Slashdot Mirror
Streaming Media - Can Linux Keep Up?
raphinou asks: "I am really worried because Microsoft is making serious progress on streaming media fields. Realplayer is available to Linux users. Windows Media isn't. How do you want Linux to succeed on the desktop if there aren't any streaming players for it? If Microsoft can convince broacasters to use Windows Media, they'll again control the desktop. It really makes me think about the Netscape thing. And I'm afraid we'll have the same end: RealNetworks dead. This is really the same thing: Microsoft is giving away for free what RealNetworks has to sell." What do you think? Have the BrowserWars become the StreamingMedia wars?
xanim supports at least one variant of H.263, but I think that right now there are a couple of variants on the file format. There's also H.263+.
.avi, I think (but won't swear to it!) that vic (one of the mbone tools) supports it also.
As far as using H.263 in a streaming context rather than in an
In the last year or so (post 2.2) hordes of companies have thrown themselves onto the Linux train. Do you really think we'd have the kind of support we have now if Linux stayed underground?
First of all there wouldn't be as many users (obviously), so the amount of information you could find online would diminish significantly. Everyone with a new Creative Labs sound card would be fsck'ed, and so would most people who want to use their 3d accelerators under Linux (NVidia & Matrox). The only way you could play Quake would be to use a 3Dfx card. And that would be about the *only* quality game you could play under Linux because I guarantee you that Loki wouldn't exist right now if Linux were still "underground".
-----------
"You can't shake the Devil's hand and say you're only kidding."
Who really cares all that much about streaming media?
... yet.
Maybe for some segments, but for those in entertainmetn, especially for those in the adult side of entertainment there is no humor in it at all.
One of our most profitable sites relies on streamed video and streamed audio to do it's thing and we have build sites for clients that serve hundreds of simultaneous audio/video feeds.
Just like porn jpegs helped drive faster modems and better video cards (not to mention bigger disk drives)...
Just like the desire for cybersex drove the rise of AOL and IRC....
The desire to watch pretty girls do bad things is driving streaming video.
The adult side of the web is, unlike the other sides consistently and very profitable. We often have uses for technology the rest of you can't see one for
Ken
--> Fight tyranny and repression.... read
You're right that rhe API needs to be in place before we can expect CODECs, whether binary or open source, to become available.
However, there *is* already a standard emerging for Linux - the "Video for Linux 2" (V4L2) API. This isn't in the release kernels yet (V4L is), but it is available, and there are drivers for it for most cards including the popular Bt848 based ones.
V4L2
One of the significant improvements in V4L2 is that it does support a CODEC API.
Mark Podlipec had had some success in getting companies to release propietary CODECs for xanim, given the standard he has established there, and I expect that an OS level API like V4L2 would with a bit of encouragement get a fuller set of CODECs released. Don't expect open source though, since the better CODECs represent significant intellectual property value for the commercial companties that developed them. But with the infrastructure in place, it'll be possible for an open source CODEC to emerge if people are interested in working on it.
Linux & OSS have a lot more people working on it that MS does
I was thinking of this today and my guess is that there actually are more people writing code for the Win32 platforms if you count the shareware writers (and you can't possibly leave them out). They produce some great programs as well.
Linux has the advantage, though, that not everyone has to start a project from scratch because all the code written before you is already there, and thus the wheel is being reinvented far less often. So the total number of programs for Linux is smaller but the good pieces are more likely to be kept alive, no matter what their original author goes to next.
Is there a standard for streaming media? If not, then one should be made, or followed. If MicroSofts format is good enough, let's use it. Let's make a Linux player for the same format. If its not good enough, let's make a windows-media-player thingie that can take care of Micro$ofts standard, and let's make our own in addition.
Personally I think that we should follow the already established standards - instead of reinventing the wheel time after time again. There is no need for a new standard, if the current is good enough.
--
"Rune Kristian Viken" - arcade@kvine-nospam.sdal.com - arcade@efnet
"Rune Kristian Viken" - http://www.nwo.no - arca
Have you tried compiling Apple's Darwin Streaming Server under Linux? It's open source, easily configurable, and Apple doesn't charge by the stream.
True, people are still waiting for a Linux QT client, but your fears about Micro$oft domination of streaming may be premature.
There are some IT types who will adopt MS streaming just because it comes from Redmond, but these folks are no different than the IT types who used to grasp at any solution IBM offered, because it came from IBM. We used to call them dinosaurs.
-Dave
It's an open standard, there are at least two open source implementations that I know of, and the quality/compression appears to be about the same as MPEG-4.
Maybe the WINE people should be making an extra special effort to target specific, needed applications (such as Media Player) and get them working asap.
I tried a build of WINE from last week and while a bunch of stuff did work properly, media player didn't.
Chris "Ng" Jones
cmsj@tenshu.net
www.tenshu.net
It would appear that the biggest problem is trying to get a corporation to open up their protocols, when they appear to have every interest in not doing it. Perhaps we, the Open Source community, should consider developing and GPL'ing our own standards for video/audio streaming and other protocols, with no interest in maintaining compatibility, but make sure that players are developed for every platform under the sun.
I know, developing a protocol and compression algorithms to rival the current competition wouldn't be even remotely easy. We might actually have to live with something that is somewhat substandard, but hopefully not so much that its unusable, until there is enough support to either pressure the competition to submit, or until we have perfected our formats to the point that they compete on even ground.
Most of our efforts are to cater to Microsoft products and protocols (samba, wine, etc). This buys us some time and helps to give linux a chance to get accepted in corporations since without compatibility with microsoft products, nobody would even give them a chance. But we must strive for a future where linux builds its own standards and microsoft is the one running to maintain compatibility with us, because it would be losing out if it didn't. This is where we will REALLY start to shine, since because of Microsoft's track record for spotting trends and running with them, theres a chance they might fall too far behind and lose significant market share in the process. This won't happen tomorrow, but if we plan right, it could be reality within the next 5 years.
