|
commit 25c2f620b6de0f47a9d7e6e7859c035b31031122
Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Jun 23 11:41:11 2015 -0700 Improve OPUS_GET_LSB_DEPTH documentation. We use at most 16 when using fixed-point (for the API or the internal implementation), but we didn't tell anyone this. commit 3bd0231be14b77ab83f3afe2785f82338a1418b8 Author: Ralph Giles <giles@mozilla.com> Date: Fri May 8 14:22:07 2015 -0700 Add release checksums. Keep a copy of the release tarball sha-256 checksums in the repository for reference. Having a copy with the code in distributed version control adds a verification path. This should match http://downloads.xiph.org/releases/opus/SHA256SUMS.txt create mode 100644 releases.sha2 commit ccfb0666852c4f636930e62a43961a41b5a1f56a Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Apr 28 10:40:46 2015 -0700 oggopus: Bump draft date and version for 07 publication. commit 5b20cb01b32139a9e918e792634d1d31caf73c58 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Apr 21 20:33:42 2015 -0400 rtp draft -11 commit 51abf4896b6373d405a7b8639e40049e944c895e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Apr 14 20:50:28 2015 -0400 rtp draft -10 (deleting text) commit 5771a672cad33150ddb324b977c32b2cc96ed625 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Apr 10 17:50:26 2015 -0400 IESG RTP draft update commit 36e0445e619a2814b05d4e48dbaee785b70aac4d Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Mar 13 16:32:23 2015 -0700 Ogg Opus draft: Address chair review comments. 1. Removed an inappropriate normative MAY. 2. Gave an explicit range of sample rates deemed to be "non-crazy". 3. Give explicit guidance on packet sizes that SHOULD and MAY be rejected. commit 6814b2c2501cbe3a5da4686146095d7aae9dc4f2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Feb 27 15:01:28 2015 -0500 Avoiding the term "mode" in opus_demo commit 2d83e7e297a11c729f8723c2f01c1c35cac0c01f Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Feb 27 11:11:57 2015 -0800 Wrap _mm_cvtepi...() intrinsics in macros on clang. We already needed these macros for gcc with optimizations disabled, but it appears clang needs them all the time. Thanks to Jonathan Lennox for the report. commit 3b74d8bd728cc5582db785bc8c1d6ac97e9874ee Author: Tristan Matthews <tmatth@videolan.org> Date: Mon Feb 23 18:22:09 2015 -0500 opus_demo: remove unused but set values commit 75d81f5bb3705135a0259be0e9602820f1c6cb5c Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Feb 20 12:44:10 2015 -0800 Document how to tell if your build is fixed-point. This way we won't break this by accident. commit 45cbdff36f0d0c2a395326d09e4dd375601e28cb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Feb 11 14:10:03 2015 -0500 Adds MIPS files to the make dist tarball commit 155f99c87a4047723c4478bab319d6d745ccc174 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Feb 6 13:34:32 2015 -0500 Addressing Gen-ART comments on the RTP draft commit 66611f136733ca7f52c26f1f1ab11d506cc6e752 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 13 09:19:24 2015 -0500 RTP draft: chair review commit fdc6d3cc1218cb164d3a2f78d1dac54e6cde619a Author: Ralph Giles <giles@mozilla.com> Date: Sat Jan 10 13:57:24 2015 -0800 Update ISO Base Media Format draft to version 0.6.6. commit af68b857a3a083494bac2c2e5725bc192588afe2 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 9 08:55:21 2015 -0800 Update ISO Base Media Format draft to version 0.6.2. commit 65fb456ce8ad96a7634c1e55a2c53af92fec3c45 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 6 02:35:31 2015 -0500 RTP: the return of maxaveragebitrate commit 7422189ab16de442554da7f73c3c6f3c15130d22 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Jan 2 15:48:54 2015 -0800 Fix silk_VQ_WMat_EC_sse4_1(). During review of c95c9a048f32, I replaced a call to _mm_cvtepi8_epi32() with the OP_CVTEPI16_EPI32_M64() macro (note the 16 instead of 8). Make a separate OP_CVTEPI8_EPI32_M32() macro and use that instead. Thaks to Wei Zhou for the report. commit 23f503ad1c388aa9171af931ccb2f114f0839e0e Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Dec 26 08:31:39 2014 -0800 Fix warnings in pedantic build. This should suppress our current issues with unused parameters, unused variables, and set-but-not-used variables. commit 0b0a2b4d066c2be3d551ba6429854c20ae147204 Author: Viswanath Puttagunta <viswanath.puttagunta@linaro.org> Date: Fri Dec 19 17:25:55 2014 -0600 armv7: celt_pitch_xcorr: Introduce ARM NEON intrinsics Optimize celt_pitch_xcorr function (for floating point) using ARM NEON intrinsics for SoCs that have NEON VFP unit. To enable this optimization, use --enable-intrinsics configure option. Compile time and runtime checks are also supported to make sure this optimization is only enabled when the compiler supports NEON intrinsics. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> create mode 100644 celt/arm/celt_neon_intr.c commit 5b712da9c2faa51ebf3336161138b2d2fa54a47d Author: Ralph Giles <giles@mozilla.com> Date: Sat Dec 20 12:37:48 2014 -0800 Document the motivation for the 120 ms repacketizer limit. See the mailing list discussion at http://lists.xiph.org/pipermail/opus/2014-December/002802.html commit 7424bb0a247c0692327001addb036f6e2313e970 Author: Ron <ron@debian.org> Date: Sat Dec 20 19:46:20 2014 +1030 Fix the quoting of AC_LANG_PROGRAM It was under-quoted in c95c9a048f3283afb2e412b10085d4f7c19e1412, which leads to a flood of warnings when it gets incorrectly expanded, like: configure.ac:354: warning: AC_LANG_CONFTEST: no AC_LANG_SOURCE call detected in body commit 3b1928ce3e6943377d9b1d96b9b95ca27e8e693c Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 8 16:25:57 2014 -0500 RTP draft: addressing comments from Martin Thompson commit aad281878d650f680a895730eaa386df7cb3e8f0 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Mon Dec 1 10:47:25 2014 -0800 Fix celt_pitch_xcorr_c signature. This should not take an arch parameter, so it can properly be used as a fallback for accelerated versions which do not. This patch instead provides a separate version which can call accelerated helpers for platforms that have taken that approach. commit 25b27a9c167302769db512a9e32c66323bc7904c Author: Mark Harris <mark.hsj@gmail.com> Date: Thu Nov 27 08:48:09 2014 -0800 multistream: improve arg check Avoid undefined behavior (signed arithmetic overflow) or implementation-defined behavior (malloc(0)) on out-of-range arguments, e.g. opus_multistream_encoder_create(48000, 2, 2147483647, 1, ...) or opus_multistream_surround_encoder_create(48000, 3, 0, ...). Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit d10755e95c49f5ec925e70598877c82ad865a189 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 19 11:07:50 2014 -0500 RTP draft in last call commit 8921f82fb9453a82f53280179f7b4fe50f4246e6 Author: Cameron Gutman <aicommander@gmail.com> Date: Wed Nov 19 03:02:11 2014 -0500 float_cast: Fix MSVC ARM build Signed-off-by: Tristan Matthews <tmatth@videolan.org> commit 8ab1968e13fcedbe5c60859b9f3b60193061d9bb Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 17:04:45 2014 -0700 oggopus: bump draft date and version for 06 publication. commit 79cce68d74cf1feb66f4f7c9c12c4be2f4f0b447 Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 17:02:02 2014 -0700 oggopus: Mention 'malformed' at the first instance of invalid TOC. This is just a should, so we want to leave the RFC 6716 reference to the second occurance, but using the same language ties them together and to the later occurances. commit ddbc768c986ffb5390b0d37a6238e5274f47899e Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 16:51:57 2014 -0700 oggopus: Refer to RFC 6716 on how to handle malformed packets. The Opus RFC doesn't really say what to do beyond rejecting a particular packet, but having the reference reinforces that we're trying to leverage the same constraints in the specific context of ogg encapsulation, and this isn't a new rule. commit ba409a420bbf836de9556f89fd5f1a9622b8e834 Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 16:35:50 2014 -0700 oggopus: use 'invalid' to match other occurances and RFC 6716. Suggestion from mark4o. commit 54f061d1598798201769ec30768a757266c552e5 Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 15:32:10 2014 -0700 oggopus: Improve wording. commit a38044e0978af97eae2607f0e774b06066f0b83b Author: Ralph Giles <giles@mozilla.com> Date: Fri Oct 17 15:08:11 2014 -0700 oggopus: Refer to 'TOC sequence' instead of byte. This avoids confusion with the number of frames being defined by more than the first byte with code 3 packets. Patch from mark4o. commit 5e0c904eb665fcc13441aba7fc89ed3cd14e9623 Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 14 18:23:32 2014 -0700 oggopus: Fix opus_encode_ctl invocations. The macros take arguments, they aren't just constants. Thanks to mark4o for the report. commit 3731a57a1b964c4e39a8b8cab9874a124f65a8b0 Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 14 17:47:10 2014 -0700 oggopus: Improve wording. commit 08aa815e0a88632d308d5517110cf8ba518fd998 Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 14 14:54:31 2014 -0700 oggopus: bump draft date and version for 05 publication. commit 7abe7d65c21d0e3ee6020d62aaa6f78554c9d3fc Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 14 14:52:31 2014 -0700 oggopus: Clarify leading zeros are ok for R128 tags. Add a limit of no more than 6 characters for these tags' values. commit fd562eefa2f7b506d7bedd42335e97256c067d6e Author: Tristan Matthews <tmatth@videolan.org> Date: Sun Oct 12 15:12:52 2014 -0400 tests: fix linking for --enable-fixed-point on ARM commit 32ff38bc8a3f1eba8cd27721cf4ce834ad12f9ad Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 7 16:05:07 2014 -0700 Update mp4 encapsulation spec to v0.5.3. Based on http://vfrmaniac.fushizen.eu/contents/opus_in_isobmff.html commit 554b349cdb50c15a227407d2845787e4c940651b Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Oct 3 21:49:57 2014 -0700 There are no tabs in source code. There is also no trailing whitespace. commit 020a0a48d5fdbfae76b2db26ad5d65c39efad6ad Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Oct 3 21:11:51 2014 -0700 Fix build bustage from c95c9a048. commit c95c9a048f3283afb2e412b10085d4f7c19e1412 Author: xiangmingzhu <xiangzhu@cisco.com> Date: Wed Apr 30 15:48:07 2014 +0800 Cisco optimization for x86 & fixed point 1. Only for fixed point on x86 platform (32bit and 64bit, uses SIMD intrinsics up to SSE4.2) 2. Use "configure --enable-fixed-point --enable-intrinsics" to enable optimization, default is disabled. 3. Official test cases are verified and passed. