Anyone used WM9 encoder for any encoding lately?

[Sgt_Strider]

How's the quality and size wise when compare to real video 9 or xvid/divx?

[bond]

I encoded a small (low bitrate) clip (matrix: the agent interrogating morpheus) with wm9 (so no real testing...) and in my opinion in that clip wm9's quality is between xvid and rv9.
Regarding to things like mosquito noise or background motion it beats xvid but it has less sharpness although it is sharper than rv9!
I dont want to make any further descriptions as i didnt really test it but altogether wm9 doesn't look that bad and it is worth trying it!

[amirm]

I work for Microsoft so please take this in context.

Compared to xvid, you should be able to get the same quality at roughly half the size, assuming similar encoding params are used for both (no fair testing one in CBR and other in VBR for example). I say "roughly" since your preference for compression artifacts may be different than mine. But you should always be able to beat it in efficiency.

As to Real, Bond's observation regarding our sharpness compared to Real is correct. We do not soften the video like they do and as such, are able to match the source much better. Needless to say, I do not accept that our quality is not as good as Real :). Sharpness is part of the equation unless you are happy with VHS tape instead of DVD :). Of course, if you prefer your video soft, you can always prefilter it (which reduces compression artifacts since there is less data to compress). But note that the other way around is not true (you can't sharpen Real's video).

Also of note is that we also have a full discreet multi-channel audio codec (WMA Pro) which goes all the way up to 24-bit/96Khz and 8 channels (7.1). We also give you full choice of encoding modes for audio including 2-pass and VBR. And if this is not good enough, we also have a full, lossless codec that goes from stereo to 24-bit/96khz 7.1 (you can even rip CDs into this using Media Player 9). Real doesn't give you any of this and neither does Dolby AC-3/DTS. For 5.1 btw, you can start at just 128kbps which gives you three times the channels of MP3 in the same space!

Amir
Microsoft

[31 Flavas]

Originally posted by amirm

Real doesn't give you any of this and neither does Dolby AC-3/DTS.Any chance for Microsoft to "open the code" like Real?

Or at least release binaries and/or servers for any os besides Windows?

[Sgt_Strider]

Release date for Wm9 encoder is January 7, 2003 right? The new microsoft codec may seem more powerful and better, but the thing is that will it have the same kind of support from the industry as mpeg-4 does? I want to be able to play my encoded videos on a standalone player one day...Also will ffdshow work on wm9?????????

[karl_lillevold]

Originally posted by amirm
I work for Microsoft so please take this in context. Compared to xvid, you should be able to get the same quality at roughly half the size.
As to Real, Bond's observation regarding our sharpness compared to Real is correct. We do not soften the video like they do and as such, are able to match the source much better. Needless to say, I do not accept that our quality is not as good as Real :). Sharpness is part of the equation unless you are happy with VHS tape instead of DVD :)
I am happy to see Amir insert a little WMV9 flavor to this forum.

MS marketing department's claim over a stock MPEG-4 codec may be true, but after having tried XviD, it is clear significant improvements have been made within the aging MPEG-4 framework. It does not sound like Amir is taking this into account, but fortunately readers of this forum know how to spot such marketing statements for what they are.

I have been and still am strongly encouraging experts on doom9 to run their own experiments instead of taking my or other's statements as true. And it sounds like they do, in fact finding great results for RV9, based on feedback here in this forum.

Snide and false Microsoft statements regarding RV9 quality are ineffective (It does not help to tag on a smiley face). Doom9 readers can themselves evaluate RV9's cababilites.

Previously I have not made any statements about WMV9 in this forum, and I would not have, if it were not for the above post. Of course sharpness is part of the equation. Sharpness was one of the parameters judged by the Keylabs test (http://www.keylabs.com/results/realnetworks/vidperf9.shtml) where RV9 thoroughly beat WMV9 (Beta). Note how clarity was in fact the parameter where RV9 outperformed WMV9 by the greatest margin.

Yes, the test was paid for by RealNetworks, but if you read the test details, you can see it was designed as fair as possible. One of the aspects taken into account was that the WMV9 encoder was not allowed to overspend bits like it normally does for all short clips (very convenient for comparisons, unless the user notices the filesize). With this overspending I am referring to much more than the claimed 5% file overhead in ASF.

RealNetworks supports freedom of choice for platforms, devices, and operating systems, as well as alternative players and encoders. RealVideo 9 runs as well on Linux and OS X as on Win32, and through Helix, binaries will be compiled for any plaform, OS, and device where there is interest.

[dattrax]

You are more likely to see WM9 on a stand alone player than MPEG4. I'll tell you the reason for this.

WM9 is a complete audio/video/system codec package. There are no .AVI/.OGG wars, no MP3/AC3/OGG for audio and no ASP/non ASP for the video. An example for this would be DivX5.0 Qpel implementation which is non standard. Do you make hardware which decodes non standard streams? I'm not sure if thats wise.

I think that MPEG4 has become too poluted with different implementations, so it just makes it more difficult and less appealing.

Now windows media has bframes and qpel, as well as some features MPEG4 hasn't got, it should be possible to get as good as, if not better than MPEG4

Jim

[trbarry]

MS marketing department's claim over a stock MPEG-4 codec may be true, but after having tried XviD, it is clear significant improvements have been made within the aging MPEG-4 framework.

Aging? Or still maturing, but with rapidly aging lawyers? ;)

Since I am already quite an Xvid fan I probably did not give WM9 a fair shake or completely objective testing. But for my own purposes compressing down-rezzed HDTV material I could not find those WME claimed improvements. So I stuck to Xvid.

But I'm still very glad to see folks from both the MS and Real video teams here.

- Tom

[Didée]

When reading the claims of RV9 and WMVV9, I must admit that I am deeply impressed:

Same quality as modern mpeg-4 codecs at half the size!

I am so deeply impressed, I can't tell you.

almost ROFL

Didée

[Christo]

I compared WMV-9 with DIVX-3 at best possible Video quality settings such az:
WMV-9 100% quality based VBR and DIVX-3 6000kbps(max) Slow Motion. The results were WMV-9 produced file size 3 times (300%) more than DIVX-3 with impeccable Video quality. Well, obviously WMV-9 100% quality based VBR was overkill…

I reduced WMV-9 quality based VBR to 98%. The results were WMV-9 produced file size 50% more than DIVX-3 with video quality again clearly superior to DIVX-3. Zero artifacts or blockiness or blurriness.

Reducing farther WMV-9 quality based VBR to 96% I mached DIVX-3 (6000kbps Slow Motion Codec) in file size. But video quality of WMV-9 was still clearly superior. However the light blockiness in a very high motion scenes occasionally was noticeable. Please note 640x480 WMV-9 was played back at full 24fps on P3 500 MHz with NO post processing.

At 90% quality based VBR, the video quality of WMV-9 was finally comparable to DIVX-3 6000kbps(max) Slow Motion. However the file size of WMV-9 was about 40% less. Funny, the main problems (blockiness) for DIVX-3 Slow Motion Codec were exactly in Slow Motion scenes while for WMV-9 the main problems (blockiness) were in High Motion.