Of course, the first step would be for someone to develop and advanced protocol and be willing to release it with no strings attached rather than sell it to a corporation for millions of $$$. I agree, this will be the hardest step. But it IS an important step, and will have to be taken.
-Restil
Play with my webcams and lights here
IMHO, there are only two solutions:
Each of these has their difficulties.
The free world has lots of brilliant people but...it is unfocused. To create a true interactive, streamming real-time audio/video experience is going to require some major work. Focused billiance.
Linux may manage to start getting to the level of Mac (at its highest) where most companies had to make a verions for it. Linux is getting closer but it still needs some muscle behind it. RH and the rest of the Linux IPO gang could be that muscle but they havn't doen anything yet. Those companies are the key to getting the big coroporations to do the dirty work for us. The same group holds the key to projects like Wine. They fight a battle with a closed stanard. After writing and maintaining device drivers for Win2k ever since it was called NT5, I can tell you that writing to a moving target is TOUGH. Some big money LinuxIPO companies could help these causes
-- soldack
That comment on the original posting rang a bell.
The author said that he was afraid that RN
will suffer the same 'death' destiny of Netscape
because Microsoft is giving away what Real
Networks is trying to sell.
In my many years using and promoting Linux
(and I have converted several people) I feel
the same about Linux. Whenever one thinks of
something that could generate some cash
(a program to sell) to subsidize one's
free projects (I have several myself) then
someone comes with a free alternative. So _we_
are doing the same and can't blame MS for
being evil because of this behaviour.
For many years, doing free software has been
a passion that I am sure many share. But the
fact is that at some point we all have to
put the bread on the table and that is nearly
impossible with Linux programming unless you
get a contract to do it.
Anyway that's my $0.04 (inflation, another
reason to earn money!). Flames will be
ignored.
Multicast will be very important in the near future. Without it, no server or server farm is going to be able to keep up, no matter how good the compression is. Every time sombody trys to webcast something vaguely popular with UDP, the server's network gets swamped instantly, and most people get choppy or no video.
I haven't tried using multicast under any OS yet. I suspect (hope) that *BSD and Linux are well ahead of NT/Windows in that area. Can anyone comment on that?
send flames > /dev/null
Only 'flamers' flame!
WMGK Windows Media Player
WMMR Windows Media Player
WSTW RealAudio
WJBR RealAudio
Y100 RealAudio
I sent a letter to WMGK, and will send a similar one to the others that use the Windows Media Player streams.
The letter:
Hopefully, someone can wake these guys up a bit.
-SsC
--
*kerchunk* *beep* "...Operator."
If you have a requirement that Linux supports streaming media, why don't you organize a streaming media project instead of writing a letter to slashdot?
-jwb
Your link didn't work but I have used QuickTime for linux. I agree that this has some hope. The only problem is that QuickTime seems pretty far behind the pack in getting content providers to use their system. Those that control content or content providers will pick the format. MS will push for their own system. The real question is what will AOL-Time push for. AOL has used MS technology (IE over netscape, even though they own netscape) but the merger may make them strong enough and brave enough to go their own way. Using MS's format puts the provider in MS's hands. They have to rely on MS to provide a high quality (or at least competitive) player and server that can reach the provider's target market. Right now MS has that so no one worries about putting all their eggs in MicroSoft Basket 2000. AOL may not want to use MS just so that they have some more control. They want all the eggs in AOL Everywhere 5. Plus, AOL and MS have cracked heads ever since MS started MSN, through the netscape purchase, and all the way to the Instant Message war.
The ball is in AOL's court and we are all waiting to see if Steve Case can indeed jump and slam dunk his competition.
-- soldack
How you people so quickly forget bringing about your own demise, and how quickly you defend companies who really aren't even your friend!
/. community would be the first people screaming against this sort of thing!
When Windows Media Player's streaming was first incarnated as Microsoft NetShow, MS worked with Xing to deliver players for damn near every single platform out there including Linux. In fact, I believe it was their second piece of software for Linux after the FrontPage extentions.
Now, we have everybody complaining about how MS is becoming more commonplace and RealPlayer is the only (large-market) solution still available under Linux.
Let me ask you this: have you ever tried to license a Real Media *SERVER* ? The licensing is absolutely ridiculous. Yes, I know they have the basic version but that's no good for any kind of commercial use. WMP's server-side application is included with IIS and has unlimited capacity. Hell, buying NT BackOffice (The entire operating system!) costs you half as much as supporting 100 streams of Real Media.
On top of that, WMP has better quality at the same bandwidth (at least for video). Just watch it or listen to it and it is not really a disputable argument. I'll give RealPlayer a chance too and say that it's more responsive when you want to skip around in a stream and it's way more fault tolerant.
Still, do you realize why you're complaining? You didnt support M$ tech when they tried to support *you*! I hate to have to defend them, but if the Linux/UNIX/Whatever people started driving and thriving in the streaming media market then we wouldnt look so second class to streaming media companies. M$ would probably still have a media player for Linux, and maybe some of their API's and CODECS would have opened up by now.
I am a bit annoyed with the defense of Real on this one. Getting the newest version of RealPlayer has always been delayed months (and close to 1 year on 5.0 and G2 -- they skipped 4.0 altogether), and until the release of at least 5.0 for Linux (which Real's web page makes nearly impossible to obtain) Real was an absolutely useless format if you wanted to do cross-platform video. Coupled with that, RealPlayer is a billion meg download that forces you to replace damn near every piece of multimedia software (at least on W32) -- I've stopped installing it on Windows. It absolutely is the most bloated, slow piece of garbage in the universe, and on top of that it (at least used to) gather up all of my personal info and beam it to RealNetworks!!
Please, stop supporting RealNetwork's trash! I would think that the
~GoRK
There is an Open standard for Streaming available for linux...both the specs and the server.
It's called Quicktime.
