Windows Media Encoder Studio Edition Beta 1 released [Archive] - Page 2

View Full Version : Windows Media Encoder Studio Edition Beta 1 released

Pages : 1 [2] 3

mpgxsvcd

13th July 2006, 03:15

mpgxsvcd:
Can you give us a short note on how who you managed to use those files as input for WMESE? I can play them, but WMESE complains about an unsupported codec.

And yes, the PQ is remarkable for a $400 cam. Too bad it isn't widescreen.

Good question. I got them to work a few times on my laptop. However, they usually give me the unkown error message. I could not get them to work at all on my desktop. I will do some more testing and let you know what I did.

Yes too bad it isn't widescreen. I bet the next round of Canon cameras will have widescreen! Nothing will come close to them then.

http://forum.doom9.org/showthread.php?t=112916

fccHandler reports that these files are 4:2:2. Is there anything different in the Studio version that would enable 4:2:2 progressive encoding that was not available in the orginal WME9?

SenorKaffee

13th July 2006, 09:03

benwaggoner

13th July 2006, 18:50

Can it be that WMESEß1 doesn´t like the Lagarith codec? It always fails to start encoding with "A valid type has not been set for this stream or a stream that it depends on."

1 Video Track, 800x600, 1:1 PAR, 25 fps, Progressive.
Source is a 25 second AVI encoded with the Lagarith Codec.
Encode Tab is completely left on default settings.
Ah, yeah, we know there are issuse in our Lagarith support. Huffyuv works fine, though.

I'm quite the fan of Lagarith myself, so I feel your pain on this one :).

SenorKaffee

18th July 2006, 08:27

I just started to really distrust HuffYUV because I have a huge pile of HuffYUV 2.2 encoded AVIs to convert to some different lossless format because the codec keeps crashing on my main video editing machine and the non-crashing HuffYUV 2.11 can´t properly decode these files.

But this is getting off-topic. ^^

foxyshadis

18th July 2006, 10:06

Try ffdshow as the huffyuv decoder (and encoder as well), might work better.

SenorKaffee

19th July 2006, 13:12

Yes, I can actually recover the video this way, but there is some weird post-processing going on somewhere (i use GrafEdit to do this). Maybe I should open another topic for that. ;)

daniel.aam

25th July 2006, 10:11

benwaggoner

25th July 2006, 17:29

Why doesn't WMSEβ support all profiles and different levels?
Or are the levels automatically set according to the resolution? It's extremely annoying as I need to do a 3rd pass and allocate more bits to a piece of PAL originated footage which is MP@ML. Any idea's anyone?
Sorry, no. WMESE Beta 1 is AP only. We intend to add other profiles in later versions. I agree that an explicit level control would be useful as well.

Vitaly Ivanov

28th July 2006, 10:10

I wonder what is the difference between VC-1 toolkit and WMESE. Do they use the same enc?

pwimmer

29th July 2006, 09:23

Here's my feedback to the WMESE, ordered by priority:

1) The encoder is too slow
2) The encoder is too slow
3) The encoder is too slow
4) Quality cannot compare with H.264
5) The GUI looks like a trainee's work. Everything is too large and doesn't adhere to Microsoft's own GUI standards. It's realy ugly.
6) I've found no way to do a side-by-side comparison of uncompressed and compressed movie.
7) There is no PSNR graph to easily identify scenes which might need a higher bitrate.

benwaggoner

10th August 2006, 17:49

Here's my feedback to the WMESE, ordered by priority:

1) The encoder is too slow
2) The encoder is too slow
3) The encoder is too slow
First thing that can help encoding speed is to avoid having to do scaling or dithering in the tool, which are very high quality but very slow. And you can always throttle back on encoder complexity, which has a huge effect on encoding speed.

]4) Quality cannot compare with H.264
What's your scenario?

5) The GUI looks like a trainee's work. Everything is too large and doesn't adhere to Microsoft's own GUI standards. It's realy ugly.
It'll be dramatically revised in 1.0

6) I've found no way to do a side-by-side comparison of uncompressed and compressed movie.
That didn't make it into the beta, but yes, we know that's a very useful feature.
7) There is no PSNR graph to easily identify scenes which might need a higher bitrate.
The QP graph is meant to serve that purpose. High QP spikes indicate regions that should be checked for reencoding.

