Return-path: Received: from smtp-vbr1.xs4all.nl ([194.109.24.21]:1663 "EHLO smtp-vbr1.xs4all.nl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751512AbaCZIuY (ORCPT ); Wed, 26 Mar 2014 04:50:24 -0400 Message-ID: <5332948E.4040506@xs4all.nl> Date: Wed, 26 Mar 2014 09:49:18 +0100 From: Hans Verkuil MIME-Version: 1.0 To: Tom CC: linux-media@vger.kernel.org Subject: Re: why frameformat instead pixelformat? References: In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-media-owner@vger.kernel.org List-ID: On 03/24/14 17:52, Tom wrote: > Hello, > > while reading into the media-api issue I found out that for configuring the > entity pads a frameformat is used. > > For that I found the negotiation rfc of that topic, but I don't really get > the relevance of a frameformat. > > http://www.spinics.net/lists/linux-media/msg10006.html > > Can anyone explain why the media-api uses the frameformat instead of the > pixelformat and what the main differences are? The pixelformat describes the image format in memory, the mbus_framefmt describes the image data as it is transferred over a hardware bus. It's as simple as that. While the two are related (i.e. to get a certain pixelformat you probably need to set up a specific mbus_framefmt), the details of this relationship are board or even use-case dependent. Regards, Hans