Possibly because four and a half hours is the best case scenario. That's playing without sound, without Wi-Fi, and without 3D.
Actually, the article's really confusing on the "without 3D" bit since it says they played without 3D "when the game allowed it" - but I think that's an editing mistake, and the "when the game allowed it" refers only to Wi-Fi being off, but I'm not sure.
In any case, if you turn on sound and turn on 3D - you know, like most people will probably do, given that the 3D is kind of the point - you get closer to three hours battery life.
Which may not be exactly horrible, but it's not exactly good. That's actually worse than the PSP, which is somewhat notorious for horrible battery life.
Even with all the "populist" tea party members in control of Congress? Don't you think they're going to look out for us Americans? Surely they believe in privacy.
They believe in privacy from the government. Any sort of privacy legislation is going to be seen as a "government interfering in private affairs." After all, the right solution to privacy concerns if the free market. If you don't like web sites spying on you, you can always browse a different Internet.
Kind of like exactly how they scuttled net neutrality.
Obama hasn't ended Don't Ask, Don't Tell. In fact, assuming I understand the law that was passed, it won't be over until Congress gives their final approval based on "proof that it will not adversely effect combat readiness" or something along those lines.
Gee, I wonder what the chances are that the current Congress will do that?
Well, let's see: There's absolutely no way the privacy legislation will make it through the current Congress. Plus, Obama only came up with this after the Democrats had already lost control of Congress. And, there's a very good chance that the copyright stuff will make it through Congress.
So, I think it's fairly safe to say that he does intend to implement the copyright stuff, and that he has no intention of allowing the privacy legislation to succeed.
It's politics, pure and simple. He's hoping to convince people who care about online privacy to vote for him next year without having to worry about actually living up to any promises he makes. He can just blame the Republican Congress.
Up until relatively recently you absolutely had to include whatever hacks were necessary to get IE6 running on your site
It's still true, to some extent. The webapp I'm working on was originally supposed to support IE6 because that's what the customer was using.
Unfortunately, infosec has declared that IE6 is a security risk and that having it installed on any machine that can connect to any network is forbidden.
So we tried to convince the customer to upgrade to version 7 while talking with infosec about allowing us to install IE6 on a virtual machine which has limited net access. (No, we can't just install the web app on the VM and use no network access.)
Since then, infosec has moved everyone to IE8, with the same "not on machines connected to any network" rule applying to IE7.
Our customer, on the other hand, is still using IE6...
I always knew those Mordac, the Preventer of Information Services cartoons from Dilbert where based in reality, but, yeesh. And I do understand why infosec forbids IE6, but sometimes, we really do need some wiggle room to work with customers.
They pop up a dialog box asking if you want to cancel or allow the package script, using wording that makes it overly confusing as to why you'd ever want to cancel.
I think it's something like "a script needs to run in order to determine if you can install this package. Do you want the script to run?"
I think later you need to provide admin rights to install software that requires it, and that requires a password.
In essence, it's very reminiscent of Microsoft's UAC. The stuff Apple made fun of in their ads.
I remember it: it was the second video game I ever played, and the first one in color. The first game was LADDER on the DEC Rainbow. (I would have been like four at the time.)
Unfortunately there don't seem to be any DEC Rainbow emulators out there, meaning that all those old Rainbow games (...both of them...) are lost to time, I guess.
If it makes you feel better, the first thing it did on Mac OS X was the exact same thing - just with the beachball instead of the Windows (Not Responding) thing.
I just figured more people would know the Windows behavior than the Mac OS X beachball.
Meanwhile, the people who knew how to design games were off doing stuff like 356/2 Days on the DS.
358/2 Days was only published by Square-Enix, the actual development was fobbed off to some third party developer. Same with DragonQuest IX, I think. (Wikipedia says it was developed by Level 5.)
While I'm looking that up, I might as well pull up the developers for 0.00207175926 hertz. Uh, thanks, Google. 358/2 Days was developed by h.a.n.d and, again, only published by Square-Enix.
Pretty much if you try and name a good Square-Enix game in the last year, you'll find that it was only published by Square-Enix but actually developed by someone else. And I say that as a bit of a Square-Enix fanboy.
