RealNetworks will release open-source software this year that will let Linux computers play Windows Media files legally. The media delivery software company and Novell made the announcement at the LinuxWorld Conference and Expo. Novell said it will include the tool in its Suse Linux Enterprise Desktop 10 in the fourth quarter.
RealNetworks to Add WMV/WMA Support to its Linux Offering
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Another scores for Novell! Keep the goodies coming…and its nice to see Real putting the results into Helix, which means it will be available for general comsumption…this will be a great additional, especially if it can keep up with streaming media formats from MS.
And if they indeed bring their music store over to the linux platform, that could be a very significant boost for the platform. I’m assuming they mean full support for Rhapsody in linux…+AAC maybe? Real could become a very integral part in the OSS community with that kind of move.
Edited 2006-08-15 23:35
AAC is already supported in the realplayer, the problem is the DRM which is used; then again, for me, screw online music, give me a physical cd.
This is really a great developement.
Linux distros often lacked on the support of media files, maybe this will come to an end – at least on some file types – in the close future.
Now if they only allowed playing CSSed DVDs …
Yet I guess we won’t see this happen anytime.
So such improvements will help the Linux movement a lot imo. As even more will work out of the box!
Imagine, you use a LiveCD on a friends PC and he could do everything – ignoring games – like watching films on YouTube, listening to his MP3s … he does on his Windows machine, but without the hassle of searching drivers, codecs, programms, cracks …
PS.: I know that you can play every file type right now, but that is really a legal grey area.
Edited 2006-08-15 23:48
//I know that you can play every file type right now, but that is really a legal grey area.//
I don’t understand how it could possibly be a grey area. Linux can use the exact same binary codecs that Windows can use.
People who are Windows users have no “contract” with a company like Real that poeple who are Linux users do not have.
Therefore, it would actually be illegal for a compnay like Real to offer a codec only to Windows users if there are Linux users who also could use it.
On what possible legal basis could Real refuse? How would Real avoid charges of arbitrarily discriminating against a particular subset of people who wanted to download one of their codecs?
This same reasoning would apply to codecs from Apple. There is no legal arguement that Apple could use to enable them to offer a codec for quicktime files to Windows users (with whom they have no contract) and at the same time refuse to offer the exact same files to Linux users (with whom they also have no contract).
Edited 2006-08-16 00:37
sounds great, less patent problems for others, which they can work on other projects . might get more windows users over to linux also
awesome. just what we need – open sourced codecs.
It’s goodies that Novell brings, certainly a good news for everyone who use Linux as desktop OS, but the best news I can think about is one free and open codec/standard rules the world, so we will no longer worry about such legal issue.
that would be nice – or for every single codec out there to be free
i am curious if this will be available for apple in universal binary (or available for apple at all)?
flip for mac does not work on intel macs yet, this would be really nice.
Perhaps when Apple makes itunes or quicktime available for Linux. Of course that would be like an endorsment for Linux, which will not happen. A port to osx would happen sooner.
Edited 2006-08-16 01:23
I had Quicktime 2.5 for UNIX years ago.
They killed it. They used to have it.
Then again, they also had MAE too…
“flip for mac does not work on intel macs yet, this would be really nice.”
So how come I have a Universal flip4mac on my macbook since I got it 3 weeks ago?
>So how come I have a Universal flip4mac on my macbook since I got it 3 weeks ago?<
Awesome, I haven’t looked in a while, I’ve had my MBP for about 3 months and I haven’t heard anything about a universal till now.
Edited 2006-08-16 01:37
Flip4Mac certainly does run on Intel Macs
Flip4Mac is now Universal and works fine on Intel Mac
Sneer though you may, this is excellent news.
People who are Windows users have no “contract” with a company like Real that poeple who are Linux users do not have.
Except the EULA that comes with those codecs (or any piece of proprietary software). And even if those EULAs are unenforceable, copyright law doesn’t allow redistribution of binary “codec packs” without permission. I have a feeling that 99% of Linux users who install Windows codecs are getting them from unauthorized sites.
//Except the EULA that comes with those codecs//
(1) If a Windows user clicks on an rm file or a mov file, and WMP opens up and downloads a suitable codec – as far as I can recall there is no EULA.
(2) A Windows user has agreed to the EULA for Windows. AFAIK he has not agreed to any EULA with Apple or Real.
(3) Just because a “requirement” is in a EULA doesn’t mean it is legal. Apple or Real have no authority to permit one class of end user (for whom they did not make the OS) but deny another (also for whom they did not make the OS). Such behaviour is akin to refusing to serve someone based on their skin colour or where they were born – see how far you get with that.
//I have a feeling that 99% of Linux users who install Windows codecs are getting them from unauthorized sites.//
… and I have just as strong a feeling that if Real or Apple refuse downloads of binary codecs based on what OS the requester is running they are on exceedingly shaky grounds and could easily get done for discrimination.
The exact same argument would apply just as well to Windows wireless drivers running under Linux and ndiswrapper. If I bought the hardware card, I can run the driver under whatever OS I please, surely? On what basis would the card manufacturer have any possible justification for trying to stop me?
Edited 2006-08-16 02:52
If a Windows user clicks on an rm file or a mov file, and WMP opens up and downloads a suitable codec – as far as I can recall there is no EULA.
…
The problem with this argument is that neither Real nor Apple have ever offered their codecs via Windows Media Player’s download mechanism. In fact, they’ve never made standard codecs available directly to Windows users, instead preferring to push their respective player applications (which do contain EULAs). The seperate codecs available on the net for those formats
are distributed without license or permission of Real or Apple. Also, with the exception of MS codecs, WMP’s auto download functionality doesn’t install codecs, it just redirects the user to http://www.wmplugins.com/ where users can download the necessary codecs if available.
//The problem with this argument is that neither Real nor Apple have ever offered their codecs via Windows Media Player’s download mechanism. In fact, they’ve never made standard codecs available directly to Windows users, instead preferring to push their respective player applications (which do contain EULAs). The seperate codecs available on the net for those formats are distributed without license or permission of Real or Apple. Also, with the exception of MS codecs, WMP’s auto download functionality doesn’t install codecs, it just redirects the user to http://www.wmplugins.com/ where users can download the necessary codecs if available.//
There is no problem at all with the original argument. If Windows users using WMP are directed to a 3rd party (ie not Microsoft) site to get separate Apple or Real codecs for WMP, then there can be no valid reason why Linux users should be disallowed from also obtaining the exact same files from the exact same website.
