Return-Path: Subject: RE: Testing SBC filtering functions From: Marcel Holtmann To: hoene@uni-tuebingen.de Cc: 'Siarhei Siamashka' , 'ext Brad Midgley' , 'Jaska Uimonen' , linux-bluetooth@vger.kernel.org In-Reply-To: <002001c969a3$fd7f1680$f87d4380$@hoene@gmx.net> References: <1227879337.20555.12.camel@esdhcp03999.research.nokia.com> <200812230130.43968.siarhei.siamashka@nokia.com> <1229994015.8047.27.camel@californication> <200812231245.14493.siarhei.siamashka@nokia.com> <1230032888.8047.34.camel@californication> <001501c96996$20018070$60048150$@hoene@gmx.net> <1230544837.13620.44.camel@californication> <002001c969a3$fd7f1680$f87d4380$@hoene@gmx.net> Content-Type: text/plain Date: Mon, 29 Dec 2008 13:03:01 +0100 Message-Id: <1230552181.15666.10.camel@californication> Mime-Version: 1.0 Sender: linux-bluetooth-owner@vger.kernel.org List-ID: Hi Christian, > > I leave the patches to the guys that actually know what they are doing. > > > > So I like the test script, but I would really prefer if we can use > > sbctester.c program to verify the results. What do you think? > > The sbctester.c works fine for the decoder and is used in the testing > script. The testing of the encoder might also work with test procedure > implemented in sbctester.c. However, the official Bluetooth testing spec > requires that > "9.2.2.1 SBC Encoder ... [E2] The subjective quality (measured by > standardized way or by objective testing > methods, see [6]and [7]) shall be equivalent to that of reference encoder." > with " [6] Rec. ITU-R BS.1116-1, "METHODS FOR THE SUBJECTIVE ASSESSMENT OF > SMALL IMPAIRMENTS IN AUDIO SYSTEMS INCLUDING MULTICHANNEL SOUND SYSTEMS", > 1994-1997 > [7] Rec. ITU-R BS.1387-1, "METHOD FOR OBJECTIVE MEASUREMENTS OF PERCEIVED > AUDIO QUALITY", 2001". > Until the encoder works fine, we should use BS.1387-1 (PEAQ). If all errors > have been fixed, then sbctester.c might can be used instead of ITU-R > BS.1387-1. However, then, we might still need the reference implementations > or some pretranslated wav and sbc files. I really prefer if we have an open source solution available to everybody for the SBC conformance testing. If sbctester.c is not good enough than we have to write one. I also don't mind using the reference encoder/decoder and the reference files. Using wine is not optimal, but I don't really mind it. Another question that I have is how you use the sbc_enc_test_01.wav and sbc_enc_test_02.wav files with Windows reference encoder? It turns into a busy loop for me. Regards Marcel