[TheXung]

:confused: This thread renewed my interest in wm9 and after giving it another shot; I don't know what to make of it. I can say, it's not the same type of comparison that we're use to making. The quality degradations are different; with the mpeg4 codecs, we're use to seeing the noise increase as size gets smaller, with wm9, it gets less detailed, not so much blurry because I don't think mpeg4 can reproduce the same image using postprocessing.

The claim that it produces a filesize of 50%, is something I've yet to see. Going into the lower bitrates, I would say that wm9 does produce more pleasing output than our traditional mpeg4 codecs, but these are bitrates what I'm not interested in. At 1 CDish bitrates, I cannot get wm9 to have the same amount of detail as xvid, though it has less noise around the edges, though at 1 CD bitrates, xvid edge noise isn't bothersome. I would rather have the detail.

To sum it up, I would say this is the vorbis of video codecs, tuned for decent quality at lower bitrates. However I am more of a videophile than an audiophile, so I prefer the MPC of video codecs.


I must admit that I am impressed by the decoding speed of this codec. It is much faster than anything that ffdshow can do, so I assume, there isn't any postprocessing being done. This leads me to think that this is either a wavelet based codec or a JVT type implementation, though the decoding speed and the fact that wm9 cannot attain the same level of detail as other mpeg4 codecs sorta rules out JVT. I guess only amirm can settle what sort of codec is this.

Now windows media has bframes and qpel, as well as some features MPEG4 hasn't got, it should be possible to get as good as, if not better than MPEG4
When was it established that wm9 has either qpel or bframes?


Perhaps, I will revisit this codec the next time I do anime.

[The Belgain]

Right. A question for amirm about WM9. I know the chances of it being cross-platform are pretty slim (apart from reverse engineering by Linux types), but will this even be truly cross-windows?

According to Microsoft's website the WMA Pro codec only plays surround sound on WinXP and downmixes to stereo in other windows. Is this true? If so, are you surprised that most on this forum don't trust Microsoft enough to use their video technology even if the codec offers very good quality?

[Sgt_Strider]

Originally posted by TheXung
:confused: This thread renewed my interest in wm9 and after giving it another shot; I don't know what to make of it. I can say, it's not the same type of comparison that we're use to making. The quality degradations are different; with the mpeg4 codecs, we're use to seeing the noise increase as size gets smaller, with wm9, it gets less detailed, not so much blurry because I don't think mpeg4 can reproduce the same image using postprocessing.

The claim that it produces a filesize of 50%, is something I've yet to see. Going into the lower bitrates, I would say that wm9 does produce more pleasing output than our traditional mpeg4 codecs, but these are bitrates what I'm not interested in. At 1 CDish bitrates, I cannot get wm9 to have the same amount of detail as xvid, though it has less noise around the edges, though at 1 CD bitrates, xvid edge noise isn't bothersome. I would rather have the detail.

To sum it up, I would say this is the vorbis of video codecs, tuned for decent quality at lower bitrates. However I am more of a videophile than an audiophile, so I prefer the MPC of video codecs.


I must admit that I am impressed by the decoding speed of this codec. It is much faster than anything that ffdshow can do, so I assume, there isn't any postprocessing being done. This leads me to think that this is either a wavelet based codec or a JVT type implementation, though the decoding speed and the fact that wm9 cannot attain the same level of detail as other mpeg4 codecs sorta rules out JVT. I guess only amirm can settle what sort of codec is this.


When was it established that wm9 has either qpel or bframes?


Perhaps, I will revisit this codec the next time I do anime.

We're almost 1 week away from the release of Wm9 series :). We'll see if it does live up to its claims.

[amirm]

Let's see if I can answer all the questions posted here:

Quality: this is really up to you all to decide. I will say this though. WMV is tuned and optimized differently for different data rates. While at lower data rates you see us trade distortion for lower resolution (a behavior that average viewer prefers), you will not see as much of this at higher data rates. Indeed, this has been key to acceptance of our codec in digital cinema. Our ability to preserve even noise (e.g.. film grain) has been super important to our success in HD encodings for this market. Double blind tests done by studios show that we outperform all leading codecs at 1080p (1920x108) and 7Mbit/sec on both resolution and subjective quality. Unfortunately, I can not disclose anything more about the test due to their confidential nature so take this for what it is worth. But I will stand behind our ability to produce better quality than any other codec at the same or even lower bit rate/resolution. Of course you should do your own testing and decide as some of you have.

Algorithm: we have not publicly disclosed WMV algorithms. So any information about what techniques we use would be a guess. While Microsoft chairs the JVT standardization effort, WMV is in no way related to JVT. They are completely different animals.

Cross-platform: We have a strong policy to make our codecs available on any platform our customers want. We do not require anyone to run our OS or any OS at all. Internally at Microsoft, we port our codecs to a wide range of hardware and OS platforms from MIPS and ARM CPUs (used in PocketPCs) to Mac OS.

We offer a full porting kit that includes specifications, verification suites, training, etc. for our partners to port our codecs to *any* platform they wish including those with competing OSes. We currently have more than 40 partners who have licensed WMV and are porting it to various chips, processors, and OS platforms.

Since Real has not opened they codecs, I don't understand the question as to whether we are as good as them in this regard. I would say that our porting program is far more broad than theirs to date (we have 200+ non-PC devices that play WMA today!).

Stand-alone players: These players will become available in 2003. In September we announced a (logical) CD-R format with Panasonic called HighMat (http://www.highmat.com) that allows fast playback of compressed content. WMV and WMA are mandatory formats (together with MP3) so you should expect to see a lot of CD and DVD players that can play WMV (many play WMA already).

Karl's comments: Your colleagues and I have been going back and forth in various forums for the last 5 years. Smart people read these things and find some good learning there between all the arrows being thrown and marketing claims. My job is to keep you honest and I am sure you will do the same in return. It is all in good fun (why else would we have this discussion during the holidays and our time off? :)) and interest of the forum.

Amir
Microsoft

[Tommy Carrot]

If the porting is so easy (and why wouldn't be?), i just hope there will be wm9 encoder on win98 platform, and the fact there wasn't such in the beta version is not the part of the strategy to force everyone to use windows xp. (wheww, what a nice sentence:))

I really think that would be useful to many. (I know about nic's little util, but it always crashed.)

[TheXung]

While at lower data rates you see us trade distortion for lower resolution (a behavior that average viewer prefers), you will not see as much of this at higher data rates

This explains a lot. This really explains a lot. The technology behind this sounds exciting. I can't wait to see it in it's finished form. It is ashame that reduce resolution encoding was somewhat dismissed or more accurately, being held back by mpeg4 profiles on the xvid mailing list. While I'm sure MS has the best in the field working on and optimizing this codec, my one wish is for the configuration of this codec to be a little more open. While there may have been a lot of testing to fine tune the quality, I would still like a little lee-way to adjust how much sharpness and distortion I prefer. I believe there is no perfect quality optimizing that can be done; on a TV, lower resolution is not as distinguishable as it is on a computer monitor or homemade projector. Without the ability to retain on par sharpness and distortion with xvid, wm9 would be utterly useless to me.