Ergo, using those separate binary codecs on Linux is no more or less of a grey area than it is using them for WMP in Windows.
It is the exact same position for either end platform.
Edited 2006-08-16 05:02
There is no problem at all with the original argument. If Windows users using WMP are directed to a 3rd party (ie not Microsoft) site to get separate Apple or Real codecs for WMP, then there can be no valid reason why Linux users should be disallowed from also obtaining the exact same files from the exact same website.
There are no seperate Apple or Real codecs for WMP distributed by those respective companies.
Ergo, using those separate binary codecs on Linux is no more or less of a grey area than it is using them for WMP in Windows.
If they existed, they could be used (i.e., the end-user downloads them) but they can’t be redistributed and Apple, Real, et al., have no obligation to support their use on platforms they didn’t distribute the codec for.
//If they existed, they could be used (i.e., the end-user downloads them) but they can’t be redistributed and Apple, Real, et al., have no obligation to support their use on platforms they didn’t distribute the codec for.//
One would not ask for support. One just asks for exactly the same rights as users of other platforms.
If I must download their entire player and run it under Wine or Win4Lin or VMware or whatever, then so be it. I have as much right to download it from them as any other user does, regardless of what OS I use. Once I have downloaded it and installed it under Wine, I can then point other media players of my preference at their codec.
There is no way that using their codec under Linux is essentially different to using it under Windows.
Ergo, my use of it when running Linux is no different to your use of it if you were running Winows. We have exactly the same legal status under the law.
Edited 2006-08-16 11:31
There is no way that using their codec under Linux is essentially different to using it under Windows.
Ergo, my use of it when running Linux is no different to your use of it if you were running Winows. We have exactly the same legal status under the law.
According to the EULA that accompanies Windows Media Player, its a Windows OS Component, not a standalone product.
This part of the EULA makes it pretty clear:
IF YOU DO NOT HAVE A VALIDLY LICENSED COPY OF THE CLIENT OS SOFTWARE, YOU ARE NOT AUTHORIZED TO INSTALL, COPY OR OTHERWISE USE THE CLIENT OS COMPONENTS AND YOU HAVE NO RIGHTS UNDER THIS SUPPLEMENTAL EULA.
// According to the EULA that accompanies Windows Media Player, its a Windows OS Component, not a standalone product.
This part of the EULA makes it pretty clear:
IF YOU DO NOT HAVE A VALIDLY LICENSED COPY OF THE CLIENT OS SOFTWARE, YOU ARE NOT AUTHORIZED TO INSTALL, COPY OR OTHERWISE USE THE CLIENT OS COMPONENTS AND YOU HAVE NO RIGHTS UNDER THIS SUPPLEMENTAL EULA.//
What has the EULA of Windows media Player got to do with my downloading a media player designed for Windiows from Real or Apple, “installing” it with Wine and then telling my Linux Xine media player where the codecs are?
Exactly nothing, that is what.
What has the EULA of Windows media Player got to do with my downloading a media player designed for Windiows from Real or Apple, “installing” it with Wine and then telling my Linux Xine media player where the codecs are?
Exactly nothing, that is what.
I looked back through the posts and please accept my apologies. I thought it was being argued that MS Windows Media Player DLLs and Codecs could be run on Linux legally.
I agree a player from Real or any other company is a different story.
//I looked back through the posts and please accept my apologies. I thought it was being argued that MS Windows Media Player DLLs and Codecs could be run on Linux legally.
I agree a player from Real or any other company is a different story.//
Exactly. The same argument applies for any number of media players from 3rd party sources designed for Windows … for example, Winamp.
Just as long as I don’t try to use the poisonware from Microsoft itself, there are any number of other media players out there that I can download and install under Wine on my Linux system and then use all of the designed-for-Windows binary codecs that it comes with or which it subsequently downloads.
Edited 2006-08-17 02:11
//Just as long as I don’t try to use the poisonware from Microsoft itself, there are any number of other media players out there that I can download and install under Wine on my Linux system and then use all of the designed-for-Windows binary codecs that it comes with or which it subsequently downloads. ///
Speaking of which …
http://microsoft.toddverbeek.com/player.html
… a few options there to choose from.
Just as long as I don’t try to use the poisonware from Microsoft itself, there are any number of other media players out there that I can download and install under Wine on my Linux system and then use all of the designed-for-Windows binary codecs that it comes with or which it subsequently downloads.
The only place I could see that might be *grey* is if the 3rd party application connects to MS and downloads MS codecs. Real Player 10.5 and Rhapsody both do this for Janus DRM and some WMA playback. Most players that use MS codecs utilize an installed copy of Windows Media in some way. I guess it might even depend on the distribution license the application developer has with MS.
I do not know of any non-ms codecs or players that could pose any licensing issue (as far as I am aware anyway).
//The only place I could see that might be *grey* is if the 3rd party application connects to MS and downloads MS codecs. Real Player 10.5 and Rhapsody both do this for Janus DRM and some WMA playback. Most players that use MS codecs utilize an installed copy of Windows Media in some way. I guess it might even depend on the distribution license the application developer has with MS.
I do not know of any non-ms codecs or players that could pose any licensing issue (as far as I am aware anyway).//
Either it says in the EULA that “you agree to run this code under Windows” or it doesn’t.
If it does say that … how can they require that I am using a product from another vendor? They can certainly recommend it … but surely they can’t require it of me. Since they are not the provider of Windows, what business is it of theirs what OS I am running? EULAs for proprietary code usually disclaim all responsibility if anything doesn’t work, anyway. In this case their EULA imposes unreasonable terms of use, and would be invalid. So I am in the clear for the use of their codecs under Linux (even if I am going against their recommendation).
If it doesn’t say that … then I am also in the clear.
PS: In reality, I would just get the native Linux version of Real or Zinf in any event … so this is rather a moot point … but nevertheless a valid point for dissipating the myth that “using codecs on Linux is illegal”.
Edited 2006-08-17 05:26
“… and I have just as strong a feeling that if Real or Apple refuse downloads of binary codecs based on what OS the requester is running they are on exceedingly shaky grounds and could easily get done for discrimination.”
