2003771 - Update opus to new version d38db32cffb00460dc52543ba3e07ab16c2bf380 from 2025-11-28 11:38:54

Status: NEW

(Core :: Audio/Video: Playback, enhancement)

Description

[Update Bot]

This update covers 157 commits, including 1 new upstream commit I've never filed a bug on before. (It's the top one.). Here are the overall diff statistics, and then the commit information.

---

media/libopus/arm2gnu.py | 9 -
media/libopus/celt/_kiss_fft_guts.h | 2 +-
media/libopus/celt/arch.h | 15 +-
media/libopus/celt/arm/celt_neon_intr.c | 3 +-
media/libopus/celt/bands.c | 467 +-
media/libopus/celt/bands.h | 4 +-
media/libopus/celt/celt.c | 35 +-
media/libopus/celt/celt.h | 69 +-
media/libopus/celt/celt_decoder.c | 369 +-
media/libopus/celt/celt_encoder.c | 458 +-
media/libopus/celt/celt_lpc.c | 18 +-
media/libopus/celt/celt_lpc.h | 6 +-
media/libopus/celt/cwrs.c | 34 +-
media/libopus/celt/entenc.c | 2 +-
media/libopus/celt/fixed_generic.h | 4 +-
media/libopus/celt/kiss_fft.c | 2 +
media/libopus/celt/mathops.c | 116 +-
media/libopus/celt/mathops.h | 165 +-
media/libopus/celt/mdct.c | 2 +-
media/libopus/celt/mips/celt_mipsr1.h | 72 +-
media/libopus/celt/mips/fixed_generic_mipsr1.h | 106 +-
media/libopus/celt/mips/kiss_fft_mipsr1.h | 120 +-
media/libopus/celt/mips/mdct_mipsr1.h | 48 +-
media/libopus/celt/mips/pitch_mipsr1.h | 328 +-
media/libopus/celt/mips/vq_mipsr1.h | 114 -
media/libopus/celt/modes.c | 74 +-
media/libopus/celt/modes.h | 10 +
media/libopus/celt/pitch.c | 41 +-
media/libopus/celt/pitch.h | 4 +-
media/libopus/celt/quant_bands.c | 46 +-
media/libopus/celt/quant_bands.h | 4 +-
media/libopus/celt/rate.c | 233 +-
media/libopus/celt/rate.h | 3 +
media/libopus/celt/static_modes_fixed.h | 3417 +++++++++-
media/libopus/celt/static_modes_float.h | 2420 +++++-
media/libopus/celt/vq.c | 497 +-
media/libopus/celt/vq.h | 25 +-
media/libopus/celt/x86/celt_lpc_sse4_1.c | 4 +-
media/libopus/celt/x86/pitch_sse2.c | 12 +-
media/libopus/celt/x86/pitch_sse4_1.c | 26 +-
media/libopus/include/opus.h | 3 +
media/libopus/include/opus_custom.h | 4 +-
media/libopus/include/opus_defines.h | 134 +-
media/libopus/include/opus_multistream.h | 8 +-
media/libopus/include/opus_projection.h | 6 +-
media/libopus/moz.build | 2 +-
media/libopus/moz.yaml | 4 +-
media/libopus/silk/API.h | 4 +-
media/libopus/silk/NSQ.h | 4 +
media/libopus/silk/NSQ_del_dec.c | 3 -
media/libopus/silk/SigProc_FIX.h | 6 +-
media/libopus/silk/check_control_input.c | 3 +
media/libopus/silk/define.h | 4 +
media/libopus/silk/enc_API.c | 25 +-
media/libopus/silk/fixed/mips/noise_shape_analysis_FIX_mipsr1.h | 336 -
media/libopus/silk/fixed/mips/warped_autocorrelation_FIX_mipsr1.h | 70 +-
media/libopus/silk/fixed/noise_shape_analysis_FIX.c | 7 -
media/libopus/silk/fixed/structs_FIX.h | 3 +-
media/libopus/silk/fixed/warped_autocorrelation_FIX.c | 2 +-
media/libopus/silk/fixed/x86/burg_modified_FIX_sse4_1.c | 18 +-
media/libopus/silk/fixed/x86/vector_ops_FIX_sse4_1.c | 2 +-
media/libopus/silk/float/structs_FLP.h | 3 +-
media/libopus/silk/macros.h | 2 +-
media/libopus/silk/mips/NSQ_del_dec_mipsr1.h | 410 -
media/libopus/silk/mips/NSQ_mips.h | 137 +
media/libopus/silk/mips/macros_mipsr1.h | 36 +-
media/libopus/silk/resampler.c | 29 +-
media/libopus/silk/resampler_structs.h | 2 +-
media/libopus/silk/typedef.h | 2 +-
media/libopus/silk/x86/NSQ_del_dec_avx2.c | 14 +-
media/libopus/src/analysis.c | 1 +
media/libopus/src/extensions.c | 6 -
media/libopus/src/opus.c | 7 +
media/libopus/src/opus_decoder.c | 87 +-
media/libopus/src/opus_encoder.c | 453 +-
media/libopus/src/opus_multistream_encoder.c | 107 +-
media/libopus/src/opus_private.h | 9 +-
77 files changed, 8866 insertions(+), 2471 deletions(-)