Don't start on the 'we don't have Soreneson thing either...because you don't need it.
What is needed is an open source *hinter* to avoid having to use Apple software if you don't want to.
Lemme elaborate:
For video, quicktime supports the following codecs, at various speeds/bandwidth:
Animation, BMP, Cinepak, Component Video, DV (NTSC & PAL), Graphics, H.261, H.263, Indeo 5 & 3.2, Indeo Raw, Motion JPEG A & B, Photo Jpeg, PNG Sorenson, TGA, Tiff and Video.
Now, out of that list, I know that either the source to these is available, or binary codecs for Xanim.
Sorenson, btw, is not the greatest codec for streaming in all cases. In fact, the final output is sometimes larger than Indeo 5.
Quicktime streams also support the following audio codecs:
IMA 4:1, 24bit Integer, 32 Bit Float & Integer,64 bit Float, ALaw 2:1, IMA 4:1, MACE 3:1 & 6:1, Meta Sound, Meta Voice, QDesign2, Qualcomm PureVoice and muLaw 2:1
Again, Linux support is there.
BTW, in case you didn't know, the audio eats up more Bandwith that the videostream...but I digress.
The Quicktime library needed for acutally reading tracks properly has been ported to linux an BSD already:; the source is available as well.
The Quicktime streaming sever has a tremendous advantage over all of the other competing technology here that a lot of folks (here and elsewhere) seem to miss:
There is no charge for streaming. You can serve as much as you want, as long as you want, to as many clients as you wish, for no charge.
MS doesn't (and won't) offer this and Real certainly will not as that is where their revenue stream lies.
Also of note: QTSS/DSS use open standard protocols for streaming...no funny stuff. RTP/RTSP over UDP and via HTTP. It also uses standard Session Description Files. It supports relaying as well.
What I suggest is that folks that are looking for a solution not recreate the wheel.
The combination of Indeo5 and IMA4:1 works *quite* well for streaming, assuming the peson putting the stream together knows what they are doing and is supported by Linux. The server is there as well.
What's missing is streaming support for a player, and a non-apple Hinter for encapsulating the stream.
Sorenson and QDesign is more hype than help...trust me on this. It really doesn't help 56kps modem connections (what does?) and for ISDN/Dual ISDN and better connections, the differnce between that combo and a Linux supported one becomes less of a big deal.
Where I work, we've spent a lot of time looking at this, as we build a Linux-based product that relies on Multimedia, and honestly, nothing out there is better than Quicktime.
The pieces are in place for streaming on Linux...they just need to be fitted together.
An Aside:
Believe it or not folks, Apple is more your friend than your enemy.
Why/how can I say this? Simple.
For anyone industrious enough to dive in, Apple is giving Unix-oriented coders a huge earning opportunity, as they will effectively be the first company in the World of History to bring Unix to the Desktop.
Read that again. Let me help. They will be the first company in the World of History to bring Unix to the Desktop.
What's worse, is they will be bringing to to what most of you seem to consider the most (ahem) stupid computer users on the planet.
Now that's a feat.
Anyway, like I said, the stuff is out there for supported cross platform streaming video...
-K
One day, you'll learn to watch what you post...
I've been using icecast for well over a year and it truly rocks. It uses a tiny amount of resources. L3enc produces excellent sound in a 24kbps stream especially if it's pre-encoded (as opposed to realtime encoding.)
I haven't tried connecting with WMP to my icecast server, however Netscape 4.7 comes with WinAmp, which allows users to connect to icecast streams without needing to install any additional software.
What might be considerred a disadvantage is that icecast uses TCP connections instead of UDP packets. If a TCP packet is dropped the music stops until the packet is re-sent and received. This causes the audio to stop (assuming it's not re-sent before the buffer runs out.) Realaudio inserts static when a UDP packet is not recieved but continues pretty much in realtime.
I said "might be considerred a disadvantage", but in practice it never has been a problem for me. I've remained connected for days with XMMS->icecast with no problems. My server is ~1000 miles and many hops away so there are plenty of opportunities for lost packets. It recovers quite well (using a 12k buffer--about 5 seconds worth of buffering--the standard for XMMS and WinAmp.)
So we've got a great solution already for streaming audio. That leaves streaming video. Other than bandwidth usage, it shouldn't require much more than an video codec to play it. I haven't really followed the XMMS project too closely but I think there is a reason it's called the X MultiMedia System and not the X Audio System. In other words I bet video will be part of XMMS before too long and it will do it well.
numb
- A.P.
--
"One World, one Web, one Program" - Microsoft promotional ad
"Remember when the U.S. had a drug problem, and then we declared a War On Drugs, and now you can't buy drugs anymore?"
You know, I knew this DOJ thing was going to start a bad trend. At what point did we become weak little powerless consumers? Do we really need to call on the government every time WE (consumers) are offered an bad product?
Look, many in the industry has had it with shoddy products coming out of Microsoft, they wanted something more stable, faster, and gennerally better. Lo and behold, linux arrived, and it has gained market share beyond what anyone predicted years ago. Not only that, but BSD is getting noticed more as well. BeOS has even become popular in various groups.
Now I ask you, how did the DOJ help all this happen? It didn't. The DOJ has done nothing yet, but we see that superior software is chipping away at the Microsoft beast. I would personally prefer to let the market decide (as it is doing) what is the best software for it to use.
Finkployd
My wine can run among other things winamp, IE, realplayer G2, sonique and seveal games. Mirc is one of the easiest and smoother things to run on it. Did you try the -managed option btw? It's very cool. Also run the 32 mirc and get the lastest wine.
--
As for how this development will happen, who knows. As far as I'm concerned, it's a myth that open source development is particularly fast. Most things take a few years longer to appear in open source form than their commercial predecessors. That doesn't bother me. And traditionally, proprietary protocols that end up predominating the market become public one way or another anyway. The major concern is patent protection, but that's only a temporary block.