I think you need to realize there are no legal protections for OS Minorities…lol…Then again what you mean by “get done” nobody but you knows…
If you take history as any example, there are no protections until a court/legislature makes it so for a minority – in other words, there is no such thing as discrimination based on OS…at least not in any effective way.
I agree though that it’s just stupid. Why can’t I use your windows codecs on Linux if someone’s written a piece of software that can use it? Release yourself from any liability or warranty, blah blah, say you DON’T claim it will function at all, and let us have at it…
Ah well. The tides are changing oh so slowly…
//Why can’t I use your windows codecs on Linux if someone’s written a piece of software that can use it?//
Exactly. Just try and stop me.
More to the point, try and write a law that would allow you to prevent me (as a Linux user) downloading a freely-offered program (from a third party) and using it even if it happens to be designed for Windows.
You can’t frame such a law and stay within the constitution. The constitution basically does not allow such discrimination.
Has anything come out as far as the exact terms of the release? Some open source releases can be quite questionable as far as user rights. I’d be very interested to see the patent implications here.
Same argument, mp3 still leads the pack in the audio real, with aac from apple a close second..Sadly OggVorbis is nowhere to be found most of the time, and gets neglected even though it is vastly superior is so many way…especially compared to mp3. AAC is great, but a world with nuthing but ogg would be better.
I searched high and low when i purchased my first higher end mp3 player, and finally settled on a U2 from Cowon Iaudio because they had OGG support.
If people could just use FLAC http://en.wikipedia.org/wiki/FLAC (for lossless) and Vorbis http://en.wikipedia.org/wiki/Vorbis (for lossy), then there would be absolutely no problem.
Two excellent and unencumbered formats that anyone may use.
Everyone wins.
Edited 2006-08-16 05:16
Agreed, surely it won’t cost the manufacturers (of mp3 players) anything to include these 2 codecs with their offerings?
The Cowon iAudio 5 supports Ogg, and the iPod can be made to support Ogg if you install the Rockbox firmware.
The problem is it costs these manufacturers something, and I suspect a lot of them feel there simply isn’t enough demand on the market to make that something worthwhile.
Agreed, surely it won’t cost the manufacturers (of mp3 players) anything to include these 2 codecs with their offerings?
Not quite true. Vorbis takes more processor power to play than an equivalent mp3. Unless a more powerful and more expensive chip is used, portable players simply don’t have the ability to play ogg.
//Vorbis takes more processor power to play than an equivalent mp3. Unless a more powerful and more expensive chip is used, portable players simply don’t have the ability to play ogg.//
So far this is just an unsupported allegation. Evidence please.
There are a multitude of players that do include the ability to play ogg vorbis files.
I have yet to see any evidence at all to indicate that these players necessarily consume more power than others that don not support ogg vorbis files.
Everyone wins.
Except Thomson of course.
Edit: and Fraunhofer as well.
Edited 2006-08-16 09:54
I recall reading that Ogg is more computationally intensive than MP3, and thus uses up battery life more quickly. I’m not very knowledgeable about these things though.
//I recall reading that Ogg is more computationally intensive than MP3, and thus uses up battery life more quickly.//
Pfft. An old wives tale, if ever there was one.
A CPU doing math on audio data uses up exactly the same juice as another one which is polling in a loop “is there anything for me to do yet? no? well is there anything for me to do now?” etc.
Same CPU, same instruction fetches, same memory … everything. Exactly the same power consumption to execute.
Edited 2006-08-16 11:11
“A CPU doing math on audio data uses up exactly the same juice as another one which is polling in a loop “is there anything for me to do yet? no? well is there anything for me to do now?” etc.
Same CPU, same instruction fetches, same memory … everything. Exactly the same power consumption to execute.”
Wrong. […], Wrong, Wrong.
//Wrong. […], Wrong, Wrong.//
Nope. Not wrong at all.
A multiply instruction takes up exactly as much power as a no-operation. Fetching instructions from memory and executing them consume as much juice no matter if they are a main idle loop or part of a matrix multiply within an FFT algorithm.
To increase the power used, one must actually activate more hardware – for example it takes more power to have a disk spinning than have it idle. This sort of thing does not come into play when playing a media file.
To decrease the power used one must power down devices and/or slow down the CPU clock. This is not done if one is playing an MP3 file compared to an Ogg Vorbis file … utter fantasy to think that playing an Ogg Vorbis file would consume more power.
Here is a challenge for you … come up with a valid reason why an Ogg Vorbis file playing in a media player would consume more juice, or a real world measurement, or some even remotely credible fact (as opposed to just blithering “wrong, wrong, wrong”) … and I will consider that you might possibly not be the complete idiot that you appear on first sight to be.
The first flaw in your argumentation is that a computationally intensive algorithm might require a *stronger* processor, which takes more power. More flaws follow:
> A multiply instruction takes up exactly as much power
> as a no-operation.
For fairness, let’s compare a multiply operation with as many NOPs as it needs to fill up the same time, otherwise the MUL loses instantly. NOPs do not use the multiplication unit. Not using a unit means that stored values do not change, and at the lowest level it means that the gate capacities of the transistors need not be (dis-)charged (reasonably assuming CMOS technology). Minor power comsumption remains through leakage currents, but this is orders of magnitude smaller.
Actually, NOPs don’t use any other computing unit either, and can even stop the controlling state machines until something happens (e.g. a timer runs out). Well-implemented NOPs can even leave the RAM alone.
> Fetching instructions from memory
> and executing them consume as much juice no matter if
> they are a main idle loop or part of a matrix
> multiply within an FFT algorithm.
First, this is wrong as explained above. Secondly, a well-implemented NOP doesn’t have to “fetch instructions” (a term which already implies a programmable processor and is thus by itself biased, but even many CPUs have a “sleep” feature that stops almost everything).
//The first flaw in your argumentation is that a computationally intensive algorithm might require a *stronger* processor, which takes more power. More flaws follow: //
hahahahahahahahaha.
Very funny.
Ogg Vorbis requires a “*stronger* processor, which takes more power” than MP3? Does it really?
Evidence, please.
PS: I can’t see any difference on my CPU usage meter. This is going to take quite some convincing on your part.
//NOPs do not use the multiplication unit. Not using a unit means that stored values do not change, and at the lowest level it means that the gate capacities of the transistors need not be (dis-)charged (reasonably assuming CMOS technology). //
You think its going to make any difference in a computer to the power drain if a few thousand more transistors somewhere in the CPU are swicthing than at other times? When the whole entire rest of the computer (all 450Watts or more worth) are doing exactly the same thing?
hahahahahahahahaha.