---

d38db32cffb00460dc52543ba3e07ab16c2bf380 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/d38db32cffb00460dc52543ba3e07ab16c2bf380
Authored: 2025-11-28 09:47:50 -0800
Committed: 2025-11-28 11:38:54 -0800

opus_demo, opus_custom_demo: Avoid integer overflow

---

57a2042d359e9d83f9bf0eea421d133e9f709083 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/57a2042d359e9d83f9bf0eea421d133e9f709083
Authored: 2025-11-24 21:10:54 -0800
Committed: 2025-11-26 07:43:03 -0800

dnn/download_model.sh: Try shasum if no sha256sum

---

226d8a5826a665918d56cb995f99a12f3db25a61 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/226d8a5826a665918d56cb995f99a12f3db25a61
Authored: 2025-11-25 18:58:23 -0800
Committed: 2025-11-26 07:30:36 -0800

test_opus_extensions.c: Fix 32-bit compile errors

---

2feff4660f0371e70bd932c35cb5ab74641b15b5 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2feff4660f0371e70bd932c35cb5ab74641b15b5
Authored: 2025-11-24 20:46:59 -0800
Committed: 2025-11-24 21:16:42 -0800

configure.ac: Fix autoconf warnings

---

9f791857ea74013908914a0ca20e57fc3b48165d by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9f791857ea74013908914a0ca20e57fc3b48165d
Authored: 2025-11-24 20:39:27 -0800
Committed: 2025-11-24 21:14:20 -0800

configure.ac: Make NE10 help text consistent

---

6f00d6d4b29b9a022d12585f6439d56784aeae61 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6f00d6d4b29b9a022d12585f6439d56784aeae61
Authored: 2025-11-21 20:35:33 -0500
Committed: 2025-11-21 20:35:33 -0500

Avoid NaNs in the analysis code

---

50c3738d0b416b4ee0ca5f78358ed5b99351fdc9 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/50c3738d0b416b4ee0ca5f78358ed5b99351fdc9
Authored: 2025-11-21 12:58:03 -0800
Committed: 2025-11-21 13:12:04 -0800

configure.ac: Make help text consistent

---

8758b029284643098a7a20d41e243b238e7c2bf5 by Mark Harris <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8758b029284643098a7a20d41e243b238e7c2bf5
Authored: 2025-11-21 12:31:55 -0800
Committed: 2025-11-21 13:10:22 -0800

configure.ac: Fix --disable-fixed-res24 help text

---

e155f9da5e1a463af9d148dca6b0273c5370a874 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e155f9da5e1a463af9d148dca6b0273c5370a874
Authored: 2025-11-21 15:12:57 -0500
Committed: 2025-11-21 15:12:57 -0500