We've got a lot of perceptual encoding features in the new codec, so PSNR isn't as useful as it might be with simpler codecs.

CruNcher

10th August 2006, 20:03

benwaggoner

10th August 2006, 20:51

benwaggoner i just have to say that Sonys Pro tools are much better Desgined and workedout than yours also the Sequential Encoding is much better layed out in Sonys Encoder Gui, WMV Encoder looks like a Semi Pro tool in that regard, you still have alot of work todo :P
Bear in mind that Studio Edition isn't the same as the encoding workflow that studios are using for HD DVD and VOD authoring. What you've got really is just a Beta 1. We appreciate all your feedback and suggestions, which is why we released the Beta 1, and didn't call it 1.0. I think you'll be very happy with what we finally release.

pwimmer

10th August 2006, 21:48

@Ben:

Speed: For some unknown reasons, the encoding complexity option doesn't have any effect. The encoding frame rate is about 1 fps, no matter what complexity I choose.

Quality: I've encoded several movies in VC-1 and H.264 and H.264 was always better, for example the Rhine Valley movie at http://www.3dtv.at/Movies. Especially, H.264 looks sharper in areas with lots of fine details. I've used the the Nero H.264 and X264 codecs and WIndows Media Encoder (not SE, but the old WME, qhick tests showed me the quality is basically the same).

Btw, what about including features for stereoscopic content in the ASF specification? For example these ones: http://www.3dtv.at/Knowhow/StereoWmvSpec_en.aspx

benwaggoner

10th August 2006, 22:16

Speed: For some unknown reasons, the encoding complexity option doesn't have any effect. The encoding frame rate is about 1 fps, no matter what complexity I choose.
Really? What's your source?

Quality: I've encoded several movies in VC-1 and H.264 and H.264 was always better, for example the Rhine Valley movie at http://www.3dtv.at/Movies. Especially, H.264 looks sharper in areas with lots of fine details. I've used the the Nero H.264 and X264 codecs and WIndows Media Encoder (not SE, but the old WME, qhick tests showed me the quality is basically the same).
What settings did you use?

Btw, what about including features for stereoscopic content in the ASF specification? For example these ones: http://www.3dtv.at/Knowhow/StereoWmvSpec_en.aspx
I think we've got a lot of other core features to get to first :).

Seriously, though, if we supported it, how do you imagine using it?

sjchmura

11th August 2006, 19:38

Ben,

For those of us just encoding HD streams from camers or OTA, I think alot of the "simplicity" of x264 and Nero Recode could be contributing to VC1 not looking as good. For example, it is quite easy in x264+MeGUI to choose "720p, fix the aspect ratio on scalling (for mod16), HQ-Slow". We get Mp4 fils that are the size we want and great quality (with dual core support etc).

I think when future betas support more inputs and are easier to configure we can get fewer "this looks better" etc. It was my thought that MS's aim was to make VC-1 as good as AVC ..

benwaggoner

11th August 2006, 19:50

For those of us just encoding HD streams from camers or OTA, I think alot of the "simplicity" of x264 and Nero Recode could be contributing to VC1 not looking as good. For example, it is quite easy in x264+MeGUI to choose "720p, fix the aspect ratio on scalling (for mod16), HQ-Slow". We get Mp4 fils that are the size we want and great quality (with dual core support etc).
Fair enough - reencoding from OTA is not a scenario we're targeting for the tool. For customers looking for that kind of solution, it probably makes more sense to use a future version of Nero that supports the new VC-1 codec implementation.
I think when future betas support more inputs and are easier to configure we can get fewer "this looks better" etc. It was my thought that MS's aim was to make VC-1 as good as AVC ..
The codec, sure. But codec<>host app - there's lots of stuff that folks want to do with codecs that aren't directly supported by Studio Edition. We're really focusing the 10bit to WVC1 workflow for cases where segment reencoding is requiered. Other workflows would be better supported by other tools using WVC1.

acidsex

11th August 2006, 22:16

Fair enough - reencoding from OTA is not a scenario we're targeting for the tool. For customers looking for that kind of solution, it probably makes more sense to use a future version of Nero that supports the new VC-1 codec implementation.