Boy, you are one one funny person.
I haven’t had a laugh like that in quite a while.
The old wives tales are coming out in force tonight.
We must REALLY be scraping the bottom of the FUD barrel by now.
Edited 2006-08-16 15:57
> Ogg Vorbis requires a “*stronger* processor, which
> takes more power” than MP3? Does it really?
>
> Evidence, please.
Learn how to read. I did not say a word about OGG or MP3, but about hardware and power consumption.
> PS: I can’t see any difference on my CPU usage meter.
> This is going to take quite some convincing on your
> part.
I highly doubt that you have measured the exact performance differences of well-optimized OGG vs MP3 decoding algorithms running on specialized hardware. More likely you’ve run Winamp or similar with one of those nice downloadable CPU meters running in background. It’s really time for you to consider that portable players do NOT use the same hardware architecture as your PC, and that measuring performance does not work with tons of other programs running at the same time (such as an operating system, let alone other gadgets).
> You think its going to make any difference in a
> computer to the power drain if a few thousand more
> transistors somewhere in the CPU are swicthing than
> at other times? When the whole entire rest of the
> computer (all 450Watts or more worth) are doing
> exactly the same thing?
You are wrong in several points here. First, you speak about a “CPU” which need not be present in a portable player, or at least not in the way you know it from your PC. Secondly, *nothing* in a portable player comes anywhere near 450 Watts. Third, and that’s the core issue here, a NOP can freeze up to the entire processor, and not “just a few thousand transistors”.
//Learn how to read. I did not say a word about OGG or MP3, but about hardware and power consumption. //
The original claim that started this discussion was the the Ogg Vorbis format was more “computationally intensive” and therefore would consume more power.
I have debunked that original old wives tale. I have no idea what you are on about.
//It’s really time for you to consider that portable players do NOT use the same hardware architecture as your PC, and that measuring performance does not work with tons of other programs running at the same time (such as an operating system, let alone other gadgets). //
A portable player still has to have a processor or sequencer of some description. Nearly all such devices have a fixed clock rate. There is no difference to power consumption if the fixed-clock processor is following an idle loop or doing a math calculation – it still runs at the same speed and still reads instructions, decodes them and executes them in its ALU.
Ergo, playing an MP3 compared to an Ogg Vorbis would consume the same power, regardless if the Ogg Vorbis codec required a few additional math sums per sample or not. All it would mean (if it were even true that Ogg Vorbis required more math, which at this point is only an unsupported assertion anyway) is that while playing an MP3, the processor of the player spent a little longer for each sample in the main idle loop and less time in the math of the decoder than it would if playing an Ogg File. There would be no appreciable difference to the power drain of the whole player.
That is the sensible, logical conclusion.
You have presented exactly zero evidence or facts to counter that conclusion. All you have done is called your opposition in the debate an idiot … which loses you credibility for being an ad hominem argument.
http://en.wikipedia.org/wiki/Ad_hominem
Please get back when you have something sensible and credible to say on the matter.
To help you, this is the type of device that would typically fill the role of processor in a portable media player:
http://en.wikipedia.org/wiki/Microcontroller
There you go. Please point out where it says: “doing math takes more power than doing event loops”.
Edited 2006-08-16 23:37
> The original claim that started this discussion was the
> the Ogg Vorbis format was more “computationally
> intensive” and therefore would consume more power.
Right. Then you presented a flawed argument which I countered. That doesn’t make the original statement either true or false, nor did I claim so. That’s why I say, learn how to read.
> A portable player still has to have a processor or
> sequencer of some description. Nearly all such
> devices have a fixed clock rate.
Agreed up to here.
> There is no difference to power consumption if the
> fixed-clock processor is following an idle loop or
> doing a math calculation – it still runs at the same
> speed and still reads instructions, decodes them and
> executes them in its ALU.
This is wrong. First, a processor need not be using instructions at all, but can contain a hard-wired algorithm. Secondly, unused parts need not switch anything, and thus not consume power, such as a multiplication unit. Third, most regular CPUs have a “halt” instruction that puts (almost) the entire processor to sleep, meaning that *nothing* happens at all until an external signal arrives (usually the interrupt line).
> Ergo, playing an MP3 compared to an Ogg Vorbis would
> consume the same power, regardless if the Ogg Vorbis
> codec required a few additional math sums per sample
> or not.
There is another fallacy here. If MP3 requires less operations for decoding, the processor can be downclocked, consuming less power.
> You have presented exactly zero evidence or facts to
> counter that conclusion.
Correct. All I have done is to counter your arguments that a more computationally intensive algorithm would not make a difference. It *would* make a difference. Whether such an algorithm is needed is another issue which we have not touched yet.
> http://en.wikipedia.org/wiki/Microcontroller
What exactly did you expect? Of course this page does not state every truth about microcontrollers word by word. There are things that you have to conclude.
> All you have done is called your opposition in the
> debate an idiot … which loses you credibility for
> being an ad hominem argument.
Wow… Not only did *you* call PowerMacX an idiot from the start and accused me of spreading FUD… now you even plain lie in accusing me of calling *you* an idiot which I never did.
This is really getting boring…
//Then you presented a flawed argument which I countered. That doesn’t make the original statement either true or false, nor did I claim so. That’s why I say, learn how to read.//
My point stands, and you did not counter it at all.
It takes virtually the same amount of power to run math subroutines as it does to run a timing loop waiting for the appropriate time to commence the next iteration of a loop. There is no power saving to be had for a “computationally less intensive” decoder algorithm.
//It *would* make a difference.//
It would not. Pure, plain and straight forward. It doesn’t matter what instructions a processor is following, if it is following them at full clock speed then the power consumed will be virtually constant. Any fluctuations due to different instruction types is incerdibly minor compared with the overall power drain.
//Not only did *you* call PowerMacX an idiot //
Read it again. I said that there would need to be presented a very solid set of facts about how certain instructions used by an Ogg Vorbis decoder algorithm used a lot of power compared to those used by an MP3 decoder algorithm in order for some poster not to appear an idiot for calling me “wrong wrong wrong” when I am not wrong at all. And of course exactly zero such evidence has been forthcoming. That is because it doesn’t exist.