Avoid DRED overflow on decoding a 1-second frame

---

7817df7908318fb3a0af2a01e4da1e25643967d0 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7817df7908318fb3a0af2a01e4da1e25643967d0
Authored: 2025-11-20 16:01:55 -0500
Committed: 2025-11-20 16:30:04 -0500

Fix use of reserved identifiers

---

ab135cdc4ad72391f959e3085b21df7b704c20a0 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ab135cdc4ad72391f959e3085b21df7b704c20a0
Authored: 2025-11-20 11:00:17 -0500
Committed: 2025-11-20 11:00:17 -0500

Don't warn about missing SSE4.1/AVX2 for lossgen

---

c80e394e1672a305984e1b85fe465641a31d6269 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/c80e394e1672a305984e1b85fe465641a31d6269
Authored: 2025-11-20 10:59:41 -0500
Committed: 2025-11-20 10:59:41 -0500

Fixing unused variable warnings

---

7ede7c18cbce106e57090b29c36440c48e44ff8f by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7ede7c18cbce106e57090b29c36440c48e44ff8f
Authored: 2025-11-19 17:49:31 -0500
Committed: 2025-11-19 17:50:50 -0500

Fix AVX detection code

---

6b151b8be1de3f779cf1f935fc2a8ce9bb4eea12 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6b151b8be1de3f779cf1f935fc2a8ce9bb4eea12
Authored: 2025-11-19 17:15:53 -0500
Committed: 2025-11-19 17:15:53 -0500

Fixes a denormal issue in DRED encoding

---

2f6660cd42e974dd2d32abf2b1eae62d76bde5e3 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2f6660cd42e974dd2d32abf2b1eae62d76bde5e3
Authored: 2025-11-19 16:43:56 -0500
Committed: 2025-11-19 16:43:56 -0500

Adjust PLC C0 lower bound to let audio go to zero

---

9fdb5703b3397743b4b0696d1cd6975c2542ca76 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9fdb5703b3397743b4b0696d1cd6975c2542ca76
Authored: 2025-11-18 18:27:54 -0500
Committed: 2025-11-18 18:27:54 -0500

Silence warnings on _mm_storel_epi64()

---

83a8274f9cc11280a9df947e135a769edf00ae6c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/83a8274f9cc11280a9df947e135a769edf00ae6c
Authored: 2025-11-18 18:16:16 -0500
Committed: 2025-11-18 18:16:16 -0500

Fix warning on assert

---

4024696fb57be5cc495c4ce175809ddbde56b3d5 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/4024696fb57be5cc495c4ce175809ddbde56b3d5
Authored: 2025-11-18 18:14:10 -0500
Committed: 2025-11-18 18:14:10 -0500

Silence cast alignment warning

---

fd68cc42e9adb24f53c0bd69a1c465fa0d5e7434 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/fd68cc42e9adb24f53c0bd69a1c465fa0d5e7434
Authored: 2025-11-18 17:59:56 -0500
Committed: 2025-11-18 17:59:56 -0500

Avoid int->float conversion warnings for RAND_MAX

---

d4f10fbe35a8fcd74c19260aef5790ce072686cb by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/d4f10fbe35a8fcd74c19260aef5790ce072686cb
Authored: 2025-11-18 16:42:47 -0500
Committed: 2025-11-18 16:43:50 -0500

Remove unused variable

---

4b59c08a12bf9d2336b3b52548f1db516691c036 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/4b59c08a12bf9d2336b3b52548f1db516691c036
Authored: 2025-11-13 11:44:05 -0500
Committed: 2025-11-18 11:15:26 -0500

Shrink encoder delay_buffer

---

f383ea8212f7f78d8c143b37a465897db72c3e26 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/f383ea8212f7f78d8c143b37a465897db72c3e26
Authored: 2025-11-13 16:54:31 -0500
Committed: 2025-11-17 17:01:30 -0500

Simplify bitrate<->bits conversion code

---

c193aee9b86a40dca5f252180fb20f0a57bf612e by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/c193aee9b86a40dca5f252180fb20f0a57bf612e
Authored: 2025-11-17 16:40:30 -0500
Committed: 2025-11-17 16:40:30 -0500