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> create mode 100644 celt/x86/celt_lpc_sse.c create mode 100644 celt/x86/celt_lpc_sse.h create mode 100644 celt/x86/pitch_sse.c create mode 100644 celt/x86/x86_celt_map.c create mode 100644 celt/x86/x86cpu.c create mode 100644 celt/x86/x86cpu.h create mode 100644 silk/fixed/x86/burg_modified_FIX_sse.c create mode 100644 silk/fixed/x86/prefilter_FIX_sse.c create mode 100644 silk/fixed/x86/vector_ops_FIX_sse.c create mode 100644 silk/x86/NSQ_del_dec_sse.c create mode 100644 silk/x86/NSQ_sse.c create mode 100644 silk/x86/SigProc_FIX_sse.h create mode 100644 silk/x86/VAD_sse.c create mode 100644 silk/x86/VQ_WMat_EC_sse.c create mode 100644 silk/x86/main_sse.h create mode 100644 silk/x86/x86_silk_map.c commit 80460334b77d70e665a390503cd8992cdad06c10 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Sep 30 14:40:13 2014 -0400 -DOPUS_WILL_BE_SLOW to disable the no-optimization warning commit 93fe3ce14f9eedff7edda494351c240721b034ec Author: Ralph Giles <giles@mozilla.com> Date: Mon Sep 29 10:53:17 2014 -0700 Add build script for mp4 spec. This is a quick script to download, rename and sanitize Yusuke Nakamura's draft for opus in the ISO Base Media File Format. create mode 100755 doc/build_isobmff.sh commit d19c314ca1519c640281a993f21f0f083e78c741 Author: Ralph Giles <giles@mozilla.com> Date: Mon Sep 29 10:51:51 2014 -0700 Update mp4 encapsulation spec to v0.4.8. Based on http://vfrmaniac.fushizen.eu/contents/opus_in_isobmff.html commit 6d5ba300e587c5940f208fefa8bf564da73fc9a2 Author: Ralph Giles <giles@mozilla.com> Date: Fri Sep 26 16:56:27 2014 -0700 Clarify that leading '+' is allowed in R128 tags. commit 39c6b8064bff54ae7a58947e72ec9b5c91f9a627 Author: Ralph Giles <giles@mozilla.com> Date: Fri Sep 26 16:34:05 2014 -0700 Remove weasel word. SHOULD or SHOULD NOT. There is no attempt. commit 958285e690bf8be738769a4e63c620563c9dcec0 Author: Ralph Giles <giles@mozilla.com> Date: Fri Sep 26 15:55:44 2014 -0700 Add wording about additional binary data in comment headers. From the codec wg mailing list thread https://www.ietf.org/mail-archive/web/codec/current/msg03107.html commit 8fd7c3fc8cede177aae1c61283d4853cc15a9a65 Author: Ralph Giles <giles@mozilla.com> Date: Fri Sep 26 10:13:05 2014 -0700 Add stylesheet for opus_in_isobmff draft. Based on http://vfrmaniac.fushizen.eu/style.css create mode 100644 doc/opus_in_isobmff.css commit ecd910518a67b46a307494a9b8ba7e06c9ab4508 Author: Ralph Giles <giles@mozilla.com> Date: Fri Sep 26 10:08:54 2014 -0700 Update mp4 encapsulation spec to v0.1.7. From http://vfrmaniac.fushizen.eu/contents/opus_in_isobmff.html commit 3c09b440b4e5140aa1f5c24ee49ac9d766b16b25 Author: Ralph Giles <giles@mozilla.com> Date: Thu Sep 25 14:21:08 2014 -0700 Import mp4 encapsulation spec v0.1.4. Based on http://vfrmaniac.fushizen.eu/contents/opus_in_isobmff.html create mode 100644 doc/opus_in_isobmff.html commit da97db1ca1f92592af3534c9a2596da0e9a009ca Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 4 02:48:21 2014 -0400 Fixes comments commit 648eb9adff253c1216a5430237cc8df30cce6751 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 4 02:44:09 2014 -0400 Takes into account a corner case with the multistream encoder Considers the case where one stream would eat up all the remaining bytes, including a 2-byte size that would then not leave enough for the even a PLC packet in the next streams. commit e1326fe6c8a246aa0b98074040542e3b8ca14c53 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 4 01:48:46 2014 -0400 Lowered the smallest packet that the multi-stream encoder can encode Limit now at 2*streams-1 and anything below that returns OPUS_BUFFER_TOO_SMALL rather than OPUS_BAD_ARG commit 370286cd4aa9c6c9d41690bbb303974321bf81c6 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Sep 3 21:52:37 2014 -0400 updating the update draft commit 741ced30bfbffa73b537bd8bcae40335c5b31991 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Jul 30 14:54:59 2014 -0400 minor language tweak to RTP draft commit 13078454c3b54db526ccfbec63b451127dfbbfc2 Author: Ralph Giles <giles@mozilla.com> Date: Sun Aug 31 13:02:35 2014 -0700 Apply further fixes from Ron. commit 1204ed1bab4b9dc69fe8a121dbd92fbb24dbc6a1 Author: Ron <ron@debian.org> Date: Mon Sep 1 05:09:59 2014 +0930 Fix mixed-case RFC 2119 keywords commit ca7bde672d26df601dfcb32756ddeed841412758 Author: Ralph Giles <giles@mozilla.com> Date: Sun Aug 31 12:20:39 2014 -0700 Fix ambiguous RFC 2119 keywords. These are normative whether capitalized or not. Capitalize the ones which make sense as normative requirements, and reword the rest. commit 0386df9c1b411b547e69d033c4b611d9cab389c7 Author: Ralph Giles <giles@mozilla.com> Date: Sun Aug 31 11:37:34 2014 -0700 Clarify ambiguity between 'streams' and 'files'. We want to general guidelines for stream-level encapsulation to apply to multiplexed as well as degenerate streams. 'Files' can therefore specific to the mine-type discussion for unmultiplexed Ogg physical streams. commit f157736ec5ef430512d08a38151626939d52367c Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Thu Aug 14 10:21:41 2014 -0700 Ogg Opus updates from Ian Nartowicz. These are the changes proposed in <http://www.ietf.org/mail-archive/web/codec/current/msg03058.html>, with some minor additions (see follow-up on the list). commit dacfd28e9abc6bf72f7ddedd3e9fb02803c620f3 Author: Tristan Matthews <le.businessman@gmail.com> Date: Sun Aug 10 13:56:31 2014 -0400 os_support: fix misleading comments cherry-picked from speexdsp 86779a06f6500d041573d6252d4971d3bfcb4b18 commit fd2992aaab93c646c5b5ef54496dd10f70635d08 Author: Ralph Giles <giles@thaumas.net> Date: Fri Aug 8 23:07:36 2014 -0700 Ogg Opus Draft: bump release date, version, and more cleanup. commit 64881eca018712cdb2a594c5e6e929c6c19f7f61 Author: Ralph Giles <giles@thaumas.net> Date: Fri Aug 8 14:16:50 2014 -0700 Ogg Opus Draft: Clean up R128 tag descriptions. Promote the new tag definitions to a separate subsection, balance tags, improve wording, and document our motivation for the R128 tags. commit 1049cada6fb74ad9b20291022e0391fefd044d2b Author: Ralph Giles <giles@thaumas.net> Date: Fri Aug 8 13:46:41 2014 -0700 Ogg Opus Draft: Fix a typo. Thanks to Calvin Walton for pointing this out. commit 2e1b043bf2ca680cba7b44e62564aefaaf0ff483 Author: Ralph Giles <giles@thaumas.net> Date: Fri Aug 8 13:41:25 2014 -0700 Ogg Opus Draft: Add R128_ALBUM_GAIN. Patch from Greg Maxwell in response to implementor feedback. https://www.ietf.org/mail-archive/web/codec/current/msg03053.html commit bb68e117fdb6c5b91c9d2dea0b27493d5e8f8172 Author: Ralph Giles <giles@thaumas.net> Date: Fri Aug 8 13:22:18 2014 -0700 Ogg Opus Draft: apply some clarifications from derf. commit 1e87fea32698ac3070ebf092d2ca08feae57373f Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Jul 25 22:33:55 2014 -0700 RTP draft edits (normative changes). Here are my proposals to resolve a few issues with the current draft. See the corresponding message to the list for details. commit f7faa90b49b8482e2847486e55ed8152cd8b753b Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Fri Jul 25 21:45:46 2014 -0700 RTP draft edits (no normative changes). This is the result of an editing pass for clarity and consistency. commit bb4b6885a139644cf3ac14e7deda9f633ec2d93c Author: Joshua Milas <josh.milas@gmail.com> Date: Sun Jun 22 20:24:48 2014 -0600 Add ~ to the arguments of genversion.bat genversion.bat would fail if an argument was passed with quotations. This is because the quotations were not being stripped before using them as variables, causing the _temp files not being created. Adding ~ to the arguments causes windows to strip the quotations, enabling it to be used with path names containing spaces and the MSVS pre-build event. Signed-off-by: Ralph Giles <giles@mozilla.com> commit 7b8ae17ac1b7cb506cd59f3a22d7cc7e2c487849 Author: Marcello Caramma (mcaramma) <mcaramma@cisco.com> Date: Tue Jul 1 10:41:14 2014 -0400 Using 64-bit acculumation for C0 in silk_burg_modified() Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 6b3d5c8a877e83fdafa6152aa9395c2e70afd480 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jun 30 17:13:16 2014 -0400 RTP draft: avoiding confusion with 2119 language commit f3d6c7a3d174b65e8e9de52d916d421fbd38050d Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jun 30 14:13:46 2014 -0400 Update Opus RTP payload Addresses comments from several months ago commit 8dc91de172b53b59e3ba6453b1ed355348419dea Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Jun 21 01:16:46 2014 -0400 multi-stream fix when setting bitrate to AUTO, MAX, or small values OPUS_AUTO should now work fine, while using OPUS_BITRATE_MAX would still be a bad idea for controlling CBR rate for multistream. The encoder should now throw an error early on if max_data_bytes is just too small to do anything. commit f662fd42eba24a3f3ba4a4f0a0f238516aa44221 Author: Koen Vos <koenvos74@gmail.com> Date: Fri Jun 20 01:42:49 2014 -0400 Fixes CNG overflows from the previous commit Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 9c7193e4f956e230d919f604df27ea4b13b62370 Author: Koen Vos <koenvos74@gmail.com> Date: Thu Jun 19 20:07:55 2014 -0400 Fixes CNG to address the DTX issue reported by Gonzalo Mariano - Bug fix: the CNG synthesis filter output consisted of only the prediction values, now added the input to that. - CNG excitation level now takes into account the level from the PLC (since both PLC and CNG can be active at the same time). - CNG is now called before PLC_glue() to smoothen the transition out of CNG. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 62351fa834571f6c7ab40d7a80458aefa6d7f6ce Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jun 19 16:16:07 2014 -0400 Whitespace fixes commit aa8c5f64d95d5ad0a36814ccee77a45467bda440 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jun 19 04:11:20 2014 -0400 Using Makefile.mips to buidl the MIPS code create mode 100644 Makefile.mips commit f133bac6f9199ca79b6e63dabf1c9839d35d207f Author: Rhishikesh Agashe <Rhishikesh.Agashe@imgtec.com> Date: Thu Jun 19 03:40:09 2014 -0400 MIPS optimizations Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> create mode 100644 celt/mips/celt_mipsr1.h create mode 100644 celt/mips/fixed_generic_mipsr1.h create mode 100644 celt/mips/kiss_fft_mipsr1.h create mode 100644 celt/mips/mdct_mipsr1.h create mode 100644 celt/mips/pitch_mipsr1.h create mode 100644 celt/mips/vq_mipsr1.h create mode 100644 silk/fixed/mips/noise_shape_analysis_FIX_mipsr1.h create mode 100644 silk/fixed/mips/prefilter_FIX_mipsr1.h create mode 100644 silk/fixed/mips/warped_autocorrelation_FIX_mipsr1.h create mode 100644 silk/mips/NSQ_del_dec_mipsr1.h create mode 100644 silk/mips/macros_mipsr1.h create mode 100644 silk/mips/sigproc_fix_mipsr1.h commit a88d8365d42e5369777afa496b43ea88a1aa9106 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Jun 18 17:06:32 2014 -0400 Fixes a bug in silk_sum_sqr_shift() that was causing issues with Burg. When an overflow happened in silk_sum_sqr_shift(), the sample pair causing the overflow would be counted twice. This would lead to the C0 in silk_burg_modified() being too large and cause problems on loud, tonal signals. Thanks to Marcello Caramma for tracking down the problem. commit f396c08f772a5ad4b364a9a2104b4af3678b7428 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Jun 18 00:52:16 2014 -0400 Adds SIG2WORD16() to fixed_debug.h commit 3996642cd731a1b609972ff5d4277f3a387cf12b Author: Ralph Giles <giles@mozilla.com> Date: Wed May 14 09:46:22 2014 -0700 Fix type conversion warning. Use abs(sweep_bps) instead of fabs(). "sweep_bps" is an integer, and on some compilers this line generates a warning. Patch from Tina le Grand. Reviewed by Jean-Marc Valin commit 6d2d5c424d2234b3665c7aa6f0f8a342d48a5ee0 Author: Gregory Maxwell <greg@xiph.org> Date: Wed Apr 16 18:29:26 2014 -0700 Suppress some coverity false positives. commit 4ae4bc61d033bd615bb6ddfe18f28994d02c51bc Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Wed Mar 26 12:58:04 2014 -0700 Fix iOS builds with assembly. The patch in 76e831d917ff got us most of the way there, but out-of-tree builds required a second Makefile.am rule, which was missing @ARM2GNU_PARAMS@. Also, the arm2gnu.pl was terminating argument processing on any argument beginning with --, rather than an argument that was just -- by itself (as is the normal convention in GNU programs). That meant it never saw the --apple flag even when it was passed. Thanks to Jonathan Lennox for the report and for testing. commit 76e831d917fffc55854ba2dc68b61f51c8ba61b7 Author: Martin Storsjo <martin@martin.st> Date: Wed Mar 19 21:36:39 2014 +0200 Make the arm2gnu.pl converter handle apple specific details This allows building the arm assembly for iOS. This checks for the __APPLE__ preprocessor built-in define to determine whether this extra handling should be enabled. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> commit ac0e294b52ecc600c05b4b67f74e0f3ccf30a030 Author: Martin Storsjo <martin@martin.st> Date: Wed Mar 19 21:36:38 2014 +0200 Add separate labels for the start of public functions This avoids having to use the public symbol name when jumping here, on platforms where the public symbols have an underscore prefix. Signed-off-by: Timothy B. Terriberry <tterribe@xiph.org> commit f5a351a3538df30fc93d3cd505545144b50545c4 Author: Bruno Randolf <br1@thinktube.com> Date: Mon Feb 24 16:08:08 2014 -0500 Fixes encoder/decoder classification of ctl() requests. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit e70faf98b0465e4f278c72ce2d7b4ad10221a2f2 Author: Martin Storsjo <martin@martin.st> Date: Sat Feb 8 12:59:18 2014 +0200 arm: Use the UAL syntax for instructions This is required in order to build using the built-in assembler in clang. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit f17623f2f9c7b129ba6b27f2c32fdfed70a5736c Author: Marcello Caramma (mcaramma) <mcaramma@cisco.com> Date: Mon Feb 24 15:06:46 2014 -0500 Fixes an aliasing bug in the MDCT when the frame size isn't a multiple of 4. In that case, the yp0 and yp1 ended up aliasing for the last element, despite being marked as restrict. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 718b10c6282c3ef9d8052e22aee5e853700c2529 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Feb 24 14:55:00 2014 -0500 Adds missing license for arm2gnu.pl commit 4e8acd5ca305f385715a2c36642fdbc91503134f Author: Ralph Giles <giles@mozilla.com> Date: Fri Feb 7 15:41:55 2014 -0800 OggOpus draft updates. Bump version and date for draft-ietf-codec-oggopus-03 submission. Move more text into figure pre/postamble to fix rendering issues in the xml2rfc html output. These need to be manually re-indented in the txt output before submission. :( Fix resampling frequency choice algorithm, which was missing a word. Fix some spelling and make some minor enphasis changes. commit e40105f4db6a45800a24606483c8cdaf2e819a80 Author: maks <maksqwe1@ukr.net> Date: Thu Feb 6 17:32:47 2014 -0500 Fixes internal check_control_input() error checking. The code was wrong, but did not have any impact on user-visible behaviour because all it did was change an internal-only error code. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit e3187444692195957eb66989622c7b1ad8448b06 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jan 31 15:38:27 2014 -0500 Skips comb filter overlap when the parameters didn't change. commit b347696a21ba26113e5a8d356347fdac1a02bced Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jan 30 11:29:35 2014 -0500 Moves opus_packet_get_samples_per_frame() to opus.c commit ad7b28ff19f2074ddd171ddea07222db1414a5f0 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jan 30 11:27:37 2014 -0500 Minor div optimization -- reducing denominators commit d9fb8a6651ea66f18f034fc70a51cfed17d3e12f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jan 30 11:16:24 2014 -0500 Optimizing divisions with a signed numerator commit 51231750178c8a4bf87157d82fb3cd2fb93b50b2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Jan 29 10:16:53 2014 -0500 Speeding up ec_tell_frac() commit 58107d896811f0e975bc1f8f75a5c1c2f6e746b0 Author: Mark Harris <mark.hsj@gmail.com> Date: Sat Jan 25 16:32:18 2014 -0800 Fix surround encoding Broken by 306d7f5a308e30ce10ba1a1dcdbeeb81de3b2872 commit c543ec1bd3d8e42a32132b139d606e2d3e67fb28 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jan 23 00:02:12 2014 -0500 Speeding up extract_collapse_mask() slightly commit 2e6876c07a2771462eff862bc6fcc423f2cd95af Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Jan 22 19:22:05 2014 -0500 Using SSAT in SIG2WORD16() on ARMv6 commit 9b3a3ad7ff3db7f43640be02faa0f869ecff5b99 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 21 16:23:24 2014 -0500 Unrolled version of the comb filter for ARM (eliminates MOVs) commit 6218cba258ae3e29af08253bebe46cfea3c3d39b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 21 15:17:01 2014 -0500 Making decode_pulses() compute the L2-norm on the fly commit 17b197837fb5bf6361e4cae7fbe0d0163e74b8a8 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 21 11:07:57 2014 -0500 Speed up the comb filter on ARM by using MAC16_32_Q16() commit 29354ff6e05c9ead9454981a7404a9b9ea203d2e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 21 10:39:33 2014 -0500 Save more integer divisions on ARM when we know the operands are positive commit ce1173c77fa2ed5e731db2bbd4c7527d57a8c335 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 20 18:45:57 2014 -0500 Fixes use of uninitialized values in dynalloc_analysis() commit cbb75ccb2635d21ab65ce4d85a9779cdeb1163e9 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 14:38:05 2014 -0800 oggopus: use https where possible. commit 5742f10c41310ff9e53e27c56e21a2ef08f7cfe9 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 13:40:36 2014 -0800 Update 'running code' reference, now RFC 6982. commit 24aa86738d25fa97dc6da2765cd88e34d21fe4bb Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 13:39:24 2014 -0800 Use xml.resource.org bibxml in the oggopus draft. These aren't available over https, but actually work, unlike the old reference urls. commit 593680a461d7c186af51c3895b7819cfe8553f20 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 12:52:38 2014 -0800 Abbreviate the first stereo matrix to fit 72 columns. Fixes a nit. commit 728938a08660aa14a421947ca4c819fb85bcd32d Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 12:47:32 2014 -0800 Add NOT RECOMMENDED to our rfc 2119 boilerplate. Fixes a nit. According to idnits, 'NOT RECOMMENDED' is optional, but we use it. Other words, like SHALL which we don't use are required in the boilerplate. commit 26e31c0ce8f967853650ec03fa607faa8af77020 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 12:33:22 2014 -0800 Capitalize Vorbis. commit 864196bff4e88c69566bf5386ba2aca6ad4121d6 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jan 20 12:32:11 2014 -0800 Use organizations authors in the ogg draft. This seems to be the correct way to represent institutional authors in the references. commit ec5d01cbe486ebc29f0c03423f9a253a1655df7f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 20 16:32:16 2014 -0500 Using a table on ARM for unsigned division by small (<=256) integers. Saves 0.6% for 64 kb/s and 1.8% for 128 kb/s when decoding on arm7tdmi. commit 379af35fd4293d34f3b0b5c15576d6d7a61324cb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jan 19 01:42:04 2014 -0500 Minor cleanup in dynalloc_analysis() commit 05548fa1c127bd4f5ae2d1143149e8428560e48b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jan 19 01:31:00 2014 -0500 Adds a median filter to make dynalloc_analysis() more conservative This should prevent extreme dynalloc behaviour in cases where some bands are heavily attenuated. commit b331167821d3cfc8190697c2a1e485c9701f5eac Author: Ron <ron@debian.org> Date: Sun Jan 19 16:03:44 2014 +1030 Don't say 'it' where being specific is better commit deb46d1f69b8a804423933411e32e62ef005c16e Author: Ron <ron@debian.org> Date: Sun Jan 19 15:52:31 2014 +1030 Fix copypasta fail commit e37262c97f0a4eef999ec78e1e192a541838efb7 Author: Ron <ron@debian.org> Date: Sun Jan 19 15:20:05 2014 +1030 Formally introduce the terms mode and configuration and use them more strictly in the rest of the text. commit 3f3cd994242814f0d8ab187d8635a626b0ea17ce Author: Ron <ron@debian.org> Date: Sun Jan 19 13:48:21 2014 +1030 Typo fix commit cf33d324404b8e11c32d789b34ee21fa16786ca8 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 16:16:46 2014 -0800 Use 'pre-skip' consistently with the hyphen. commit 7c67888988b1e48bbcbd3190b9324407e319b0b5 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 16:16:15 2014 -0800 Missed another 'preskip flag'. commit ce72229e337089aa4825271d1b6d609fa738223b Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 16:13:49 2014 -0800 Preskip isn't a flag. Comment from mark4o. commit d9f9e705f2b4debde36d441b60fc3359c4b9b194 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 16:07:59 2014 -0800 Add 'surname' to WAVE reference. commit 785a21fee93105c995c1d38a0892bf9051b89234 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 16:02:52 2014 -0800 Fix markup and note needing to renable-prediction. commit e0220e08f526f9ca41b12d640ac52d26e1ccec05 Author: Ron <ron@debian.org> Date: Sat Jan 18 09:10:18 2014 +1030 Show how to turn off prediction with libopus commit d3f0867395074011a51ce983b41393a20b38c460 Author: Ron <ron@debian.org> Date: Sat Jan 18 10:24:02 2014 +1030 SILK and CELT mode are singular commit 74f6a017aedb5f82fceaf0e57c234406a3b0e614 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 15:47:08 2014 -0800 Try to improve wikipedia citation formatting. It wants a surname, and to show it as Surname, Initial. This leaves an extra comma in the xml2rfc2 output, but at least mentions wikipedia. commit 078775bee5683e3c0a74ab5b5313d11fa2ee0229 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 15:44:08 2014 -0800 Reference vorbis source code for lpc coefficient calculation. commit c42c6db46f79df5ea281c97ea3421740855fc964 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 15:36:03 2014 -0800 Reference wikipedia for LPC prediction. commit b30b2ba21e511f61b718cdd321a21360ffe9dad8 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 15:22:57 2014 -0800 Try to clarify frame packing. Marko was concerned that RFC 6716 section 3.2.1 narrowly describes zero-length DTX frames _only_ for code 2 and 3 packets, and therefore wanted this sentence to state clearly that code 0 and 1 can be used with zero-byte frames as well. I've tried to do that. commit e26ed59ad0bdac36c1b9138a2e9dee1bf2644780 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 14:33:54 2014 -0800 Use SILK/CELT everywhere describing modes. There was some attempt to use LP/MDCT instead, to avoid confusion of the Opus modes with the earlier codecs of the same name, but Jean-Marc says they gave up on doing that in the Opus RFC, and in particular the tables a reader would need to reference from RFC 6716 Section 3.2 mentions SILK and CELT, so I think it's important to use the same terms here. commit af807b28c3cab0ef2ad76c571609ab4dbfbd324a Author: Ron <ron@debian.org> Date: Sat Jan 18 08:35:38 2014 +1030 Drop the oddball hyphenation of MDCT-mode It's not hyphenated anywhere else. commit 64f558557eadbea9125f5e9405fc7e1da5e77fdf Author: Ron <ron@debian.org> Date: Sat Jan 18 08:03:31 2014 +1030 Clarify that MB->WB is better than MB->NB I hope. Now that it's been clarified for me what this originally meant to say, it's tricky to know if this actually contains the missing clue for other readers. Hopefully it should be obvious to anyone actually implementing it once they look at what modes they have to choose from. commit 6d27d40c798ddde41b07fca55c93d807a7cf9eea Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 14:26:58 2014 -0800 Remove line about undefined speaker configurations. Ron's suggestion. We think it's clear enough as is and removing it avoids confusion about this specification