Just out of curiousity, what sort of computer is needed to decode 7Mbit video at HDTV resolutions? Is this something that a single computer can do?

[trbarry]

Stand-alone players: These players will become available in 2003. In September we announced a (logical) CD-R format with Panasonic called HighMat that allows fast playback of compressed content. WMV and WMA are mandatory formats (together with MP3) so you should expect to see a lot of CD and DVD players that can play WMV (many play WMA already).

If this included the ability to play HDTV resolutions it would certainly be enough to make me give it another try.

I'd love to start encoding my own 1080p (do now) but don't currently have any way to play them except maybe hardware accelerated MPEG-2.

But if the players were limited to 480 lines then I'd probably have much less interest. Will they likely do HD?

- Tom

[The Belgain]

Well, I just installed the release candidate for WM9 and guess what? In WMP9, when I rip a CD and choose the compression I get a dialog box informing me that WMA now includes a lossless codec, available only on Windows XP!

Will this be the case in WM9 Final? Is there any good technical reason why this is so, or is this just another "incentive" to get peoplpe to give up older MS OSes and move on to XP?

I don't particularly care about the lack of a lossless audio codec since I'm not planning on using wma (flac's fine for lossless, but I don't use lossless anyway). I did a quick wma test and the audio quality was very poor. Still the same old artefacts and very bad stereo image - much worse than Vorbis (probably better than mp3 at this bitrate, though that's not really surprising). This was on a Metallica track, using VBR giving a bitrate of 81, comparing to Vorbis -q 1 which was 82 kbit.

[amirm]

Originally posted by TheXung
Just out of curiousity, what sort of computer is needed to decode 7Mbit video at HDTV resolutions? Is this something that a single computer can do?

720p (1280x720@24fps) can be decoded on anything faster than P4 2.2 Ghz with a decent graphics card (any modern 3-d card should do -- the main thing we need is high write throughput (fast-write) and nothing else). As a matter of fact, the dell laptop that I am typing this on can easily decode any 720p content. It is nice to have a few HD movies with you when you travel :).

For 1080p (1920x1080), the demands go up. While some content can be decoded on 2.8 Ghz PC, you really need a 3 Ghz system to decode anything you throw at it.

Amir
Microsoft

[amirm]

Originally posted by Tommy Carrot
If the porting is so easy (and why wouldn't be?), i just hope there will be wm9 encoder on win98 platform, and the fact there wasn't such in the beta version is not the part of the strategy to force everyone to use windows xp. (wheww, what a nice sentence:))


We don't support Win98 with our encoder because the bulk of our (traditional streaming) customers use Win2K or XP. These platforms are more reliable and hence the choice. However, the Format SDK itself (which handles decoding and encoding of our format/streams) is supported on Win98. So anyone can develop a Win98 encoder/player if they like.

Amir
Microsoft

[trbarry]

No comments about whether the stand-alone players will likely handle any HDTV resolutions?

- Tom

[amirm]

Originally posted by The Belgain
Well, I just installed the release candidate for WM9 and guess what? In WMP9, when I rip a CD and choose the compression I get a dialog box informing me that WMA now includes a lossless codec, available only on Windows XP!
We live in the Windows division. Put another way, our salary comes from the sales of Windows. So yes, we do have to provide an incentive for folks to use our latest OS once in a while. Without it, there would be no fund to develop such technology. Having said this, many times we shy away from supporting our older OS platforms simply because it lengthens our development cycle. The permutations of codecs, SDKs, players, encoders, servers and OS versions can go crazy if we don't watch it. So once in a while we take shortcuts on platforms that are older. We realize not everyone is happy with these choices and I apologize for that.

Note that WMA Lossless content will playback everywhere. The only thing that is XP specific is the ripping capability.

BTW, we just announced a new version of Plus! (http://www.microsoft.com/windows/plus/default.asp)for XP (called Digital Media Edition) that has a general purpose format converter to and from WMA/WMA Pro and MP3. This way, you can rip everything in WMA Lossless and then target other formats at will. Of course, you can also build your own free converter using our SDK. But Plus has a bunch of other features such as Photostory that creates efficient "movies" (based on WMV) out of your still pictures.

Originally posted by The Belgain
I don't particularly care about the lack of a lossless audio codec since I'm not planning on using wma (flac's fine for lossless, but I don't use lossless anyway).[/B]

I appreciate the point of view. But there is something great about not ever having to rip your content again because there is yet another new audio codec. With WMA Lossless (which we call the "never rip" codec) you can maintain your entire library in bit-exact form and covert it to other formats if needed. With hard disks now hitting 200 Gigabytes, it is a shame to leave all that disk space unused :). I know most people in this forum are using the space for video but if you are an audiophile, then a lossless codec is a good thing to consider. For the first time, you can put away the jewel box and be done with the shiny disks. Being an avid audio guy, this is what I do anyway.

As to other lossless codecs, they do exist of course. The advantage of WMA Lossless is that it is one click away in the Media Player 9 instead of having to rip into wav and then converting the files, etc.

Originally posted by The Belgain
I did a quick wma test and the audio quality was very poor. Still the same old artefacts and very bad stereo image - much worse than Vorbis (probably better than mp3 at this bitrate, though that's not really surprising). This was on a Metallica track, using VBR giving a bitrate of 81, comparing to Vorbis -q 1 which was 82 kbit. [/B]

Discussing audio codecs is a bit like arguing whether a blue car is prettier than a black one :). Everyone will have their opinions and preferences. I will say a few things though. Codecs behave differently on different content so I would not necessarily draw any conclusions on one or two clips or at specific settings. While I am pleased to see that you matched the encoding settings by using VBR, it is unlikely that the settings in our VBR match what Vorbis uses even if the file sizes are similar on one or two clips. The extra 1000 bytes that Vorbis had could make a huge difference in how transients sound btw.

Also note that for VBR content, the "bit rate" display in the media player can be inaccurate. So be sure to look at the file sizes and compute it that way rather than trusting what the player reports.

Finally, audio codecs testing can be a strange animal. For example, once you like one, you can easily convince yourself that any other codec doesn't sound as good (even in blind tests!). Let me tell you a funny story about this.

I used to help out with tuning of WMA. The routine was that I would be given multiple versions of difficult clips and be asked which one I liked better. I once had a situation where I was convinced we already had the optimal parameters and that our architect was just making things worse, not better. I was given yet another variation which I was quick to pronounce as being much worse than my "golden reference." It sounded tinnier and bright. To my surprise, the architect announced that the files were identical! Not believing him, I tested the clips again and thought once more that they sounded different. To prove it to him, I did a bit for bit comparison and to my amazement, found the files to be identical :(. The most shocking thing was that when I then listened to them after knowing they were identical, they sounded the same this time! The "big" difference had just vanished. It is amazing what tricks your mind plays on you when it comes to audio :).