Of course, Microsoft does try to dominate media delivery: their on-line reading efforts and Windows Media are examples. We do need to be vigilant, but I don't think their strategy will work out in the long run.
Why people want streaming media other than MP3 and MPEG-2, however, is still beyond me. I don't think there is a lot of interesting content out there. For the most interesting content available only in WMF and Real, like news, reading the article is faster than watching a video.
The answer is: it depends.
Working against us:
Working for us:
What will decide this battle:
Summary: streaming media is an *extremely* young technology. Of the 10% or so of the population with net access, probably only 5% (the DSL/Cable/University crowd) or so of those can even use streaming media effectively. (Most people can't deal with static images very well.) We have some time before streaming media becomes the 'killer app'. Even if we don't win on streaming media, we're the 'small, nimble competitor'. Microsoft is the 'large, entrenched industry leader'. Call it manifest destiny, if you like. We're bound to win one of these days. :)
Check http://www.fefe.de/rtp/ for a realtime MP3 encoding and multicast streaming solution for Linux. It's based on LAME, and open standards such as RTP. We're working on video. Relax, everything will be good.
I'm going to risk (some) karma here, and say you're wrong about Wine.
"The Win32 API is a moving target."
Sure. Each time they release it, though, it does not change. How much "spontaneous evolution" does Win98 do once installed?
"Microsoft will attempt to break WINE compatibility with every new release of Windows."
People have started to realise there's no reason to upgrade. Why use Office 2k, when Office 97 works fine? Why get Windows "Millennium" when Win98 works fine? And once thet break the Win32 API again, they have a whole slew of apps that no longer work.
Here's what Wine should do -- allow you to run different Win32 APIs in emulation. Want to run Win 3.1 apps? Use the Win16 emulation layer. Gaming, and general apps? The Win98 with DirectX 6.1 or 7.0 emulation layer would likely be fine -- as long that they don't mix Win16 code in.. But the tricky thunking could be setup with the two emulation layers cooperating. Win NT 4.0? Heck, it'd be easier than Win98 as NT assumes the Win16 stuff will be emulated in its own VDM. Emulating the NT 4.0 API for apps that need NT, the Win98 API for normal apps, and the Win16 API for Win3.1 apps would work fine.
Yet they still continue to make a huge, combined Win32 API emulator that must act differently depending on a huge list of variables. That's why they have failed.
---
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
The current state of streaming media seems to be a joke. I have found some radio programs I would like to listen to, but the "high speed" feeds are for 28.8 modems and stream at 16kbps. The quality is horrid. Its a shame, considering I have ADSL. I want a 128kbps feed for audio, until then I think it is just a novelty and to painful to listen to. I wonder if the broadcasting industry will legally prevent high quality streaming media.
:)
This remains sadly true for the majority of commerical audio sources on the net. But you have heard of shoutcast and icecast? Both sites have a directory service listing mp3 streams, an number of which are 128kbps or greater. That offers excellent quality, better than fm radio, if I can compare apples and oranges.
I live in an area with poor broadcast radio coverage, so this has been an invaluable service for me in finding new music. And it's really nice not to have to listen to commercials.
Which is why there are now an awful lot of people streaming mp3 with Icecast and shoutcast.
;-)
... that's right - it's Mpeg2 video .... a universally accepted format
If you're interested in audio then there's one very persuasive argument to *not* got for Windows Media. A case of market penetration. Real Will argue that they have the largest potential listener base out there - after all their clients are available for an awful lot of platforms. Windows Media if of course only available for Windows and Mac.... no unix clients.
But... Icecast offers even more... because Icecast is as open as possible we can boast a wider potential audience than either WMP or Real.
So if you're going for audio then why bother with Real or WMP....
Well there is the bandwidth argument.... but at the rate badwidth is now increasing that's only going to be important for a couple of years - mp3 is more then Good enough to be usuable at modem speeds and excellent at DSL speeds. Why bother developing proprietory codecs?
OK... we're still working on Live Video, - but we can do static video in many formats (we can even stream windows media video files via icecast
Perhpas people should look back at the technologies which have made the wired world what it is.... all the bits that make up a streaming technology. Lets go right to the place where traditional media is still holding out against the internet.
Digital TV in europe has been one of the most successful hardware launches
ever, people talk about DVD being an example of a great hardware launch but
this is eclipsed by Digital TV, either satellite or terrestrial.
Now.... I should maybe ask everyone who is promoting Microsoft's media server
as being the technology of the future to tell me which video codec is being
used
which has implementations available for any platform.
What about proprietory streaming technologies which have been launched in the
past... after all this is a streaming media list.
What's the most popular audio format available on the internet, do people go
searching for Real Audio? Asf? Wma? All of these formats have hyp machines
telling us that they are the next generation - in fact they;ve been telling us
this for a while. And yet in the past few years it's been Mpeg Layer 3 that's
taken off - the VHS of audio formats. It may not be patent free, but it's free
enough that every platform has players and encoders...
What about the actual method of delivery?
Remember 5 or six years ago, Microsoft was launching windows 95, and at the
same time decided that they needed an 'online' serivce, something like AOL or
Compuserve. Everyone else at the time was talking about the Internet as being
the future, but MS wanted to have the Microsoft Network. A closed system
available to the users of Windows, using it's own networks, its own protocols
- after all - the internet was based on 25 year old technology - why would
users upgrading to windows 95 want to use something so outdated? We all know
that microsoft got the whole network thing waaaaay wrong.
5 years on... what's teh standard medium for the exchange of computer data? Is
it's AOL's network? Compuserve? Or MSN.... nope, nope, nope.... It's the
internet - if you don't speak TCP/IP then you're not in the party. Plus
there's all the protocols and formats which form the backbone of internet
content - HTTP, FTP, NNTP, HTML, JPEG, GIF.