versus future extensions. commit 1474e710e6ccc648f774e4c3bff339df9d547162 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 12:10:50 2014 -0800 Bump Ogg draft version and date. commit 3ba1bb01dba72c269c25b4ab91a12d185b293868 Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 12:09:45 2014 -0800 Replace MAY with SHOULD in the gap-filling guidelines. This is really the best implementations can do as far as we understand the possibilies, so it's better to be straightforward. commit 0f9c40499fd58cbb51626818b2ad169c6acde70e Author: Ralph Giles <giles@mozilla.com> Date: Fri Jan 17 11:15:34 2014 -0800 Update Ralph's phone number. commit 49e6c050a38af44cee539599756f992eab21d581 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jan 17 14:05:37 2014 -0500 Update authors in RTP draft commit 283853539d14c1a56f4d39d2aba8d44eabc6a879 Author: Mark Harris <mark.hsj@gmail.com> Date: Mon Jan 13 16:53:21 2014 -0500 Another nit in the update draft Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 221a8b70c7a608ae43a294d223ae5269b85a2e26 Author: Ralph Giles <giles@mozilla.com> Date: Wed Jan 15 09:08:08 2014 -0800 Bump the static version string in version.mk. We failed to do this before the 1.1 release, so that went out still marked at 1.1 beta. The plan was to get rid of this file but it's still used as a fallback by the msvc build. commit bfcc1dd4f4b4f6ea4fe7da9bb54aef78536c3880 Author: Ralph Giles <giles@mozilla.com> Date: Wed Jan 15 09:02:01 2014 -0800 Reword framing bit ommission, again. commit d63b0dd7b4eeafb44b5dabe0578eb83292f771a4 Author: Ron <ron@debian.org> Date: Wed Jan 15 19:41:21 2014 +1030 Be a bit less ambiguous about the comment framing bit commit 1e0b6fd9f7b4dfeec36ea57c1d682f9211250f64 Author: Ralph Giles <giles@mozilla.com> Date: Tue Jan 14 17:23:00 2014 -0800 Rewrite gap filling section. Incorporate list feedback from Mark Harris, Tim and Jean-Marc and try to improve clarity. commit 998e9e00fd01e0eff8cca0e752c48f2286804c13 Author: Ralph Giles <giles@mozilla.com> Date: Tue Jan 14 15:40:16 2014 -0800 Add Tim's gap repair text. From http://www.ietf.org/mail-archive/web/codec/current/msg03030.html commit 4784e0717e48cd10040b29805152e84d125158c4 Author: Ralph Giles <giles@mozilla.com> Date: Tue Jan 14 11:07:37 2014 -0800 Fix malloc_hook warning on glibc 2.17 and later. In glib 2.17 the __malloc_ptr define was removed in favour of using void* directly. Our declaration of mhook using this type for the second argument therefore rightly generates a warning on modern systems, since the type is assumed to be an int, which is too narrow to hold a pointer on 64 bit architectures. Since it was only ever a define we could use an #ifndef to define __malloc_ptr ourselves. However we only use it once, so using void* in the signature directly is cleaner. This should cause no problems on older systems where it will match either the void* or the char* (for non __STDC__ code) the define resolved to. commit 2c7eb787f0ff0632ff81254066b53572860f02f9 Author: Mark Harris <mark.hsj@gmail.com> Date: Mon Jan 13 16:30:55 2014 -0500 Minor nits on update draft Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 0b644be506f44a78228bd8698946e54fa36d7b47 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 13 15:31:01 2014 -0500 Update draft accepted as WG document create mode 100644 doc/draft-ietf-codec-opus-update.xml delete mode 100644 doc/draft-valin-codec-opus-update.xml commit a65db56e540585d02b724b162c9c923c208a7c95 Author: Gregory Maxwell <greg@xiph.org> Date: Wed Jan 8 11:48:38 2014 -0800 Fix declaration after statement in fixed point. commit 5b215d8423785bd1a18a7c497a568eefe1976e0c Author: Gregory Maxwell <greg@xiph.org> Date: Wed Jan 8 11:04:29 2014 -0800 Add tests/test_ to Makefile.unix. commit fe69528133bf4daebbae9260d0934fdc4963a942 Author: Mark Harris <mark.hsj@gmail.com> Date: Tue Jan 7 21:33:03 2014 -0500 Fixes the decode_fec case for b76888d Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit e7750904270bbcad218f05bd7ca4313c6f080fbb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 21:32:41 2014 -0500 pseudostack instrumentation (off by default) commit 0b9f7f994402df32d6b69f9aca3412f0f75bd43a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 21:25:09 2014 -0500 Minor refactoring of the SILK PLC code to save more stack commit b76888dc866868c073e5f94a63ccb80912a5be41 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 21:00:18 2014 -0500 Preventing unnecessary stack use when using a large decode buffer This was causing pseudostack builds to fail because opus_demo uses a 2-second buffer. commit 9134e96cb24fa3a4c5249e3940a163d889fe692d Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 17:50:46 2014 -0500 Fixes SMALL_FOOTPRINT for float commit 81af25950f061aa18339c23ae07846d72a49e3d3 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 16:26:27 2014 -0500 Another aliasing hack in the SILK PLC with SMALL_FOOTPRINT commit e17ca25617009ce0ec7701c08bdb0e367da94ca0 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 15:27:02 2014 -0500 Don't allocate pulses on the stack when calling the SILK PLC. Also minor C89 fix for the previous commit commit b63e7110cb72f48c1319e8848ccf670bcfbe4f63 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 15:02:43 2014 -0500 Moves CELT PLC pitch search to a separate function to reduce peak stack commit c25836205da1f2163064c7d819b3ac1bf48c40a0 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 13:46:06 2014 -0500 Delaying allocation of the SILK temporary output buffer to reduce peak stack commit 5f807c176fd9bc395f5e1cd030dfca3f0b14ba9f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 04:48:42 2014 -0500 Adds SMALL_FOOTPRINT hack to the CELT PLC too commit 9d1b6fef2a8a8cc4a57af6bf8bed2fed645dea80 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jan 7 04:32:41 2014 -0500 Moves deemphasis() call out of celt_decode_lost() to reduce peak stack commit dce69d2b9b4f33ecb482023f0fe1230ef5b53235 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 21:59:48 2014 -0500 Store decoded SILK pulses as 16-bit vector This saves 640 bytes on the peak stack usage. commit ad8371d172e930c28bc66da165bf8498c5d16d15 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 17:45:57 2014 -0500 Cleaning up leftovers of "freq" in celt_decode_with_ec() commit 4d07b1357e1993d7fb69d15f6e5205d7633629f5 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 17:43:20 2014 -0500 Reduces the decoder stack use by removing the pcm_silk buffer in fixed-point We only keep when concealing less than 10ms with SILK. commit 14ca4ed682b1db8a1ee5d858fd647a2955ef2f85 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 09:31:09 2014 -0500 Moves the remains of compute_inv_mdcts() to celt_synthesis() commit 32454dcadc8efe61deb92dc586c69f65cf22e07a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 09:11:52 2014 -0500 Hack that makes the SMALL_FOOTPRINT CELT decoder use only 4.25 kB of stack. commit bdc7b9335875a089ea61023cbe6e772599d9deac Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jan 6 08:58:38 2014 -0500 Reduces decoder stack usage by only storing one channel of denormalized MDCT commit 4a6744a4467c58c5c848f2819cd05edeef3ef5ff Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jan 5 21:40:02 2014 -0500 Some cleaning up of the synthesis code. commit ed01a596dc525801ec9d601da29ded8a82a2ccd6 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Jan 4 21:06:24 2014 -0500 Making exp_rotation1() use MAC16_16(), which saves a few cycles on ARM commit ccec752a384cf6ad110ec141f8bb035d14328eee Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Jan 4 00:49:46 2014 -0500 Silences unused parameter warning commit ef0eac497f762e1dce7f338a573c27aa6a94dea7 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jan 3 23:55:52 2014 -0500 Moving the radix-2 to expose trivial twiddle factors commit c8f62a4aedc1e6fd6ab3091f95df301f38d1488b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 31 00:00:37 2013 -0500 Improving the accuracy of the fixed-point radix-3 and radix-5 commit e1f846208e5d584673fc171009ada085538c7c56 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 29 18:45:49 2013 -0500 Minor cleanup -- nothing to see here commit 05291fd6a699cf384e578368231be928abbc5c52 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 29 18:31:17 2013 -0500 Fixed-point: slight accuracy improvement in the comb filter commit 30f52cbe2d036dc88a16c0e9e4ccf38a2074f808 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 29 16:21:06 2013 -0500 Remove a SAVE_STACK that was pasted accidentally in the previous commit commit e1dc1e22384e6b632d7a8f9d54a0e693ebc77624 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 29 13:34:17 2013 -0500 Unifying scaling of fixed-point and float FFT commit dbb96ab5cc42491167e063cadc90e420c53b08be Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 29 00:09:06 2013 -0500 Fixes C89 issue commit 4c1a90a847a2af528cbfe6924a85ba8173e5c4f9 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 28 23:14:26 2013 -0500 Getting rid of some negations Since we're doing two rotations, we can invert the sign on both. Also adding a few comments for optimizing the FFT. commit cc344fb8ff9649a13d7628c38137f2f65cd65ec4 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 28 19:10:44 2013 -0500 Slightly improving the accuracy of the fixed-point MDCT downscale Also simplifying the code commit e0c00e27d8decde15560b35d1ec5139fceb53a81 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Dec 27 03:16:34 2013 -0500 Commit 99968ab was causing us to allocate too much stack in the MDCT commit e43a0abe0a908603a71b0c35e8c2307a77a7211f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Dec 27 00:10:54 2013 -0500 Removes the separate 1/8N rotation in the (I)MDCT and unmerges the MDCT sizes Undoes commits f7547a4e and 72513f3c commit a5e3c8a6a62208a6853a39ce07e05e829152139e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 23 02:26:03 2013 -0500 Inverse MDCT no longer requires any scratch space commit e2bcb3fe9b64388f39248af9aefc515c875540d9 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 22 02:17:24 2013 -0500 Reverse the ordering of the FFT stage to optimize a degenerate radix-4 case. This also happens to increase the accuracy since it appears that the new ordering is optimal (at least for 20 ms frames), whereas the previous ordering was pessimal. commit c8f4e1608af20d2d29d34176debcd15c0dbec685 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 21 16:30:49 2013 -0500 Merges the FFT scaling with the MDCT pre-rotate commit 153def288491a4e5ecc4939bdd438006b2eda7b3 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 21 15:45:17 2013 -0500 Getting rid of the inverse FFT entirely IMDCT now uses the forward FFT. commit 99968abba8f6d488e8ea331b5c9e31aa61212f93 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 21 14:29:41 2013 -0500 Moving bitrev step to forward MDCT too commit bc13bbaad7deb1c81dfab7a7c0d0bd3bbd9884ef Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 21 02:33:22 2013 -0500 Applying the forward FFT gain up-front for fixed-point too This makes us lose a bit of precision in the first steps, but our gain is more precise because it's only rounded once. Overall, SNR is slightly improved. commit 2e26b82ec24060f61bc553d1f36efad49d97167e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Dec 20 23:13:29 2013 -0500 Moves the bitrev step to the IMDCT pre-rotation commit 8e02afcfff4d1e4ddbf06a5b6878d49d6c3fccb5 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 17 13:56:00 2013 -0500 Fixes a warning about "conversion from '__int64' to 'int'" on MSVC commit 253e15f7e840b8598c36779a6931d79289b7a2bc Author: Peter Korsgaard <peter@korsgaard.com> Date: Sun Dec 8 11:05:41 2013 +0100 configure.ac: fix bashism in ARM optimization handling Breaks configure when /bin/sh isn't bash with: configure: Trying to force-enable ARMv6 media instructions... checking if assembler supports ARMv6 media instructions on ARM... yes configure: Trying to force-enable NEON instructions... checking if assembler supports NEON instructions on ARM... yes ./configure.lineno: 12799: Bad substitution Fix it by using the %% expansion to remove everything from the first space instead. Signed-off-by: Peter Korsgaard <peter@korsgaard.com> Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 306d7f5a308e30ce10ba1a1dcdbeeb81de3b2872 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 16 01:08:21 2013 -0500 fixed-point: slight (but free) accuracy improvement in compute_band_energies() Also moves the VSHR32() condition outside the loop just in case some compilers don't optimize that properly. commit e0f26246b08122cef31c1ac85f7ed228e2daca61 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Dec 14 11:07:13 2013 -0500 fixed-point: adds rounding to some shifts to eliminate bias This reduces the peak decoding error by removing small (inaudible) spikes in the error at the frame boundaries. These were due to the frequency-domain bias ending up as a small pulse in the middle of the IMDCT overlap. None of this was ever audible, but fixing it is still cleaner. commit 4a168eb343af95be4b3a00ef2c5cf40da1d6b7e0 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Dec 11 01:34:06 2013 -0500 Remove useless code in alloc_trim_analysis() commit 5752d659fda9194a95542093498e669d7059d725 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Dec 11 00:21:38 2013 -0500 Minor fixed-point accuracy improvements that were completely free commit 91f8010108a88e7578a177dde0f800bef2e0f00b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 10 18:38:44 2013 -0500 Removing indirections commit 5607d5d1c8f0d7110927bb145754017a84059e3a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 10 18:10:27 2013 -0500 Annotating pointer arguments with OPUS_RESTRICT and const commit 32ada84b5b5a147c38c428fb0863d419dd802b37 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 10 16:41:44 2013 -0500 Fixes code that was doing arithmetic on a void pointer. Thanks to Kevin O'Connor for reporting this. commit 122971b8cc0ab46f0ab85927f109f89606800c3c Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 10 13:55:35 2013 -0500 More NaN hardening in the analysis code commit fa85e3bf2823da90a031cbbb051022f346589da4 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Dec 10 08:16:51 2013 -0800 Add a reminder to bump the version.mk version. We released 1.1 with 1.1-beta as the hard-coded version. Hopefully this only affects people doing their own windows builds without the git tools available. Thanks to Ron Lee for the report. commit d5553e8aca32a7bd25e3bcea3c1fb00c68453225 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 10 02:32:26 2013 -0500 Using OPUS_COPY()/OPUS_CLEAR() in the decoder too commit 15edb78b3e12d2b5c94e2354c91f354c881627de Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 9 21:56:21 2013 -0500 Making NaN detection more robust to -ffast-math. commit 4fda6b01424aab5f6c5b13b1bd01697c899b6fcd Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 9 18:05:50 2013 -0500 Using celt_inner_prod() in compute_band_energies() commit 3039c9bde55eef9b343123a533fc2199c8793f36 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 9 16:31:07 2013 -0500 Makes silk_inner_prod_aligned() use celt_inner_prod() in fixed-point. Should make it easier when writing assembly. commit a9b7def9f5b6e6ee64747572d678735595a897fd Author: Pedro Becerra <pbecerr@gmail.com> Date: Mon Dec 9 16:08:29 2013 -0500 s/MAX16/MAX32/ in transient_analysis() Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 66b7fe0a5cb2777a72b3bec6309f9d43daa51fea Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 9 13:54:41 2013 -0500 Using celt_inner_prod() for input sanity check. commit 57cd849cf71e6abdfedfea1d381d4e06581015d5 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 9 02:33:42 2013 -0500 Defining celt_inner_prod() and using it instead of explicit loops. Also adds an SSE-optimized celt_inner_prod(). commit ff072009fe5bdd3540ac6ac331e9961e83da722a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 8 23:31:30 2013 -0500 Replaces inline copies and initialization with OPUS_*() macros. This is a bit faster at -O2 because memcpy()/memmove()/memset() are vectorized. The code is also cleaner. commit 0f869cba0f271f3d9aee1289d490c44ce1e12975 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 8 03:54:18 2013 -0500 Changes ABS16() and ABS32() to use fabs() in the float build gcc is better at optimizing it than the ?: version commit c94e4bb103e6a989bcd2677c9178ee6ef461912c Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Dec 8 03:31:50 2013 -0500 Optimizes encoder NaN detection and clipping by only running them when needed NaN detection should now be able to catch values that would create NaNs further down. commit 5626908ec3c96a79640b26bee42549c82eb26689 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Dec 5 16:40:59 2013 -0500 Fixed-point fast-path for normal 48 kHz encoding in celt_preemphasis() commit aed1009df97a0e116c276751d83455bbbf47d3fb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Dec 5 13:36:48 2013 -0500 Turns a 16x32 multiply into a 16x16 one in celt_preemphasis(). commit 65471dd567beb6b1156dc292858c0a28ca55ca3e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Dec 4 18:25:34 2013 -0500 Bump soname commit 6fb985515808c161fbc4936d434de4eb8d32c712 Author: Ralph Giles <giles@mozilla.com> Date: Wed Dec 4 15:18:41 2013 -0800 Mention updating the website examples. commit 6dc50f48e7014f841c1a8e13312ed6e249f56908 Author: Ralph Giles <giles@mozilla.com> Date: Wed Dec 4 15:16:12 2013 -0800 Start a release checklist. create mode 100644 doc/release.txt commit 17dcb270267ac20bbf96a27c26549bfac5034d02 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Dec 3 13:31:24 2013 -0500 Revert "Disables assembly optimizations by default because of several problems reported" This reverts commit 2446445b15e0a1961e1566006b34c586fd3162c4. commit 2446445b15e0a1961e1566006b34c586fd3162c4 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 2 22:17:58 2013 -0500 Disables assembly optimizations by default because of several problems reported commit 96408b62958bb7b337311103a5bbb6cd5613cf5a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Dec 2 20:02:37 2013 -0500 Fixes Opus-custom preemphasis Also adds denorm avoidance for non-48 kHz rates commit 086ea7c7e567b68e01ea15561e8b825ecfc78606 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 29 16:47:52 2013 -0500 Whoa! This fixes a pretty bad fixed-point stereo regression commit 35a44c6100648f62955e5389b2ffdcff77fccffc Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 29 12:21:34 2013 -0500 Fixes OPUS_SET_GAIN (was a bug in MULT16_32_P16()) commit c16a96b45886283d23f8bfda5bad115397f21689 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Wed Nov 27 11:06:59 2013 -0800 Minor speedups to celt_pitch_xcorr_edsp(). This moves some stuff out of the main loop for the unaligned case. commit a1740d9ad176ed3716b5c35cba6c47da733b789c Author: Ron <ron@debian.org> Date: Thu Nov 28 04:18:45 2013 +1030 Drop the use of --always to git-describe This is a partial revert of c435f06b308e52d7a5d2aad9eec98a1541832b47, since it didn't actually fix the original problem with shallow clones. It returns just a hash, not something we (currently) can extract as a version string in the form things are expecting. commit a8fa0cb629f076f546748673378694b8416e6c53 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Nov 26 21:56:50 2013 -0800 Remove the celt_pitch_xcorr() test code. Revision control will remember it for us forever. commit e98816e9c10c9e0567f57e06b9b23e515783502c Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Nov 26 21:55:29 2013 -0800 Actually use my NEON code. I'd commented out mine and stuck azanelli's code in for testing, and then accidentally committed it like that. commit 5c02c5ffb58a08c1b705e6fa2ab73f631f658823 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Nov 26 21:51:39 2013 -0800 Make celt_pitch_xcorr_edsp() work on ARMv5TE. We were assuming that LDR, LDRD, and STRD could be used on unaligned addresses, but this turns out not to be true on really old hardware. commit aa8ff201b5f8c489c52f00d83cd188bb168c310b Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Tue Nov 26 17:10:37 2013 -0800 Clean up celt_pitch_xcorr() test code. This will be removed shortly, but at least it should be correct before it goes. Makes the code get used unconditionally (when uncommented), and actually frees the buffer it allocates (now using the libopus allocation functions). commit d6eb9c468f3343768755550c250c975f405aeeaf Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 25 22:33:43 2013 -0500 Silences a warning in fixed-point commit 2924af4db0898e3bdb4572432bf9dcc69ca975c3 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 25 21:24:30 2013 -0500 Merges surround and stereo intensity stereo thresholds again. The different results for surround were mostly due to a bug in squishyball. commit 93e1096093b219e41835866a904d8ebc6a43a0ee Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 25 01:52:08 2013 -0500 Makes SILK surround masking a bit more aggressive in both directions This also splits the rate adjustment between SILK and CELT (previous code only changed SILK) commit a47d6f308ffe7a002cb9d903b2d5466854fa05fb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 24 23:59:09 2013 -0500 Intensity stereo update, partially based on Monty's listening Intensity stereo is now based on equiv_rate and should no longer oscillate in hybrid mode. Also, there is now a separate set of thresholds for surround tuned by Monty. The non-surround thresholds were also updated at low bitrate to be closer to what Monty found for surround. commit f4d2c034a629726cf5978fdcac7f60c0af2e3add Author: Gregory Maxwell <greg@xiph.org> Date: Sun Nov 24 06:49:01 2013 -0800 Put custom_encoder_init decl behind CUSTOM_MODES. More softclip tests. This removes an unused definition warning when compiling without custom modes. commit 1bf1f51c0a2b4e8bef93a1e949d57b33232b6455 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 24 02:01:55 2013 -0500 Using a more reasonably SILK surround calibration value Still needs more tuning commit d94ed205c08bddb5af0b67c09bcb07ca1c27592a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 24 01:59:43 2013 -0500 Reducing the thresholds for enabling stereo and higher audio bandwidths commit ccf7d9b18c39089a7ac363568709168ce2b7f7fa Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 24 01:57:52 2013 -0500 Surround low-passing thresholds set to much lower bitrates commit abf91fe9d019257b536264120e2da69dbc52c4f9 Author: Gregory Maxwell <greg@xiph.org> Date: Fri Nov 22 11:32:32 2013 -0800 Minor opus_pcm_soft_clip API hardening and tests. commit f6ec71925a3d1b7e783c509bc57b8dc66ca43785 Author: Gregory Maxwell <greg@xiph.org> Date: Fri Nov 22 10:30:30 2013 -0800 Add some multistream resets to the tests. commit 21b131fbe52a4e5e5c0376ceb0da627ce04e715f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Nov 21 14:42:00 2013 -0500 Applies padding to all CBR packets, including PLC and CELT silence commit 1bf1da3c962a8c8bf397918c82102c2a157cb6dd Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Nov 21 14:36:41 2013 -0500 Adds trivial_example.c to make dist commit 574247f79c5befa97905a92d477a14c8ba6016dd Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Nov 21 14:33:42 2013 -0500 Adds a simple example of Opus encoding+decoding create mode 100644 doc/trivial_example.c commit 41e8906cd65023ef76492098a0d9ed78a7c1c554 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 20 19:34:14 2013 -0500 opus_custom_encoder_init() now only needed for custom modes. commit 28b49e04b789ec7c8fe7523449be7ba1395470d9 Author: Ralph Giles <giles@mozilla.com> Date: Thu Nov 21 11:31:09 2013 +1300 Add copyright header to test_opus_padding.c. commit 9864c51592078072a84bf8a4776760c850dd875a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 20 17:11:15 2013 -0500 Makes pitch gain control less aggressive commit 854e9bd499dcb368c0a25e2b41ae7ebef0498547 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 20 13:39:23 2013 -0500 Using the maximum frequency response of pitch taps instead of maxabs commit c63fb978be59fd09179abe9d714ee598d888e789 Author: Koen Vos <koenvos74@gmail.com> Date: Wed Nov 20 08:32:32 2013 -0500 Constrains accumulated pitch gain to avoid potential instability. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit ac31a0af4ace081404de39342eed41cf59f91e2d Author: Gregory Maxwell <greg@xiph.org> Date: Tue Nov 19 08:52:49 2013 -0800 Additional opus_multistream_packet_(un)pad tests. commit 4fa2b548d42b00645f2685e8eb4d572cda18bce7 Author: Gregory Maxwell <greg@xiph.org> Date: Tue Nov 19 08:35:26 2013 -0800 Validate merge argument in repacketizer_demo. commit d8302fc093430a342297caae7ee1906bfa6202b0 Author: Gregory Maxwell <greg@xiph.org> Date: Tue Nov 19 08:29:06 2013 -0800 Validate channel count in opus_demo. commit 733b47f31cfa72eb31323dc8b3dd1ba16074a106 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Nov 18 23:25:20 2013 -0800 Add opus_packet_(un)pad tests. commit 135d1c9848cb5114f485c02d969fd1717045dcef Author: Ron <ron@debian.org> Date: Tue Nov 19 17:34:12 2013 +1030 More make dist fixes Don't include the generated celt/arm/celt_pitch_xcorr_arm-gnu.S in the dist tarball, but do include its celt_pitch_xcorr_arm.s source file. commit d5cb06305b6707a51d976cf79ae513c2fe023382 Author: Ralph Giles <giles@mozilla.com> Date: Tue Nov 19 17:50:05 2013 +1300 Fix 'make distclean'. Automake's dependency tracking seems to be confused by our asm generation double-indirection. It's sufficient to have just CLEANFILES = $(CELT_AM_SOURCES_ARM_ASM:.s.in=-gnu.S) to fix the error about 'armopts-gnu.S' not being cleaned, but celt_pitch_xcorr_arm-gnu.S is also not cleaned without raising a warning. Therefore we explicitly clean both sets of ARM_ASM. commit ca5bf5b0496137bc57e3a0dfd0fa829a53736728 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 18 23:00:36 2013 -0500 Don't attenuate hybrid high-band for surround commit 43a287fd758b1e67235f3c1054095749249ff13d Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 18 21:22:30 2013 -0500 Fixes unpad for multi-stream commit 2ce30358ecb646720f8612bef64ed73d4a6d64ca Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Mon Nov 18 20:02:08 2013 -0500 Unpad asserts Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 3d09c5cb35ac3a3050179dbf9c202521d6a51525 Author: Mark Harris <mark.hsj@gmail.com> Date: Mon Nov 18 16:52:22 2013 -0500 Fixes previous commit Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit f18356675295587d41bc7cf2aa173af8c1619e0d Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 18 16:46:38 2013 -0500 Adds functions for multistream padding/unpadding and single-stream unpadding These are all completely untested. commit 7a0b68233f61b48e1980ca3aa1268f91955806e8 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Mon Nov 18 14:10:57 2013 -0500 ASM build fix Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit f50df8224601b3496e416f72514c176ab27e0e2d Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 18 13:49:34 2013 -0500 Fixes permissions on arm2gnu.pl mode change 100644 => 100755 celt/arm/arm2gnu.pl commit 39386e0b85ec0f978aa104d312604badb9047d58 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Mon Nov 18 13:30:13 2013 -0500 Adds Neon assembly for correlation/convolution Optimizing celt_pitch_xcorr()/xcorr_kernel() which also speeds up FIRs, IIRs and auto-correlations Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> create mode 100644 celt/arm/arm2gnu.pl create mode 100644 celt/arm/arm_celt_map.c create mode 100644 celt/arm/armopts.s.in create mode 100644 celt/arm/celt_pitch_xcorr_arm.s create mode 100644 celt/arm/pitch_arm.h commit 530198f955e49571b3f890b4da4d933a4cd5df4e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 17 22:42:42 2013 -0500 Adds SATURATE16() to the fixed-point debug build too commit 6bc3e3cff7c81674d62455a5a55a83998e0dbcf6 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Nov 17 22:35:15 2013 -0500 Fixes SILK surround calibration while fixing other MSVC warnings commit 2632ef0396189e40c6415abad52ccbb774edf53f Author: Gregory Maxwell <greg@xiph.org> Date: Sun Nov 17 08:52:00 2013 -0800 Add some basic testing for OPUS_{GET|SET}_PREDICTION_DISABLED. commit 0c7c4254313f49ad6c409baaee460787652286b4 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Nov 16 13:33:38 2013 -0500 oops, fix previous commit commit 2c37846249fdaa81f781b8a5026afc7d80ec49cb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Nov 16 03:54:58 2013 -0500 Make CELT_PVQ_U_ROW static commit cbe93e23be564195d8bcc9842aeb764c32e3fe48 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 15 13:50:38 2013 -0500 Adds OPUS_SET_PREDICTION_DISABLED() ctl to force "independent" frames Works by turning off pitch and energy prediction in CELT, while setting first_frame_after_reset in SILK to disable pitch and LSF interpolation and reduce LPC gain. commit aad4117d780cee87e1be03e4a108b9e601367dc8 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 15 13:35:04 2013 -0500 Adds a simple padding test to opus_demo (disabled by default) commit 1b28e0cae55f66f76fdd7fae433c392a039565ff Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 15 01:57:25 2013 -0500 Oops, missing semicolon on RESTORE_STACK in previous commit commit e83d2aa3f7e61cdea146f580c99fda7d3593a3ff Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 15 01:52:28 2013 -0500 Adds missing RESTORE_STACKs in celt_encode_with_ec() commit d7aadd808fd2d6f2d8f4a5daf8b3ec32a0968e8e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 15 01:41:12 2013 -0500 Fixes a bug where the encoder was trying to use redundancy in CELT mode The problem was that forcing CELT-mode for low bitrate CBR was done too late, after the encoder had decided to use SILK. This was causing redundancy to be allocated because the encoder didn't realize it was going to keep using CELT. commit c5635d284b489c7705c2357cc4f0601587416b78 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 14:08:22 2013 -0500 Adds packet padding that works for all codes and fixes 40/60 ms CBR. Padding is now handled by the repacketizer. commit 8bbdf5fa0a98d686df6a3bdc5553dbdaa82b1019 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 17:50:39 2013 -0500 Fixes max_redundancy so that hybrid CBR can fill all bytes commit 3a4659a622a8353b2e2815c1dc44f32972b45660 Author: Mark Harris <mark.hsj@gmail.com> Date: Wed Nov 13 19:31:59 2013 -0500 opus_multistream_packet_validate() now called with the total number of streams Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 2dc27dfa13eb4f2b9cd291852b4e9c9508cd9cb2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 19:35:43 2013 -0500 Rename preemphasis() to celt_preemphasis() to avoid amrnb symbol clash http://lists.xiph.org/pipermail/opus/2013-November/002372.html commit ab86a9cca7f891044d64e4ca6c055bc95224d773 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 23:06:25 2013 -0500 Fixes more warnings commit 87ca6c08eae6b5353c451775528d3f4bcf2cc91a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 22:58:10 2013 -0500 Increase surround allocation offset for smaller frame sizes commit 0d584b905a89405b6cfc8c3f8d74e3a5b9976639 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 22:42:04 2013 -0500 Prevents LFE from busting at really low bitrate commit a71c9adf526685024baa46a1e6266264a1aeb866 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 12:07:01 2013 -0500 Fixes MSVC conversion warnings commit 8848171b2fd695ae1db362286db40b5588005c37 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 11:57:31 2013 -0500 Variable frame size fixes (still not exposed in the API) This fixes an actual error in the downmix (using the float version even for the int API), as well as a bunch of conversion warnings. commit 8ea01eed109b3621f3b9e0186ac017119509a5dc Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Nov 13 09:39:05 2013 -0500 Making the CELT fixed-point decoder a bit more robust to extreme signals denormalise_bands() can now produce signals close to the max MDCT amplitude. commit db5b19455f39c521998d521d9d91960ceeebacaa Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Nov 12 14:20:00 2013 -0500 Fixes an initialization issue in SILK prefill found by Coverity commit 1d72d383ac602987882569f362838cd866c5c199 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 11 18:24:56 2013 -0500 oops, don't need RESTORE_STACK when there's no stack commit 77a596365854482a3233335f450148e8e375cd94 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 11 17:39:02 2013 -0500 gcc -pedantic had "comparison of unsigned expression < 0 is always false" commit f6066df2b9f8231ba3d5df27962e05cb658aac1f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 11 13:06:54 2013 -0500 More size-zero VLA fixes and making opus_decode* return BAD_ARG on framesize<0 commit ca6fac041bc6ffe4a8b21eb6ee429dd32d803764 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Nov 9 18:28:40 2013 -0500 Fixes some minor issues found by scan build commit a599ccd2172c159cb690a8b446bb4b3986d594fb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 8 21:47:44 2013 -0500 stupid uninitialized variable in the mlp training caught by cppcheck commit 5174817d3627fe740d3903b341dafa7109d44be8 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 8 19:54:03 2013 -0500 Hides OPUS_FRAMESIZE_VARIABLE from the API until it actually works commit 420231ab1956ba64191ae4b6c46462060048ed52 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 8 19:53:27 2013 -0500 Makes surround bandwidth decision based on the number of channels (duh!) commit c1959e7dc346fac747ca29efa63ff2a6ddb44f44 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Nov 8 19:24:10 2013 -0500 Taking into account the frame size in more encoder decisions Stereo mode, stereo width, min bandwidth, VBR damping commit 9ffce06c773b53e4d509420227df7bd88e0f2e74 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 4 21:16:00 2013 -0500 This should be less confusing for static analyzers Code behaviour is unchanged commit d814c5d2832e721d1740631ae41fc4268445b372 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Nov 4 12:26:50 2013 -0500 Exposes --disable-float-api in autoconf commit 8f466274d3bfc1e14503fef02be1a5ffc6ed6d64 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 28 21:50:10 2013 -0400 Removes a float var that shouldn't have been there Defining FIXED_POINT and DISABLE_FLOAT_API now leaves no float instruction commit c2b34412113a56eefc602782dd59482c083c3173 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 28 21:48:50 2013 -0400 Fixes DISABLE_FLOAT_API build commit 0fb0fd715cc85a7c4d710a94d740a1ece11ad275 