Simplifying qext_oldBandE storage

---

69508c5c001956be38f60034a31932a674de11b2 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/69508c5c001956be38f60034a31932a674de11b2
Authored: 2025-11-17 16:19:35 -0500
Committed: 2025-11-17 16:19:35 -0500

Reorder extra fields in CELT decoder

---

0a1b5d0c101854ff1fcb207b18b9247d87166b62 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0a1b5d0c101854ff1fcb207b18b9247d87166b62
Authored: 2025-11-17 12:50:24 -0500
Committed: 2025-11-17 12:50:24 -0500

Silence warnings on unaligned load/store

---

2c0b1187258dd9631ebc0a2f39f2ac547cefc784 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2c0b1187258dd9631ebc0a2f39f2ac547cefc784
Authored: 2025-11-14 17:33:42 -0500
Committed: 2025-11-15 16:51:14 -0500

Fix more warnings

---

2093947ccd5182d38b249a1b89306cb196e9c492 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2093947ccd5182d38b249a1b89306cb196e9c492
Authored: 2025-11-14 17:28:37 -0500
Committed: 2025-11-14 17:28:37 -0500

Avoid warnings when disabling assertions

---

bfc2dafa26cd9cd6e041f8d25c310d14fa0f8cb3 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/bfc2dafa26cd9cd6e041f8d25c310d14fa0f8cb3
Authored: 2025-11-14 15:05:07 -0500
Committed: 2025-11-14 15:05:07 -0500

Silence some warnings

---

7b46ecd7164f2ec9bf4521efb3e6afae621c8229 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7b46ecd7164f2ec9bf4521efb3e6afae621c8229
Authored: 2025-11-14 14:06:30 -0500
Committed: 2025-11-14 14:06:30 -0500

Fix fixed-point debug error on 64-bit shift

---

285ef69edd68a3799e6759d8f2d455e3bcffcbb9 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/285ef69edd68a3799e6759d8f2d455e3bcffcbb9
Authored: 2025-11-07 20:55:44 -0500
Committed: 2025-11-11 16:26:38 -0500

Use silence detection even with float_api off

---

ad854445cf914ad5224386b9eeeb1f8de35ec1d8 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ad854445cf914ad5224386b9eeeb1f8de35ec1d8
Authored: 2025-11-07 20:55:44 -0500
Committed: 2025-11-11 16:26:37 -0500

Fix activity when no analysis (fixes DRED and DTX)

---

7f097e0fbf31f0e1b679f863646b3746646ab2a0 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7f097e0fbf31f0e1b679f863646b3746646ab2a0
Authored: 2025-11-07 19:02:42 -0500
Committed: 2025-11-07 19:02:42 -0500

Disable DRED decoding in CELT at 96 kHz for now

---

970a8a2f1dd7a0b50e4619772dd1acd4ccacb681 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/970a8a2f1dd7a0b50e4619772dd1acd4ccacb681
Authored: 2025-11-07 15:42:16 -0500
Committed: 2025-11-07 15:43:02 -0500

Fix DRED for 96-kHz encoders

---

dde422e430f2c876170690df08667336f09e0bb2 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/dde422e430f2c876170690df08667336f09e0bb2
Authored: 2025-11-06 14:01:34 -0500
Committed: 2025-11-07 12:58:39 -0500

Cap st->bitrate_bps at max_data_bytes

---

71056d73fb39dad9b27b641a2683baed09dd4ecc by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/71056d73fb39dad9b27b641a2683baed09dd4ecc
Authored: 2025-11-05 15:50:52 -0500
Committed: 2025-11-05 15:50:52 -0500

update DRED testing script

---

940d4e5af64351ca8ba8390df3f555484c567fbb by Jan Buethe <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/940d4e5af64351ca8ba8390df3f555484c567fbb
Authored: 2025-11-04 16:12:51 -0800
Committed: 2025-11-04 18:28:05 -0800

extending mathops for use in osce

---

2f03101027039f0039568bde59207c7610f5eb93 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2f03101027039f0039568bde59207c7610f5eb93
Authored: 2025-11-04 12:21:26 -0500
Committed: 2025-11-04 12:21:26 -0500

oops

---

5607d7d736fb721c69a225ab69010132a7b63283 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5607d7d736fb721c69a225ab69010132a7b63283
Authored: 2025-11-03 23:37:32 -0500
Committed: 2025-11-03 23:37:32 -0500

