Age | Commit message (Collapse) | Author | |
---|---|---|---|
2022-07-22 | Using saturating round to fix some wrap-arounds | Jean-Marc Valin | |
Reviewed by Mark Harris | |||
2021-06-10 | Revert relaxing comparison to 0 for fixed point only | Felicia Lim | |
2021-06-08 | Relax comparison to 0 to avoid a floating point divide-by-zero error. | Felicia Lim | |
2021-03-03 | celt_lpc: avoid overflows when computing lpcs in fixed point | Felicia Lim | |
The LPCs are computed in 32-bit, so increase the allowed range from +/-8 to +/-64 to avoid overflows caught during fuzzing. Before downshifting back down to the +/-8 range in the final 16-bit output, perform bandwidth extension to avoid any additional overflow issues. | |||
2018-07-26 | Fix comma that should have been a semicolon | Jean-Marc Valin | |
Didn't cause any problem, but still good to fix. Reported by keithchugg on github. | |||
2017-10-07 | Fixes CELT PLC | Jean-Marc Valin | |
The regression was introduced in 783ad76766e1f6b6aaca5d6eb415ac8a8269e1f2 by changes to celt_fir() that make it no longer in-place. | |||
2017-02-19 | Fix sum initializer in celt_fir() | Jean-Marc Valin | |
2017-02-18 | Remove SMALL_FOOTPRINT code for celt_fir() | Jean-Marc Valin | |
Wasn't worth it given the small code size of the alternative, which also got refactored a little (still bit exact). | |||
2017-01-18 | Revise celt_fir_c() to not pass in argument "mem" | Linfeng Zhang | |
The "mem" in celt_fir_c() either is contained in the head of input "x" in reverse order already, or can be easily attached to the head of "x" before calling the function. Removing argument "mem" can eliminate the redundant buffer copies inside. Update celt_fir_sse4_1() accordingly. | |||
2016-07-22 | fixes overflows in celt_iir() by doing proper saturation | Jean-Marc Valin | |
2016-06-20 | Fixes minor code quality issues in CELT | Jean-Marc Valin | |
Reported by Durandal. | |||
2014-10-04 | Fix build bustage from c95c9a048. | Timothy B. Terriberry | |
2014-10-04 | Cisco optimization for x86 & fixed point | xiangmingzhu | |
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> | |||
2013-11-18 | Adds Neon assembly for correlation/convolution | Timothy B. Terriberry | |
Optimizing celt_pitch_xcorr()/xcorr_kernel() which also speeds up FIRs, IIRs and auto-correlations Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | |||
2013-08-29 | Implements fixed-point silk_LPC_analysis_filter() in terms of celt_fir() | Jean-Marc Valin | |
Saves 2.5% on ARM without any asm. The float build still uses the old code because celt_fir() then becomes a float function. | |||
2013-08-29 | Share auto-correlation code between SILK and CELT | Jean-Marc Valin | |
2013-06-17 | Avoids symbol clashes with Speex (pitch_xcorr) and libm (y1) | Jean-Marc Valin | |
2013-06-07 | Improved SSE version of xcorr_kernel() | John Ridges | |
The loop no longer reads past its buffer and is slightly faster. Also fixes RESTORE_STACK in celt_iir(). | |||
2013-05-27 | C89 fix | Jean-Marc Valin | |
2013-05-27 | Creates xcorr_kernel() that gets used by pitch_xcorr, celt_fir and celt_iir. | Jean-Marc Valin | |
2013-05-26 | Adds missing RESTORE_STACK calls | Jean-Marc Valin | |
2013-05-25 | Speeds up celt_iir() by more than a factor of two. | Jean-Marc Valin | |
Again, this only impacts the PLC and we assume the order is a multiple of 4. | |||
2013-05-25 | Speeds up celt_fir by more than a factor of two. | Jean-Marc Valin | |
Only impacts the PLC. We now assume that the order is a multiple of 4. | |||
2013-05-25 | Optimizes _celt_autocorr() by using pitch_xcorr() | Jean-Marc Valin | |
Computes most of the auto-correlation by reusing pitch_xcorr(). We only need lag*(lag-1)/2 MACs to complete the calculations. To do this, pitch_xcorr() was modified so that it no longer truncates the length to a multiple of 4. Also, the xcorr didn't need the floor at -1. As a side benefit, this speeds up the PLC, which uses a higher order LPC filter. | |||
2013-05-24 | Remove an unused variable added in 85ede2c6. | Timothy B. Terriberry | |
Thanks to John Ridges for the report. | |||
2013-05-23 | Use more MAC16_16's and unroll a loop. | Timothy B. Terriberry | |
This splits out the non-arch-specific portions of a patch written by Aurélien Zanelli <aurelien.zanelli@parrot.com http://lists.xiph.org/pipermail/opus/2013-May/002088.html I also added support for odd n, for custom modes. 0.25% speedup on 96 kbps stereo encode+decode on a Cortex A8. | |||
2012-04-21 | s/FOUNDATION/COPYRIGHT OWNER/ in CELT code and "glue code" | Jean-Marc Valin | |
Also added 3rd clause to "master" COPYING file | |||
2011-11-29 | Some minor (non-bitstream-affecting) changes to help us have better test vectors | Jean-Marc Valin | |
These fix corner cases discovered during the latest fuzzing tests. | |||
2011-10-04 | Renaming the CELT plc.c file to celt_lpc.c to avoid MSVC conflicts with ↵ | Jean-Marc Valin | |
Skype's PLC.c | |||
2011-04-29 | Moved all SILK source code to the silk/ directory | Jean-Marc Valin | |
2011-04-25 | Loss rate adaptation for the CELT layer | Jean-Marc Valin | |
2011-03-31 | draft update | Jean-Marc Valin | |
2011-03-23 | Build fixes | Jean-Marc Valin | |
2011-03-21 | Making mode switching use the same window as CELT (squared) | Jean-Marc Valin | |
2011-03-18 | CELT update | Jean-Marc Valin | |
With minor fixes | |||
2011-03-18 | More project files update | Jean-Marc Valin | |
2011-03-17 | Fixes a few PLC/DTX bugs due to the recent decode API change | Jean-Marc Valin | |
2011-03-14 | Adding constrained VBR mode | Jean-Marc Valin | |
2011-03-14 | Draft update (allocation | Gregory Maxwell | |
2011-03-12 | Disable newly introduced CELT signalling | Jean-Marc Valin | |
2011-03-10 | Nothing to see here | Jean-Marc Valin | |
2011-03-09 | CELT update | Jean-Marc Valin | |
2011-03-08 | Support for glitchles mode switching | Jean-Marc Valin | |
Uses a 5ms redundant CELT frame embedded into the SILK or hybrid packet to handle the switching. It's still possible to use the PLC-based method when no redundant packet is included. | |||
2011-03-01 | SILK/CELT update | Jean-Marc Valin | |
2011-02-15 | preparing the next version | Jean-Marc Valin | |
2011-02-15 | missing files | Jean-Marc Valin | |
2011-02-15 | SILK and CELT updates | Jean-Marc Valin | |
2011-02-14 | Update for in-band FEC | Koen Vos | |
2011-02-04 | draft date | Jean-Marc Valin | |
2011-02-04 | CELT update | Jean-Marc Valin | |