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 28 16:41:26 2013 -0400 Moves opus_packet_parse_impl() from opus_decoder.c to opus.c Because it's indirectly used in the encoder (through the repackerizer). commit 811db62ead34b114244892d1d594bb3564211e6e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 28 16:11:53 2013 -0400 Implements OPUS_RESET_STATE for multi-stream encoder commit 60429d366fba6f3176d91af7d1d0a67c15b897d7 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Oct 28 12:55:41 2013 -0700 Avoid a bogus uninitialized warning and simplify some code. commit 1750f568d8ea60b1aabcbc4b6b985fc5f5d5eaf0 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Oct 28 11:40:00 2013 -0700 Remove now-unused check_decoder_option. commit 9c23f5cdca89b9ef7af158d19f436360bb404837 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 28 14:15:18 2013 -0400 Makes stereo savings still less aggressive Caps the savings at 1 bit per (coded) sample. This doesn't really increase the bitrate and fixes some issues, e.g. with IgorC's "twilight" sample. commit 5484a280634011425dbdb5d5d676420a7aaaeaf1 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Oct 28 11:04:57 2013 -0700 Add a little missive when compiling without optimization. The library really depends on all the little fixed point math functions being inlined in order to get acceptable performance. It turns out that it's very easy for someone to compile with optimization disable when twiddling cflags or cooking up their own build system. commit 7830cf1bd2c71bad9aa296254cf5c5f9842e8b8e Author: Gregory Maxwell <greg@xiph.org> Date: Thu Oct 17 15:56:52 2013 -0700 Replace "inline" with OPUS_INLINE. Newer versions of MSVC are unhappy with the strategy of the build environment redefining "inline" (even though they don't support the actual keyword). Instead we define OPUS_INLINE to the right thing in opus_defines.h. This is the same approach we use for restrict. commit 2891d852a38b3acb1810fad9f26ba7e0b181cfb2 Author: Stefan Beller <stefanbeller@googlemail.com> Date: Wed Oct 9 00:01:36 2013 +0200 Remove dead code The compiler warned: silk/float/pitch_analysis_core_FLP.c: In function 'silk_P_Ana_calc_corr_st3': silk/float/pitch_analysis_core_FLP.c:499:36: warning: variable 'basis_ptr' set but not used [-Wunused-but-set-variable] const silk_float *target_ptr, *basis_ptr; This means we can safely remove it. Signed-off-by: Stefan Beller <stefanbeller@googlemail.com> Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit d6b56793d84490e78c91afcb3af96071094b7292 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 21 17:53:48 2013 -0400 Fixes a potential crash when encoding NaNs This fixes tansig_approx() to avoid crashing when the input is NaN. The problem could only be triggered when calling the float API with a float build at a complexity of 7 or more (i.e. analysis called). Since the crash was due to an out-of-bound read (typically the index is INT_MIN), it's unlikely to be exploitable in any other way than causing a crash. commit e8f18c403cdafd012cc90df172ad9f7d6d07b2ab Author: Ralph Giles <giles@mozilla.com> Date: Thu Oct 17 14:52:56 2013 -0700 Remove restrict definition from the unix Makefile. This define was replaces by OPUS_RESTRICT in opus_defines.h in de0b5324b687 and subsequent commits. It is no longer necessary to include it. commit 2e653a7b47b72d3569a78615b0a9d8aa378f7cab Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 14 17:47:18 2013 -0400 Fixes multistream CBR encoding and multistream surround for >20 ms commit c6d0c43df7836e59e34b07ea5c8077718593e708 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 14 17:33:53 2013 -0400 Fixes weird LFE bug The LFE encoder would first decide to switch to SILK, and allocated redundancy bytes. Then, it would force CELT because of LFE, but keep the redundancy allocation, and end up with too few bytes, that would end up being padded to the desired size for CBR. commit eab134c813725976181d6d66608d55597cea1e6f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 14 15:01:36 2013 -0400 Surround encoder can now produce hard CBR streams again. Even when using SILK/hybrid. commit dabdb32ce6d33ab8998defd317e8e2813202b3d1 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 14 13:58:51 2013 -0400 Oops, thanks to Mark Harris for spotting this! commit 58042adc1943a75fd39659d090ed75c896e4d4d5 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Oct 14 13:45:58 2013 -0400 opus_packet_parse_impl() now computes the packet size with padding This should fix decoding of padded multistream packets and (hopefully) multistream fec. commit 2a82908062b9396f8fcf9c020b14a55adc17c583 Author: Mark Harris <mark.hsj@gmail.com> Date: Sun Oct 13 01:23:58 2013 -0400 Rejects bad multistream frame length Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 101c8c874e11f5fc0d52319de62af251a6a3fb81 Author: Mark Harris <mark.hsj@gmail.com> Date: Sun Oct 13 01:12:35 2013 -0400 Fixes PLC for sizes that don't match basic Opus frame sizes. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 98a8baa4ef9def4358a8678fca7819a344ecee6c Author: Mark Harris <mark.hsj@gmail.com> Date: Mon Oct 7 22:25:06 2013 -0700 Make -loss valid for opus_demo -e In addition to its decoding use, it is the value used for OPUS_SET_PACKET_LOSS_PERC. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit b88a3ad39672ffa8ccf5ea73a0bffa1dd76a93aa Author: Mark Harris <mark.hsj@gmail.com> Date: Mon Oct 7 18:50:59 2013 -0700 Fix 40/60ms zero-length frame decode failure Decoding failed with OPUS_BAD_ARG on a packet containing a 40ms or 60ms zero-length frame when it followed a hybrid or MDCT frame. It now invokes the PLC for the duration of the packet as expected. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit dbc83d316c2c221f9790e6162c9942082db8362a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Oct 11 20:41:30 2013 -0400 Cleaning up multistream packet validation commit 970d0529eeaa97f7c06fb37d6bd1b5658ad1e68f Author: Mark Harris <mark.hsj@gmail.com> Date: Fri Oct 11 17:22:00 2013 -0700 Correct opus_packet_parse code 0 packet tests. commit ed4632345ef8edd8ce07403bb4a4996d8c8f440a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Oct 11 18:06:00 2013 -0400 Do up-front validation of multistream packets Prevents the decoder from being out-of-sync on an invalid packet. Also returns OPUS_INVALID_PACKET on a corrupted FEC packet. commit aee4d8057632ea0cfc1d55d88acf8466b47b7b4b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Oct 1 19:46:48 2013 -0400 Removes useless VARDECL() of "out" commit 942fc816d12be9e42025f3b0e473660b4e5e89bf Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Oct 1 19:27:30 2013 -0400 Replaces log2() with celt_log2() and fixes a potential divide-by-zero commit 1ad6f6d557d06289a58abb726f987e5e79688924 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Oct 1 19:25:40 2013 -0400 Fixes scaling of downmix_float() for fixed-point. The previous version simply produced zeros for fixed-point. commit e8e5ecb3f896a72cd484d6c41dbee5679ac0f1b9 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Oct 1 17:16:33 2013 -0400 Fixes missing VARDECL() commit c9e2014f4bd8e247970c6b05b6e43d0c53ed5484 Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 1 14:07:17 2013 -0700 Don't use --dirty in genversion.bat. This script updates version.mk, which is under version control, so the tree is always dirty. Until the script is updated to write package_version instead, incorporating version.mk as a fallback the way update_version does, it's better not to use --dirty. commit c435f06b308e52d7a5d2aad9eec98a1541832b47 Author: Ralph Giles <giles@mozilla.com> Date: Tue Oct 1 13:05:33 2013 -0700 Pass --always to git describe. This generates a useful version string even for --depth=1 checkouts. The option was added in git v1.6.6. commit a32fa319d3e300d80df648e7edb2157b599003ab Author: Stefan Hacker <dd0t@users.sourceforge.net> Date: Fri Sep 20 04:07:53 2013 +0200 Avoid using C++ keyword 'new' Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit 6b9087aa6efbb98731216d2b03cc977826fd6062 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Sep 28 23:51:11 2013 -0400 Avoids unnecessary copying in opus_encode for delayed decision commit 292544453fc89a9a2d14781328f8b1a517e756bc Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Sep 28 19:29:23 2013 -0400 Makes analysis work when encoding more than 20 ms at a time commit a4c2512aa4fbd49142e2026ab5cdeacaa1e1e394 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Sep 28 17:22:41 2013 -0400 run_analysis() doesn't need to return the frame size anymore commit f2446c25c6519bae190152f7a579310b83dc43fd Author: Ralph Giles <giles@mozilla.com> Date: Mon Sep 16 14:40:04 2013 -0700 Remove trailing whitespace from the license headers. commit d517586141200d80bdbe1f1b1fa8eae753219189 Author: Ralph Giles <giles@mozilla.com> Date: Mon Sep 16 14:29:24 2013 -0700 Remove trailing whitespace. commit f5645f5c72fc214d60c297881d771d27af56c4bf Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Sep 16 14:36:53 2013 -0400 Makes the fixed-point build only run the float analysis at complexity 10 Default complexity is now 9. commit b90e63b5bab727bc33dce854b6a4d7e325fef3ab Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Sep 16 13:08:52 2013 -0400 Moves frame size selection back out of opus_encode_native() Should fix delayed decision on surround, but not well tested. commit 36a21ed53636ed4471aa62a8d06dcf175d6ff395 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Sep 14 15:46:09 2013 -0400 Very basic surround rate calibration commit 6fbfed6a0b927f0b54340081675c116955dd3705 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Sep 13 17:16:38 2013 -0400 Allowing surround to use lower bandwidth than fullband commit ae7dc8a13fb61a50204c5b28aac3c1c4c2e1c0dc Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Sep 13 15:05:50 2013 -0400 More surround masking tuning commit 039e9ab155f3527e0e4343d7989c26a4e50a271e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 12 03:05:43 2013 -0400 Tuning the surround masking We're now a bit more conservative and only take into account the bands that are being codec. Also weighting the rate saving by the band width. commit 54bddf0f1a83f37cfeb8519627beaa2422a86418 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Sep 11 23:34:51 2013 -0400 Surround analysis should now work for non-20ms frame sizes. Not well tested. commit 978e4cb507b1fdb90d62e2b15537ebd56087b542 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Sep 11 00:51:22 2013 -0400 Code for handling upsampling in surround_analysis() Still untested. commit 8824fdb1c242b450f981a8ba1f23c4dadeaf6558 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Sep 10 01:15:19 2013 -0400 Adds surround masking to SILK commit 7317de4fbb90407825a0d916d4582abc41dfab55 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Sep 9 17:11:25 2013 -0400 Fixes delayed decision for fixed-point commit ba7dbb365ebb1acff445a017d31e1bb859f0a9fb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Sep 9 16:39:19 2013 -0400 Analysis scaling fixes This should make the scaling the same for fixed and float. It changes the float scaling too by normalizing by the number of channels, which matters for bandwidth detection. commit fc1b1f9b441c2ffe090046c968e5d95f2eafb038 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Sep 6 16:32:50 2013 -0400 Makes speech/music detection work with FIXED_POINT (code still float) commit 3ab03e05561f81f0eacac3ee724baac33b89683b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Sep 