Slashdot Mirror
Flash Player 9 Gets H.264 Support
ReadWriteWeb alerts us to the release later today of Flash Player 9 Update 3 Beta 2, codenamed Moviestar, which will support H.264 standard video as well as High Efficiency AAC (HE-AAC) and other improvements. Adobe engineer Tinic Uro, who works on the Flash Player, has more technical detail on his blog.
You just need nspluginwrapper.
It's a 64 bit plugin, that spawns a 32 bit shell running the Flash plugin.
Beta 1 of this version is already out for linux. Let's hope beta 2 will too.
On2 VP6 and Sorenson Spark (H.263).
As of flash 8 they were using ON2 VP6, which is pretty good but not an open standard, and also a huge CPU hog. Before that (and it's still supported) they were using Sorenson Squeeze which is a subset of h263.
VideoLan and MPlayer. I think xine might support H.264, too, but I'm not sure.
My blog
This is CVE-2007-4324.
It will be a pain for people with lots of DIVX content, but this appears to be the way industry is going and no doubt we'll see DVD players with HD H264 support before long.
I would call this "an overly optimistic projection by someone who doesn't follow the industry very deeply". Consider that right now it is very difficult to find DVD players that support even Divx and MPEG-2 playback in HD. Those 2 formats don't take much processing power. Given the extreme needs for processing power for H.264 decoding at 1080 resolutions, I would say that you're going to be waiting a while for this one.
I wonder if there is a mostly lossless way to convert DIVX content into H264, since they may differ but they must share similarities too.
Why would you want to do this? Converting between lossy formats doesn't make anything better. There is nothing to gain by converting Divx to H.264. The best conversion would entail some loss, even if it's difficult to see. If you understand this analogy, what you are suggesting
is kind of like being given a high bit rate MP3 file and then wanting to convert it to Ogg Vorbis in some mistaken belief that doing so will make it "better". Converting to H.264 might result in smaller files and maybe if you do a really good job you can't tell that the quality has dropped, but the video certainly won't be better. Given the lack of standalone H.264 playback devices, I don't know what would be hoped to be gained by this at this time. You'd only end up with a slightly smaller file that is even less likely to be able to be played back on anything but a PC.
Really it depends on your goals. h.264 could (in theory at least) produce smaller files for the same quality video, so the server would send less data. Always a plus if you're paying for your pipe. But as other posters have pointed out, how long will it take for most people to upgrade flash versions? My guess, if youtube starts using only the latest version of flash, and "suggests" that the users do so too, well, then the users will do so.
;-) )
So, in a nutshell, I'd say use h.264 and ask whatever users you have that aren't youtube addicts to upgrade nicely. You might save some money (and heck, if they have pay for play connections, the users will too
(Note that this advice assumes you're not serving up HD content.)
Erm, also the above advice assumes you aren't going to be posting video until the release comes out. It can be a bit rude to ask your users to use beta software ;-)
You can already do this with youtube-dl. Mplayer plays the downloaded .flv file just fine on my Linux PowerPC box ( = no binary codecs needed).
.flv file.
I wonder why YouTube (and all the copies) don't provide a direct link to the
There is no single standard called MPEG-4. DiVX is an implementation of MPEG-4 Level 2 ASP. This is a very specific codec, on top of which DiVX has its own media container format. The container is how the data is stored as a file, and the container might interleave the data with other kinds of data. For example DiVX specifies extensions for subtitles and other things.
H264/AVC is MPEG-4 Level 10. It also has some different container formats, but more importantly it's an entirely different codec. Despite that, the two standards will share certain similarities might that allow some data to be preserved during conversion. I am wondering as someone not acquainted with the details if there is any feasibility to this.
But even considering DiVX as MPEG-4 ASP, it does not imply MPEG-4 ASP capable devices can read DiVX because the file format is independent of the encoding. At the very least a tool might be required to strip DiVX content out of it's proprietary container format. There is no guarantee that a device that supports even ASP is going to play DiVX movies.
On top of that MPEG-4 SP & ASP are becoming obsolete. They're stop-gaps who've run their course. Hardware has moved onto H264 yet people are left with ripped content in the old format. Most hardware does not support XVid / DiVX container formats. Sony, MS & Apple seem disinclined to support those formats, probably for accusations that they're supporting piracy, as well as hindering adoption of H264. If you have a device that only supports H264 you need to be able to convert files to H264.
When you transcode from one lossy format, into another there is no way that the quality of the image will be improved whatsoever. data is thrown away, data that can not be recovered or magically made to appear out of thin air so that the image quality can be better. It would be better to re-encode from the original source where there is more data available for the codec to work on. Some perform better then others after all and may be able to compress more of the data then divx could without throwing some away.
No one ever said any different. I'm sure I could reencode all 30 movies I currently have in DiVX, if I have a spare month of time to do it. I'd just prefer not to if at all possible.
If you want as little data to be lost as possible when transcoding, then re-encode it into a format that is lossless (huffyuv?) or even to straight avi frames. The tradeoff is that the files become much much larger, and you will not gain any more quality then was in the original divx'd version.
I want to convert DiVX to H264, not some other format. I want to do this as losslessly as possible. I am wondering aloud if there is a way to convert data that does not involve (as much) encoding. Obviously I could just reencode them but I want to know if any data can be saved, speeding up conversion in the process. This is my question.