Amir
Microsoft

[Rash]

Ignore me! :(

[amirm]

Originally posted by trbarry
No comments about whether the stand-alone players will likely handle any HDTV resolutions?
Sorry Tom. I lost track of all the questions :).

We showed a stand-alone reference design by Equator (who makes "VLIW" video processors) at International Broadcasters Conference (IBC) in September that was decoding WMV at 720p. It is up to our hardware partners to decide whether they want to productize such a solution.

Amir
Microsoft

[Sgt_Strider]

Hey Amir, I'm wondering how well do you think WM9 will compare to xvid when it comes to anamoprhic encodes?

[Neo Neko]

Originally posted by amirm

Cross-platform: We have a strong policy to make our codecs available on any platform our customers want. We do not require anyone to run our OS or any OS at all. Internally at Microsoft, we port our codecs to a wide range of hardware and OS platforms from MIPS and ARM CPUs (used in PocketPCs) to Mac OS.

We offer a full porting kit that includes specifications, verification suites, training, etc. for our partners to port our codecs to *any* platform they wish including those with competing OSes. We currently have more than 40 partners who have licensed WMV and are porting it to various chips, processors, and OS platforms.

So basically most OS will have a snow balls chance in hell of ever seeing an implementation. Linux, BSD, ONX, BeOS and the decendants of BeOS don't likely have enough dough between the lot of em' to fork over for the development kit and Microsoft knows it. Or might we see unsupported binaries being snuck out from Microsoft insiders? It's great and all to know that it could be ported. Just sickening to know that it likely never will.

[temporance]

Probably one of the main reasons that MS would never open up the WM specs is that they open themselves up to patent infringement suits. Note: I'm not saying or even insinuating that MS would deliberately infringe any patent, but there are just so many in audio/video compression that they may have stepped on someone's toes without even knowing it.

Another reason could be that they would lose control of the standard - you can't protect a standard itself with patents, only the ideas and inventions used by that standard. If WM was to become successful and remain closed, then hackers would open the standard themselves (as they have before).

So which is it?

[amirm]

Originally posted by Neo Neko
So basically most OS will have a snow balls chance in hell of ever seeing an implementation. Linux, BSD, ONX, BeOS and the decendants of BeOS don't likely have enough dough between the lot of em' to fork over for the development kit and Microsoft knows it.

I am happy to say this is not true. We have some of the smallest hardware vendors in the world shipping WMA players. Surely if the technology had been expensive these vendors would not have been able to afford the fees.

BTW, the fee for the porting kit itself is trivial (comparable to the salary of a good engineer for a month or two :)). For this fee, you get reference source code that is in ANSI C (making it easy to use simpler compilers for specialized processors). The code is 100% OS neutral and is ready for porting (i.e. it is available in integer-only form). We also provide a ton of verification tests to make sure the implementation is correct. And we provide on-site training by the engineers who designed the codec. In other words, this is quite a bargain when you consider what you get.

Likewise, the royalties for the codecs themselves is less than any MPEG codec such as MPEG-4, and MP3 and we do NOT charge any "use" or content fees (as does MPEG-4 for example). The royalties also cap at a very low amount so you could pay one fee and ship all the devices you want. And of course, on Windows there is no cost to use our technology.

Put another way, if cost is an issue, we rank way on top, not the other way around. Yes, the technology is not free but it is the next best thing :).

Amir
Microsoft

[Ramirez]

Does MS plans to implement some kind of forced DRM Policy with WM9?
Or it's really a matter of distance future to worry about?

Thanks.

[amirm]

We have always taken a hard stance on DRM being optional. It is totally up to the content owners to decide to protect their content or not (we provide them the tools, they get to decide when and how to use them). We have no plans currently to change this policy -- with WM9 or otherwise.

If you are encoding your own personal content, the output will be unprotected unless you explicity choose to DRM it. In other words, you can pretend DRM does not even exist.

Happy new year everyone!

Amir

[Neo Neko]

Originally posted by amirm
BTW, the fee for the porting kit itself is trivial (comparable to the salary of a good engineer for a month or two :)).

As I thought. Snow balls chance in hell.

Originally posted by amirm
For this fee, you get reference source code that is in ANSI C (making it easy to use simpler compilers for specialized processors). The code is 100% OS neutral and is ready for porting (i.e. it is available in integer-only form). We also provide a ton of verification tests to make sure the implementation is correct. And we provide on-site training by the engineers who designed the codec. In other words, this is quite a bargain when you consider what you get.

Likewise, the royalties for the codecs themselves is less than any MPEG codec such as MPEG-4, and MP3 and we do NOT charge any "use" or content fees (as does MPEG-4 for example). The royalties also cap at a very low amount so you could pay one fee and ship all the devices you want. And of course, on Windows there is no cost to use our technology.

What if I told you we did not necessarily care that much about encoding on this issue. More about decoding. Is it to much to expect Microsoft to put out even a reduced functionality decoder only module like everyone else in the world? While my windows computer is outnumbered 3 to 1 I would use that one windows computer to encode in this forsaken format and codec if I had the ability to simply watch it elsewhere. To date that is why no microsoft format or codec has ever gotten any regular use by me and many others. Sure I test em for sh1t5 and giggles every once in a while. But they are of no real use to me. Even Real Networks for all their problems allow playback on other OS.

Originally posted by amirm
Put another way, if cost is an issue, we rank way on top, not the other way around. Yes, the technology is not free but it is the next best thing :).

Amir
Microsoft

Xiph already has you beat in one area. And they are hard at work in the other areas.

What I would really like to know is why in the last 3 versions of Windows Media Player has Microsoft reduced it's usability and configurability as a media player. And in it's place put in stupid futile features such as encoding? What ever happened to the Windows Media Encoder? Who came up with the bright idea to make a player that encodes? Granted it is a boon to mankind just like the ability to send email from MS-Solitare would be. But is it to much for Microsoft to make a player with more player related features and better player features than previous versions? Microsoft so often misses the big picture. Well maybe not so much that. More that they have a very different big picture. One no one but Microsoft wants.

Nothing against you man. Just the Microsoft fud.

[Neo Neko]

Originally posted by amirm
We have always taken a hard stance on DRM being optional.

Which is why it was enabled by default in many previous versions of windows media player that offered such functionality? Bit devicive.

[Sirber]

Will WM9 be avalible as a VFW codec?

[ChristianHJW]

Originally posted by Sirber
Will WM9 be avalible as a VFW codec? ... if hell freezes over maybe ;) ...

[amirm]

Originally posted by Neo Neko
Which is why it was enabled by default in many previous versions of windows media player that offered such functionality? Bit devicive.

"On by default" is not the same as mandatory which was the question posed. However, you are correct in that where possible, we try to balance the needs of the consumers and commercial content providers. But we do draw the line somewhere and that line is to give the user the option to turn off DRM. Other technologies go to one extreme or the other. We feel that by taking the middle road, we get the benefits of both approaches. Our customers get access to content they wouldn't have any other way (e.g. Peter Gabriel's "Up" album in 5.1 WMA Pro before becoming available in any other surround format) while not limiting their choice.