I could continue to cite other computer technologies which have gone the same
way - the IBM PC - technologicall inferior to other systems at the time - but
it was easy to copy and so the clone industry was born and created the
standard PC that can run Windows95/98/NT, Linux, Beos, Gnu HURD and several
varieties of BSD.
The technologies which are successful and end up winning are all either open
technologies, or technologies which are open enough that anyone can get in.
And the same will likely be true in the next few years as bandwidth continues
to rise and streaming media applications *really* get going.
(So - if anyone wants to help me write a live video encoder/streamer for iceast we'll have a complete package....)
The trouble here is that tech. developments come from hardcore reaserch into signal analysis. OSS imho (please give counterexamples!) most often suceeds in making widely understood technology stable and full of features.
Even if we have the Norwegians on our side, MS is pushing its money into research to stay ahead. If you check our Microsoft Research, specifically the internet media section (http://research.microsoft.com/research/china/imed ia/) you'll see what they're up to.
MP3 codecs rely as far as i know on Fourier analysis and, while encoding hard to do well, there are a few open source MPG decoder implementation.
The next version of the MPEG codecs, much like the new JPG2000, is going to be relying on wavelet technology. MPEG-4 uses things like 'shape-adaptive wavelet transforms and scalable shape coding'. Now this stuff is outside the realm of comprehension of a self net-taught hacker like myself (for now! i got me some books to read!), and i suspect it will be a while before this latest research gets implemented in OSS.
Which brings another interesting question: how is the opensource community going to lead in scientific research? Perhaps some notion of open knowledge community, where the knowledge is published, documented and organized for most efficient absorbtion into the neural tissue of the crowds of OSS coders that make this all possible.
meanwhile i'll be getting friendly with Dr.Daubechies and her orthogonal bases.
flip -out.
This type of market domination killed the browser market, as it raised the barrier to entry so high that competition was impossible. Now MS is threatening domination of the streaming market as well. With the recent explosion of broadband access, streaming media is going to become increasingly pervasive. Allowing any one company to dominate the field would be a disaster.
I can't help but wonder what would happen to MS Media Technologies if the whole division split off and became their own corporate entity. Instinct tells me that it would be in the best interests of this new corporation to offer media products and technologies across all platforms. There might be some bias to remain tied to Windows (after all, the programmers in the group probably know Windows best, and like developing for it.) However, their market share could only grow by moving on to more diverse platforms; thus, I believe that eventually we would see compatible products on all platforms. This would in turn create pressure for Apple to embrace a truly cross-platform solution, further improving the quality of product available to the consumer.
It seems as if the only barrier to this happening is the fact that the media group is still a part of Microsoft. If the group were separated, the technology would be free to grow, as the incentive/limitation to work only on Windows would be lessened, if not altogether removed.
THIS is why the DOJ needs to act.
It's not just about the players. It's about the software used to create the streams. Real do have some free tools, but they charge for most of it.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
I personally use windows media player because it's small, un-obtrusive, and more flexible than the others. Apple's quicktime 4 player is a great example of bad design, and real-player is quite simply a bloated, ugly, advertisement ridden, difficult to use pig. Of course, Windows Media player can (and does) support Quicktime, MP3, etc... (but not RealNetworks, at least on my setup). It doesn't really matter what player people are using, as long as the player supports different media drivers. I think it is useful to avoid players that aren't extensible (only support 1 format). If people only use easily extensible players, the most successful streaming format will hopefully be the best one. I am also somewhat biased agains Real Networks - their codecs look and sound awful, their player is ugly and obtrusive. What less could you want?
why not use a modified version of MPEG-2 for streaming with Linux. MPEG is good because surrent video cards have hardware support for the format and more of those card vendors are opening specs on their cards. Not only that but the compression is great, it would beat the pants off .rm and .asf movies for streaming purposes. It could be coupled with a streaming server and software encoder available in any flavour you wanted. Such a project would convince big name websites to produce their content in the format so everyone could watch it. Remember, the content people are in it for the eyeballs, not the codecs.
I'm a loner Dottie, a Rebel.
But it would pay to step back and think about what you are supporting when you are working on streaming video. To me, streaming video allows the same people who have run television to carry that tradition of poor programming to the 'net. Do we really want to do this?
Streaming video on Linux will take computers that might otherwise be used for creating new tools or solving real problems and turn them into TVs. Isn't that really a terrible waste? Should it be aided by your efforts?
thad
I love Mondays. On a Monday, anything is possible.
53,000,000 people spammed every time they did a mailing run.
You betcha we want them dead.
How about we all go support free stuff like free-expression.org, which is supposed to create a compatible but *free* streaming media server and player?
My blog: http://www.seebs.net/log/ --- My iPhone/iPad app: http://www.seebs.net/seebsfrac/
Bravo. /me wishes he had modpoints.
--
I think there is a world market for maybe five personal web logs.
At the risk of being a heretic, I would like to say that this is a symptom of a much larger problem than whether Linux users can play some streaming media or not. Basically, the whole key to the freedom and success of the Internet has been that the protocols, which date back to the pre-billion-dollar-dot-com-marketcap days, have been open and standardized. It is very concerning that this is not the case for many of the protocols that will be important on the Internet of tomorrow.
Streaming media may or may not be a big deal (on the one hand Real Networks and Microsoft are both evil as sin, on the other, how long until we can just stream with the semi-open mpeg standard instead?) but there are certainly other protocols that are. Is there any standard for Voice over IP? Are these open? Or just look at instant messaging. Flash. Secure communications.
Sooner or later the services that make up the Internet today are going to fade into obscurity and be replaced by whatever comes next. However, it seems that ever since commercial interest came to the Internet, they have not been able to agree on one single standard. Is the future of the Internet going to be one perpetual standards war because everyone believes that a monopoly is the only way to do bussiness?
I believe very firmly that this has already hurt the Internet and it's developement. Why has there not been a single new standard service since the WWW? Why has the last ten years seen the least developement of new innovations on the Internet although more money has been spent on it then every before?