Chrome is better in just about every way with the exception of extensions. There are basically two killer features that work better under Firefox than Chrome: script blocking (NoScript) and ad blocking (AdBlock Plus). There are ad blocking extensions for Chrome, but they don't work quite as well as AdBlock Plus does.
There is no real equivalent to NoScript for Chrome. There are a bunch of things that kinda provide script blocking functionality, but nothing that's anywhere near as good as NoScript.
Beyond that it's much faster and more memory efficient. It doesn't like being left open long periods of time, though. I can get away with leaving Firefox open for like a week or so, Chrome pretty much demands that you kill it and restart it every day. Not really a huge deal.
The only thing I really miss in Chrome is NoScript. The ad blocking is mostly good enough.
When I tried the new beta, the first thing that happened was that it popped up a "welcome" page touting how fast the new beta was.
Then it froze long enough to get Windows to mark it (Not Responding) in the title bar. (I reenabled the menu while using an earlier beta so I guess I'm missing out on "tabs in title bar." Somehow, I don't care.)
To their credit, it doesn't always do this, but it does it enough to be annoying. I don't care how fast Firefox can run JavaScript - really all I care about is that running JavaScript doesn't make the browser completely non-responsive.
Wouldn't be the first time: they've already released FFX 2.
Apparently it was better than FFX, but I never got around to playing it. Somehow learning that they focused the game on the two most annoying characters killed it for me. The idea of having to listen to more Yuna is just... not appealing.
They shoulda just called it "Final Fantasy 14: We did better."
FFXIV already exists. Actually, never mind that. Given that I played it on launch, I'd rather forget that it exists. In fact, let's all just pretend there is no FFXIV, and that they've already done their "giant UI overhaul" and it changed nothing.
I mean, there is no FFXIV. Deep breaths, deep breaths...
I think the number one thing about XIII that is really awful is it's absurdly linear gameplay; unlike any other FF, there is no "freedom" to do anything other than follow a single path through the zones - you can't even take two routes to the same place.
At one point, the tutorial mentions that since you only have two party members, you should avoid tougher fights and come back when you're back up to three party members. Of course, being FFXIII, you can't ever backtrack to that point, and you have to fight the tougher monsters to progress anyway.
I'm not really sure what my point is, I guess that I think at one point it was going to be less linear, but they ran out of time or something.
I did actually complete the game, so I'm not really expecting anything amazing out of a sequel to it. If anything it more calls for a prequel to explain what actually happened between Cocoon and Pulse. Throughout the game you get this sense that there's this really amazing world here - that you're completely forbidden from seeing.
b) click to specifically allow an application to access them, too fucking bad if something bad happens.
You don't click to allow it. It just automatically happens because it's included amongst the various "basic settings" that you're not allowed to disable. Essentially you're told "do you want to allow this app access to your information" and that covers a shitload of stuff, which has been applied retroactively: your "basic information" (name), your profile information (random BS that you've "liked," basically), your contact information (this story), all pictures of you (includes pictures you're tagged in), and your friend's information.
All of this is required for every single application, and none of it can be disabled. The only way around this is to simply never use Facebook applications. Which works for me as a solution, I guess, but I'd rather be allowed to opt out and use certain apps that have no legitimate reason to access my contact information.
Even then it doesn't matter. Contact information currently appears limited to apps you install, but basically everything else on that list is also covered by any apps your friends install. So if your friend installs an app, that app automatically has access to everything an app you use does. You can opt out of this, by turning off each and every category individually.
Again, contact information being given to third party apps is retroactive. I never gave permission, yet for some reason, they just silently added it without any notification that I can recall. I certainly wasn't re-asked to give permission on loading an app.
What HTML5 clearly needs is a fullscreen button which can't be activated by a script, but which maybe can be styled with CSS.
It needs some form of fullscreen function. I think that the popup blocking rules that Firefox uses are adequate to prevent abuses: only allow fullscreen on a user-initiated event that's intended to enable functionality, specifically "mouse click or key press," and do something to make it clear the browser has transitioned to fullscreen mode. I've posted that before, just not here, because it's not really related to my main point.
