Age | Commit message (Collapse) | Author | |
---|---|---|---|
2016-03-18 | fix for divide by zerosilk_tunings | Felicia Lim | |
2016-02-21 | set back to floating point by default | Koen Vos | |
2016-02-21 | Clean up: alignment of comments | Koen Vos | |
2016-02-21 | Clean up: replace tabs by spaces | Koen Vos | |
2016-02-21 | take advantage of more efficient NLSF quantization by reducing number of ↵ | Koen Vos | |
survivors in search | |||
2016-02-21 | NLSF_del_dec_quant minor clean up | Koen Vos | |
2016-02-21 | NLSF_encode now also uses NLSF weights from table | Koen Vos | |
2016-02-21 | NLSF VQ now uses absolute error, and prediction | Koen Vos | |
2016-02-21 | NLSF decoding now uses tables for NLSF weights | Koen Vos | |
2016-02-21 | removed prefilter | Koen Vos | |
2016-02-21 | restores decoder bit exactness | Koen Vos | |
2016-02-20 | slight clean up | Koen Vos | |
2016-02-19 | undo some loop unrolling | Koen Vos | |
2016-02-19 | fixes bug in sum_sqr_shift where overflow could happen for the last sample ↵ | Koen Vos | |
of an odd-length frame | |||
2016-02-19 | undo experimental setting from previous commit | Koen Vos | |
2016-02-19 | pitch estimator: added saturation to low-pass filter and unrolled loop. also ↵ | Koen Vos | |
fixes read-beyond-array bug in previous commit | |||
2016-02-19 | pitch analysis (in FIX) now scales the input down at the start, instead of ↵ | Koen Vos | |
at every stage | |||
2016-02-19 | minor clean up of pointer offsets | Koen Vos | |
2016-02-19 | replaced 64-bit correlations with scaled 32-bit ones, in Burg | Koen Vos | |
2016-02-19 | sum_sqr_shift: reduced headroom (ie more preserved resolution); shift ↵ | Koen Vos | |
increments by one instead of two | |||
2016-02-18 | simplified computation of LTP coefs | Koen Vos | |
2016-01-13 | faster Burg implementationexp_lbr_tune4 | Koen Vos | |
2016-01-11 | Quality: Retrained classifier MLP with better data | Jean-Marc Valin | |
New MLP doesn't attempt to classify silence as speech/music | |||
2016-01-11 | Using "hybrid" flag instead of "start!=0" | Jean-Marc Valin | |
2016-01-11 | Quality: Making the encoder more aggressive at using wider bandwidths | Jean-Marc Valin | |
2016-01-11 | Quality: Removes high-band attenuation in hybrid mode -- at least for now | Jean-Marc Valin | |
2016-01-11 | Quality: Adds SILK rate interpolation function | Jean-Marc Valin | |
2016-01-11 | Quality: Makes real CELT VBR work for hybrid too | Jean-Marc Valin | |
2016-01-11 | Quality: Forces trim to 5 on hybrid mode | Jean-Marc Valin | |
This saves bits and makes more sense since alloc_trim_analysis() mostly looks at the lower bands that are coded with SILK | |||
2016-01-11 | Quality: Encoding music in full-band stereo from 32 kb/s | Jean-Marc Valin | |
Gradual downmix now happens in the 24 kb/s to 32 kb/s range | |||
2016-01-11 | Quality: removes VBR attenuation at low bitrate | Jean-Marc Valin | |
Turns out that even low bitrates benefit from VBR | |||
2016-01-04 | Add a check target to the static makefiles. r=ron | Ralph Giles | |
2016-01-04 | Add 'force' to the phony target list. r=ron | Ralph Giles | |
2015-12-31 | Renamed MAX_PULSES to SILK_MAX_PULSES and CELT_MAX_PULSES | Jean-Marc Valin | |
Avoids accidental name collisions | |||
2015-12-31 | Actually fix pedantic build bustage. | Timothy B. Terriberry | |
2015-12-31 | Fix pedantic build bustage. | Timothy B. Terriberry | |
ISO C90 forbids mixed declarations and code, r=bustage | |||
2015-12-31 | Add tonality_analysis_reset(). | Ralph Giles | |
Add a reset function for the TonalityAnalysisState struct and call it on encoder reset. Move the state struct above the clear line in OpusEncoder so reset doesn't clobber reusable fields. Currently this is only the arch field, which is moved to to top of the struct so we can use the same memset-to-the-end pattern as OpusEncoder. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | |||
2015-12-31 | Add an initializer for TonalityAnalysisState. | Ralph Giles | |
This interns the asm flags parameter in the state struct so we don't need to pass it with every call. It can be expensive, so we don't want to query every run_analysis() call, but since this (private) api is used by webrtc code we need to provide a supportable interface for filling in the correct value. Note the initialization code is partially duplicated between opus_encoder_init and the OPUS_RESET_STATE switch case, so we must re-initialize it there. Signed-off-by: Jean-Marc Valin <jmvalin@jmvalin.ca> | |||
2015-12-30 | Remove unused encoder field analysis_offset | Mark Harris | |
2015-12-30 | Avoids having the Opus-level "arch" be set to 0 on encoder/decoder reset | Jean-Marc Valin | |
2015-12-28 | oggopus: More updates for AD review comments. | Timothy B. Terriberry | |
Removed 2119 language for general Ogg requirements. Added IANA registry for channel mapping families. Adjusted additional copyright grant to match RFC 6716. Additional comments addressed (see the CODEC mailing list). | |||
2015-12-24 | Fixes patch_transient_decision() for hybrid mode | Jean-Marc Valin | |
...and also make it not ignore the right channel | |||
2015-12-24 | Oops, fixes max_decay for LFE in fixed-point | Jean-Marc Valin | |
2015-12-12 | oggopus: Revert SHOULD->MUST for reserved mappings. | Timothy B. Terriberry | |
Mark Harris convinced me that the significant delay between "WG consensus" and "RFC" means we shouldn't rely on RFC updates to give people permission to start deploying new things. | |||
2015-12-11 | oggopus: Fix typo. | Timothy B. Terriberry | |
Thanks to Mark Harris for the report. | |||
2015-12-11 | oggopus: First pass updates for AD review comments. | Timothy B. Terriberry | |
2015-12-05 | Make the CELT background noise estimator adapt more quickly on DTX update | Jean-Marc Valin | |
2015-12-04 | opus_demo: fixes potential division by zero on files with no active frame | Jean-Marc Valin | |
2015-12-04 | Fixes memory leak in some CELT unit tests | Jean-Marc Valin | |
Reported in https://trac.xiph.org/ticket/2241 | |||
2015-12-04 | Always use the max of the decaying PLC energy and the background noise | Jean-Marc Valin | |
Should no longer cause discontinuities in the noise after 5 packets |