, and while some of the changes aren’t as good as they could be (like filter becomming -ie-filter instead of rgba, opacity, and transform), overall there hasn’t been any moves in bad directions.
Huh? how can you say with a straight face that you haven’t seen Microsoft extend and corrupt standards?
Further proof that web “developers” don’t know what the heck are talking about most of the time. Good grief.
good examples to back up your claim there…. Oh wait… I was thinking of the parent post, not your’s.
used to be…
Reality calling… you can’t mess up XML… it has a single set of conventions and no standard tags. If it is well formed, that is all that is required. So try something else pal.
Actually MS messed up XML…
In XML you can encode characters using one of two formats; however, MS created a third one.
Can’t find the references right now…
I’ll believe you when you get me that reference.
I’m sure you will get right on that.
I should also probably send you here:
http://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish
Granted, it is Wikipedia, so read it with any of the Wikipedia warnings you prefer.
Maybe that’s cuz they’re still trying to “embrace”
so… now MS is playing by the standards and you are criticizing them because “they are just going to embrace and extend”
Are you kidding me!!
I simply don’t see them doing it because they’ve come to a realisation; HTML5 isn’t a threat to them because ultimately people who want to use Silverlight will use it, just as those who want to use HTML5 will use HTML5. They provide development tools for both types of developers and they come out winners at the other end. So it isn’t an all or nothing game, it isn’t either all or HTML5 or all Silverlight.
For Silverlight, I find that the people it is being pitched to are those whom Adobe AIR is being pitched to; so expect to start seeing the VB applications gradually die out as they’re replaced with Silverlight applications where the one application can be run on the desktop and hand held. A large corporation creating a nice database front end for employee’s to access that is accessible via a desktop or via hand held WM7 devices without the need of any funky recompiling, refactoring, recoding etc.
The Microsoft of today, as ruthlessly competitive, is a totally different beast than it was 10 years ago or even 5 years ago. What I would love to see from Microsoft is Silverlight development tools for Mac – if we must have a platform like Silverlight, I would love to see Microsoft make it an open standard that is royalty free to implement and development tools for Windows and Mac.
In particular they are adding a non-compatible video tag (doesn’t support royalty free theora codec). The average “incompatibility rate” of IE will still remain high.
They are not doing “same markup, same results” for the new elements, e.g. video element.
As much as I dislike h264, this is just nonsense. There is, sadly, nothing non-standard about using h264 as the video codec for HTML5 video.
And the beauty is – if you have the Theora codec installed on Windows (through the mega codec pack, for instance), Theora video will most likely work fine as well – the wonders of DirectShow (or whatever it’s called in Vista/7).
You’re really grasping at straws here.
EDIT: Oh, it’s called Media Foundation in Vista/7 – DirectShow is still there though for compatibility reasons.
Edited 2010-03-16 23:18 UTC
I think this long-overdue move to support for web standards in IE is very welcome, and I applaud it. Only good can come from this.
I have but one quibble. If Microsoft can include h264 support by default (and charge their users for the royalties), what would it hurt to include Theora support by default as well? It wouldn’t hurt Microsoft, and it wouldn’t cost Microsoft or Microsoft’s users anything.
With the latest development of Theora in the experimental Ptalarbvorm build, Theora is looking as if it may be able to surpass h264 in performance as a codec for use on the web.
http://xiphmont.livejournal.com/48207.html
If Microsoft included support for Theora as well as h264 by default, then IE would support the leading codecs for HTML5, and there would be an even playing field between them.
Web page providers could choose either video format and be confident that most browsers could render the video. Those browsers that couldn’t (*cough* Safari) would soon have to comply to remain competitive. This would make it viable for low-budget and community web sites to still host video without having to pay license fees to MPEG LA for h264. We would enable participation for all on the web (not just corporations), as it was originally designed for.
If Microsoft leave out default support for Theora, then they are still trying to use their position to suppress open formats on the web, IMO.
Edited 2010-03-17 00:03 UTC
What I don’t get in this context is that supporting ogg would be even a competitive advantage for IE over Firefox and Opera as IE then would support more codecs.
“It takes about 70 seconds to identify a 300ms difference between browsers.”
No, no, no… You guys are wrong. People can tell what the speed difference between firefox and chrome is in less than a minute. It feels extraordinary faster. The human eye can hardly detect lags of 10-20ms (remember those flickering CRTs ?), but you’re talking about delays that are an order of magnitude longer.