Of course, as always our hope lies in that the Open Source revolution can convince companies that terms like "proprietary" and "patented" are everything but marketing catch phrases, and that fostering freedom is the only way to be successful on the Internet. But as long as Steve Case is looked up to as the archetype Internet executive, I wouldn't hold my breath.
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
Open systems and the open Internet don't have anything to fear from commercial products, in my opinion, but only from commercially imposed lock-ins of various types, like patent restrictions and trade secrets.
The most dangerous thing that's been happening in recent years is probably the development of proprietary "standards" which are then imposed on the world by the power of cartel. We've seen lots of examples of this highlighted here right from the earliest days of Slashdot, but of course it's been happening forever. The difference is that before there wasn't really a single world for most people, whereas now there is, and it's a world that's held together by open communication and open information. Tie it up in the red tape of proprietary restrictions and we've got problems.
Some might call for enforcement to ensure that world-adopted standards are never proprietary, but that is easier said than done. The main problem is that the world's most obvious enforcers (governments) are in the hands and pockets of the big corporations that are of course creating the standards to their own advantage. I doubt if the IETF, EFF and others could get themselves onto the relevant forums even if they wanted to. In any event, they wouldn't be welcome even if the show weren't effectively invite-only, because corporations focus on control and profits, not openness.
Where does this leave us? Probably in a perpetual war against oppression by the corporate machine, but that isn't as bad as it sounds. Remember that they need us since we're the source of their profits, so there is a limit to how nasty they can be without losing money. To put it in other terms (control theory?), their success is dampened by negative feedback, whereas the growth and chaotic direction of the Internet is very much in the exponential grip of positive feedback as everyone builds on the work of everyone else. It'll be a bumpy ride, but I reckon we'll come out on top.
"The question of whether machines can think is no more interesting than [] whether submarines can swim" - Dijkstra
I have done a little bit of reading about video standards and codecs. It seems there are several issues here.
1.) Windows media is centered around the ASF file format. The documents I have read at Microsoft's web site give lip service to ASF being a new, open standard. There certainly seems to be an oppurtunity to pressure Microsoft, and see just how open they are willing to be. Maybe they would be willing to hand over the specs and not sue the daylights out of the open source/Linux community (hold on... gotta stop laughing). Personally I don't like ASF. It embeds a GUID into all ASF files. The GUID contains the MAC address and other information about the PC creating the files. Maybe open source tools would give back privacy.
2.) Codecs. Even if Microsoft allowed open development of the ASF format, that really solves very little. ASF is just a wrapper. The appropriate codecs will need to be available in order to play any given ASF file. The most popular video codec seems to be the three versions of Microsoft's MPEG 4 codec. According to Microsoft the codec is based on the proposed MPEG 4 standard. That could be bad news. That sounds like MS Speak for "proprietary". It may be difficult, costly or impossible to make a legal codec. Such a codec would threaten MS dominance.
3.) Workstation apps and streaming media server apps are needed to support the ASF/ASX pseudo standard. I don't know what exists in the Linux world, but it may have to be updated to support windows media. Workstations require the codec to view streams. Does a streaming server app need to actually have the codec to transmit the data, or is understanding the ASF format sufficient?
MP3 seems to be a seperate issue. Real and MS have the streaming video/audio market locked up for now. I think MS could take over, and it is forward planning on there part. It seems that nothing exists to threaten MP3, the cat is out of the bag on that one. Perhaps something similar would happen if a high performance, open source, free MPEG4 codec existed.
The current state of streaming media seems to be a joke. I have found some radio programs I would like to listen to, but the "high speed" feeds are for 28.8 modems and stream at 16kbps. The quality is horrid. Its a shame, considering I have ADSL. I want a 128kbps feed for audio, until then I think it is just a novelty and to painful to listen to. I wonder if the broadcasting industry will legally prevent high quality streaming media.
I am a streaming media professional, and here is the scoop. Windows Media Technology (WMT) is an excellent streaming platform. Microsoft has spent a lot of money and time to develop the software and to promote it.
Real Networks, who have tried to keep the Linux side of the RealServer and RealPlayer going, do not have funds from selling an OS to support them like Microsoft does. As a result, their server costs money. I've paid out a lot of money for their server to be able to serve Linux users. They also need to have advertising money driven by the "player portal." It sucks, but that's the way it is.
On the codec side, it is my opinion that there is no video codec that can compete with Real or WMT at the 20kbps department. That's usually 5-6 kbps audio, 14-15 kbps video. This is your typical 28.8kbps connection, and many people with 56k modems still need to use this data speed because of oversold ISPs and such.
Personally I don't see the server as a difficult piece of software to write...it just moves data. The encoder and player are the tough parts.
I am calling on the IPOed Linux companies to look seriously into ensuring that there is a low bitrate video system for Linux. Much like Microsoft, they are the people with the bucks to make this happen.
Imagine an open-source low-bitrate video system. Videoconferencing could be combined with broadcast capability to provide incredibly interactive new mechanisms of global communication. The possibilities are endless if the Linux community has access to the codecs.
If patent free codecs can't be created, they could be licensed by the big companies supporting Linux.
Let's do this the Right Way... Open Source right from the beginning. The OSSMAPI can serve as a reference API... we invite you to help us define data structures and the programming interface.
Wouldn't it be nice to just #include "ossmapi.h" and be able to stream a mp3 or mpeg file from your little application without having to worry about streaming works or what client your "partner" is using?
Please take a look and mail us your ideas!
Open Source Streaming Media API (OSSMAPI)
yeah, though the throughput at RPI was better pre-firewall, and pre-freshmen with laptops (hey, I've graduated and am off far away now...). They finally put in some new switches around campus so the local mp3/mpegvid traffic wouldn't slow everything down quite so much... Helped some though. Oh well... the other side was usually the limiting factor, though...
"It's tough to be bilingual when you get hit in the head."