6 16:00:39 2013 -0400 First pass at making the analysis code run with FIXED_POINT Code is still float, but at least tonality esitmation seems to work. Speech/music analysis is still disabled. commit 91904a4c91eb19abcebe843045f4599ee67adbee Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 5 21:34:43 2013 -0400 Moves the analysis back to opus_encode_native() to it runs for all streams. I checked that it's not *completely* busted, but it probably needs more testing. commit 0f68696a97ca2e42ec452fef1dc421faa3aa2092 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Sep 5 12:49:55 2013 -0400 Adds trim and dynalloc control to surround masking commit a714ae98a849d03ab9948720a65b9ec57f1c9523 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sat Aug 31 02:05:32 2013 -0400 Makes surround_analysis() work in fixed-point commit fdceae89bff3bbdcb8cb720ef6f6c735da1aae7b Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Aug 30 21:58:02 2013 -0400 Surround masking rewrite The channel energy is now computed using an MDCT in the multi-stream encoder rather than computing the energy of the stereo downmix. commit 02fed471a4568852d6618e041c4f2af0d7730ee2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Aug 29 15:29:02 2013 -0400 Implements fixed-point silk_LPC_analysis_filter() in terms of celt_fir() Saves 2.5% on ARM without any asm. The float build still uses the old code because celt_fir() then becomes a float function. commit 00a1b3e0da9337939f528c71dd4be4bb6872668f Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Aug 28 21:38:15 2013 -0400 Converts the fixed-point silk_burg_modified() to use celt_pitch_xcorr() bit-exact with original version commit 6a7ee7fb55b4a41fca2dea19624600fe42d3b7f7 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Wed Aug 28 17:55:34 2013 -0400 Share auto-correlation code between SILK and CELT commit a156c5ece7133383468d4cba33f067595d9da391 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Aug 26 18:54:39 2013 -0400 Makes the SILK pitch search use celt_pitch_xcorr() Should gives us ARM/SSE optimizations for free. commit 568de0a17b7b8e49fe3fa4d2a3cc4ebd62400cab Author: Ralph Giles <giles@mozilla.com> Date: Wed Aug 21 16:15:57 2013 -0700 Fix more VS2010 header paths. commit 2af12f3c8a682e96432e062587b1d3899b0e3cb9 Author: Ralph Giles <giles@mozilla.com> Date: Wed Aug 21 16:06:30 2013 -0700 Fix relative paths the header file in the VS2010 projects. This prevented opening the files from the 'Solution Explorer' interface where the source listing is displayed. Thanks to Koen Vos for reporting the issue. commit 03e82037170b2226c2c7b60d9280bdf26dad9ebb Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Aug 12 14:47:14 2013 -0400 Fixes a denorm issue in the VoIP mode high-pass filter commit 5771b5a2e5c2c639c2e32d564ff0c71938cdf493 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Aug 2 12:04:50 2013 -0400 Update the RTP draft for resubmit commit 0aa8b42db0b4fd3910cfe49a37d3fa4a40cbf6a4 Author: Timothy B. Terriberry <tterribe@xiph.org> Date: Sun Aug 4 22:47:07 2013 -0700 Fix broken URL in entdec.c. Thanks to thinqa for the report. Fixes #1968. commit e45509294c0f965762cbb8e050157895480a28b7 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Tue Jul 30 05:12:46 2013 -0400 Better handling of the dynalloc cap for CBR/CVBR commit 8259492a7db58cd9c24f514546215098765caa26 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Thu Jul 25 13:52:42 2013 -0400 Fixes an integer wrap-around due to an exceedingly large QCONST16 commit c0c0ef677d64458e0ef34d3ce8dc05e15f4406cc Author: Daniel Jensen <jensend@iname.com> Date: Mon Jul 22 16:31:31 2013 -0600 Add SET_MAX_BANDWIDTH to multistream encoder ctl list Signed-off-by: Gregory Maxwell <greg@xiph.org> commit ecc81ee6299dfe0f96f0d4fed3d24917c786fce2 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jul 15 16:17:02 2013 -0700 Disable MSVC posix security warnings in the project file. By default Visual Studio warns about various C stdlib calls, recommending non-portable replacements instead. We disable this warning in a number of places in the source. Since they're specific to the Visual Studio build, it's better to disable them just in the project files where they always apply to the correct toolchain. I have only added the disable setting to project files which need it currently: opus, test_opus_encode, and test_opus_decode. commit 25eca1c33e240c39e5af4a055c5f00200bf359b4 Author: Ralph Giles <giles@mozilla.com> Date: Mon Jul 15 16:04:14 2013 -0700 Conditionalize #pragma warn for MSVC. This warning disable was conditional on _WIN32, which mingw also defines, resulting in a warning about the unknown pragma on gcc. Instead make it depend on something only the Visual Studio compiler defines. commit 5280c71883e4a2d072d6f69ab7c9b6dc7ace66dc Author: Gregory Maxwell <greg@xiph.org> Date: Mon Jul 15 15:51:24 2013 -0700 Revert "Fixes a number of double promotions and missing casts." This reverts commit 43279728cd39947f95dbf77557a0a698f164fdd1. commit 4effd308828dd6bbbd1ad210c63d6518161b640d Author: Gregory Maxwell <greg@xiph.org> Date: Mon Jul 15 15:51:14 2013 -0700 Revert "Additional MSVC warning missed in the last commit." This reverts commit 715b37f6d65b855c2ac2677ecafef6f9f7290de1. commit 715b37f6d65b855c2ac2677ecafef6f9f7290de1 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Jul 15 15:31:44 2013 -0700 Additional MSVC warning missed in the last commit. commit 43279728cd39947f95dbf77557a0a698f164fdd1 Author: Gregory Maxwell <greg@xiph.org> Date: Mon Jul 15 15:23:03 2013 -0700 Fixes a number of double promotions and missing casts. At a minimum MSVC warns on some of this stuff. commit cfe8e2b86d778da6c3d4521ae6f3bd46311399d8 Author: Ron <ron@debian.org> Date: Sat Jul 13 17:23:04 2013 +0930 Always try to update the version when autogen.sh is run This avoids at least one case where ./autogen.sh && ./configure && make will re-run configure because the makefile rules updated something that it depends upon. Pulling a new version from git will change the version so we should update that at the first step rather than iterating after the last one. commit ae521b0fe56edb19dacad72851eb996f2b2d5ae9 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jul 12 23:52:00 2013 -0400 Update draft for the Opus RFC create mode 100644 doc/draft-valin-codec-opus-update.xml commit b29ad007e5f287bd16c780b778128b3a3bbafd13 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jul 12 22:59:29 2013 -0400 Fixes MSVC warnings commit ea95c533c8c3e8c6ec901fd6d34b8751e2c2243e Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jul 12 10:08:52 2013 -0400 Updates version.mk for 1.1-beta until we can get rid of it commit 260474fb81463420adf6acbcfdcac892a4b7c912 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Fri Jul 12 01:22:09 2013 -0400 Fixes a denorm problem when the input goes silent after active audio commit 1af7f9562be00a5ca9a27859e5b66f0c2d056abc Author: Pedro Becerra <pbecerr@gmail.com> Date: Thu Jul 11 00:00:47 2013 -0400 Fixes build when DISABLE_FLOAT_API is defined Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> commit cd373b54de9e5f8b29510f27d522ff303917dba6 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jul 7 02:51:07 2013 -0400 Surround masking tuning More conservative surround masking Disables temporal VBR on surround Disables tonality on LFE commit 4958751ef4eaccf2624368a50cccd76acf0a6186 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jul 7 02:50:18 2013 -0400 Better handling of centre for surround masking, bitrate tuning Centre channel now contributes sqrt(.5) to both channels Reduces the stream offset at low bitrate commit 2f7a315101754f1becca9c7b253711222af4407a Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jul 7 01:11:23 2013 -0400 Fixes silly fixed-point bug with the surround downmix commit 4365498d48a324e8e7de0b9be90bbb843c933a74 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Sun Jul 7 00:39:35 2013 -0400 Prevents zero-energy on LFE commit 45106b7e4d6ec3d3f13690cafd018ea0dfb601a4 Author: Ron <ron@debian.org> Date: Tue Jul 2 17:01:23 2013 +0930 Link libopus itself with libm too commit 4eb399d7f4d97f5f20081db6196dba04d22113a4 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jul 1 20:19:24 2013 -0400 More speech/music detection comments (delayed decision) commit 0f5ff80bacdde3f68e746c5eeede9590149b6ac2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jul 1 18:39:34 2013 -0400 Commenting the speech/music Markov code commit 92a06f51a7821c4399004ebf9bf4a48173b72a73 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jul 1 16:27:07 2013 -0400 Disables temporal VBR for LFE commit 337f34cf5b6d880ed66a1f1c69e96441693d3fba Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jul 1 16:17:01 2013 -0400 Adds some checks to the surround API commit 8e3a1cbf4848567eb4c5b79d2855a00f87c7b9c2 Author: Jean-Marc Valin <jmvalin@jmvalin.ca> Date: Mon Jul 1 16:12:27 2013 -0400 Temporal VBR tuning Applies temporal VBR up to higher bitrates after IgorC reported it being useful at 64 kb/s for amount=0.1 commit f92286b3dab0b59c984eb12ce32e7f40da3d671a Author: Ron <ron@debian.org> Date: Tue Jul 2 02:35:53 2013 +0930 Get rid of PC_LIBM It only existed to not include -lm in the .pc for fixed point builds, but that is still needed since the float API is still enabled and will use at least lrint. commit a9daf9fec3b43573e164e0486b4efca4272934bf Author: Gregory Maxwell <greg@xiph.org> Date: Sat Jun 29 22:20:54 2013 -0700 Add a couple minor additional decoder tests. commit a0d096f18f13b94ca1d804dfd9e21b42b6bcd702 Author: Gregory Maxwell <greg@xiph.org> Date: Sat Jun 29 20:33:32 2013 -0700 Braces go on the next line. commit b271dae70cca6b6434f717e4a24c97b0d3735084 Author: Gregory Maxwell <greg@xiph.org> Date: Sat Jun 29 20:25:55 2013 -0700 Brace a number of if statements instead of one-lining them. commit dd7b0dac3ba85cda7e314eaa867b4f0d716c9ac4 Author: Gregory Maxwell <greg@xiph.org> Date: Sat Jun 29 20:06:07 2013 -0700 Fixes some return without va_end in the api, adds tests. Also makes the CTL bad argument handling more consistent to avoid mistakes like that in the future. Also updates the variable duration docs. commit 90a39039a46040faa944b4690c16452a088016e2 Author: Ron <ron@debian.org> Date: Sat Jun 29 16:36:24 2013 +0930 Make the parallel test mode of automake 1.13 useable commit 64b398573da586e67ef266ece6e8ef147c2e0b66 Author: Ron <ron@debian.org> Date: Sat Jun 29 15:40:27 2013 +0930 Replace INCLUDES with AM_CPPFLAGS Automake 1.13 complains that INCLUDES is obsoleted now. commit 5bac4f26cbf1745ad6564c6ac43bc653e8772a45 Author: Ron <ron@debian.org> Date: Sat Jun 29 02:06:53 2013 +0930 More autoconf housekeeping Drop the test for getopt, it's not used anywhere anymore. Switch the last uses of AC_TRY_COMPILE to AC_COMPILE_IFELSE now. The former is marked as obsolete, and this will leave no confusion about which to cut and paste if new tests are added. Double quote all the parameters to AC_LANG_SOURCE and AC_LANG_PROGRAM. This is actually required, even if you can get away with not doing it sometimes, so again set a good example for future changes to follow, to hopefully avoid people getting bitten harder than they need to be. Don't bother checking for alloca if we're never going to use it (ie. if we have C99 variable-size array support). The test for this is a bit sketchy anyway ... we separately test for HAVE_ALLOCA_H and USE_ALLOCA, but the test for USE_ALLOCA depends upon having alloca.h present, yet the use of these macros in stack_alloc.h only tests for HAVE_ALLOCA_H in ... |