And then, they pretend to have invented multi-threaded javascript engines… Wow…
For the standard stuff, I don’t see Canvas, < video > and < audio >, and so on… Hardware acceleration is great, but if you’re only accelerating basic things, it’s no use…
I hope that IE 9 performs really well in order to compensate this, because such bad communication is not impressing me at all.
Edited 2010-03-16 17:54 UTC
55/100 is a good step up from IE8 crashing at 13/100.. and i heard HTML5, CSS3 and most surprisingly SVG!
But will it help upgrade all those win98/2k/xp IE6 installs or will it mend the pain from the last decade of neglect? not one bit..
I know they are only doing it because their marketshare is flushing down the toilet..
Win98/2k are out of support. XP is in extended support although 10 years old. You can get IE8 for it.
This is for Vista SP2 or newer only as per Microsoft site.
In all honesty, why are you still using Windows 98/2000/XP? if you’re using such systems it is highly unlikely that you’re interested in upgrading to Internet Explorer 9. They aren’t back porting to anything older than Windows Vista because it depends on API’s that aren’t present in version of Windows prior to Windows Vista.
But honestly, if you’re still running Windows XP today, other than on a NetBook (which can easily be upgraded to Windows 7 without much fuss and bother), I really have to ask where your priorities lay, especially when I see businesses plead poverty when it comes to Windows 7 upgrades but are quite happy to splash out on private jets, expensive meals, obscene bonus’s and other ‘creature comforts of office’ that add zilch to the bottom line in the long run (or is thinking about the long run these days not a ‘cool’ thing for management to do any longer?).
Canvas is pretty much the only thing that is sticking out as a sore thumb to me as missing. I wasn’t expecting video this soon, that and the GPU performance stuff was a nice surprise.
It would be sort of interesting to see where IE9 sits when it comes to actual javascript conformance. ACID3 is nice and everything, but an ACID3 number under 100 is not really reflective of much, since both the real world use cases and the extreme corner cases are weighted the same. Probably going to give this a run through http://sputnik.googlelabs.com/ later tonight.
Canvas is nice, but given that SVG finally is there, I can wait longer. The biggest pain for me was that almost every browser had a decent SVG engine, but not Microsoft.
Thing is, I haven’t seen anything interesting or compelling done at all with svg… anywhere. What little svg work I have seen on the web is stuff like this http://emacsformacosx.com/, which just seems like a fairly mediocre web page.
Overall, svg on the web seems like a slower, less functional flash with a bigger file size. The only selling point I can see is that it is an easily parsed format.
Seems like every few months there is someone doing cool things with canvas, like http://mrdoob.com/projects/harmony/, or the javascript cloth simulator by andrew hoyer that is down at the moment)
http://raphaeljs.com
So thats the sort of thing I am talking about, flash from five years ago
What’s flash from today, then?
flash charts: http://flycharts.net/gallery.html
javascript + canvas charts: http://people.iola.dk/olau/flot/examples/selection.html
svg charting with raphael: http://raphaeljs.com/analytics.html
first two are more functional and look way better. that is just charting, but i have yet to see an interactive svg demo that doesn’t look dated.
But that appears to require canvas as well…
for some other good canvas apps/games/demos I like to point people here:
http://www.canvasdemos.com/type/applications/
Many of these even work in Haiku’s new Web+ browser
Ignore that, I didn’t actually look at the page source, just saw the word “canvas” and assumed
Actually SVG is important because it becomes more or less the vector graphics format for the web now, a thing which has been missing for years. Sure there have been others but none of them worked over all browsers. That you can do a lot more is a nice addon, but vector graphics were severely missed and all that had to be worked around constantly.
In what way? The only thing I can think of is having gradients that scale properly, but we have background-size and gradient now too, so even for that I wouldn’t turn to svg.
Is there something I’m missing?
Microsoft moving to standards means Flash will become irrelevant in a few years. Excellent.
That would also mean that Silverlight would become irrelevant. Do you really believe that’s Microsoft’s goal?
They’d have to be really afraid of becoming a speed bump in history to change that dramatically.
Not really, as silverlight seems to be their preferred platform for their new phone.
that’s not possible.. silverlight is not relevant to begin with..
Not really. Flash will rule as long as it has great content editing software and html5 doesn’t. Visual designers don’t want to deal with too much markup.