Avoid integer overflows on very large bitrates

---

c2ae376d6ea9be83fb69b341e882188aa84167d3 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/c2ae376d6ea9be83fb69b341e882188aa84167d3
Authored: 2025-11-03 13:44:36 -0500
Committed: 2025-11-03 13:44:36 -0500

Model update: aug3b26d

---

5e6ec3f54c34a3af0ef7256ac687bcba46be9c8f by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5e6ec3f54c34a3af0ef7256ac687bcba46be9c8f
Authored: 2025-11-03 13:36:36 -0500
Committed: 2025-11-03 13:36:36 -0500

Fix from Mark Harris

---

52dfc510f39fca9653fd5eaee08fc97869e4b1c1 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/52dfc510f39fca9653fd5eaee08fc97869e4b1c1
Authored: 2025-10-10 17:28:18 -0700
Committed: 2025-11-03 12:30:33 -0500

Implement mode forcing to save state memory

---

898f42522a15f3d1f27e3c2d6f2d78f0c16e78e5 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/898f42522a15f3d1f27e3c2d6f2d78f0c16e78e5
Authored: 2025-10-10 16:03:07 -0700
Committed: 2025-11-02 12:40:59 -0500

opus_encoder_init(NULL,...) now returns the size

---

3563bfb4ecf19f053755014f6b08dfb4c98ef6c5 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/3563bfb4ecf19f053755014f6b08dfb4c98ef6c5
Authored: 2025-11-02 12:39:04 -0500
Committed: 2025-11-02 12:40:59 -0500

Fix multistream error checking

---

e0dfb5c3bdccebb2a8ac0e6b8b10f20cff2fe675 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e0dfb5c3bdccebb2a8ac0e6b8b10f20cff2fe675
Authored: 2025-10-27 11:16:00 -0400
Committed: 2025-10-31 13:49:44 -0400

Implement variable allocation of the QEXT bits

---

bd5f1450b8f28dbff67303cd0cc59adf559d8b47 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/bd5f1450b8f28dbff67303cd0cc59adf559d8b47
Authored: 2025-09-12 22:16:06 -0400
Committed: 2025-10-30 16:35:42 -0400

Augment latents with quantizer features

---

57d1bc33f978e8328a5a812e1f3edc515f79e7f9 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/57d1bc33f978e8328a5a812e1f3edc515f79e7f9
Authored: 2025-09-07 10:17:58 -0400
Committed: 2025-10-30 16:35:42 -0400

Train model with new loudness loss function

---

6190099f5557969dc3e0ef48d5b2c592521d838c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6190099f5557969dc3e0ef48d5b2c592521d838c
Authored: 2025-08-23 15:29:12 -0400
Committed: 2025-10-30 16:35:42 -0400

New DRED loss based on loudness and error^4

---

964bf2d843c44c16e23b96cf2bda997eadd1fa30 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/964bf2d843c44c16e23b96cf2bda997eadd1fa30
Authored: 2025-08-23 15:26:30 -0400
Committed: 2025-10-30 16:35:41 -0400

sparser GRUs

---

1457b5c6e5df8a92b80b80d2bed1296f23e34ef4 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/1457b5c6e5df8a92b80b80d2bed1296f23e34ef4
Authored: 2025-07-26 13:59:34 -0400
Committed: 2025-10-30 16:35:41 -0400

sparsify gru_init layer

---

2f3ee1d007fbe4dc3579d558ad1d2af36fbfc59b by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2f3ee1d007fbe4dc3579d558ad1d2af36fbfc59b
Authored: 2025-07-07 17:41:14 -0400
Committed: 2025-10-30 16:35:41 -0400

