Re: [rtcweb] Input to Video Codec Selection

Harald Alvestrand <> Sat, 02 March 2013 09:57 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id F1F9A21F8FEA for <>; Sat, 2 Mar 2013 01:57:58 -0800 (PST)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -111.444
X-Spam-Status: No, score=-111.444 tagged_above=-999 required=5 tests=[AWL=1.154, BAYES_00=-2.599, GB_I_LETTER=-2, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-8, USER_IN_WHITELIST=-100]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id 8HtHN1ksoo2i for <>; Sat, 2 Mar 2013 01:57:56 -0800 (PST)
Received: from ( []) by (Postfix) with ESMTP id 1732421F8FFA for <>; Sat, 2 Mar 2013 01:57:56 -0800 (PST)
Received: from localhost (localhost []) by (Postfix) with ESMTP id 40F9439E0FE; Sat, 2 Mar 2013 10:57:55 +0100 (CET)
X-Virus-Scanned: Debian amavisd-new at
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id 4tnauKb5VkJT; Sat, 2 Mar 2013 10:57:52 +0100 (CET)
Received: from [IPv6:2001:470:de0a:27:b136:6ef7:2913:b4cd] (unknown [IPv6:2001:470:de0a:27:b136:6ef7:2913:b4cd]) by (Postfix) with ESMTPSA id 7D98739E0A7; Sat, 2 Mar 2013 10:57:52 +0100 (CET)
Message-ID: <>
Date: Sat, 02 Mar 2013 10:57:52 +0100
From: Harald Alvestrand <>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130221 Thunderbird/17.0.3
MIME-Version: 1.0
To: Gaelle Martin-Cocher <>
References: <> <>
In-Reply-To: <>
Content-Type: multipart/alternative; boundary="------------070805080201080207090603"
Cc: "" <>
Subject: Re: [rtcweb] Input to Video Codec Selection
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: Real-Time Communication in WEB-browsers working group list <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Sat, 02 Mar 2013 09:57:59 -0000

On 03/01/2013 11:21 PM, Gaelle Martin-Cocher wrote:
> Dear All,
> Further to Magnus email, while I assume there might not be "something new" to learn at the meeting, I believe the below requested clarifications on existing information would be useful.  Implementers should clearly know which license they can pick or get when it comes to VP8 and by which groups.  I believe answers in advance of the meeting would help the discussion at the meeting.
> Questions 1 & 2:
> It is assumed that in  the case of choosing VP8, RTCWeb would reference  the informational RFC 6386.
Yes, that is the intent.
> Q1: Is there an intent to move that RFC to the standard track at a point in time?
No. I don't personally see any benefit in doing so at this time.
> Q2: Would that change the rule of "who" is obliged to make an IPR declaration?
Speaking with my IETF-amateur-lawyer hat on (and as a former chair of 
the IPR WG): No, it does not change the rule. The rule depends on 
whether the technology in question is discussed in the IETF, not on the 
nature of the contribution. RFC 3979 section 6.1.2 refers to 
"Contribution", the definition of that term in RFC 3979 section 1 letter 
j makes it completely explicit that RFC Editor Contributions are covered 
by the term "Contribution".
> Question 3:
> The IPR disclosure was made on the draft 2 of draft-bankoski-vp8-bitstream-02" as per
> Draft 3 and onward contains the copyright license and the additional IP rights grant.
> Q3: Is the initial IPR disclosure still valid?
Yes. RFC 3979 section 6.2.1.

    The IPR disclosure required pursuant to section 6.1.1 must be made as
    soon as reasonably possible after the Contribution is published in an
    Internet Draft unless the required disclosure is already on file.
    For example, if the Contribution is an update to a Contribution for
    which an IPR disclosure has already been made and the applicability
    of the disclosure is not changed by the new Contribution, then no new
    disclosure is required.