However, upon reviewing the agreement I still stand by my previous rant. It might be free in cost, but not in obligation. All the agreement allows me to do is to offer Real for download from one of our own servers. I still must not install it on behalf of the user because I'd be violating their section 1c saying I can't disable the user from seeing the EULA when they first INSTALL Real.
I also must report to them quarterly how many of my users downloaded the real player. Yack...
I have a very fast cable connection at home, and I have a very fast T3 line at work. Streaming Media doesn't work so well at either place. Its frustrating, tedious and annoying. I haven't clicked on a media clip in at least 6 months. I cant even imagine how bad it is for folks using analog modems.
The bigger issue is, IMHO, the issue of focus. To beat Microsoft, IMHO, don't get into a war over features. They can churn out features faster than anyone. Look at the software we have as a result. What Linux should be concerning itself with is superior stability with less features...initially. Tackle the features that actually have to get done later, after a need has been established, relying on your reputation as superior software craftpeople. Nobody has ever beat Microsoft going toe to to on features. Its a a distraction at best.
I'm still working on a clever footer.
First, at least in regards to streaming media, we must realize that there are no standards for it, unlike something like HTML, and thus we'll have to follow what the current trend is (and that is going the way of Windows Media). While some would probably agrue that Mpeg is the real standard, how many popular web sites use it?
Next, a group of programs need to get together with some commercial entity with a good interest in Linux (RedHat, Corel, Sun, etc..). The commercial interest is necessary, as I doubt that Microsoft would want to work with just a set of Linux programmers.
With that in mind, this collective group should approach the company that holds the defacto standard in question, and ask if they can have the APIs to the library that handled the translation of the streaming media to audio and visual elements. I suspect parts of the library itself are trade secrets or patented or something along those lines, and may include encryption and other details. Don't ask for the source to this library, just the APIs.
Then comes the tougher part. While the APIs can be easily used to make the client wrapper (and the GPL'd part), you'd also have to convience the company to recompile their library into the appropriate OS format. Now, take streaming media; if it's done properly by Microsoft (haha), the GUI code will be nowhere within the library, and the library should be nearly cross-platform, relying only on TCP/IP and data decryption algorythms. If this is not the case, then the commercial company should offer to go into a non-disclosure agreement, and work on making a cross-platform version of it available; the library would still be closed source, of course. (This is another reason why as just programmers, we can't do this alone).
FInally, once the library and the GUI wrapper is finished, you have a program that parts of can be released under the GPL, while the library is distributed as a binary and can only be redistrubted as a library.
The key thing is getting some commercial company into it. Money speaks louder than words in this case, although the above transactions should require no money to be spent. It has to be made clear to the company with the controlling standard that doing this would result in a wider acceptance of their standard, which means more money for them in the end (in the case of Windows Media, more WM users; in the case of DVDs, more DVDs bought, etc). In addition, any 'secret' parts of the library would be easier to get if there was another company with an interest in it, as opposed to a 'random group of hackers'.
"Pinky, you've left the lens cap of your mind on again." - P&TB
"I can see my house from here!" - ST:
for use in non-commercial stuff. For more than 25 streams, they have a few options; you can buy a 40-stream server for $600, or a 100-stream server for several thousand dollars.
WMBC freeform/independent online radio.
If you are running MS Windows you've got it all: mp3, real audio, quicktime, MS media, etc. As soon as you move away from windows you get less choice.
That's why MS Media won't gain much popularity because who'd want to provide streams for only part of the potential audience? A succesfull standard (propietary or not) needs to have a free client that is available on most platforms. Especially since small webdevices (without MS software) are becoming increasingly popular, non availability of MS Media on those platforms will work against MS.
Both quicktime and realaudio & video are already available on other platforms. If you want to provide streaming audio/video for non windows users, you'll have to go with either of those.
Jilles
Accidentally, we're three Computer Science students working on what we think will be an intuitive and flexiable Streaming Media API. In fact, it's the final project in our education.
We know codecs are evolving. The codecs are not really the issue, the programming interface to the codecs is. What we need to do is to establish an Open Source alternative to the Win32 Streaming Media API. Our project is still in the planning period but we've already decided to work with the Open Source community.
Check out my URL or search for streaming on SourceForge if you want to help out with the API design or have other ideas for the future of streaming media! Think C++/Java classes, well-defined objects and so on.
Searching for MBONE only turned up 3 hits on freshmeat, but http://www-mice.cs.ucl.ac.uk/multim edia/software is a site that has updated versions of some of the tools I used in 1994. It looks like a lot of them are covered by the Berkeley license, but there are some precompiled linux binaries.
the good ground has been paved over by suicidal maniacs
Yes, that's right. Their license agreement specifically states that redistribution is not permitted, only end users are permitted to download and install their player. There is nothing on their web site stating of a way this restriction can be removed either.
Somehow they got it in their mind that corporate business users are permitted to freely download and install apps on their managed PCs. Plus, as a college, we have to keep lab PCs orderly by locking down permissions so students can install stuff on the computers either.
(Yes, they all run NT, not Linux).
I wrote to Real and complained and they told me to send my request to client_redistribution@real.com but they never respond to my e-mails. (My latest attempt to contact them was the middle of January 2000)
So -- flock() 'em. I installed WMP along with IE, which is permitted (license wise) if done via the IEAK.
Free is useless to me if I am forbidden by license to freely copy the software onto the client machines I maintain. If Real thinks all users manage their own PCs, they are horribly out of touch with reality.
You may hate Microsoft, but at least they understand the business environment. Real can shrivel up and die for all I care.
BTW, did we all forget already Real's huge intentional privacy violation regarding their players sending player listening info back to Real?
I don't mean to get off on a rant here, but Microsoft bashing is getting old. Yeah, I love Linux. Where I control the decisions (my home network of 6 machines), I have everything run by Linux clients and servers and use UNIX on servers at work to run everything from Apache to Samba. But the desktop corporate world still revolves around Microsoft and I can make no sane business case to have students to use anything other than that.