Did I miss an announcement somewhere on this? I wonder if this is one of those "big" features in the Sept update that John V was touting not too long ago.

benwaggoner

11th August 2006, 22:23

Did I miss an announcement somewhere on this? I wonder if this is one of those "big" features in the Sept update that John V was touting not too long ago.
No, no announcement, sorry. But as always, we intend to get our codec technology supported in every tool. The codec is still in beta, so I doubt anyone has announced a schedule to ship with it.

acidsex

12th August 2006, 17:59

understood but you can see where someone can read your comment about a future version of Nero with VC1 as something that is already being discussed or worked out which gives the impression that its not a matter of if but more of a matter of when.

One other question. If VC1 is still beta, any idea why are companies releasing content for purchase (HD DVD for example) with a beta codec? Or are they using release version implementation and MS is still working on beta code?

benwaggoner

12th August 2006, 20:57

understood but you can see where someone can read your comment about a future version of Nero with VC1 as something that is already being discussed or worked out which gives the impression that its not a matter of if but more of a matter of when.
I expect it's a matter of when for all WMV encoding tools to support WVC1. I don't have anything to report on Nero in specifc.

One other question. If VC1 is still beta, any idea why are companies releasing content for purchase (HD DVD for example) with a beta codec? Or are they using release version implementation and MS is still working on beta code?
The implementation is in beta, but the bitstream is locked down. All the HD DVD titles shipping have passed bitstream validation, so we know they're good to go.

There are betas and there are betas - the codec beta is much more mature than the Studio Edition beta - it's really an enhancement of mature codec, not a ground-up new tool.

sjchmura

15th August 2006, 13:38

bond

15th August 2006, 18:15

for vc-1 incompatibilities of wmv9 read this:
http://forum.doom9.org/showthread.php?t=113968

benwaggoner

15th August 2006, 19:18

Ben,

For those of us compfortable with our (mostly) consumer archiving efforts using WMEncoder,

1. will there be a way in the future to "convert" WMV stuff we encode to VC-1 without the need to "re-encode" and simply change the container or is the stream fundementally different?
We've looked at how to enable this, but it's up in the air if it's reliably possible. The buffer model is different enough between the two formats that you'd probably need a conservative peak rate. And of course, GOP size, resolution, etcetera will need to be HD DVD compliant.

2. If #1 is true then are most of the reigstry "hacks" posted by MS to enable more advanced features after installing WMPlayer 11 in the codec "safe" to use for VC1 streams?
Those are mainly encoding mode features - they shouldn't have any effect on HD DVD compatibilty one way or another.

From what I have read, I ASSume #1 is incorrect - that the streams are slightly different would at least need transcoding etc further degrading quality. I really want to archive a bunch of HD streams (some great high speed skiing stuff in Vail from last year looks amazing from the my camcorder) and some low light data for work, but be sure in the future I can burn these to HD-DVD or Blueray or HD-Ray or whatever is the future. I am scared that my patient data will be lost (that would suck if you were a patient) in 10 years if the WMV format is not "readable" sicne it is not VC-1 complaint with Windows 2012 :)
Well, both VC-1 and the ASF full specifications have been published, so you're in the same boat as with other open standards as far as future compatibility goes - anyone with a compiler and sufficient motivation can build a decoder.

sjchmura

15th August 2006, 19:45

Sorry ben, I meant encoding to WMV using WMencoder 9 and NOT using the VC-1. That is why I am saying VC-1 would be great to have today for alot of smaller work (home also) since I know VC1 (like MP3 will be around forever).

I am sure MS will support VC-1 in 2012 :) I am a bit scared of WMV (non-vc1) will go away from native support in Media Center Ultimate 2012 :)

benwaggoner

15th August 2006, 19:58

Sorry ben, I meant encoding to WMV using WMencoder 9 and NOT using the VC-1. That is why I am saying VC-1 would be great to have today for alot of smaller work (home also) since I know VC1 (like MP3 will be around forever).
If you're using WMV9-Advanced Profile with either the WMP11 beta, Vista, or the beta codec installed, you're making fully SMPTE compliant VC-1 today.

I am sure MS will support VC-1 in 2012 :) I am a bit scared of WMV (non-vc1) will go away from native support in Media Center Ultimate 2012 :)
Well, all those MS MPEG-4v1 with WMA still play, right :)?