Just a quick test
http://ie.microsoft.com/testdrive/Performance/01FlyingImages/Defaul…
Ubuntu 9.10 x64, Nvidia 8400 GS, Nvidia drivers 195.17 (those are the ones required for CUDA Beta 3.0), Xeon E5405, 8 GB
Under Google Chrome 5.0.342.3 dev it runs miserably (4 fps with the default image size )
but with Firefox 3.5.8 it runs quite smooth (55 fps with the zoom in activated, no matter how fast is the spinning)
The pic in the blog shows 60 fps and it is supposed to be GPU-accelerated. So 55 FPS under Linux looks quite good.
I just took a quick look of the code and it is just javascript scaling the pictures so this looks more a test of how fast the javascript machine and the render system can work together.
I’m wondering how Firefox achieved such good performance, my understanding is that 2D acceleration in Firefox is not yet implemented under Linux, in principle 2D acceleration is in the pipeline and this only for Windows.
My guess this is due only to javascript performance to render the whole scene before sending to the graphic subsystem.
Another interesting test, using Vmplayer to run Win XP 32 bits and Firefox 3.6 shows stills a good 40 fps. And for this one the rendering passes several layers between Windows, vmplayer and finally Xorg. IE8 shows only 3 fps and Chrome even less with 2 fps.
I wonder then what exactly are we looking on the 60 fps shown with IE9: purely GPU acceleration or just the improvement of their javascript subsystem?. Anyways, this is a very simplistic example but still interesting to see the degree of performance of the javascript engines to render an scene composed of few flying images.
Same here (on Linux too). Remarkable how much Chrome sucks on that test.
Here’s the odd thing (under linux of course):
Chromium 5, 5 fps
Konqueror 4.41, 40 fps
That’s just messed up.
For those interested:
Opera 10.10 6 fps
firefox 3.58 49 fps
I’m getting 61 fps on Firefox 3.6.
firefox number seem to be pretty high, but not Dependant on the version. There was someone down below that reported 49 on ff 3.7. It’s probably more hardware dependent than version dependent.
.. So what’s under the hood? i7/i5? My numbers were on a quad core 3 gHz phenom II machine. I’m guessing the raw cpu frequency is the bottleneck. Here at work with a machine with identical software, but less Ghz I get only 45 fps.
Or simply just where you place the mouse… since it rotates faster/slower, and skews differently depending on where your mouse is in relation to the page. I get vastly different FPS numbers from FF 3.6 depending on how fast I have it spinning and whether it is high, low, or in the middle of the page.
A naive question is How SVG rendering is GPU accelerated? I think only the canvas operations are accelerated. I also believe that 2D acceleration is necessary anyway to have a good experience n the desktop (I hate my VESA supported V3 XGI, have to test xf86-xgi). So 2D acceleration of canvas != GPU acceleration. Am i right?
Fianlly I hope to see a DirectX/OpenCL/OpenGL whatever API at the hardware level. If GPU rendering will become the norm OSes must have it anyway with specs and standards, without proprietary drivers, just like VESA.
“Fianlly I hope to see a DirectX/OpenCL/OpenGL whatever API at the hardware level. If GPU rendering will become the norm OSes must have it anyway with specs and standards, without proprietary drivers, just like VESA.”
Not sure that I agree with that.
Norms like VESA appeared because at the time, it was thought that hardware should offer advanced capabilities to all devs, get compatible with software.
But times have changed. Now, for the sake of reduced hardware price, the trend is that a lot of hardware (including very widespread one, like wireless) would use proprietary technology instead of norms, and that the provider would provide a driver for Windows only.
3D acceleration is already everywhere. It is on more than 95% of the computers due to Windows Vista/7 specs. I don’t think pushing it forward even further and making it mandatory for normal internet operation will make standardization less painful. Instead, it’ll probably make thing worse for indie OS devs who only get beta-quality driver or no driver at all from the manufacturer.
Edited 2010-03-16 20:35 UTC
hardware rendering is a completely implementation specific optimization, specs and standards for it would be silly.
With intel graphics on winxp, the FPS varies quite a bit in FF 3.6 depending on where you have the mouse located in relation to the images, and how fast they’re moving. I can get it to go as low as 20fps, and as fast as 66fps.
And yes, Chrome 4.x performance is horrible here as well.