If you read my post history, you can see me ranting about the often horrible cruft Microsoft shovels out too. But in this case, WMP beats Real in the Windows world and beyond that, there are no other viable alternatives (Quicktime install methods and redistribution crap deserves a separate rant...)
I agree with most what you say, but there's a difference between standards for network transfer and video compression.
MSN and their proprietary protocols was offering absolutely no advantages, it was done just for strategic reasons.
OTOH, there are huge differences from a performance point of view with video codecs. You must regard CPU power necessary to decode, memory requirements and most important, what bitrate does your codec require to deliver a certain quality (with lossy codecs you obviously can reach any grade of compression, so it's more useful to compare quality of two codecs at the same bitrate). MS is doing pretty well here (I think they're using MPEG-4 for low bitrates), and it'll be hard to come up with a free and patent-free decoder that delivers the same quality.
There already is a set of standards!
l e /.
Check out the homepage of the AVT working group of the IETF at http://www.ietf.org/html.charters/avt-charter.htm
and also check out the set of tools (video conferencing, audio etc.) already available on
http://www-mice.cs.ucl.ac.uk/multimedia/softwar
We already do have streaming video and audio on UNIX.
You make it sound as if making a Windows product available for Linux is a matter of taking a tarball of the source over to a Linux box and running make. Well, it isn't. The technical reason there's no Office for Linux suite is that it simply can't be done without either stabilizing the Win32 API long enough to develop a good compatiblity library or ending up with two completely separate code bases for the same product. Neither is desirable for Microsoft because the former would stifle their God-given right to (Ahem!) innovate and the latter would simply be a big mess.
Microsoft likes to brag about the low average age of its software staff (the figure I heard was around 25). That explains why their products are of low technical quality: they're being built by people without the experience to know better. Before you reach for the flamethrower, I'm not saying that younger people aren't any good at doing software, because there are plenty that are. I'm saying that a horde of inexperienced people developing software without the leadership of people who've been there, done that and got the tee shirt is a bad thing. Rick Downes did an interesting analysis of Microsoft's RegClean app in the RISKS-FORUM digests Volume 35 and Volume 37. The long and the short of it is that he found tons of unnecessary left-overs in the program that go a long way to prove that someone smart at Microsoft built an app template and people are boilerplating apps from it without taking the time to understand what they were doing.
Lest anyone think this scores more points for the open source movement, it happens on this side of the fence, too. The difference is that others have the opportunity to find these problems and correct them.
This isn't about audio, it's about video.
But if you want to talk about audio, the thing Microsoft can offer people to make them use a MS audio format instead of MP3 is copy protection. We don't care, but content providers love that idea, remember.
Please keep this crap out of slashdot. By posting a hundred character word, you made the whole page be as wide as that line, and caused most of us to get horizontal scrollbars as a result.
That's your code for ``loser.''
This smacks of a double standard to me by slashdot readers (similar to the how slashdotters agree with the sale of the linux.net sale but railed against the seriousdomains.com auction) For a a forum that is constantly complaining about government intrusion into our lives, slashdotters seem to see nothing wrong with using the government as a personal attack dog when the mood suits them.
The reason the DOJ got involved in the browser wars is because MSFT used their position as the maker of Windows(tm) to force OEMs to preload Internet Explorer and charged higher licensing fees to those that disagreed until they toed the line. I have not seen or heard of any OEM being forced not to preload Real Player by MSFT and thus I cannot see how the browser wars are a good precedent for involving the DOJ. The giving away of IE is also different from the distribution of Windows(tm) Media player for free because besides the fact that they gave away IE to undercut Netscape, it can be argued rightly that in the industry today it is regular practice to give away content viewers to gather eyeballs so as to charge an arm and a leg for content creators/server software. This seems to be the business model of Real and Apple...should we launch a class action lawsuit against them for giving away software and thus stopping me from charging for my Carnage Player, I hope the answer is no.
Secondly involving the government in every little tiff in the software industry can only be a bad thing. The animosity of Sun reached distasteful levels during the MS-DOJ case and several statements made by Scott McNealy during the case are clearly products of envy. It would be sad indeed if the software industry is reduced to calling on the government for help every time a market leader emerges like angry school children paying the school bully to beat up the smartest kids in class.
I also dislike the premptive strike nature of the above post as displayed by this line.. With the recent explosion of broadband access, streaming media is going to become increasingly pervasive. Allowing any one company to dominate the field would be a disaster. This seems to indicate that it is OK for the DOJ to punish MSFT for having better technology technology than the rest of the current industry. Streaming media support is NOT an issue to anyone I have ever spoken to about a computer purchase and I am very sure that the current industry landscape will change before it ever does. Asking for an attack on MSFT now by the DOJ is premature and is only justifiable by twisted anti-MSFt logic. Why not ask the DOJ to sue Winamp or ICQ (wow just realized AOL owns both of them) since once computers become cheap enough and high bandwith is ubiquitous they are set to dominate their fields also?
Finally the entire above post smacks of an intense feeling of sour grapes and misconceptions. It seems that the poster is implying that Windows Media is so good that MSFT should be forced to share... (I'd rather they shared IE first, because I'm tired of Netscape's bugginess) but does not realize this has never been a reason for the DOJ to get involved in an issue. MSFT is allowed to develop cool software for Windows after all Windows is their principal product and they should make it as attractive as possible by writing cool apps for it. What is illegal is forcing people to use their product or else. Instead of bitching to the DOJ about issues that do not concern them maybe the richer Linux community (VA, Andover, Redhat) can fund research into open codecs or work on free (as in beer and speech) media players, servers and file formats. Instead of bitching to the government maybe the answer lies within us as a community.
It's a nice little server, easy configuration, seems to work quite well. It's also an RTSP server, so I would guess it is independant of media and players, right?
Everyone go try it out: http://www.apple.com/publicsource