Fixes a seed issue introduced in 5367dac
Remove condition in pre/de-emphasis when not building with custom modes
seed and remaining_bits moved to band context
Using a band context to reduce the number of arguments being passed around.
more bands.c cleaning up
Removes useless parameters and re-indents the code
Splitting off the recursion in quant_partition()
stereo part of quant_band() moved to quant_band_stereo()
Moves the split angle code to compute_theta()
First step in quant_band() cleanup: N=1 case.
Fixes MSVC warnings for double->float and float->int conversion
Add float literal suffixes to mlp_data.c.
Add float suffix to tansig_table literals.
Add tansig_table.h to the MSVC project files.
missing comment in win32/config.h
Only use automatic bandwidth detection when the user doesn't force bandwidth
Fixes bandwidth detection by not relying on uninitialized data
skip bands with no energy
Capping lsb_depth to 16 unless using the float API of a float build
New bandwidth detection code
Update ogg draft based on comments from Tina le Grand.
Update the oggopus script to use xml2rfc v2.
Minor tuining of complexity thresholds
Cleanup: tf_estimate now ranges from 0 to 1
Dynalloc rate calibration
Makes "stereo savings" less aggressive on mono or near-mono.
Makes tonality boost less aggressive
Disables the new spread+tapset estimator
Document that opus_packet_get_nb_frames, etc. can return OPUS_BAD_ARG.
Numerous PLC cleanups.
Remove the obsolete CELT_BUILD define from win32/config.h.
Add tests for GET_LAST_PACKET_DURATION.
Fixes OPUS_GET_LAST_PACKET_DURATION
Add missing packet_get_nb_samples test coverage and fix test output.
Bump soname version to match the 1.0.x branch.
Link to the new Microsoft patent license.
Adds assert to catch bug from previous commit
Fixes a multi-frame FEC issue that was caught by valgrind
Adds explicit valgrind checks when ENABLE_VALGRIND is defined
Implements opus_packet_get_nb_samples()
Implements OPUS_GET_LAST_FRAME_DURATION decoder ctl()
Don't update the internal decoder state until we know the packet is valid
Changes the PLC behaviour and fixes the FEC behaviour on concatenated packets
Remove MSVC incremental linking directives.
Do imdct post-rotate and deshuffle in-place.
Fixes forward MDCT for overlaps that aren't a multiple of 4
Fix packet length handling for 16-bit machines (makes no difference on 32-bit)
More comments in the PLC
Place output of imdct post-rotate and deshuffle in out buffer to avoid copying and simplify the code.
Window both sides of overlap when adding them.
Add a regression test for the padding issue.
Fixes an out-of-bounds read issue with the padding handling code
RTP draft updates addressing the rest of Tina's comments
Better comments in the PLC
Remove obsolete check for endianness.
Fixes an overflow in the fixed-point celt_sqrt() for large values.
Fixes many many issues in the CELT PLC
Addressing some of Tina's comments on the RTP draft
Improve the !OPUS_BUILD #error.
Allow the build files to override OPUS_EXPORT.
Remove obsolete #ifdef OPUS_BUILD.
Guard _BitScanReverse on MSVC so that MSVC 6 doesn't break.
More minor updates to RTP draft
Misc updates to the RTP draft
Reword 239e9a33 for better clarity.
Add support for RFC5576 source-level attributes.
Update Ogg draft to make it a WG item.
Tweaks to the new transient detector
Enable maintainer mode by default on automake 1.11
Follow up to commit 8b906c102b37ce74571898c8879c01b2a9088fb3
Fix the draft build.
Should make celt_maxabs*() slightly faster
RTP draft: Using sampling rates instead of bandwidths
Oops, fixes transient_analysis() for the right channel
Fixes a VBR bug caused by maxDepth not to be computed without dynalloc
Fix line endings on MSVC project files.
Add the new source files to the MSVC projects.
Better description of the the stereo signalling behaviour
Remove long-outdated comment
Fixes a quant_coarse_energy() hybrid mode bug
Fixes a bug where transient frames would use previous spreading
Disables tf_analysis() for hybrid mode
Fixes issues with multiple files defining CELT_C
Splits out the CELT encoder and decoder
Splits out the Opus multi-stream encoder and decoder
Changed denormalize_bands() to only work between "start" and "end"
Reduces stack in quant_all_bands() for hybrid mode
More stack usage reduction in quant_all_bands()
Removes the need for lowband_scratch in quant_all_bands()
Moved up the second encoder mdct to reduce stack usage
Reduces decoder stack usage
Further cleanup of the MDCT code, fixes PLC bug
Avoid copying imdct output