And yes, the approach is not without criticism and the amount of time it takes to explain our position on new year's day :).

Amir
Microsoft
Microsoft

[amirm]

Originally posted by ChristianHJW
... if hell freezes over maybe ;) ...

Be careful what you wish for :). Also funny to see your signature next to this comment:

"...; Make sure AVI and VfW have seen the end of their days ..."

I feel like I am being bashed by a "wet noodle" :).

Amir
Microsoft

[temporance]

Originally posted by amirm
And of course, on Windows there is no cost to use our technology.

Really? I had no idea that Windows was a freeware OS. Where can I download my free copy of Windows XP?

[amirm]

Originally posted by Neo Neko

...And in it's place put in stupid futile features such as encoding? What ever happened to the Windows Media Encoder? Who came up with the bright idea to make a player that encodes? Granted it is a boon to mankind just like the ability to send email from MS-Solitare would be. But is it to much for Microsoft to make a player with more player related features and better player features than previous versions? Microsoft so often misses the big picture. Well maybe not so much that. More that they have a very different big picture. One no one but Microsoft wants.

Nothing against you man. Just the Microsoft fud.

I appreciate the comments Neo but I don't understand the reference to "encoding". Yes, the media player can encode WMA files it rips from CDs. Are you suggesting that we remove CD ripping? I assume not since millions of people like this feature. Other than ripping, Media Player has no other "encoding" function. Perhaps you are confusing us with QT player which encodes video. BTW, Windows Media Encoder 9 is alive and well. You can download it today.

As to usability, it would be hard for me to say that a single media player will please everyone in the world. But I will say that with version 9, we are doing better than we have ever done in getting people to like the player. We are faster in start up than just about any player now (including Winamp), have unique features such as lossless CD ripping, and lots of other usability improvements (such as one click user rating in the media library).

As to your other comments regarding the decoder, I appreciate some of the points you are making. But I think the situation is not as bad as you describe. From what I recall, the source code cost is just $10K. Yes, it is too much for a hobbyist but not for any commercial entity including those shipping Linux, etc.

As to Xiph source being free, this is true but they are not offering optimized reference source code as we do. There is tremendous cost and expense involved in taking a decoder and making it suitable to run on any integer processor with very little memory (the reference code uses less than 32k RAM btw -- we are talking *kilobytes* here). And I am sure you know that we are not getting rich from $10K licenses :). The other thing to note is that until a big company with deep pockets ships their codec, I would say the jury is out as to whether the technology is really "free" or not. Note that I am NOT saying there are any issues here but that there is a test that we have passed that they have not yet. And of course, they don't have a video codec.

As to decoders for other boxes, I can tell you that as Microsoft, we won't be putting out a Linux decoder. It would be like Sony making an Xbox game :). But we provide the tools for anyone else to do the work. This is the best we can do. Other OS platforms are fine and as I mentioned, we already have players for Mac OS X.

Amir
Microsoft

[ChristianHJW]

Originally posted by amirm Be careful what you wish for :). Also funny to see your signature next to this comment:
"...; Make sure AVI and VfW have seen the end of their days ..." I feel like I am being bashed by a "wet noodle" :).
Amir Microsoft

:) ... sorry, i am not a native english speaker, so i have problems understanding what this means.

In any case i know 2 things :

1. VfW is old and outdated and needs to be replaced, and if ever possible by a real open standard like UCI or gstreamer codec API. We should never ever base the future of our open source constraints on an API created by a comany with financial interests, whoever this is ... they might decide one day they dont need it anymore and abandon it ;) .

2. WMV9 will never be released in either a VfW or a UCI version. Microsoft will prevent this with all their might, even if somebody would leak the internal API they would forbid use of the codec outside of Microsoft tools, even with legal measures. Same is valid for Real IMHO, even if i admit i dont understand the full scope of Helix yet ... maybe Karl could tell if its possible to use Real codecs with other encoding tools ( i seem to remember he is still waiting for a response from his company if a VfW codec would be within the Helix license or not ;) .. ) ....

[Sgt_Strider]

Originally posted by amirm
I appreciate the comments Neo but I don't understand the reference to "encoding". Yes, the media player can encode WMA files it rips from CDs. Are you suggesting that we remove CD ripping? I assume not since millions of people like this feature. Other than ripping, Media Player has no other "encoding" function. Perhaps you are confusing us with QT player which encodes video. BTW, Windows Media Encoder 9 is alive and well. You can download it today.

As to usability, it would be hard for me to say that a single media player will please everyone in the world. But I will say that with version 9, we are doing better than we have ever done in getting people to like the player. We are faster in start up than just about any player now (including Winamp), have unique features such as lossless CD ripping, and lots of other usability improvements (such as one click user rating in the media library).

As to your other comments regarding the decoder, I appreciate some of the points you are making. But I think the situation is not as bad as you describe. From what I recall, the source code cost is just $10K. Yes, it is too much for a hobbyist but not for any commercial entity including those shipping Linux, etc.

As to Xiph source being free, this is true but they are not offering optimized reference source code as we do. There is tremendous cost and expense involved in taking a decoder and making it suitable to run on any integer processor with very little memory (the reference code uses less than 32k RAM btw -- we are talking *kilobytes* here). And I am sure you know that we are not getting rich from $10K licenses :). The other thing to note is that until a big company with deep pockets ships their codec, I would say the jury is out as to whether the technology is really "free" or not. Note that I am NOT saying there are any issues here but that there is a test that we have passed that they have not yet. And of course, they don't have a video codec.

As to decoders for other boxes, I can tell you that as Microsoft, we won't be putting out a Linux decoder. It would be like Sony making an Xbox game :). But we provide the tools for anyone else to do the work. This is the best we can do. Other OS platforms are fine and as I mentioned, we already have players for Mac OS X.

Amir
Microsoft

When you mean you can download the codec today, you mean the release candiate and not the final right???

[The Belgain]

Ok. First of all I'd like to thank amir for all his posting despite being bashed quite a lot. It's just there is quite a lot of resentment towards Microsft in general, he's been very helpful despite this and I hope he doesn't leave the forums because of it.

Secondly, I have the latest version of Windows Media Encoder, and it crashes when I try to open any input file (avi, mpeg, ...). I'm running Win2k SP3. I dunno what the problem is. Ok, it turns out WMP 9 also crashes on opening many of these files (they all work fine in all other players). Any idea what this bug is?

Also, can the Encoder open avs scripts?

[Rash]

Originally posted by Sgt_Strider
When you mean you can download the codec today, you mean the release candiate and not the final right???

Come on! They'll be really "stupid" if they make the final version paid.

You know, let me tell you something I don't get. The Intel C++ compiler. Intel says it can compile any code and make it optimized for Pentium 4 SSE2 instructions. That it would look you code for what it can optmize and do so. But this compiler is paid! (Very well paid, if you know whay I mean) So what's the point? Wouldn't it be better for Intel if any programmer developed P4 optimized programs?

