From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mail-vk0-f52.google.com ([209.85.213.52]:35834 "EHLO mail-vk0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751432AbcF3N4S (ORCPT ); Thu, 30 Jun 2016 09:56:18 -0400 Received: by mail-vk0-f52.google.com with SMTP id u68so67834946vkf.2 for ; Thu, 30 Jun 2016 06:55:26 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <790b5e1664c84e806a13143eff1c79b95fb4bf63.1467240152.git.mchehab@s-opensource.com> References: <0003e025f7664aae1500f084bbd6f7aa5d92d47f.1467240152.git.mchehab@s-opensource.com> <790b5e1664c84e806a13143eff1c79b95fb4bf63.1467240152.git.mchehab@s-opensource.com> From: Steven Toth Date: Thu, 30 Jun 2016 09:48:30 -0400 Message-ID: Subject: Re: [PATCH 08/10] dvb_frontend: create a new ops to help returning signals in dB To: Mauro Carvalho Chehab Cc: Linux Media Mailing List , Mauro Carvalho Chehab , Jonathan Corbet , Michael Ira Krufky Content-Type: text/plain; charset=UTF-8 Sender: linux-media-owner@vger.kernel.org List-ID: > add a new ops that will allow tuners to better report the > dB level of its AGC logic to the demod drivers. As the maximum > gain may vary from tuner to tuner, we'll be reversing the > logic here: instead of reporting the gain, let's report the > attenuation. This way, converting from it to the legacy DVBv3 > way is trivial. It is also easy to adjust the level of > the received signal to dBm, as it is just a matter of adding > an offset at the demod and/or at the bridge driver. Mauro, Have you verified this work with a detailed spectrum analysis study? If so then please share. For example, by measuring the I/F out of various tuners in a mix of use cases, with and without the AGC being driven by any downstream demodulator? Also, taking into consideration any external LNA variance. I'm concerned that a tuner AGC Gain is a meaningless measurement and in practice demodulators don't actually care, and tuners don't implement their gain reporting capabilities correctly at all. This feels like a solution to a problem that doesn't exist. -- Steven Toth - Kernel Labs http://www.kernellabs.com +1.646.355.8490