[whatwg] MPEG-1 subset proposal for HTML5 video codec

[jjcogliati-whatwg at yahoo.com]

Thank you for a very informative reply.  Inline comments follow.

--- On Sun, 5/31/09, Gregory Maxwell <gmaxwell at gmail.com> wrote:

> From: Gregory Maxwell <gmaxwell at gmail.com>
> Subject: Re: [whatwg] MPEG-1 subset proposal for HTML5 video codec
> To: whatwg at lists.whatwg.org
> Date: Sunday, May 31, 2009, 2:17 PM
> 2009/5/31  <jjcogliati-whatwg
> at yahoo.com>:
> > Since the near complete MPEG-1 committee draft was
> publicly available in December 1991,
> [snip]
> 
> You keep repeating this particular piece of misinformation,
> so I'm
> worried that people are going to take your word for it and
> get into
> trouble.
> 
> What you are claiming with respect to the inventors
> disclosure and
> patent duration is correct for patents filed and granted
> today but it
> not true for patents from the mid-1990s.
> 
> Prior to mid-1995 was possible to use application
> extensions to defer
> the grant date of a patent indefinitely.  You could
> begin an
> application in 1988, publicly expose your invention in
> 1991, all the
> while filing extensions only to have the patent granted in
> 1995.
> 
> I am somewhat surprised that you are unaware of this
> issue,
> considering that you mentioned it specifically by name
> (submarine
> patent).

Yes, I agree and I was not making this clear in my reply posts.  The first email I sent I did detail this.  

> I'm more familiar with the area of audio coding than video,
> so I don't
> have a ready list of patents that read on mpeg1 video.
> However, There
> are mid-90s patents which read on both layer-2 (e.g.
> 5,214,678) and
> layer-3 audio which followed the 'submarine patent' style
> of prolonged
> application and late disclosure times.

That is interesting that 5,214,678 is considered to read on Layer-2 since
AudioMPEG says that they are not doing licensing for Video-CD, which uses
MPEG-1 Layer 2 audio.  It was granted in May 25, 1993 and filed on May 31, 
1990, so it barely made it in three years (and will not expire till 
May 31, 2010).  

http://patft1.uspto.gov/netacgi/nph-Parser?patentnumber=5,214,678
http://www.audiompeg.com/

> Additionally, Theora avoids some techniques used in MPEG1
> which have
> been believed to be patented.  For example, the
> differential coding of
> motion vectors. While I don't have the knowledge needed to
> provide a
> detailed analysis, even I know enough to point out at least
> a few
> engineering reasons why Theora has less patent exposure
> surface than
> MPEG1.

I can certainly believe that MPEG-1 Video might be non-royalty free and 
Theora might be.  I haven't really looked at the exact coding of Theora motion vectors.  That is an interesting thing to look at. 

> Without the benefit of mpeg layer-3 audio MPEG1 is left
> enormously
> handicapped compared to Theora+Vorbis. 16kHz 16bit stereo
> PCM is
> 512kbit/sec on it own, which is comparable to the total
> bitrate 'high
> quality' option delivered by sites like Youtube. And 16kHz
> audio is
> pretty poor for anything that needs to carry music. 

Layer-2 audio can certainly beat PCM for compression, since it can reduce
the bit rate for coding the quieter frequency bands.  Typical stereo bit 
rates for stereo Layer 2 audio are probably more on the order of 256 
kbit/s.  Vorbis and MPEG-1 Audio Layer 3 certainly can beat this rate.    

> While you could
> argue for using MPEG1+Vorbis, none of the few parties who
> indicated
> that they would not ship Theora have stated they would (or
> are
> already) shipping Vorbis. (For example, Nokia does not ship
> Vorbis on
> their Linux tables)  Everyone shipping Vorbis already
> seems to have no
> issue with Theora.

I am not going to argue for MPEG-1 video plus Vorbis audio.  

> Even if you pay fairly low prices for transit the cost of
> sending PCM
> audio vs Vorbis is likely enough to pay for the H.264+AAC
> licensing no
> matter what it turns out to be in 2010.  A 'free'
> format which has an
> effective price much higher than the 'non-free' stuff would
> be
> something of a hollow victory.

Interesting point.  I think that transit costs will decrease 
faster than H.264+AAC licensing costs, (unless Theora and Vorbis 
start causing serious competion.)

> And really, now that we see multiple large companies with
> experienced
> legal teams and non-trivial exposure committed to shipping
> Theora I
> think we're kidding ourselves when we attempt to analyze
> this as a
> legal issue. It's not. It's a business/political decision.
> The market
> is now going to battle it out.  Enjoy the show.

I agree.  I was not aware that Google planned on shipping Theora support when I made the first email last week (Wikipedia article since updated). 
If Ogg Theora and Vorbis become the defacto standard, that is 
fine with me.  

Right now, the best video codec/audio codec that works with Gstreamer good 
plugins (i.e. Linux), Quicktime and Media Player is Motion JPEG with PCM 
audio, which I have used for shipping videos on CDs and USB drives, but is 
impractical for online transfers.  I am hoping for a better outcome with 
the video tag.  

Josh Cogliati