OK, I know this is not of anyone's business around here. I know you Microsoft and Helix guys wouldn't like to comment this because I'm talking about another company. But I just don't understand Intel strategy. Just don't forget you guys have great free competitors. ;)

[Neo Neko]

Originally posted by amirm
I appreciate the comments Neo but I don't understand the reference to "encoding". Yes, the media player can encode WMA files it rips from CDs.

A player that encodes is missleading. If that is the goal call it a media center.

Originally posted by amirm
Are you suggesting that we remove CD ripping? I assume not since millions of people like this feature. Other than ripping, Media Player has no other "encoding" function.

I am not so adverse to the rippig feature. And if people use it and like it good for them. I can think of much better software for them to use. But getting people to use the right and best software for the job is almost an impossibility. My main caveat with all WMP versions after 6.4 is that there is no menu item like the 6.4 propperties menu which allows you to adjust filter properties in the directshow chain. I have no idea how many times I use that feature alone under 6.4. A lot does not seem to adequatly describe it. The ripping features are just an example of the little things I did not need that we gaind at the major loss of alot of functionality.

Originally posted by amirm
Perhaps you are confusing us with QT player which encodes video. BTW, Windows Media Encoder 9 is alive and well. You can download it today.

Aaaaah QT player. If it were not out of nescessity it is another player I would never use. But now that MPEG4 is hitting hard the QT player will likely be depredicated as there will be many players capable of the new format out of the box. Perhaps except for WMP. As for WMP9..... it will never touch my PC without great protest unless of course it re-introduces the capability to moddify the filter chain and other newer usefull features.

Originally posted by amirm
As to usability, it would be hard for me to say that a single media player will please everyone in the world.

True.

Originally posted by amirm
But I will say that with version 9, we are doing better than we have ever done in getting people to like the player.

There are people and then there are people. I know I am not alone in wondering what the real use in upgrading to WMP9 is. I could get and use the codecs without it. But I preffer not to use them on a regular basis either. Out of curiosity though. How is it judged "how much people like the player"? It would be interesting to see how the numbers were calculated. I see lots of room for skewed data here. If WMP8 was not forced on my system with the OS I would stick to 6.4 alone. I guess at least I can be thankfull that Microsoft has still afforded us this small reffuge. But I wonder how long it will last.

Originally posted by amirm
We are faster in start up than just about any player now (including Winamp),

Start up speed means little to me. Even if a program took several times longer to start up but offered better features/functionality I would use it in a heart beat. But how is load speed determined when a large portion of the program and it's resources are always constantly loaded? And since that is often the case it would be quite absurd for it to be otherwise. Just an unfair advantage. :p

Originally posted by amirm
have unique features such as lossless CD ripping,

I have tons of software that can do that though. And with just about any codec I choose. Whether it be monkies, shorten, optimfrog, LPAC, etc, etc, etc. Plus Winamp can play all of them. And now FooBar2k can play many of them as well. WMP can't play any of em IIRC.

Originally posted by amirm
and lots of other usability improvements (such as one click user rating in the media library).

Libraries are all good an well. I like them to organise my albums which I have encoded to Vorbis on my HD. But I fail to find the use in a rating for the library. I know what I like, and as long as the library lets me find it then I am happy. I don't want to take the time to go file by file etc rating the songs.

Originally posted by amirm
As to your other comments regarding the decoder, I appreciate some of the points you are making. But I think the situation is not as bad as you describe. From what I recall, the source code cost is just $10K. Yes, it is too much for a hobbyist but not for any commercial entity including those shipping Linux, etc.

You seem to miss a key issue here. Most of the commercial entities dealing with Linux are mostly server based. Where is the use in such thing for a server? As has been said before, and has held true. Microsoft very much has a Monopoly on the desktop. The use of such things is on the desktop/workstation. With no hopes of really being able to penetrate a market in which the main player is so deeply entrenched and quick to dirty pool to hold that position where is the incentive to do this? Something like Linux is more than a corporation. Lets say for fun redhat bit on Microsoft's line. They are the Icon of corporate Linux. When you say Linux most non Linux users shout redhat! Redhat has built their market on providing server grade OS systems that run rings around any Microsoft system for speed/stability/configurability/upgradability etc. How could they use an investment in such a kit to their advantage? In the bigger picture keep in mind that Linux is much more than redhat. Each of my 3 PC that run Linux run a different distribution. Only one of which is remotely commercial. Microsoft knows exactly what they are doing. :devil: They extend a line that they know any possible competitors could not really avail themselves of even if they wanted to. When someone cries foul Microsoft can sit back and flash a big toothy grin saying "Hey we are not excluding anyone. No monopoly like business here. They just could not fork over the dough.". Lets say all us users of all non MS OS got together and took up a collection on paypal or some such to get the needed funds and then had guys make libraries for decoding on *NIX varriants, BeOS varriants, QNX, etc. How could we then legally distribute such decoders and what would Microsoft's reaction be if we just gave them away for free? :confused: Is it to much to ask for rudamentary playback libraries? Microsoft knows what will happen if they don't provide such. The actual Microsoft products will be reverse engineered(My won't Microsoft love that) and the growing resentment to Microsoft would continue to bloom. Such a simple concession as a rudamentary decoder library would go a looooooooooooooooooooong way. Microsoft was providing WMP binaries for Solaris. Why not others. ;)

Originally posted by amirm
As to Xiph source being free, this is true but they are not offering optimized reference source code as we do. There is tremendous cost and expense involved in taking a decoder and making it suitable to run on any integer processor with very little memory (the reference code uses less than 32k RAM btw -- we are talking *kilobytes* here).

Ah you missed it. Xiph has plenty of optimised code and integerised decoders. Perhaps not as optimised as possible. But it is progressing at a rapid pace. But the code compiles easily on pretty much any ARM, MIPS, Alpha, Amiga, PPC, general Intel, etc processor on the market. Hey I was playing Vorbis on my Dreamcast just the other day.

Originally posted by amirm
And I am sure you know that we are not getting rich from $10K licenses :).

Nor are you coming up short for the lease on those luxury sports cars either. ;)

Originally posted by amirm
The other thing to note is that until a big company with deep pockets ships their codec, I would say the jury is out as to whether the technology is really "free" or not.

*cough*AOL*cough* :D It is verry free indeed. AOL spent quite alot of time investigating this along with Xiph themselves. And to keep you abreast Emmett and his crew are currently in mid Alpha testing with their first video codec that will be out of beta before the year is out. Shortly after July if the schedual he disclosed to me is able to be adhered to. It is being built on an existing/prooven codec that produces results at least as good as the current WMV codec in stable distro. So unless they severly screw something up Theora will be a contender.

Originally posted by amirm
Note that I am NOT saying there are any issues here but that there is a test that we have passed that they have not yet.

Which one?

Originally posted by amirm
And of course, they don't have a video codec.

ON2 gave it to them.

Originally posted by amirm
As to decoders for other boxes, I can tell you that as Microsoft, we won't be putting out a Linux decoder.