Seriously, WMV9 Main Profile (aka WMV3) is also VC-1 Main Profile. You've got decades of support coming for those files.

bond

15th August 2006, 20:38

Seriously, WMV9 Main Profile (aka WMV3) is also VC-1 Main Profile. You've got decades of support coming for those files.not necessarily 100% true, as i pointed out in the thread i linked to

benwaggoner

15th August 2006, 21:06

not necessarily 100% true, as i pointed out in the thread i linked to
Right, stay away from Complex mode.

bond

15th August 2006, 21:26

Right, stay away from Complex mode.you read my thread already, didnt you? complex mode and old advanced profile are not the only issues with wmv9 being incompliant to vc-1

sjchmura

16th August 2006, 19:08

BIt off topic - no more...

So since I have WMP11 installed (actually this was the only WGA problem I ever had but whatever) is there a nice article on how to hack WMEncoder to use the WMV-Advanced Profile that will be VC-1 compliant ... (meaning, do you still need to do those old registry hacks - I assume not right?)

Thanks.... back to the topic :)

benwaggoner

16th August 2006, 22:28

benwaggoner

16th August 2006, 22:29

you read my thread already, didnt you? complex mode and old advanced profile are not the only issues with wmv9 being incompliant to vc-1
It sounds like you've already got Alex on the case there - all I'd do is ask him what was going on :).

zambelli

17th August 2006, 02:34

You're correct. If you have WMP11 installed, any WMV-AP encode will be VC-1 compliant.
Assuming you don't run into any of our beta codec bugs... ;)

Everyone should keep in mind that what's currently available of WMV Advanced Profile codec is all in beta form. Until we release a non-beta version (i.e. with Vista final), everyone should probably refrain from creating any WMV AP videos for "posterity". With any codec that's in beta stage, whether it's WMV, AVC or MPEG-2, there's always the possibility that the encoder might have bugs which result in illegal bitstreams.

Main Profile should be safe though, especially if you're using v9 or v10 bits. Like Bond mentioned, just stay clear of Complex Profile and Main Profile interlaced mode.

bond

17th August 2006, 18:24

Assuming you don't run into any of our beta codec bugs... ;)

Everyone should keep in mind that what's currently available of WMV Advanced Profile codec is all in beta form. Until we release a non-beta version (i.e. with Vista final), everyone should probably refrain from creating any WMV AP videos for "posterity". With any codec that's in beta stage, whether it's WMV, AVC or MPEG-2, there's always the possibility that the encoder might have bugs which result in illegal bitstreams.dont you use the same codec for the studios for hddvd productions?

Main Profile should be safe though, especially if you're using v9 or v10 bits. Like Bond mentioned, just stay clear of Complex Profile and Main Profile interlaced mode.yepa and this strange rtm thingie noone knows where it comes from

benwaggoner

17th August 2006, 19:37

dont you use the same codec for the studios for hddvd productions?
They're actually different binaries at this point.

zambelli

17th August 2006, 20:39

yepa and this strange rtm thingie noone knows where it comes from
There's no way to set that flag "accidentally", so nobody needs to worry about it.

bond

17th August 2006, 20:43

There's no way to set that flag "accidentally", so nobody needs to worry about it.microsoft offers clips having the flag set wrongly on their webpage and there are also other clips floating around having it, being created with the microsoft encoder for sure

so there must be a way to get an incompliant stream "accidentially"

zambelli

17th August 2006, 22:51

microsoft offers clips having the flag set wrongly on their webpage and there are also other clips floating around having it, being created with the microsoft encoder for sure
so there must be a way to get an incompliant stream "accidentially"
I should've been clearer: there is no way to set that flag accidentally by using any available release version of a Microsoft codec.

I'll explain more in the other thread. In the meantime, could you point me to some of the videos hosted on MS web that have that flag set? I should probably check them for VC-1 compliance.

pwimmer

27th August 2006, 21:36

benwaggoner

28th August 2006, 17:31

I've tried to encode a 2880 x 1080 pixel video file today, but noticed that the resolution in WMESE is still restricted to 2048 pixels (thats only 48 pixels better than the old WME).

Please remove this restriction in beta 2!