Ubuntu 10.04a3 Amd Athlon 2500+ Nvida 7600GS Firefox 3.7 a4pre => 47 FPS
With a crappy Intel C2D T5550 plus an even crappier Intel GM965, I can get 60 FPS from Opera 10.51, Firefox 3.6, and Konqueror 4.4.1. Only Firefox has noticeable lag on page load, and the FPS number stays high when the lag happens.
What about offering a full-fledged extension framework, not just useless “addons” that allow you to install an AOL toolbar?
At least the IE9 test preview does not support Windows XP bummer
Does someone actually believe MS will continue to follow standards because they supposedly “learned their lesson?”. I hope no one is actually believing that.
I’d say, from my POV, it doesn’t matter at all how good IE9 is as long as there is Firefox or similar around which guarantees a future without vendor lockin issues.
Can we please just learn from the past?
You’ve got some good points but to be honest it doesn’t matter if IE gets 80% market share on the desktop anymore. More and more of web browsing is being done on portable devices and there is no way in hell developers are going to code for solely one platform anymore.
Windows Phone 7 sounds great and all but it would have to be amazing to take over the phone space. RIM in the enterprise space and Apple for consumers, with Android making steady headway, it would be crazy to develop a website for a non standards complaint browser anymore.
If we can learn from the past, why can’t MS?
Can we have a more sober style when it comes to MS announcements please?
How is telling the world what you are going to do someday worth such headlines? After all, this is all about MS trying to close the gap in some distant future release(s) and IMHO not worth mentioning in such big letters. That skews the big picture: That most other browser vendors are at a point where MS want to go tomorrow.
The factual news is that MS is still not at the same level as others and doing marketing for a faster JS engine etc. is somewhat revealing after MS tried the “get the facts” stunts how IE8 is not really slower etc. compared with others (so why do they need to improve the JS engine?).
There is nothing big or surprising: MS just does not have any chance to play any significant role in the future browser market if they do not at least reach “good enough” level compared with the concurrence. Losing market share like IE despite being the default, preinstalled browser on most desktops tells MS exactly how customers feel about their product ..
(And yes: I really hate the MS style of “announce today, deliver someday or never” when others are running circles around them. Even more despicable however are “news sites” that trumpet such marketing announcements without a grain of reflection .. )
Except, the linked article does provide a download link to grab the current progress for developers and users to test against if they so desire, which is actually news. Last I checked just about every other browser preview/beta in the recent years gets widespread PR just the same.
Uhm, you didn’t really read, did you?
You can download the test build today. If you had read the article, you would’ve known.
I’m happy with the improvement on IE 9, but still Microsoft have to do a lot on their browser, nowly they wokeup after IE 6. Thanks to Magic browser FireFox for wakeing them up
Yes, definitely they are doing catching up job in many areas of browser and they are coming up at good speed to beat the rivals.
My guess, the JavaScript engine speed has improved a lot to match the demand of hardware accelerator, not because of competitor.
JavaScript speed alone doesn’t determain the performance of the browser. The real performance of a browser should be evaluted in all of the followings
(1) Secure & Privacy
(2) Stability
(3) Speed – not only JavaScript, but also rendering, downloading.
(4) Perfect rendering- Pixel level accuracy in rendering.
(5) Memory consumption
(6) Battery life -for netbook/laptop/slate scenarios
(7) End User friendly – like better printing support,huge favorites management,
(8) Developer friendly – standard compliance, better API to develop plugins,
..
..
There might be some more to this list, but my worry is no browser reviewer or nuteral person is doing this kind of testing/review. Most of the reviews on browser are one side or reviews on only one aspect of a browser.
For example, Opera says they have the fastest java script engine. But they consume more memory than IE 8 in most of the site. Now you cannot say Opera better than IE 8, both has plus and minus.
We need one complete review of all the leading browsers for better judgement!
HTML5. You keep using this word. I don^aEURTMt think it means what you think it means.
None of the demos actually use HTML5, and it scores the same as IE8 in the HTML5 test: http://rakaz.nl/2010/03/microsoft-talks-big-about-html5-but-shows-v…
What they^aEURTMve shown is disappointing. But I will give them the benefit of the doubt. It^aEURTMs an early preview and it will be built up regularly.
This news is just a hype-building exercise. I expect to see the goods delivered in due time now.