Adds a dense layer in front of convolutions

---

269e82a3534c57154966fbe16d39804c3ed253b6 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/269e82a3534c57154966fbe16d39804c3ed253b6
Authored: 2025-07-04 17:45:21 -0400
Committed: 2025-10-30 16:35:41 -0400

DRED: Don't reset sparsification when adapting

---

e526333872a25a1e95e164ace90285571cc37df0 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e526333872a25a1e95e164ace90285571cc37df0
Authored: 2025-07-04 14:15:32 -0400
Committed: 2025-10-30 16:35:41 -0400

Shrink DRED model

---

264a5c700ee4bfc23f224aba9920ace94293a003 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/264a5c700ee4bfc23f224aba9920ace94293a003
Authored: 2025-05-20 21:31:58 -0400
Committed: 2025-10-30 16:35:41 -0400

More augmentation for DRED training

---

43efb99765aeed17eb9066f4eaedcf4f4081eb3b by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/43efb99765aeed17eb9066f4eaedcf4f4081eb3b
Authored: 2025-10-07 18:40:46 -0700
Committed: 2025-10-21 16:53:27 -0400

Reduce SILK encoder state size for mono

---

23b38056d9cc983ba52e2288f9d97840214fbbcc by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/23b38056d9cc983ba52e2288f9d97840214fbbcc
Authored: 2025-08-13 14:05:59 -0400
Committed: 2025-10-20 15:35:03 -0400

Decimate QEXT PLC pitch search by 4 instead of 2

---

913cac3dad3e9d0dc2ad3a481df39f9b710102f9 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/913cac3dad3e9d0dc2ad3a481df39f9b710102f9
Authored: 2025-07-22 14:27:05 -0400
Committed: 2025-10-20 15:35:03 -0400

Opus HD testvector script

---

9bae4aa4f419f15b8644760d488f1d321de87a8c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9bae4aa4f419f15b8644760d488f1d321de87a8c
Authored: 2025-07-02 13:42:49 -0400
Committed: 2025-10-20 15:35:02 -0400

qext_compare: add -skip argument

---

e2e0e3b289b2eb04a10645ecdf321d0a6d22bfe5 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e2e0e3b289b2eb04a10645ecdf321d0a6d22bfe5
Authored: 2025-05-26 16:44:18 -0400
Committed: 2025-10-20 15:35:02 -0400

qext_compare: add comparison thresholds

---

bcd71866badf28a0bae7914d40b0a0d31229ab55 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/bcd71866badf28a0bae7914d40b0a0d31229ab55
Authored: 2025-05-09 15:44:28 -0400
Committed: 2025-10-20 15:35:01 -0400

qext_compare: shape the noise floor

---

f2540346ca236592440b992a0af48eda3780c9a9 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/f2540346ca236592440b992a0af48eda3780c9a9
Authored: 2025-04-25 15:59:13 -0400
Committed: 2025-10-20 15:35:01 -0400

qext_compare: lower sensitivity at low power

---

435b185ab32fd9b30123d8f552cf623098f558c6 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/435b185ab32fd9b30123d8f552cf623098f558c6
Authored: 2025-04-24 18:44:18 -0400
Committed: 2025-10-20 15:35:00 -0400

qext_compare: Add RMS calculation

---

1d4099325a24ad3b93751fa1955ca7beb1eee019 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/1d4099325a24ad3b93751fa1955ca7beb1eee019
Authored: 2025-04-23 13:05:07 -0400
Committed: 2025-10-20 15:35:00 -0400

qext_compare: use trimmed-down version of kiss-fft

---

a7ad267502a6667193ff06b5f80ef67992b8480e by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/a7ad267502a6667193ff06b5f80ef67992b8480e
Authored: 2025-04-22 16:31:14 -0400
Committed: 2025-10-20 15:34:59 -0400

qext_compare: Blackman-Harris window, weighting

---

6ecb478abb8353aa2d81b5e26e5c7df0d13e4e7c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6ecb478abb8353aa2d81b5e26e5c7df0d13e4e7c
Authored: 2025-04-22 14:10:50 -0400
Committed: 2025-10-20 15:34:59 -0400

qext_compare tuning

---

cbc7327815099f1b63cf39e4fec2cba1580e0fa4 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/cbc7327815099f1b63cf39e4fec2cba1580e0fa4
Authored: 2025-04-11 16:33:30 -0400
Committed: 2025-10-20 15:34:58 -0400