I suppose I should have known better to say "JavaScript" instead of just "embeddable control" or something on Slashdot, I just don't really want to talk about why implementing a fullscreen button should be possible without really any security holes.
Where doesn't it work?
Safari (at all, as far as I can tell), and Mac OS X, where it's Shift-Command-F in Chrome and Firefox. F11 under Mac OS X is the hotkey for the Dashboard.
If by "complete mess" you mean "usable today", sure. Exactly which part of it is a mess?
As I understand it, there are multiple APIs that can be used depending on which graphics driver you're using. Specifically I think it's nVidia being as asshole and doing it a different way than everyone else, but I'm not sure, since I've never tried. I just know that someone explained it somewhere and there was a good reason that Adobe didn't bother supporting hardware acceleration under Linux.
And which source would that be? Is it actually a raw source?
Sort of: it's a FRAPS recording, which means it first has to be converted to YUV and is ever-so-slightly lossy. But it's the same original lossless source for all three encodes. And Theora always comes out looking horrid (relatively speaking) at identical bitrates.
The important thing is that it's not a transcode from an original H.264 source, which a lot of people wind up doing and then comparing the results, and discover that amazingly enough encoding something twice looks worse than encoding it once.
I know H.264 has some sort of proprietary ties, but they're pretty weak, and introducing something completely new (instead of, say, enhancing and throwing their weight behind Ogg/Vorbis, which in itself would be somewhat irresponsible if less so) seems really crazy.
My understanding is that its patents that sink H.264 on the "open" front. Specifically the lack of open licensing terms.
Of course, my understanding is that Theora and VP8 infringe on a subset of the H.264 patents anyway, although no one is entirely sure which ones and no one really wants to go looking.
Incidentally, WebM uses Vorbis for its audio, so in a way, Google is backing Vorbis. Just not Ogg, instead they've invented a new container format that I think is either Matryoshka directly or a modified version of it.
This got modded up? This is just completely wrong on all levels.
Bullshit. Chrome has always supported Theora, as far as I can tell, and Firefox is about to support WebM. In fact, IE is going to support WebM soon, which means by this time next year, Safari will be the only HTML5-compliant browser without H.264.
You obviously mean without WebM, and that's all nice, but like you say yourself, that's next year. My post is about right now, and right now, if you want to use HTML5 video, you need to do three encodes. Two if you're willing to put up with Theora, but Theora looks like ass.
Flash forbids allowing ActionScript to fullscreen, either.
But it doesn't forbid fullscreen entirely. Since there are half a million Flash apps that do fullscreen right now and telling people to just fullscreen their browser when they're used to just clicking the little button below the video is a nonstarter. And F11 doesn't work for all browsers on all OSes.
Speaking of H.264, I've got an H.264 decoder in hardware, in my fucking video card. Where is that feature in Flash?
Standard as of Flash 10 for Windows, and Flash 10.1 for Mac OS X. Since hardware decoding in Linux is a complete mess, who knows when it'll be available under Linux. Wait, didn't you just claim I didn't bother looking up simple facts? This isn't exactly unknown.
So what you're saying is you suck at encoding?
Unless there's a hidden "--suck=no" option in ffmpeg2theora, creating a Theora file at equivalent bitrate from the same source to either WebM or H.264 looks horrid. And, yes, ffmpeg2theora is just a frontend to libtheora, so it's not just a random crappy Theora implementation, it uses the official implementation. As far as I can tell, there are no quality options to trade off encoding time for a better encode. Note that the "video quality" flag in ffmpeg2theora is actually a shortcut to predefined bitrates, as far as I can tell.
I'm not sure, because as I've also mentioned somewhere, the Theora tools are completely horrible, and Xiph apparently has no interest in improving the situation.
So if there's some magic way to make Theora not look like crap, I'm all ears. As far as I can tell, WebM is miles ahead in terms of both tools to create them and in quality.
Using Theora alone is enough to support all the WEBM browsers in all versions
So what? WebM is better, if you have to encode to H.264 and Theora anyway (and you do, since despite what someone else said basically no one is going to have a Theora codec installed for Safari), you might as well throw WebM in as well, since it produces much better results than Theora.