(N.B. The preview doesn’t work with my site or OSn5 because they didn^aEURTMt know^aEUR”or ignored^aEUR”that the body tag is optional, blaming my valid markup instead)
edit Oh^aEUR”and they^aEURTMve finally supported application/xhtml+xml just as it^aEURTMs now dying out. Way to go, real finger on the pulse there Microsoft.
Edited 2010-03-17 11:21 UTC
Honestly, I was expecting them to hold their ground at least for another version on the “we aren’t going to put effort into an evolving standard” thing. It may not support html5, but the beginnings of css3 support is more then I hoped for.
After more snidey backchat from me (I^aEURTMm touchy over these issues) Microsoft acknowledged my problem. It^aEURTMs good that they^aEURTMre listening and being decent about the whole thing^aEUR”there^aEURTMs a whole lot of hate being radiated at them for the years and years of neglect web developers have gone through and they cannot afford to screw up the IE9 release or they will see their market share continue to free fall.
CSS3 selectors probably makes the biggest difference. That^aEURTMs what^aEURTMll get rid of tag-soup.
All of this is _just_ catchup though. I would still like to see CSS gradients, CSS variables and other experimental ideas being explored with IE9. What about WebGL? Two browser already have experimental support for this.
It^aEURTMs going to take a long, hard slog for MS to get up to speed with the web again. It seems that they are beginning to understand that with this new preview. I^aEURTMm not too fussed when IE9 comes out, just that they do a proper job of it.
From the Ars article, it sounds to me like they still have a bad attitude towards compliance. The announcement of IE8 was the same story. Everyone said, “finally, they are listening” yet we’re still writing workarounds for IE!
I understand little mismatches between browsers & OS’s, but CSS3 should be a done deal. Like you said, where’s my rounded corners and shadows? Fonts are another issue for me. I have to convert to EOT fonts just for IE while everyone else can handle OTF and others. In the article they dismiss compliance here too.
Someone more savvy than I (or less lazy) should write up a list of elements, selectors, and other features that are implemented in Gecko, Webkit, and Presto, but not IE and say, “Here’s what developers want. Give us that or keep your browser.”
I assumed that was the case, but I just now installed the IE9 Platform Preview (which I’ll refer to as “IE9” below) and ran all of the tests, comparing it with Firefox 3.6 and IE8, and IE9 bests and even blows away Firefox 3.6 in many of the tests.
I do give Firefox props for overall rendering speed; it’s pretty fast even without using the graphics chip acceleration. Certainly blows Chrome away (as others stated above).
But IE9 is still faster (particularly for SVG, as well demonstrated by playing the SVG Astroids game in “nebula” mode), and handles all of the tests as well as, better than, or even far better than FF (everything from borders to text rendering). FF outright fails multiple of the tests. For example, the “Widen Justify Box” test, FF totally fails, even worse than IE8 does. (Well, IE8 doesn’t “fail” that test, actually, it’s just painfully slow. Incidentally, many of the tests are grayed out for IE8, since IE8 can’t perform them at all.)
To be fair, IE9 hasn’t been released. By the time it’s released, maybe a version of FF will have been released that performs these tests as well as IE9 does. (Microsoft has handed over the tests to WC3 (or whatever) so other browser makers can make use of them.) But the current FF does not perform these tests nearly as well as IE9, which is why I don’t think it’s fair to dismiss IE9 as “just catch up”. It clearly leapfrogs FF in the various functionality.
Note: I don’t have other browsers installed at the moment. But I gather that Safari and Chrome (both webkit based, I believe) would do better than FF at actually passing the tests, but would have extremely poor speed at some of them (according to posts by others above). As for Opera, I assume it would perform well on these tests (historically it’s been “fast” and supports advanced HTML features more than or better than other browsers, generally speaking), but I don’t really know.
Edited 2010-03-18 07:04 UTC
Oh, and I forgot, the IE9 platform build demoed at MIX10 (which is a later build than the one currently available on the web) blew away Chrome in HTML 5 video rendering, both in speed and handling of opacity (Chrome’s stuttered all over the place for hi-def HTML 5 video and didn’t handle opacity at all (except for the border surronding the video)). So again, this isn’t “just catch up”.
considering HTML 5 is 10 years from ratification, supporting the most popular tags is good enough at the moment.
what took them so long?
They’ve shown themselves to be pretty nimble, once they pull their heads out of their rears. This is going to get very interesting in the next few years. Everyone will benefit.