Or is there actually a limitation in the codec?
That one was my fault, sorry :).
It's a limitation in the published profiles and levels. That makes validating streams with parameters outside those a lot more complex, which we just didn't want to tackle in Beta 1.

So, what would you like o use >2 KiB widths for?

pwimmer

28th August 2006, 20:46

So, what would you like o use >2 KiB widths for?

Stereoscopic 3D @ 1080p. There is already a 2 x 1440 x 1080p sample on my website (http://www.3dtv.at/Movies) and it plays fine on a Pentium D. It's encoded as two separate streams in a single file.

To avoid the overhead of decoding two streams, I'd like to try a single stream instead. Previous decoder versions (I've tried up to 2000 pixels) didn't work because they were not able to use both CPU cores (that's why I created the dual stream format), but I hope to be able to use single stream files soon.

I'm aware that >2k is beyond any existing profile, but why not simply display a warning message but lets users continue encoding?

benwaggoner

28th August 2006, 20:57

Stereoscopic 3D @ 1080p. There is already a 2 x 1440 x 1080p sample on my website (http://www.3dtv.at/Movies) and it plays fine on a Pentium D. It's encoded as two separate streams in a single file.

To avoid the overhead of decoding two streams, I'd like to try a single stream instead. Previous decoder versions (I've tried up to 2000 pixels) didn't work because they were not able to use both CPU cores (that's why I created the dual stream format), but I hope to be able to use single stream files soon.

I'm aware that >2k is beyond any existing profile, but why not simply display a warning message but lets users continue encoding?
That's a reasonable suggestion - I'll pass it on.

Cool project!

Bathrone

3rd September 2006, 06:40

zambelli on Vista Ultimate RC1, it fails to install claiming there isnt enough disc space -when there is. Then it complains about service pack levels that arent relevant for Vista.

Would appreciate your thoughts in my WVC1 encoding thread for the xbox 360 under media center extender.

http://forum.doom9.org/showthread.php?t=115479

benwaggoner

3rd September 2006, 08:15

The current beta of Studio Edition isn't supported on Vista.

Obviously something we're inclined to rectify.

Of course, the new codec is certainly available for Vista, via WME or many other tools - Studio Edition is only needed if you want the advanced preprocessing or segment reencoding.

Sagittaire

4th November 2006, 12:26

Well you are certainely bug because quality is incredibily poor if I compare with VC1 from WME9 ... ???

popper

5th November 2006, 07:31

If you're using WMV9-Advanced Profile with either the WMP11 beta, Vista, or the beta codec installed, you're making fully SMPTE compliant VC-1 today.

Well, all those MS MPEG-4v1 with WMA still play, right :)?

Seriously, WMV9 Main Profile (aka WMV3) is also VC-1 Main Profile. You've got decades of support coming for those files.

well, apparently not if you intend using the MS Xbox360 as your content player....

perhaps as a turnkey re-box job?.

it seems theres going to be a serious NEED for a free MS converter RSN or theres a risk of loosing sales, would you be working on such a converter now perhaps?.

http://www.360-hq.com/article1187.html
"Can I get more technical details about the kinds of files you support?
- WMV video files with the extension ".wmv" are stored in a generic file container format known as ASF. ASF files can contain many different types of codecs within them. Xbox 360 supports most common WMV video codecs, as seen in the following chart:
Name (FourCC)
Windows Media Video 7 (WMV1)
Windows Media Video 8 (WMV2)
Windows Media Video 9 (WMV3)
Windows Media Video 9 Advanced Profile (WVC1)

- There are some files with ".wmv" extension that may play on a PC, but are unsupported on the Xbox 360 console. Windows Media Screen, Windows Media 9 Image (Photostory), and some older Microsoft codecs are unsupported, as seen in the following chart:
Name (FourCC)
Microsoft MPEG-4 Video Codec V3 (MP43)
Microsoft MPEG-4 Video Codec V2 (MP42)
Microsoft MPEG-4 Video Codec V1 (MP41)
Windows Media Screen V9 (MSS2)
Windows Media Screen V7 (MSS1)
Windows Media Video 9 Image (WMVP)
Windows Media Video 9 Image v2 (WVP2)
ISO MPEG-4 Video V1 (MP4S)

Anamorphic content with nonsquare pixels is currently unsupported on Xbox 360 console, and will be displayed as if the pixels were square. If you want to play such content on your console, it is recommended that you re-encode the content in a square pixel format.