It turns out that in some portable media players there is hardware support for the MP3 decoder. It is done in a separate dedicated chip. The processor chip can be clocked way down … because it is not doing any “computation” at all but merely feeding undecoded MP3 data straight out of the file and into the hardware decoder chip.
There is no such hardware decoders for Ogg Vorbis.
That is the power saving right there.
It has absolutely nothing to do with “computationally less intensive”. That fantasy was bunk, is bunk, and will always be bunk.
//This is really getting boring…//
Agreed. How many times can you be shown to be so dead wrong about this?
Edited 2006-08-17 13:10
Obviously *reading* my replies is too time-consuming for you, otherwise you would pick up my arguments and try to tear them apart instead of just repeating your initial point over and over again. I do not care – I have presented my arguments and hopefully provoked other readers in this forum to think about the issue. There’s no point in continuing this discussion.
Morin: Obviously *reading* my replies is too time-consuming for you, otherwise you would pick up my arguments and try to tear them apart instead of just repeating your initial point over and over again. I do not care – I have presented my arguments and hopefully provoked other readers in this forum to think about the issue. There’s no point in continuing this discussion.
Translation of Morinspeak: “Oops! I’m wrong. I read a report about one player using a lot of battery power when playing Ogg Vorbis, and it never occurred to me that this was because the Ogg Vorbis decoding was done in software compared to the MP3 decoding done in hardware. I just assumed some BS all by myself about the “Ogg Vorbis math must be a lot harder”. Now I have got called on it. Oh hell. I know what to do … I’ll just claim victory anyway, and just pretend the other guy was wrong.”
Edited 2006-08-17 22:34
Just because something doesn’t appear on Wikipedia doesn’t mean it isn’t true. I suggest you look at this:
http://www2.imm.dtu.dk/courses/02206/java/power.html
Every gate that changes from 0 to 1 or 1 to 0 in CMOS logic (used in most digital circuits) leaks a small amount of voltage across the power leads. NOPs minimize these changes and therefor take less power than more complex instructions.
Edit: Meant to reply to RE[8]: No different than MP3 vs OGG.
Edited 2006-08-17 00:09
//Just because something doesn’t appear on Wikipedia doesn’t mean it isn’t true. I suggest you look at this:
http://www2.imm.dtu.dk/courses/02206/java/power.html
Every gate that changes from 0 to 1 or 1 to 0 in CMOS logic (used in most digital circuits) leaks a small amount of voltage across the power leads. NOPs minimize these changes and therefor take less power than more complex instructions. //
Nothing at all there to indicate that playing an Ogg Vorbis file requires more power on exactly the same architecture as playing an MP3 would take.
Nada. Zilch, Zero. Squat. Diddly.
Gates will transition “from 0 to 1 or 1 to 0 in CMOS logic” regardless of the instructions followed. Memory will be fetched, ALU gates will be exercised … etc, etc, etc … regardless of what instructions are followed.
You will only see a difference if entire new areas of the chip are “invoked” … such as a math co-processor. You would be going some to find a math co-processor in a portable medi player device.
lets go further – how much power to run the CPU versus the memory or the LCD display – or more importantly, driving the earphones?
I think you find that the vast majority of battery power would be consumed driving the earphones. This of course is no different for an Ogg Vorbis file or for an MP3 file.
Any other differences in playing each file type on a portable player (constrained as they are to the microcontroller interior) would pale into insignificance compared to the drain taken by driving the earphones.
This argument is getting very silly indeed. Face it people, there is no difference in battery power used when playing an Ogg Vorbis file in a portable player compared to playing an MP3 file in the same device at the same volume.
Nada. Zilch, Zero. Squat. Diddly.
Edited 2006-08-17 00:59
//Gates will transition “from 0 to 1 or 1 to 0 in CMOS logic” regardless of the instructions followed. Memory will be fetched, ALU gates will be exercised … etc, etc, etc … regardless of what instructions are followed. You will only see a difference if entire new areas of the chip are “invoked” … such as a math co-processor. You would be going some to find a math co-processor in a portable medi player device. lets go further – how much power to run the CPU versus the memory or the LCD display – or more importantly, driving the earphones? I think you find that the vast majority of battery power would be consumed driving the earphones. This of course is no different for an Ogg Vorbis file or for an MP3 file. Any other differences in playing each file type on a portable player (constrained as they are to the microcontroller interior) would pale into insignificance compared to the drain taken by driving the earphones. This argument is getting very silly indeed. Face it people, there is no difference in battery power used when playing an Ogg Vorbis file in a portable player compared to playing an MP3 file in the same device at the same volume. Nada. Zilch, Zero. Squat. Diddly. //
I have been looking back over this thread, and I note with considerable surprise that several of my posts on this topic – debunking the old wives tale myth about Ogg Vorbis requiring more battery power – have been modded down.
What for exactly? For being right?
>Not using a unit means that stored values do not change,
>and at the lowest level it means that the gate >capacities of the transistors need not be (dis-)charged
>(reasonably assuming CMOS technology). Minor power >comsumption remains through leakage currents, but this
>is orders of magnitude smaller.
You’re not telling the whole story here. Typically, logic units are clocked, so they consume basically the same power per cycle whether or not the logic values change. However, newer processors can shut off units that are not in use and save power that way.
> You’re not telling the whole story here. Typically,
> logic units are clocked, so they consume basically the
> same power per cycle whether or not the logic values
> change.
Hmm, that would be new to me. To my knowledge, if a clocked register tries to store the same value again, or receives a clock edge while clock-enable is low, no transistors are (dis-)charged.
You are the idiot.
A large part of power consumption comes from transitor switching. nops keep the transistors for boolean and sequental logic in a static state and thus consumes far less power than when changing every clock cycle.
//A large part of power consumption comes from transitor switching. nops keep the transistors for boolean and sequental logic in a static state and thus consumes far less power than when changing every clock cycle.//
Sigh! The transistor in a processor will switch regardless if the processor is running software in an “idle” loop waiting for an event or it is in the middle of a math calculation. It makes no difference to the processor. It can’t even tell if the instructions it is following at any given time are part of “main application event loop” or “guts of the codec”.
//You are the idiot.//
Not at all.
Please find the text which says: doing a math operation takes more watts than a loop instruction:
http://en.wikipedia.org/wiki/Cpu_power_consumption
Put up or shut up, bozo.