The prohibition against script initiated fullscreen has been removed from the HTML5 spec-- too bad from a security perspective, since a rogue page can take over your screen and totally give you a trojan UI
Bullshit. While that's theoretically true, I challenge anyone to pull that off in a way that isn't immediately obvious. Plus, Flash has had full screen support for ages, and no one's managed to pull of any sort of trojan UI using Flash.
Theora might not be amazing, but it's perfectly reasonable compared to codecs of a similar vintage. It usually beats xvid at the same rates.
Too bad Theora is up against H.264 in the HTML5 battle and not Xvid, then.
And since I've encoded multiple videos in all three formats from a single original source and compared the results, it's fairly safe to say the H.264 looks best, followed by VP8, followed (way behind) by Theora. Also, Theora's other major issue is the lack of tools to encode to it. I wound up having to encode to lossless and then use ffmpeg2theora to get a Theora video, rather than going straight from source, although that's because I'm using AviSynth to create the videos in the first place.
I will say that WebM produces results that are good enough. Once Firefox 4 gets released, I intend to limit my encodes to only H.264 and WebM, since while H.264 is better, WebM is good enough for streaming web video.
I must have edited it out, but at one point I had mentioned that only Firefox 4 supports WebM. Current non-beta versions of Firefox only support Theora.
Plus I'm fairly certain that some of the smaller open source browsers (Konqueror, for example) only support Theora. But it's Firefox 3.6 that really kills WebM. Technically all you need is H.264/Theora in order to cover all browsers, but if you're already encoding multiple times you might as well throw WebM in there anyway, since it looks much better than Theora.
Back? Have you ever tried using HTML5 video? It's completely fucking useless.
No, really, it is. OK, first off, we have the codec issue. If you want to support all browsers, you need to encode to the following formats: H.264+AAC, VP8+Vorbis, and Theora+Vorbis. You're stuck with all three if you want to hit all browsers.
Then there's the part where the HTML5 spec forbids allowing JavaScript to fullscreen the video. Which means that you're stuck with either using the lousy solution YouTube uses (blow up the video to screen size, and assume the user can figure out how to fullscreen their browser on their own), or just dropping the feature all together.
Both suck. Users are used to being able to fullscreen the video, and they do NOT want to jump through the two-step hoop just to get fullscreen video.
Of course, most browsers allow the user to fullscreen the video on the context menu. But that's still really a two-step process: right click on the video, and then click on "Full screen." And to add insult to injury, most HTML5 video toolkits manage to block this option anyway by the way they generate their UI. (Including YouTube, in fact.)
So instead, you just use H.264 and a Flash-based player. Now you hit every major browser including IE, you don't have to encode your video three fucking times, and you don't have to have continuously explain the hoops required to fullscreen the video.
But what all this also means is that by ditching H.264, Google really doesn't lose anything anyway: if you were trying to support more than just Chrome and Safari with HTML5, you were already encoding to at least Theora anyway. So all this does is mean that Chrome will now be stuck with the same crappy, blurry Theora video you already had to encode to anyway to support Firefox. Or maybe, if they're lucky, they'll get the WebM video, which while worse than H.264 at the same bitrates, is still better than Theora.
It looks like that survey is missing two important questions:
1. Would you have bothered to answer this thing if we hadn't bribed you with something?
2. At what point did your eyes glaze over and you started answering randomly to get the thing over with?
I mean, shit, it's a 43-question survey, and that's ignoring things like the 11-part Q7 and the demographic questions. I don't know if anyone's done an actual study on how long a survey can be before the participants stop caring about providing accurate answers, but I think this survey probably exceeds that. I know that, personally, I'd have stopped caring somewhere around the never-ending question 7.
Slashdot Mirror
Possibly because four and a half hours is the best case scenario. That's playing without sound, without Wi-Fi, and without 3D.
Actually, the article's really confusing on the "without 3D" bit since it says they played without 3D "when the game allowed it" - but I think that's an editing mistake, and the "when the game allowed it" refers only to Wi-Fi being off, but I'm not sure.