Add 24-bit and float32 support to qext_compare

---

5ddaee4ae7067843b716792fa24f692e508727bd by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5ddaee4ae7067843b716792fa24f692e508727bd
Authored: 2025-04-11 14:28:08 -0400
Committed: 2025-10-20 15:34:58 -0400

Add 48 kHz support to qext_compare

---

f456ef7cebc55bdbd893f84cefa8191534506d56 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/f456ef7cebc55bdbd893f84cefa8191534506d56
Authored: 2025-04-09 16:34:44 -0400
Committed: 2025-10-20 15:34:57 -0400

First version of qext_compare tool

---

8486ee950ba2236ea8d7798adb219fc5c6836184 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8486ee950ba2236ea8d7798adb219fc5c6836184
Authored: 2025-03-21 16:17:14 -0400
Committed: 2025-10-20 15:34:57 -0400

Use a follower to better distribute extra bits

---

ae7f5a201d1177d3387ffcee47fac697627a4270 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ae7f5a201d1177d3387ffcee47fac697627a4270
Authored: 2025-03-03 10:24:25 -0500
Committed: 2025-10-20 15:34:57 -0400

Better accuracy for fixed-point pitch on tones

---

8a9980fe97aac3a56bfb03a6962d4816375a0a1d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8a9980fe97aac3a56bfb03a6962d4816375a0a1d
Authored: 2025-02-21 14:12:34 -0500
Committed: 2025-10-20 15:34:56 -0400

Better intensity_stereo() fixed-point accuracy

---

866d8609eab135c30ba40f0f45deeeffe5fd0bd7 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/866d8609eab135c30ba40f0f45deeeffe5fd0bd7
Authored: 2025-02-20 23:49:08 -0500
Committed: 2025-10-20 15:34:56 -0400

Convert cubic_quant() to fixed-point

---

6c2ecc4e0952c696313544cb782f64211b9e172b by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6c2ecc4e0952c696313544cb782f64211b9e172b
Authored: 2025-05-08 14:40:39 -0400
Committed: 2025-10-20 15:34:55 -0400

Improve fixed-point PVQ search accuracy

---

8f1a7f9da5899dd4db7c9ebb61ccac37285d8546 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8f1a7f9da5899dd4db7c9ebb61ccac37285d8546
Authored: 2025-02-20 23:13:39 -0500
Committed: 2025-10-20 15:34:55 -0400

Better PVQ search fixed-point accuracy

---

de33a1f31b8bfab7e049c997f2f0dfc3b79b7b25 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/de33a1f31b8bfab7e049c997f2f0dfc3b79b7b25
Authored: 2025-02-20 23:12:25 -0500
Committed: 2025-10-20 15:34:54 -0400

Better normalise_bands() fixed-point accuracy

---

e183475000b5cee079c51f4a7a392e4eced81c58 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e183475000b5cee079c51f4a7a392e4eced81c58
Authored: 2025-02-20 22:54:57 +0000
Committed: 2025-10-20 15:34:54 -0400

Implement a 32-bit norm-reciprocal function in fixed-point arithmetic.

---

2d08af0c1c36b1c44d78dcf1d01e3bb7deb6d992 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2d08af0c1c36b1c44d78dcf1d01e3bb7deb6d992
Authored: 2025-02-19 13:16:30 -0500
Committed: 2025-10-20 15:34:53 -0400

Better stereo_itheta() fixed-point accuracy

---

8b6139fe6e72e460424e7a05b8371c3b4dd1bd8f by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8b6139fe6e72e460424e7a05b8371c3b4dd1bd8f
Authored: 2025-02-19 13:10:21 -0500
Committed: 2025-10-20 15:34:53 -0400

Better compute_band_energies fixed-point accuracy

---

0a1be8e210975d8b81d5cafc2a6808fcd7d401a2 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0a1be8e210975d8b81d5cafc2a6808fcd7d401a2
Authored: 2024-12-12 00:03:52 +0000
Committed: 2025-10-20 15:34:52 -0400

Implement a 32-bit fixed-point sqrt function.