> It makes no difference to the processor. It can’t even
> tell if the instructions it is following at any given
> time are part of “main application event loop” or “guts
> of the codec”.
Except that a waiting loop can execute without executing instructions. You should read about about “halt” or “wait” instructions.
//Except that a waiting loop can execute without executing instructions. You should read about about “halt” or “wait” instructions.//
Sigh.
This question has been settled. You have come in very late to this party.
The source of the belief that it takes more battery power to play an Ogg Vorbis file on a portbale media player than it does an MP3 file derives from the fact that apparently a lot of portable media players have hardware support for MP3 decoding. All the processor needs to do in this case is feed the MP3 data straight into the hardware codec – at the bitrate of the audio data – which is glacial compared to the full clock spped of a processor. The processor can be clocked way down and still be able to handle this task … or indeed wait states or halt instructions could be used instead (given the hardware decoder) on a per-sample basis to slow the processor right down.
Most media players, OTOH, have to decode Ogg Vorbis in software. This requires running the processor at full speed, and yes, undoubtedly, that would consume a great deal more power. There is no opportunity to use wait states or halt instructions in a software-based media player.
However … my original point statnds. If an MP3 file is played using the same processor resources as an Ogg Vorbis file then the power consumed would be virtually identical … virtually regardless of the amount of math used in each cycle of the decoder.
One only achieves “low power MP3 play” in a portable media device by using hardware assistance and putting the MP3 decoder algorithm in dedicated hardware logic chips.
> The source of the belief that it takes more battery
> power to play an Ogg Vorbis file on a portbale media
> player than it does an MP3 file derives from the fact
> that apparently a lot of portable media players have
> hardware support for MP3 decoding.
Every single one of my arguments assumed an optimal combined hardware/software solution for both OGG and MP3. So my arguments still hold.
//Every single one of my arguments assumed an optimal combined hardware/software solution for both OGG and MP3. So my arguments still hold.//
Your argument does not hold at all.
The difference in power drain that is apparent between Ogg and MP3 that is seen on some portable media players is due entirely to the fact that the MP3 decoder function is not done by a software computation at all – it is done by an external hardware MP3 decode chip.
If both Ogg and MP3 decoders are provided with “an optimal combined hardware/software solution” then they will consume virtually exactly the same amount of power to run … less than 1% difference, probably much less different than that.
Edited 2006-08-17 13:22
“To decrease the power used one must power down devices and/or slow down the CPU clock. This is not done if one is playing an MP3 file compared to an Ogg Vorbis file … utter fantasy to think that playing an Ogg Vorbis file would consume more power.
Here is a challenge for you … come up with a valid reason why an Ogg Vorbis file playing in a media player would consume more juice, or a real world measurement, or some even remotely credible fact (as opposed to just blithering “wrong, wrong, wrong”) … and I will consider that you might possibly not be the complete idiot that you appear on first sight to be.”
First: According to PortalPlayer’s brief specs for the 5020 & 5022 chips (used in the iPods or, in other words, something like 3/4 of all DAPs):
“Power Management
The PP5022 features advanced power management capabilities that enable shutdown of most functional modules when not in use, providing significant power savings and longer battery life.
[…]
– Clock frequencies programmable from 32 KHz to
100 MHz for optimal performance and power
consumption
– Integrated 8-bit, 4-channel ADC for battery level
monitoring
– Ultra-low 1.2V core supply voltage and system
enabled dynamic voltage scaling”
http://www.portalplayer.com/products/documents/5020_Brief_0108_Publ…
(note that Ogg vs. MP3 wasn’t the main reason I posted “Wrong.” – it was your assertion that a CPU under heavy load used exactly the same energy than even a mostly idle one – do you still believe that?)
So, do all those power management capabilities make any difference when playing OGG vs. MP3? Let see:
Second:
This is post by one of the developers of Rockbox, from a thread discussing Ogg playback ( http://www.rockbox.org/ ):
“Yeah, anytime the vorbis stream goes over 300kbps for extended periods, it can’t keep up. This is because 1) the Tremor decoder is not yet well optimized and 2) rockbox on the H340 only runs at 90mhz boost currently. The other side effect of this is that when playing OGG/Vorbis files is BRUTAL on battery life currently. Hopefully once we get some better profiling information from the decoder (see my thread on profiling for my progress in that arena), the decoder can be better optimized so that even at 90mhz it won’t be skippy.”
Thread link: http://forums.rockbox.org/index.php?topic=2052.0
Third:
Stop insulting people, it doesn’t add anything to the discussion.
//(note that Ogg vs. MP3 wasn’t the main reason I posted “Wrong.” – it was your assertion that a CPU under heavy load used exactly the same energy than even a mostly idle one – do you still believe that?)//
(1) A CPU is not “idle” when it is playing an MP3 file or an Ogg file.
(2) I didn’t say anything about “idle”. I was debunking the myth that an allegedly “more computationally intensive” real-time task would take more power than and allegedly “less computationally intensive” doing an equivalent real-time task. There is no “idle” anywhere in that. For the real-time task of playing a media file (of any format), the CPU is either doing the math for a particular sample, or it is waiting (in a loop) for it to be time for the next sample. In either state it is not idle. Neither state allows an opportunity to slow the clock.
//This is post by one of the developers of Rockbox//
What does the inefficiency of one particular experimental implementation of an algorithm on one particular architecture by a programmer have to do withthis argument?
Show me a production player (with a finished algorithm) that supports both Ogg & MP3, and show me where it uses more juice when playing the Ogg file – or give up on your myth.
//Stop insulting people, it doesn’t add anything to the discussion.//
Agreed. I’ll do it if you do it first, Mr wrong, wrong, wrong.
Edited 2006-08-17 03:53
(2) I didn’t say anything about “idle”. I was debunking the myth that an allegedly “more computationally intensive” real-time task […]
?
You said: “A CPU doing math on audio data uses up exactly the same juice as another one which is polling in a loop “is there anything for me to do yet? no? well is there anything for me to do now?” etc.
Same CPU, same instruction fetches, same memory … everything. Exactly the same power consumption to execute. ”
What does the inefficiency of one particular experimaental implementation of an algorithm on one particular architecture by a programmer have to do withthis arguemt?
Show me a production player (a finished algorithm) that supports both Ogg & MP3, and show me where it uses more juice when playing the Ogg file – or give up on your myth.