In any case, if you turn on sound and turn on 3D - you know, like most people will probably do, given that the 3D is kind of the point - you get closer to three hours battery life.
Which may not be exactly horrible, but it's not exactly good. That's actually worse than the PSP, which is somewhat notorious for horrible battery life.
Even with all the "populist" tea party members in control of Congress? Don't you think they're going to look out for us Americans? Surely they believe in privacy.
They believe in privacy from the government. Any sort of privacy legislation is going to be seen as a "government interfering in private affairs." After all, the right solution to privacy concerns if the free market. If you don't like web sites spying on you, you can always browse a different Internet.
Kind of like exactly how they scuttled net neutrality.
Obama hasn't ended Don't Ask, Don't Tell. In fact, assuming I understand the law that was passed, it won't be over until Congress gives their final approval based on "proof that it will not adversely effect combat readiness" or something along those lines.
Gee, I wonder what the chances are that the current Congress will do that?
Well, let's see: There's absolutely no way the privacy legislation will make it through the current Congress. Plus, Obama only came up with this after the Democrats had already lost control of Congress. And, there's a very good chance that the copyright stuff will make it through Congress.
So, I think it's fairly safe to say that he does intend to implement the copyright stuff, and that he has no intention of allowing the privacy legislation to succeed.
It's politics, pure and simple. He's hoping to convince people who care about online privacy to vote for him next year without having to worry about actually living up to any promises he makes. He can just blame the Republican Congress.
Up until relatively recently you absolutely had to include whatever hacks were necessary to get IE6 running on your site
It's still true, to some extent. The webapp I'm working on was originally supposed to support IE6 because that's what the customer was using.
Unfortunately, infosec has declared that IE6 is a security risk and that having it installed on any machine that can connect to any network is forbidden.
So we tried to convince the customer to upgrade to version 7 while talking with infosec about allowing us to install IE6 on a virtual machine which has limited net access. (No, we can't just install the web app on the VM and use no network access.)
Since then, infosec has moved everyone to IE8, with the same "not on machines connected to any network" rule applying to IE7.
Our customer, on the other hand, is still using IE6...
I always knew those Mordac, the Preventer of Information Services cartoons from Dilbert where based in reality, but, yeesh. And I do understand why infosec forbids IE6, but sometimes, we really do need some wiggle room to work with customers.
They pop up a dialog box asking if you want to cancel or allow the package script, using wording that makes it overly confusing as to why you'd ever want to cancel.
I think it's something like "a script needs to run in order to determine if you can install this package. Do you want the script to run?"
I think later you need to provide admin rights to install software that requires it, and that requires a password.
In essence, it's very reminiscent of Microsoft's UAC. The stuff Apple made fun of in their ads.
Anyone remember "SCRAM"?
I remember it: it was the second video game I ever played, and the first one in color. The first game was LADDER on the DEC Rainbow. (I would have been like four at the time.)
Unfortunately there don't seem to be any DEC Rainbow emulators out there, meaning that all those old Rainbow games (...both of them...) are lost to time, I guess.
If it makes you feel better, the first thing it did on Mac OS X was the exact same thing - just with the beachball instead of the Windows (Not Responding) thing.
I just figured more people would know the Windows behavior than the Mac OS X beachball.
Not trolling, but I may have been trolled: I've never actually played it past the opening.
I do use it, there are some corner cases where it fails to block ads that AdBlock Plus on Firefox does, mostly involving video sites.
For most browsing though, it works well enough. And it does (attempt) to prevent the ad from being downloaded.
If I don't restart it, the tabs will start to sad-tab after a couple of days.
...Actually, wait. I do know what's wrong with the tabs: Flash. This only happens after viewing Flash content.
So, Flash is wrong what's wrong with my Chrome.
Meanwhile, the people who knew how to design games were off doing stuff like 356/2 Days on the DS.
358/2 Days was only published by Square-Enix, the actual development was fobbed off to some third party developer. Same with DragonQuest IX, I think. (Wikipedia says it was developed by Level 5.)
