Robert McLaws: Windows Edition

Blogging about Windows since before Vista became a bad word

Microsoft Successfully Fools Industry on Video Codecs

I was puzzled why so many news sites were referring to Windows Media Video as Microsoft's implementation of VC-1. Then I looked on Microsoft's Silverlight website, which has a self-contradicting statement:

VC-1 is an industry-standard video format, recognized by the Society of Motion Picture and Television Engineers (SMPTE), and most notably ships in all HD-DVD and Blu-ray Disc–certified electronics, hardware, and tools. Windows Media Video 9 (WMV-9) is the Microsoft implementation of the SMPTE VC-1 standard video codec. Microsoft initiated development of the standard with the release of WMV-9 to SMPTE.

Then I looked on Wikipedia, and even their first paragraph contains a similar contradictory statement:

VC-1 is the informal name of the SMPTE 421M video codec standard initially developed by Microsoft. WMV3, better known as Windows Media Video 9 codec, served as the basis for development of the VC-1 codec specification. On April 3, 2006, SMPTE announced the formal release of the VC-1 standard as SMPTE 421M. Its most popular implementation is Windows Media Video 9.

Wanting to make sure I wasn't in a parallel universe or something, I went on to Microsoft's WMV9 site... and there it is again!

The Simple and Main profiles have been complete for several years, and existing implementations such as WMV 9 have long supported the creation and playback of content using these profiles, as well as an early implementation of the Advanced profile.


Windows Media Video 9 is the Microsoft implementation of the VC-1 SMPTE standard. It supports Simple, Main, and Advanced profiles.

So, while it may be a relatively minor point of contention, in light of the whole Office Open XML debate, I think it's important that we be clear on where this stuff came from. I'm not saying that this was an intentional, coordinated effort on Microsoft's part to mislead people, even though my title might suggest otherwise (hey, I gotta get people to read it, right?) but if it was, it sure worked.

WMV9 is not Microsoft's implementation of VC-1. That would insinuate that VC-1 came first, which is not the case. WMV9 was submitted by Microsoft to the SMPTE as a standard, which 15 companies contributed to. This process was similar to the way Microsoft submitted OOXML as a standard. The result of that standardization process was given the name VC-1.

So WMV9 is not Microsoft's implementation of the standard, it's the codec from which the standard was based. Maybe I'm just being nitpicky and argumentative today, and maybe the industry is better-served if people don't see the MS stamp on the codec, but I think there is a difference.



  • tf said:

    I think you are being nitpicky and even misguided. Microsoft essentially ripped off mpeg-4 part 2 for WMV, they used a slightly altered blocking algorithm. When they standardized it, they realized it was all existing technology (mpeg-4) and that all that belonged to Microsoft were 2 blocking algorithms. Considering it is a ripoff of a preexisting standard and then it regained its "standard"-ness, I think it's most appropriate to refer back to VC-1 as the standard. (The question of whether or not VC-1 codecs can digest WMV remains a big one, which leads me to suggest that Microsoft dump WMV entirely for the standard.)

    You can't take credit for a standard if your implementation is nonstandard.

    April 16, 2007 5:06 PM
  • Charles said:

    "So WMV9 is not Microsoft's implementation of the standard, it's the codec from which the standard was based. Maybe I'm just being nitpicky..." No, I'm afraid you're just wrong. You're assuming there haven't been multiple releases of WMV9, which is incorrect.

    April 16, 2007 5:34 PM
  • Shawn Oster said:

    "So WMV9 is not Microsoft's implementation of the standard, it's the codec from which the standard was based."

    You're just running into an origination issue, the classic chicken and the egg problem.  The standard may have come from WMV9 yet it *is* a standard, meaning anyone can impliment it.  This is in fact how these things usually happen, standards rarely just appear without an implimentation that goes with it.  

    Take Java for example.  When it first came out you may have had the same issues with Sun saying "Sun's version of Java" yet today no one questions it since there are at least three different versions.  The only reason it seems weird to you is because no one else has an VC-1 encoder, once someone else makes one you're damn well going to want to know whose version of the VC-1 encoder you're using.

    April 16, 2007 6:45 PM
  • tf said:

    Shawn, there are other VC-1 encoders. Most of the stuff being released at NAB are including it.

    April 16, 2007 7:52 PM
  • bill said:

    Sorry Mr. McClaws but you are wrong on this one. The statement on all those sites is spot on accurate.

    Think about other industries. Volvo originated many of the safety devices/processes that are not required by U.S. law. And Volvo's current vehicles ADHERE to those new laws. But using your logic... they couldnt claim this as the laws were the outgrowth of Volvo's hard work intially.

    In this example. Volvo=Microsoft. US Government = SMPTE

    April 17, 2007 1:11 AM
  • Tom Servo said:

    WMV9 Advanced Profile is VC-1.

    April 17, 2007 7:35 AM
  • Yeah i'm putting my eggs in the nitpicky basket :)

    Whichever way you look at it WMV9 is an implementation of VC-1. Doesn't matter that it came first, it's still an implementation of it.

    April 17, 2007 5:44 PM
  • Andrew Simpson said:

    Yup, have to agree with the other commenters.

    It's also kinda like how Microsoft's implementation of the CLI standard is .net. Sure, .net came first and then was standardised, but you also have the Rotor reference implementation and the Mono open-source implementation of the standard.

    It's only a contradiction if you take it as an axiom that standards must be developed before implementations.

    April 17, 2007 6:04 PM
  • Neil Smith said:

    Charles 'rebuttal of Robert's final point saying "No, I'm afraid you're just wrong. You're assuming there haven't been multiple releases of WMV9, which is incorrect" is probably better explained by this post from Waggoner detailing the release order towards VC-1 :

    It's always nice to refer to primary documentation, but this is as close as we get  ;-)


    Cheers - Neil Smith

    MVP Digital Media

    April 17, 2007 6:24 PM