And then what? What would prevent you from giving me a similar answer? (the Ogg decoder is not as optimized as it could be!)
Rockbox is the most widely used firmware replacement to allow the playing of Ogg files on the vast majority of DATs which do not have that capability by default.
That doesn’t count? OK. To *me* it does.
//You said: “A CPU doing math on audio data uses up exactly the same juice as another one which is polling in a loop “is there anything for me to do yet? no? well is there anything for me to do now?” etc.
Same CPU, same instruction fetches, same memory … everything. Exactly the same power consumption to execute. ” //
Yes. Exactly. It is either doing the math for one sample, or it is timing until the next sample is due. In neither instance is it idle. For the task of playing an Ogg Vorbis file or an MP3 file in a portable media player the samples are required to be output to the output DAC at (typically) a rate of what? 44 KHz or so? Where is the idle time? We are in the middle of a real time decode & output task.
That is wildly different state to … well, that playlist is done, so slow the clock, go into semi slumber, “screensaver” mode if you will, there is nothing for us to do until the operator comes back and presses a button to wake us up.
Edited 2006-08-17 04:33
//So, do all those power management capabilities make any difference when playing OGG vs. MP3?//
Lets see. Let’s stumble on the very first device on google that matches both “portable media player” and “ogg”:
http://reviews-zdnet.com.com/Kingston_K_PEX_Portable_Media_Player_2…
Looking at the “bad” part of the review – nope. No mention whatsoever of “uses more juice playing ogg files”.
What about this page: http://wiki.xiph.org/index.php/PortablePlayers
… where “low power consumption” is mentioned in the exact same breath as Ogg Vorbis capability (Pretec’s Allegro).
Edited 2006-08-17 04:26
There exist many mp3 decoders in silicon, and a lot fewer decoders that support ogg.
The linux based ipod firmware takes up a lot more power to decode ogg vorbis through the arm cpu then the dedicated decoder chip.
//There exist many mp3 decoders in silicon, and a lot fewer decoders that support ogg.
The linux based ipod firmware takes up a lot more power to decode ogg vorbis through the arm cpu then the dedicated decoder chip.//
Now that is credible.
That is an instance of one portable media player using two entirely different parts of its architecture for playing the different media file types.
But that also happens to be an entirely different thing to “ogg vorbis decode algorithm takes more juice than an MP3 decode algorithm” … if the same computational engine is being used for both.
Edited 2006-08-17 04:50
I really love Real Networks… These types of announcements are the only ones I ever look forward to coming out of LinuxWorld. I mean this one’s for consumers, not every company distributing to everyone but consumers.
If they ever do bring Rhapsody/RealMusic Store to Linux wouldn’t they bring Harmony (WMA/M4P) for portable device support/conversion aswell?
I would assume they would also bring it to the Mac around the same time.
Good Job Real, I’m glad…
This is another step for linux, i will enjoy being able to watch more videos on my linux box next year
This is of course a good move but it is anemic for the amount of codecs we need from linux.
Linux must fight a real war with Hardware and Software giants to take its rightious share of support according to the antitrust legal rules approved here in US.
After all, almost all companies practice illegal actions towards competitors, and thus no embarracement to sue them. God Bless our legal system!
Don’t forget native 64bit support using the new codecs… the current sollution for using 32bit binary codecs on a 64bit system ain’t that clean :p This prevented me from using a 64bit linux distro because i don’t have the the time to get all the 32bit multimedia emulation to work… chrooting, emulating, linux32,… bah
//Don’t forget native 64bit support using the new codecs//
All the more reason to use a format with an open-source codec.
Plenty to choose from:
http://en.wikipedia.org/wiki/Xiph.org
Just re-compile it for the 64-bit architecture, surely?
Edited 2006-08-16 11:33
//Just re-compile it for the 64-bit architecture, surely? //
Only if it’s written in portable code from the start.
There is a reason, why applications like OpenOffice.org doesn’t run native amd64.
it will be only for linux x86?
no other os?
no other processors?
TIA
I’m hoping it works in freebsd
What is Real getting out of this? Does this hurt MicroSoft in some way? Or does Real figure this is a way of getting it’s software on more computers? If any of you have loaded their software on a Windows box recently you know they try to take over your entire computer (for audio/video). Will they try the same thing for linux? All I want is the codec.
In a perfect world everyone would use FLAC (home) and Ogg (portable), but I doubt anything will beat MP3 for the forseeable future. I just wish people would vote with their wallets and only buy products that follow some kind of standard. That is why I am boycotting Sony. The world doesn’t need another audio format (minidisc), video format (beta), memory card format (memory stick), CD format (with rootkit), or exploding batteries (Dell).
Granted MP3 prevents Ogg and FLAC from making inroads as the defacto DRM-free codec, but there are two bright sides:
* its patent will expire in only 4 years
* it is so popular that AAC and WMA aren’t able to dominate the market.
Very good I wonder how much Real is paying MS though. Now I hope they make Kaffeine faster and look better in the browser.
//Very good I wonder how much Real is paying MS though. Now I hope they make Kaffeine faster and look better in the browser.//
Apart from an ability to utilise the codecs, what does Kaffeine have to do with Real?
http://en.wikipedia.org/wiki/Kaffeine
Kaffeine is at least as fast and as full-featured on my system (under KDE 3.5.3) as any other media player.
See for yourself:
http://kaffeine.sourceforge.net/index.php?page=features
AmaroK is better for audio, however. AmaroK roKs.
In Konq browser it is very slow to load the media compared to Win/WMP, for instance. http://www.tomgreen.com
Full featured? Why doesnt it expand full-screen in the browser without having to rev up the main kaffeine interface (which is very slow) compared to instantly with Windows WMP. Where are the full screen controls. Plus no themes so it looks generic.
Open codes will fuel better Linux media players.
//In Konq browser it is very slow to load the media compared to Win/WMP, for instance. http://www.tomgreen.com //
Not in my experience.
In KDE 3.5.3 or KDE 3.5.4 things are very snappy indeed.
If things are a bit slow for you, you might try pre-loading an instance of Konqueror (this is what Windows does for explorer). For my KDE setup, the configuration is under: Menu -> Configuration -> KDE -> System -> KDE performance. On the “Konqueror” tab, check the box marked: “Preload an instance after KDE startup”. I personally find I don’t have to do that, but you might feel differently about it.