News-Source: Official Team Xbox Blog"
http://blogs.msdn.com/xboxteam/archive/2006/10/31/fall-06-supported-video-formats.aspx
Tuesday, October 31, 2006 4:16 PM by L0G0
Fall '06 Supported Video Formats

Bathrone

5th November 2006, 07:38

popper

5th November 2006, 07:58

Those details are not fully correct.

E.g. WMVP, WVP2 and WMVA are playable on the 360 in cases where vista ultimate with media centre extender is used. There maybe others as well, Ive only tested those codecs so far.

i see, i took it that the official Team Xbox Blog url would be a good reference....
http://blogs.msdn.com/xboxteam/archive/2006/10/31/fall-06-supported-video-formats.aspx

but then they did say 'but are unsupported on the Xbox 360' rather than un-playable.

i dont know what to make of this latest blog then.

http://blogs.msdn.com/xboxteam/default.aspx
"Shaheen Gandhi
http://blogs.msdn.com/user/Profile.aspx?UserID=17268
Member since 2/1/2006 8:02:49 PM
Last visited 11/5/2006 1:46:10 AM
Timezone
The timezone the user resides within. -8 GMT
Location Redmond, WA
Occupation Software Engineer "

"Emergence Day
Posted Friday, November 03, 2006 5:15 PM by shaheeng
What takes 4.7 million lines of code, partner teams from all over Microsoft, and millions of dollars to create? The Xbox 360 system software? Nope. This is just the HD DVD player"

"The Xbox 360 HD DVD Player, for the most part, is an entirely software based implementation. Other players on the market have specialized chips (called DSPs) that decode things like H.264, MPEG, VC1, DTS, Dolby Digital, and other codecs. Much like how backwards compatibility for Xbox 1 works on Xbox 360, the heavy parts of HD DVD are all done on Xbox 360's triple-core CPU.

If DVD is an audio/video pipeline with some navigation data (go to the menu, start playing, etc.), HD DVD can be considered a runtime environment where audio/video playback is just one major feature. So let's break down that 4.7 million lines of code. I don't have the numbers for each component, but each of these is a very significant chunk:

Video Codecs: H.264, MPEG-2, VC1
Audio Codecs: Dolby Digital+, DTS, TrueHD, LPCM, MPEG

HDi: The HD DVD runtime engine.
GDI: Drawing stuff like menus

AACS: Cryptography/DRM stuff
MF: Audio/Video pipeline
That's a lot of stuff."

of course, the addition of a simple cheap and low power re-programable chip such as the KiloCore FPGA
might have been a very good thing to add for the future expansion rather than some 'specialized chips (called DSPs)' but thats OT so ill just mention it in passing interest.

http://www.rapportincorporated.com/
http://www.rapportincorporated.com/kilocore/kc256.html
"Rapport’s initial chip, the KC256, includes 256 processing elements in a 16x16 array, with a peak of 25 Giga Byte Operations per second (GBops) at 100 MHz while using less than 500 mW. For even higher performance, multiple KC256 chips can be clustered.

The KC256High performance at low power consumption makes Kilocore™ ideal for demanding applications like advanced security. A KC256 runs the IDEA digital security decryption algorithm 10 times faster than an ARM7®. And the ARM7 consumes 4 times the power (120 mW vs. 500 mW), giving the KC256 a 40x power/performance advantage. Compared to a 1.8 GHz Pentium, the KC256 offers a 500x power/performance advantage. "

http://en.wikipedia.org/wiki/Kilocore
"Kilocore, from Rapport Inc. and IBM, is a high-performance, low-power multicore processor, with 1025 cores. It contains a single PowerPC processing core, and 1024 8-bit Processing Elements running at 125 MHz each, which can be dynamically reconfigured, connected by a shared interconnect. It allows high performance parallel processing."

travisbell

10th November 2006, 05:39

Hey guys, just curious... it's been 4 months since beta 1 was posted... any updates as to when a beta 2 will arrive?

Thanks!

photoguy123

21st December 2006, 06:17

The studio edtion has vanished from the MS site with no explaination.

No more beta 1, no beta anything, no release.

Would be kind of nice if some on could have at least posted what the heck happened.

Was the product scrapped?