While I'm looking that up, I might as well pull up the developers for 0.00207175926 hertz. Uh, thanks, Google. 358/2 Days was developed by h.a.n.d and, again, only published by Square-Enix.
Pretty much if you try and name a good Square-Enix game in the last year, you'll find that it was only published by Square-Enix but actually developed by someone else. And I say that as a bit of a Square-Enix fanboy.
Chrome is better in just about every way with the exception of extensions. There are basically two killer features that work better under Firefox than Chrome: script blocking (NoScript) and ad blocking (AdBlock Plus). There are ad blocking extensions for Chrome, but they don't work quite as well as AdBlock Plus does.
There is no real equivalent to NoScript for Chrome. There are a bunch of things that kinda provide script blocking functionality, but nothing that's anywhere near as good as NoScript.
Beyond that it's much faster and more memory efficient. It doesn't like being left open long periods of time, though. I can get away with leaving Firefox open for like a week or so, Chrome pretty much demands that you kill it and restart it every day. Not really a huge deal.
The only thing I really miss in Chrome is NoScript. The ad blocking is mostly good enough.
Oh, man, isn't that the case.
When I tried the new beta, the first thing that happened was that it popped up a "welcome" page touting how fast the new beta was.
Then it froze long enough to get Windows to mark it (Not Responding) in the title bar. (I reenabled the menu while using an earlier beta so I guess I'm missing out on "tabs in title bar." Somehow, I don't care.)
To their credit, it doesn't always do this, but it does it enough to be annoying. I don't care how fast Firefox can run JavaScript - really all I care about is that running JavaScript doesn't make the browser completely non-responsive.
Wouldn't be the first time: they've already released FFX 2.
Apparently it was better than FFX, but I never got around to playing it. Somehow learning that they focused the game on the two most annoying characters killed it for me. The idea of having to listen to more Yuna is just... not appealing.
They shoulda just called it "Final Fantasy 14: We did better."
FFXIV already exists. Actually, never mind that. Given that I played it on launch, I'd rather forget that it exists. In fact, let's all just pretend there is no FFXIV, and that they've already done their "giant UI overhaul" and it changed nothing.
I mean, there is no FFXIV. Deep breaths, deep breaths...
I think the number one thing about XIII that is really awful is it's absurdly linear gameplay; unlike any other FF, there is no "freedom" to do anything other than follow a single path through the zones - you can't even take two routes to the same place.
At one point, the tutorial mentions that since you only have two party members, you should avoid tougher fights and come back when you're back up to three party members. Of course, being FFXIII, you can't ever backtrack to that point, and you have to fight the tougher monsters to progress anyway.
I'm not really sure what my point is, I guess that I think at one point it was going to be less linear, but they ran out of time or something.
I did actually complete the game, so I'm not really expecting anything amazing out of a sequel to it. If anything it more calls for a prequel to explain what actually happened between Cocoon and Pulse. Throughout the game you get this sense that there's this really amazing world here - that you're completely forbidden from seeing.
b) click to specifically allow an application to access them, too fucking bad if something bad happens.
You don't click to allow it. It just automatically happens because it's included amongst the various "basic settings" that you're not allowed to disable. Essentially you're told "do you want to allow this app access to your information" and that covers a shitload of stuff, which has been applied retroactively: your "basic information" (name), your profile information (random BS that you've "liked," basically), your contact information (this story), all pictures of you (includes pictures you're tagged in), and your friend's information.
All of this is required for every single application, and none of it can be disabled. The only way around this is to simply never use Facebook applications. Which works for me as a solution, I guess, but I'd rather be allowed to opt out and use certain apps that have no legitimate reason to access my contact information.
Even then it doesn't matter. Contact information currently appears limited to apps you install, but basically everything else on that list is also covered by any apps your friends install. So if your friend installs an app, that app automatically has access to everything an app you use does. You can opt out of this, by turning off each and every category individually.
Again, contact information being given to third party apps is retroactive. I never gave permission, yet for some reason, they just silently added it without any notification that I can recall. I certainly wasn't re-asked to give permission on loading an app.
What HTML5 clearly needs is a fullscreen button which can't be activated by a script, but which maybe can be styled with CSS.