Then my decision has been made for me. I can't in good concience no matter how nice it may be use a codec of that nature. I would not care about encoding. Just an unsupported decoder library. Something that would not put any real strain on Microsoft and would go along way to increasing the codecs useage as well as the companies immage. Even if I had to use a statically linked binary player ala Quicktime it would definatly make such codecs a viable option to me. It's not like I am asking for a complete Directshow implementation for Linux/Solaris/BeOS/Mac OS etc.

Originally posted by amirm
It would be like Sony making an Xbox game :).

Or Sega making a Nintendo game. Everyone knows plumbers and hedgehogs are mortal enemies and could never co-exist. Or could they? :D I have a nice entertainment center setup. My only gripe is that ATM if I want to play Sonic Adventure 1 I have to power up my Dreamcast, but to play Sonic Adventure 2 I have to power up my Gamecube. Both of which sit on top of my PS2. Just no room or desire left over for an Xbox though. :D I suppose I could find the dreamcast ISO on line if I was desparate enough. But that sort of things would tend to go against my ethics and morals. ;)

Originally posted by amirm
But we provide the tools for anyone else to do the work. This is the best we can do. Other OS platforms are fine and as I mentioned, we already have players for Mac OS X.

Then why not BeOS, Linux, QNX, or BSD for x86 chipsets at least. I already know your answer and the real answer. But I mean honestly. Mac OSX is BSD based *NIX. The ammount of porting required would be almost nill. The only real work would be the GUI interface. A play/pause/stop button and a seek bar in a rectangle is all I need. It's not like I'm asking for Microsoft Linux here. <Dr. Evil voice>"Somebody throw me a frickin bone here. Ok?"</Dr. Evil voice> One simple x86 BSD compatable decoder library would allow playback on BSD, Linux, BeOS, and QNX no porting for each nescessary.:sly:

[Neo Neko]

Originally posted by The Belgain
Ok. First of all I'd like to thank amir for all his posting despite being bashed quite a lot.

The only one here who could remotely be considdered as being bashed is Microsoft. Amir's presence here is very welcome. :)

Originally posted by The Belgain
It's just there is quite a lot of resentment towards Microsft in general,

Indeed.

Originally posted by The Belgain
he's been very helpful despite this and I hope he doesn't leave the forums because of it.

Yes indeed.

Please do not construe my quizzitive nature to many of Microsoft's dubious practices as an insult to Amir. Microsoft employs many bright individuals who are often stiffled by the company themselves. WMV9 is good and all. But some people like myself have to take portability into account as to the codec of their choice. I would go beyond testing WMV9 and actually use it if only I could decode it. ;)

[amirm]

Neo,

Our goal for Media Player and compression technologies is to provide technology for the masses. When someone sticks a CD in their XP machine, they expect it to play and nowadays, rip into a music library. They like rating systems because that allows them to find things quickly (I always rate my songs after ripping so that I can create a dynamic query for “play anything with 4 stars” and have it just play my favorites). While you and I might like the Dshow filter properties feature of 6.4 player, 99.999% of our customers wouldn’t know what a Dshow filter is let alone want to set its properties :). So for good or bad, the masses do not have the same requirements or preferences that you state.

If people don’t like our player, they can always take our free SDK and build their own as have Real, AOL Winamp, MusicMatch, etc. That’s the beauty of the PC/Windows. It is a programmable device.

As for Xiph using the old stuff from ON2 to make a competitor to WMV 9, let me tell you a story. A guy goes to the butcher and asks how much the steak is. The butcher says $10/pound. The customer says the store across the street has them for $6/pound. The butcher asks why he wouldn’t go and buy it from the guy across the street. The customer says that the other guy doesn’t have any! The thing is, we have a state-of-the-art video codec that is shipping now, and is carefully designed to be implementable on many architectures. What someone will have someday is not relevant to what is going on now. They have a claim, we have the real thing. Would you stop buying an AMD machine just because Intel says they will have a 5 Ghz CPU one day? The only thing that matters is what they both have today.

If I can answer any other questions about our technology, just go ahead and ask. I am happy to answer them. I worry that arguing about philosophical topics is not adding much value to the forum.

Thanks,
Amir Majidimehr
Microsoft

P.S. Just in case you think we are just a bunch of anti-Linux/Unix guys, do a search on my name in Amazon.com...

[charact3r]

Amir (and others),

I guess there's no point arguing about how open WMV9 is. In time, if it is proven useful, source code will be released by some insomniac teenage hacker, or by a disenfranchised employee of one of MS's 40 partners. That's assuming MS doesn't do the right thing and GPL it first. ;-)

The real question that remains is the performance of WMV9. No-one on this forum has yet said "WOW, WMV9 ROCKS!!". People are saying that RM is good for anime (due to its H.264-style in-loop filter), but no-one's advocating WMV9, yet.

Can any independent person provide annecdotal or better evidence that Microsoft's technology is any good? We need it to be benchmarked against something like DivX 5 Pro which is a good, well-known implementation of MPEG-4 SP / ASP. (Or a known version of xvid for that matter).

From my point of view I can only say that WMV9 seems to causes a different type of degradation to other DCT-based codecs. Its quantisation noise appears to have a wider frequency distribution, so instead of conventional blocking or ringing you see blocks with a raised white-noise level. Or perhaps this is how their postprocessing works. Either way, this could be argued to be more pleasing to the eye.

[TheXung]

From my point of view I can only say that WMV9 seems to causes a different type of degradation to other DCT-based codecs. Its quantisation noise appears to have a wider frequency distribution, so instead of conventional blocking or ringing you see blocks with a raised white-noise level. Or perhaps this is how their postprocessing works. Either way, this could be argued to be more pleasing to the eye.

Sounds like you've already done your own testing and conclusion drawing. Why the need for other's opinions? Others can't draw any more accurate conclusions.


However, if everyone remembers, there was a significant quality difference between Vorbis 1.0 and it's release candidates, so I would reserve my judgement for the final release.

[Neo Neko]

Originally posted by amirm
While you and I might like the Dshow filter properties feature of 6.4 player, 99.999% of our customers wouldn’t know what a Dshow filter is let alone want to set its properties :). So for good or bad, the masses do not have the same requirements or preferences that you state.

When it comes to MPEG4 codecs accessing the filter properties can often be usefull. My whole point is why did they have to remove it in the first place? Did it cause WMA ripping to crash?

Originally posted by amirm
If people don’t like our player, they can always take our free SDK and build their own as have Real, AOL Winamp, MusicMatch, etc. That’s the beauty of the PC/Windows. It is a programmable device.

I will not go in on that scenario. There are flaws in it though. When it comes to almost everything on the x86 PC architechture MS always has a conflict of interests. But nothing that need be discussed here.

Originally posted by amirm
As for Xiph using the old stuff from ON2 to make a competitor to WMV 9, let me tell you a story. A guy goes to the butcher and asks how much the steak is. The butcher says $10/pound. The customer says the store across the street has them for $6/pound. The butcher asks why he wouldn’t go and buy it from the guy across the street. The customer says that the other guy doesn’t have any!