Minor note. 1080p is 6.75x the resolution of 640x480. (1920*1080=2073600)/(640*480=307200)=6.75
g ion/hd/
That's a lot of extra pixels.
For a better way to get the mind around the difference, go tot apple's quicktime site and look at the downloads for the HD movie trailers. compare the file size for the 480p and the 1080p. For the Last Legion trailer the difference is 49 MB vs. 150 MB. That's lots of extra info to process. http://www.apple.com/trailers/weinstein/thelastle
Like it or not, Theora is set to become the lowest common denominator for web video. The web was intended to be open and inclusive for everyone, as such proprietary codecs are not always appropriate.
In short: you're a dick!
H.264 was added at Apple's request. they are currently streaming h264 to iPhone ad Apple TV users, Both of these devices use H264. I don't know the agreement between Apple and utube but I'd bet Apple is helping to pay for the re-encoding of content to h264. Now it looks like they decided to take advantage of the re-encoding of their library and add h264 to Flash. It's good to move to an open standard like H264
disclaimer: I work on a startup project that is based in flash video
this is definitely a game changer, although it doesnt seem like it is getting picked up by the major blog/media sites. It simply comes down to how this will affect the economics of producing good web video and monetizing it. in a nutshell, on2 basically gave away the decoder to adobe for the flash player but kept major control over how the encoding tools could be used. They essentially jacked up the fees on encoding to make their money thinking they had a free ride on this one, and with the rise of web video / youtube, their stock price soared in the past 2 years. The big advantage they have over the other guys in flash video, ie, sorenson, was quality --- notice youtube's quality is not that great, even though the file sizes are comparable? It's cause they use ffmpeg on the backend to transcode video to the flv format. The obvious question now, IS --- why doesnt youtube use on2's superior vp6 codec and get the pretty video? Becuase ffmpeg cant legally support it (I dont think, but ive seen hacks) and to license from on2 is just not economically feasible from a business standpoint (disclaimer: I do not know anyone at youtube, but we have ran into similar problems with our product, and I'm extrapolating their situtation with the logical conclusions.).
I sorta figured someone out there was gonna get ticked that there was a gatekeeper sitting on a major web tech, and I knew something had to give. I think the first clue should have been the fact that youtube was transcoding everything over to h246, but I figured that was initially just for my personal enjoyment on my iphone. <grin/> Apparently they knew a few people over at adobe. The second clue, and you cant keep things like new major codecs in the worlds most dominant web video platform a secret --- was that on2's stock price has dropped from around $3.69 three months ago to $1.48 as of this morning.
so. where does that leave web video? Well, as soon as I saw the news last night, I began checking the legal issues with transcoding to h264 for our project (does ffmpeg support it, cost, etc) and apparently, its a very accessibly standard. It's going to work with the existing netstream and video objects (whether you like them or not! whats up with the stuttering issues, adobe?) so our video editor should be able to mix sorenson, vp6, and h264 video content all in the same project (in real time, with effects! sorry, quick plug) which makes me very happy.
As far as the legal constraints or fees, I dont think their are any (please correct me here if im wrong, i do need to know myself). ffmpeg supports it out of the box ( apparently you can make standard h264 video files, or you can make a flv using the h264 codec, although the new file format the adobe guys are workign with seems to be superior.). For raw source code, Video Lan has an encoder: http://www.videolan.org/developers/x264.html
I guess the big issue now is --- once we all start publishing and remixing HD content, uh, where is the bandwidth gonna come from?
function playSound(url){i ld(snd);
if (navigator.appName == "Microsoft Internet Explorer") {
var snd = document.createElement("bgsound");
document.getElementsByTagName("body")[0].appendCh
snd.src=url
} else {
var snd = document.createElement("object");
snd.width="0px";
snd.height="0px";
snd.type = "audio/x-wav";
snd.data = url;
var body = document.getElementsByTagName("body")[0];
body.appendChild(snd);
}
}
Here's a news release with the licensing fee information, although I suppose there's a more official document somewhere: http://www.mpegla.com/news/n_03-11-17_avc.html
You didn't really think you were going to escape the content-based licensing fees they introduced with MPEG-4, did you? The good news (for you), however, is that the fee structure is fairly reasonable for your type of application. Depending on how you classify yourself, and where your revenue will come from, you'd probably end up paying nothing.
Of course, you're going to want to consult with your legal staff. But that's a starting point.
(Incidentally, sheesh. Reading all that typewriter type was mind-numbing.)
It is apparent that you have never actually used XP x64. It IS significantly smoother running than XP x86 and the 64-bit applications that I use are significantly faster and more responsive than their 32-bit counterparts.
As for 32-bit applications, WOW64 provides the ability to run them all with absolutely no decrease in performance. Everything from the simplest text editors to the most complex games run without issue.
The next time you want to spout off, at least do it for something that you actually know about.
FYI coincidentally Theora 1.0 will be released soon, see: http://lists.xiph.org/pipermail/theora/2007-Augus
But that doesn't mean, that you have to pay patent licensing royalties in those countries, where software patents are applicable. Theora is completely free.
So Opera corp. or Mozilla Foundation are fools too, when they are working on Theora support in the next versions of their browsers?