It needs some form of fullscreen function. I think that the popup blocking rules that Firefox uses are adequate to prevent abuses: only allow fullscreen on a user-initiated event that's intended to enable functionality, specifically "mouse click or key press," and do something to make it clear the browser has transitioned to fullscreen mode. I've posted that before, just not here, because it's not really related to my main point.
I suppose I should have known better to say "JavaScript" instead of just "embeddable control" or something on Slashdot, I just don't really want to talk about why implementing a fullscreen button should be possible without really any security holes.
Where doesn't it work?
Safari (at all, as far as I can tell), and Mac OS X, where it's Shift-Command-F in Chrome and Firefox. F11 under Mac OS X is the hotkey for the Dashboard.
If by "complete mess" you mean "usable today", sure. Exactly which part of it is a mess?
As I understand it, there are multiple APIs that can be used depending on which graphics driver you're using. Specifically I think it's nVidia being as asshole and doing it a different way than everyone else, but I'm not sure, since I've never tried. I just know that someone explained it somewhere and there was a good reason that Adobe didn't bother supporting hardware acceleration under Linux.
And which source would that be? Is it actually a raw source?
Sort of: it's a FRAPS recording, which means it first has to be converted to YUV and is ever-so-slightly lossy. But it's the same original lossless source for all three encodes. And Theora always comes out looking horrid (relatively speaking) at identical bitrates.
The important thing is that it's not a transcode from an original H.264 source, which a lot of people wind up doing and then comparing the results, and discover that amazingly enough encoding something twice looks worse than encoding it once.
I know H.264 has some sort of proprietary ties, but they're pretty weak, and introducing something completely new (instead of, say, enhancing and throwing their weight behind Ogg/Vorbis, which in itself would be somewhat irresponsible if less so) seems really crazy.
My understanding is that its patents that sink H.264 on the "open" front. Specifically the lack of open licensing terms.
Of course, my understanding is that Theora and VP8 infringe on a subset of the H.264 patents anyway, although no one is entirely sure which ones and no one really wants to go looking.
Incidentally, WebM uses Vorbis for its audio, so in a way, Google is backing Vorbis. Just not Ogg, instead they've invented a new container format that I think is either Matryoshka directly or a modified version of it.
This got modded up? This is just completely wrong on all levels.
Bullshit. Chrome has always supported Theora, as far as I can tell, and Firefox is about to support WebM. In fact, IE is going to support WebM soon, which means by this time next year, Safari will be the only HTML5-compliant browser without H.264.
You obviously mean without WebM, and that's all nice, but like you say yourself, that's next year. My post is about right now, and right now, if you want to use HTML5 video, you need to do three encodes. Two if you're willing to put up with Theora, but Theora looks like ass.
Flash forbids allowing ActionScript to fullscreen, either.
But it doesn't forbid fullscreen entirely. Since there are half a million Flash apps that do fullscreen right now and telling people to just fullscreen their browser when they're used to just clicking the little button below the video is a nonstarter. And F11 doesn't work for all browsers on all OSes.
Speaking of H.264, I've got an H.264 decoder in hardware, in my fucking video card. Where is that feature in Flash?
Standard as of Flash 10 for Windows, and Flash 10.1 for Mac OS X. Since hardware decoding in Linux is a complete mess, who knows when it'll be available under Linux. Wait, didn't you just claim I didn't bother looking up simple facts? This isn't exactly unknown.
So what you're saying is you suck at encoding?
Unless there's a hidden "--suck=no" option in ffmpeg2theora, creating a Theora file at equivalent bitrate from the same source to either WebM or H.264 looks horrid. And, yes, ffmpeg2theora is just a frontend to libtheora, so it's not just a random crappy Theora implementation, it uses the official implementation. As far as I can tell, there are no quality options to trade off encoding time for a better encode. Note that the "video quality" flag in ffmpeg2theora is actually a shortcut to predefined bitrates, as far as I can tell.
I'm not sure, because as I've also mentioned somewhere, the Theora tools are completely horrible, and Xiph apparently has no interest in improving the situation.
So if there's some magic way to make Theora not look like crap, I'm all ears. As far as I can tell, WebM is miles ahead in terms of both tools to create them and in quality.