Hey ON2's VP3 could easily hold it's own with WM8 in everything but encoding speed. It is still avalible today. And that is saying something.

Originally posted by amirm
The thing is, we have a state-of-the-art video codec that is shipping now, and is carefully designed to be implementable on many architectures.

Yes WMV8. And eventually WMV9. But from what I have seen the only place they are clearly better than the aging VP3 is in the arena of encoding speed. WMV9 is ok but to date it is not some groud breaking earth shattering improvement. In all other areas VP3 has a nice advantage. I can decode it on just about any OS I want for example. ;)

Originally posted by amirm
What someone will have someday is not relevant to what is going on now. They have a claim, we have the real thing. Would you stop buying an AMD machine just because Intel says they will have a 5 Ghz CPU one day? The only thing that matters is what they both have today.

Yes I would. It depends on time frame. But aprox. 6 months is not an outrageous length of time to wait. Now I will not be holding my breath waiting for Tarkin because I would die of asphixiation by the time it is released. But the revamped VP3 just by virtue of it's previous performance is almost 100% guaronteed a good thing.

Originally posted by amirm
If I can answer any other questions about our technology, just go ahead and ask. I am happy to answer them. I worry that arguing about philosophical topics is not adding much value to the forum.

Philosiphy is not my goal here. I would just really like some fresh insight on why Microsoft the corporation(not the people) feel that support for any OS/system they do not have a major stake in will be the death of them. Something so tiny as decoding.

Originally posted by amirm
Thanks,
Amir Majidimehr
Microsoft

P.S. Just in case you think we are just a bunch of anti-Linux/Unix guys, do a search on my name in Amazon.com...

I know of many Microsoft related/employed people who use/dabble in BSD/Linux/Qnx/BeOS/Unix/etc. I never said it was because of you. Rather the policies of the company. You unfortunatly were just a handy guy to toss some questions at. ;) If it were up to the majority of Microsoft's employees/coders we would have WMV9 decoders for every OS under the sun and such things as MS Office for BSD/Unix/Linux/BeOS/etc. The reason there will be a player for WM9 on MacOSX is due to the fact that Microsoft has a large stake in apple and sales of products like MS Office for Macintosh pull in a nice bit of revenu for Microsoft. And the reason that there is no Windows for Mac is that at least on the OS front Microsoft has relied on Apple to provide innovation for them for years. Apple along with BSD.

http://www.xiph.org/archives/vorbis/200106/0396.html
http://david.weekly.org/audio/amir.php3
http://www.wave-report.com/1996%20Wave%20issues/wave612.html
One Thousand Questions and Answers About Unix Systems
Amir Majidimehr / PB, 1998 (NYP On Order)

:sly: I know who you are. :sly: I have seen you around before. And that is why I say you are welcome here. As a member of Microsoft I do not hold you responsible. Don't take my stabs at Microsoft as anything personal. But can you blame me for making the case to someone in your position that I would like the ability to simply play back WMV9 encoded content under OS other than Windows or Mac OSX. I would not even care if it would not play DRM content. Since I would never use such a feature myself. Just so that I could play the material. It's not that much to ask. Just give them libWM-9.x.x.so and let them write the GUI for it. All I ask is to be able to play it without some major wikkid voodoo hacking or convoluted virtual machine or Windows API emulation setup.

[temporance]

Originally posted by Neo Neko
Just give them libWM-9.x.x.so and let them write the GUI for it. All I ask is to be able to play it without some major wikkid voodoo hacking or convoluted virtual machine or Windows API emulation setup.
Now where's the fun in that? Some people love the challenge and thrill of illegitimacy!

Originally posted by TheXung
Sounds like you've already done your own testing and conclusion drawing. Why the need for other's opinions? Others can't draw any more accurate conclusions.

However, if everyone remembers, there was a significant quality difference between Vorbis 1.0 and it's release candidates, so I would reserve my judgement for the final release.
Good points, but like a lot of others, I'm too impatient to wait for the jury's verdict on wmv9. My own conclusions are based on recent demos by MS and its partners who were using the latest internal code.

I know for a fact that lurkers and contributors to this forum have seen the wmv9 standard and are clever enough to see the potential for this codec. These are the people who can help us all with their opinions.

For example, dattrax (Jim) opined,
Now windows media has bframes and qpel, as well as some features MPEG4 hasn't got, it should be possible to get as good as, if not better than MPEG4.
This is the most interesting comment so far in this discussion.

[Neo Neko]

Originally posted by temporance
Now where's the fun in that? Some people love the challenge and thrill of illegitimacy!

Groan. But when you are right you are right. It is what will happen if some sort of decoder is not provided. An illegitamate hack will be found. Like using the MS DLLs under linux to decode the content. It has been done before. And doubtless will be done again. But what I would not give for a legitimate decoder. Of which I doubt Amir would have any major qualms with. But there is corporate policy that dictates otherwise. :(


Originally posted by temporance
Good points, but like a lot of others, I'm too impatient to wait for the jury's verdict on wmv9. My own conclusions are based on recent demos by MS and its partners who were using the latest internal code.

I know for a fact that lurkers and contributors to this forum have seen the wmv9 standard and are clever enough to see the potential for this codec. These are the people who can help us all with their opinions.

For example, dattrax (Jim) opined,

This is the most interesting comment so far in this discussion.

With any lossy codec there are at least two things that must be considdered. Technology and tuning. Doubtless Microsoft has put up lots of money to buy or get the latest technology inside. That is the first thing they always do. But my main concern lies in tuning. WMA for an example seems slow to progress in this arena. It might have all that technology under the hood. But a company like Xiphophorus using public domain and discarded algorythms have succedded to produce a codec that holds it's own easily even with the best of the big boys. And it is in large part due to an insane ammount of tuning. I also subbmit for your parousle MPC or MP+. MPC produces quality superior to pretty much every lossless codec at the mid bitrate range. But the technology it is using is pre-MP3! MPEG1 Layer 2 mostly to be more precise. Yet for this range it outperforms MPEG1 layer 1-3, MPEG2 Layer 1-3, and MPEG2 AAC which use the exact same technology or better!!

Xvid and Divx are yet another example of this. Divx has the technology front as far as MPEG4 goes covered in spades. Xvid on the other hand produces output as good or quite often better with less technology fluff and more hard core attention to detail. The fine tunings. Things like Modulated quantizers or individual quantizer controls. Better motion estimation, curve controls, etc.

WMV has always been optimised for the low low end much like WMA. Often achieving acceptable results in that range. But the over all codec performance has always been lacking. When WMV9 final is out we shall have to see if all these high-rez tunigs come through for Microsoft. It has always been my oppinion that on many things Microsoft's over all attention to detail is often lacking. Again no fault of Amir. It extends far beyond the media division and is most often found in the general software division.

Here is a question Amir might be able to answer though. I have heard from people I trust that at least on internal code there are switches that can be used that cause WMV9 to output ISO-MPEG4 streams. Is this true. If so is it something we might see though possibly not publicised on the final release version? Also will this final version be format locked as all previous versions.(Another peeve I have always had against Microsoft)