Re: [codec] it MUST NOT exceed 1275 bytes?

Gregory Maxwell <> Mon, 25 July 2011 22:01 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 8395821F8A7E for <>; Mon, 25 Jul 2011 15:01:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id r2D6+ZPJ6wSW for <>; Mon, 25 Jul 2011 15:01:34 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id BB3F421F8A57 for <>; Mon, 25 Jul 2011 15:01:33 -0700 (PDT)
Received: from ([]) (using TLSv1) by ([]) with SMTP ID; Mon, 25 Jul 2011 15:01:33 PDT
Received: from ([fe80::c821:7c81:f21f:8bc7]) by ([::1]) with mapi; Mon, 25 Jul 2011 14:59:54 -0700
From: Gregory Maxwell <>
To: Christian Hoene <>, "" <>
Date: Mon, 25 Jul 2011 14:59:53 -0700
Thread-Topic: [codec] it MUST NOT exceed 1275 bytes?
Thread-Index: AQJtwoAskKc2oLEEloJVQeG4M1e3jQHD8BXzk6xO0MCAACpa2g==
Message-ID: <>
References: <007101cc4aee$e622bd50$b26837f0$> <>, <00d701cc4afb$24ef21c0$6ecd6540$>
In-Reply-To: <00d701cc4afb$24ef21c0$6ecd6540$>
Accept-Language: en-US
Content-Language: en-US
acceptlanguage: en-US
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [codec] it MUST NOT exceed 1275 bytes?
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: Codec WG <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Mon, 25 Jul 2011 22:01:34 -0000

Christian Hoene [] wrote:
> [Christian Hoene] Uuh, be careful, there is a Fraunhofer claim on switching
> frame sizes depending on signal information.
> The algorithm that you describe is not part of the reference implementation,
> isn't it?

The reference implementation does not— though the format allows
for an encoder to choose do this.  (And could not prohibit it without
also breaking switching for other purposes) 

The reasons the reference implementation does are primarily for
encoder complexity  and latency (it's a decision better made with
more look-ahead) not IPR. 

And, to avoid sounding like I agree with your blanket statement should
someone cite this post later for other reasons:

Frame size switching is something which _can_ be done without IP
encumbrance as demonstrated by the established codecs which do this,
(e.g. Vorbis, though there was substantial care taken to do it in a manner
which prevented risk) and the fact that the most commonly
cited patent in this area, though not the one you're talking about,
 is now expired (and, thus, arguably provides strong prior art for some
 implementation approaches).