//Full featured? Why doesnt it expand full-screen in the browser without having to rev up the main kaffeine interface (which is very slow) compared to instantly with Windows WMP. Where are the full screen controls. Plus no themes so it looks generic. //
The GUI preferences are different for every user, but I do not understand your point here. The main interface is not slow at all. Fullscreen mode is the first item under the view menu. Shortcut key is Ctrl-Shift-F. A “Fullscreen” button can be placed on any toolbar, and this is easily configurable.
You want themes? Use just the normal Xine GUI player, rather than the Kaffeine front end.
Sigh! You can’t please everyone, I guess, but do please try to come up with actual real valid criticisms.
//Open codes will fuel better Linux media players.//
The better players are already here for Linux. It is simply that you just missed them.
Edited 2006-08-16 13:23
When I right click on the Kaffeine window in the Konqeror browser it only gives me the option to “play in Kaffeine player externally.” No option for fullsccreen and when I press CTRL-SHIFT-F it only makes the browser bigger not the Kaffeine window fullscreen.
I dont see any fullscreen coltrols embedded into the player like in Totem.
//When I right click on the Kaffeine window in the Konqeror browser//
Say what?
Right-click on a media file in Konqeror … is that what you mean? (Hint: there is no such thing as a “Kaffeine window in the Konqeror browser”).
Right-clicking on a media file in Konqeror gives two options for me … (1) “Open with” -> Kaffeine or gxine, and (2) “Preview in” -> Kaffeine of Gstreamer part.
A preview (option 2 above) will open the file for viewing within Konqueror. There is no fullscreen for a preview.
Opening the file in the full player (option 1 above), however, allows the full use of the player. As I said, fullscreen is under the view menu once Kaffiene has opened. Or one can use Ctrl-Shift-F. Or one can put a “Fullscreen” button on one of the toolbars. Your choice.
Alternatively, one can start the full player from the menus.
For me, it is: Menu -> Multimedia -> Video -> Kaffeine.
Edited 2006-08-16 13:58
I am talking about a browser plugin where the media player is embedded into the webpage.
Sorry if we got our wires crossed.
Noone wants to have an external player pop up in today’s Web 2.0 environment. They want it to play within the browser (embedded)
Edited 2006-08-16 14:00
//I am talking about a browser plugin where the media player is embedded into the webpage.
Sorry if we got our wires crossed.
Noone wants to have an external player pop up in today’s Web 2.0 environment. They want it to play within the browser (embedded) //
If a media file is playing embedded within the webpage, then presumably the webpage is dictating the size of the display window, not the browser nor the media player part.
“If a media file is playing embedded within the webpage, then presumably the webpage is dictating the size of the display window, not the browser nor the media player part.”
Ok and when I double click onto the media player box in Windows WMP goes fullscreen, not with Kaffeine.
//Ok and when I double click onto the media player box in Windows WMP goes fullscreen, not with Kaffeine.//
That is just a GUI difference.
I think what you really want is the mplayerplugin, and mplayer. Try that.
Thanks. It seems mplayer works well enough but would like to add a snazzy theme to it.
Edited 2006-08-16 14:34
//Thanks. It seems mplayer works well enough but would like to add a snazzy theme to it. //
I just tried it myself. There is a fullscreen button in the lower-left corner of the plugin window. Right-click on the plugin window gives you an option to configure the plugin.
As far as “snazzy” goes – I am just now downloading “mplayer-skins”. I’ll let you know how that goes.
//As far as “snazzy” goes – I am just now downloading “mplayer-skins”. I’ll let you know how that goes.//
Hmmm, config problem.
I can see a lot of skins in a directory, but mplayer itself cannot.
Mainly it’s rediculous that Firefox still has no in browser extension support for popular Linux media players in it’s etensions GUI. Open codecs might help to get Firefox to bother I guess
Edited 2006-08-16 14:01
//Mainly it’s rediculous that Firecfox still has no in browser extension support for popular Linux media player//
Try this (or something similar for your setup):
Menu -> Configuration -> Packaging -> Synaptic package manager.
Once you have Synaptic up and running, press the “Search” button.
Enter the Search term “Mozilla”.
Install one of either:
(1) gxine-mozilla, or
(2) kaffeine-mozilla, or
(3) mozilla-plugin-vlc.
Is that for embedded players embedded into the browser. I am not talking about external players.
Thanks
//Is that for embedded players embedded into the browser. I am not talking about external players. //
OK, the best option then would be:
(1) start Synaptic.
(2) Press the Search button.
(3) enter the search term “mplayer”
(4) select for installation all three of the following items: mplayer, mplayer-gui and mplayerplugin.
(5) press “apply”.
The description for “mplayerplugin” reads: “Mplayerplug-in is a Netscape/Mozilla browser plugin to allow playing embedded movies on web pages.”
I believe that is what you want.
Do a search for the “MediaPlayerConnectivity” extension. It lets you select exactly what player to use for each file format.
While you’re at it, if you use KMPlayer (no doubt there are others with the same feature), you can record the stream as it comes down.
KMplayer is nice. I was a little weirded on its raw playlist screen last time though but has a nice interface. I hope wit would install embedded and unistall mplayer when I install it.
Thanks
Windows Media is not exactly a format that has swept the internet, and Novell and Real want to help Microsoft’s share? Oh dear.
It’s already too late. When you get to any major website for movie/game trailers, all of them offer WMV because this comes with default Windows installation. This tendency is not going to change, since most corporations target the most common denminator in the user base which is WMV. It’s is same like with IE – a website that doesn’t render correcltly in Firefox/Opera is acceptable, but opposite is a no go. The only hope is in governments that would force MS not to include their ‘standards’ in the most popular OS in the world, and only after that 3rd party vendors will target the best product/format instead of what comes with Windows by default.
What I don’t understand is why MPEG-4 isn’t used more. It’s standard, widely supported, and streamable.
In Soviet Russia, the power and CPU used in OGG and MP3 consume YOU!
It is a great news to all the Linux users and Suse lovers. I hope Novell continue like this even faster than before. I paid for Suse enterprise 10 only because I wanted to support this company as a leader in Linux world and guess what ? yes my money worth every single peny.
Edited 2006-08-17 10:10
I compiled today Mplayer from Subversion on my FC5 x86_64 and WMV9 plays perfectly :-))