> Question 4:
> The informational RFC 6386 contains the decoder code and some piece of encoder code.
> Though the IP rights grant mentioned in the RFC is offered against:
>     "This implementation" means the copyrightable works distributed by
>     Google as part of the WebM Project."
> Q4: As such the  IP rights grant does not seem to apply to the RFC itself or to an implementation of the code contained in the RFC.  Should that be corrected or is that the intent?
Speaking with WEBM hat on:

There are two grants - the grant of license to copyrighted works, and 
the grant of license to patented technology.

Software license: - 
classical 3-clause BSD.
Patent license 1: - covers 
any implementation that produces or consumes VP8 bitstreams.
Patent license 2: - 
covers usage of the implementation.

> Question 5:
> The additional IP grant is applied to a particular implementation (namely the WebM VP8 code) without modifications.
> Any derivative work either:
> - produced from the reference code in WebM (that is a possible optimized version of it); or
> - produced from the RFC text or the code provided within the RFC (while not using the WebM code)
> does not have the benefit of the additional IP grant.
> In other words a conformant implementation does not necessarily have the benefit of the additional IP grant.
> I am not confident that the VP8 code can be used "as is" for certain platforms. I would think that the code might need some modification to provide the desired performance. In other words, it should be clear that those implementers would not necessarily receive the benefit of that grant.
> If the answer to Q3 is negative, then there is no IP license statement at all that applies to a "conformant implementation of the RFC" (aka a derivative work).
> If the answer to Q3 is positive, it is not clear  how to reconcile the declaration inside the RFC and the declaration that is attached to the the RFC draft for implementers that would not modify the code.
> Q5: Can this be clarified or confirmed?
All 3 of the pages referred to above permit the production of derivative 
works. Quoth:

- bitstream: " ... license to make, have made, use, offer to sell, sell, 
import, and otherwise transfer implementations of this specification" 
(whether derived from the example code or not)
- copyright: ".... Redistribution and use in source and binary forms, 
with or without modification, are permitted provided that.."
- patent: "... patent license to make, have made, use, offer to sell, 
sell, import, transfer, and otherwise run, modify and propagate the 
contents of this implementation"

I believe there should be no issue here; modification is permitted.
> Question 6:
> The IPR disclosure in IETF is different than the IPR statement made in MPEG (see document sent by Harald earlier).
> Q6: the differences in license statement and IP grant referring to WebM code are rather confusing. Can it be clarified which license, copyright, grant are provided for RFC 6386?
The statement we made in MPEG was crafted to be as similar to others' 
statements made in MPEG as possible, in order to respect MPEG's legal 
language traditions - which in turn should minimize the need for 
clarification of what was granted when discussing with people used to 
the MPEG language tradition.

We believe that the statement made in MPEG is wholly within the 
statements made on the WEBM website - all permissions implied by the 
statement in MPEG should also be permitted by the statements on the WEBM 
website. We haven't tried to analyze whether there are cases where 
someone can do something within the permissions granted on the WEBM 
website that would not be permitted under the MPEG statement - the MPEG 
statement was aimed to allow the document to progress within MPEG; 
people who want to read the
> In conclusion, before advancing this draft, or considering it as a candidate for RTCWeb,  consistency and clarity should be ensured between the IPR grant associated with the IETF draft, the IPR grants within the IETF draft document itself, the IPR grant given for MPEG, and any IPR grant given in connection with the WEBM project for this same work.  Otherwise, the IPR status of the work that is undertaken is indeterminate, and likely will not produce a result that will be useful.

Speaking with my Google hat on:

We will of course seek maximum clarity of the statements we make. 
Unfortunately, different organizations have different traditions on how 
these things should be worded, and we cannot guarantee that there can't 
be differences in interpretation.

However, I (speaking with my personal hat on) think the current 
statements on the WEBM website are pretty clear. I have yet to see a 
concrete scenario where there would be any reasonable doubt about 
whether usage of VP8 is permitted by Google or not - and in all cases 
except for those that fall within the defensive suspension exceptions, 
it is permitted.

Hope that helps!