Using Theora alone is enough to support all the WEBM browsers in all versions
So what? WebM is better, if you have to encode to H.264 and Theora anyway (and you do, since despite what someone else said basically no one is going to have a Theora codec installed for Safari), you might as well throw WebM in as well, since it produces much better results than Theora.
The prohibition against script initiated fullscreen has been removed from the HTML5 spec-- too bad from a security perspective, since a rogue page can take over your screen and totally give you a trojan UI
Bullshit. While that's theoretically true, I challenge anyone to pull that off in a way that isn't immediately obvious. Plus, Flash has had full screen support for ages, and no one's managed to pull of any sort of trojan UI using Flash.
Theora might not be amazing, but it's perfectly reasonable compared to codecs of a similar vintage. It usually beats xvid at the same rates.
Too bad Theora is up against H.264 in the HTML5 battle and not Xvid, then.
And since I've encoded multiple videos in all three formats from a single original source and compared the results, it's fairly safe to say the H.264 looks best, followed by VP8, followed (way behind) by Theora. Also, Theora's other major issue is the lack of tools to encode to it. I wound up having to encode to lossless and then use ffmpeg2theora to get a Theora video, rather than going straight from source, although that's because I'm using AviSynth to create the videos in the first place.
I will say that WebM produces results that are good enough. Once Firefox 4 gets released, I intend to limit my encodes to only H.264 and WebM, since while H.264 is better, WebM is good enough for streaming web video.
I must have edited it out, but at one point I had mentioned that only Firefox 4 supports WebM. Current non-beta versions of Firefox only support Theora.
Plus I'm fairly certain that some of the smaller open source browsers (Konqueror, for example) only support Theora. But it's Firefox 3.6 that really kills WebM. Technically all you need is H.264/Theora in order to cover all browsers, but if you're already encoding multiple times you might as well throw WebM in there anyway, since it looks much better than Theora.
Back? Have you ever tried using HTML5 video? It's completely fucking useless.
No, really, it is. OK, first off, we have the codec issue. If you want to support all browsers, you need to encode to the following formats: H.264+AAC, VP8+Vorbis, and Theora+Vorbis. You're stuck with all three if you want to hit all browsers.
Then there's the part where the HTML5 spec forbids allowing JavaScript to fullscreen the video. Which means that you're stuck with either using the lousy solution YouTube uses (blow up the video to screen size, and assume the user can figure out how to fullscreen their browser on their own), or just dropping the feature all together.
Both suck. Users are used to being able to fullscreen the video, and they do NOT want to jump through the two-step hoop just to get fullscreen video.
Of course, most browsers allow the user to fullscreen the video on the context menu. But that's still really a two-step process: right click on the video, and then click on "Full screen." And to add insult to injury, most HTML5 video toolkits manage to block this option anyway by the way they generate their UI. (Including YouTube, in fact.)
So instead, you just use H.264 and a Flash-based player. Now you hit every major browser including IE, you don't have to encode your video three fucking times, and you don't have to have continuously explain the hoops required to fullscreen the video.
But what all this also means is that by ditching H.264, Google really doesn't lose anything anyway: if you were trying to support more than just Chrome and Safari with HTML5, you were already encoding to at least Theora anyway. So all this does is mean that Chrome will now be stuck with the same crappy, blurry Theora video you already had to encode to anyway to support Firefox. Or maybe, if they're lucky, they'll get the WebM video, which while worse than H.264 at the same bitrates, is still better than Theora.
I would have donated, but apparently I'm not notable enough, and so my donation was speedily deleted.
It looks like that survey is missing two important questions:
1. Would you have bothered to answer this thing if we hadn't bribed you with something?
2. At what point did your eyes glaze over and you started answering randomly to get the thing over with?
I mean, shit, it's a 43-question survey, and that's ignoring things like the 11-part Q7 and the demographic questions. I don't know if anyone's done an actual study on how long a survey can be before the participants stop caring about providing accurate answers, but I think this survey probably exceeds that. I know that, personally, I'd have stopped caring somewhere around the never-ending question 7.