---

a80d72f525cb54dedebba8f95b9c94efe912e9a7 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/a80d72f525cb54dedebba8f95b9c94efe912e9a7
Authored: 2025-02-18 15:04:27 -0500
Committed: 2025-10-20 15:34:52 -0400

Enable comb filter at complexity<5 for tones

---

77e7eb05950be67d366449ca5c101157d47662b7 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/77e7eb05950be67d366449ca5c101157d47662b7
Authored: 2025-01-28 19:47:24 +0000
Committed: 2025-10-20 15:34:51 -0400

Refactor prototypes to use the ARG_QEXT macro

---

b67472e1b71a69a8724fd209b4d262a724de3b47 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/b67472e1b71a69a8724fd209b4d262a724de3b47
Authored: 2025-01-28 18:15:33 +0000
Committed: 2025-10-20 15:34:51 -0400

Add opus_custom_demo ctl's error handling

---

3fb39c1d83cef132dc5e6f86b5ca8d124cd894c4 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/3fb39c1d83cef132dc5e6f86b5ca8d124cd894c4
Authored: 2025-04-02 14:33:23 -0400
Committed: 2025-10-20 15:34:51 -0400

Add QEXT support to test_opus_encode

---

28d129a3a30a5d4c186ed7ce1b5c683583c279b1 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/28d129a3a30a5d4c186ed7ce1b5c683583c279b1
Authored: 2025-04-16 11:29:51 -0400
Committed: 2025-10-20 15:34:50 -0400

Increase qext max bitrate

---

cc4ed9c22d4573263e7f12b9afe248524d37924c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/cc4ed9c22d4573263e7f12b9afe248524d37924c
Authored: 2025-01-31 09:59:29 -0500
Committed: 2025-10-20 15:34:50 -0400

Add QEXT support in encoder API

---

a82512c2cb5784189849f2ec23106f7cb69e54ac by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/a82512c2cb5784189849f2ec23106f7cb69e54ac
Authored: 2025-01-29 10:33:08 -0500
Committed: 2025-10-20 15:34:49 -0400

Add 96 kHz support to Opus encoder

---

8161640db03727aa9a0d76377d16e5288b7b2342 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8161640db03727aa9a0d76377d16e5288b7b2342
Authored: 2025-01-24 16:31:41 -0500
Committed: 2025-10-20 15:34:49 -0400

Support for QEXT decoding in Opus API

---

0e8456aee3ad12293b84fb1f60f53e8c1c4eec6d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0e8456aee3ad12293b84fb1f60f53e8c1c4eec6d
Authored: 2025-01-20 15:48:09 -0500
Committed: 2025-10-20 15:34:48 -0400

QEXT: More precision in pre-emphasis

---

2a50df5714bbdee4410681d013b5da56a0dbc037 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2a50df5714bbdee4410681d013b5da56a0dbc037
Authored: 2025-01-17 17:06:26 -0500
Committed: 2025-10-20 15:34:48 -0400

Add --enable-opus-custom-api for QEXT

---

6ca750ecc6b6501eaf87b3c161af6799a801d30f by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6ca750ecc6b6501eaf87b3c161af6799a801d30f
Authored: 2025-01-17 14:29:44 -0500
Committed: 2025-10-20 15:34:47 -0400

Avoid undefined shift in endianness conversion

---

9954dfca48b427e164a3981a8bf75e8fb9091491 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9954dfca48b427e164a3981a8bf75e8fb9091491
Authored: 2025-01-14 13:23:47 -0500
Committed: 2025-10-20 15:34:47 -0400

Using 32 bits (Q24) for celt_norm

---

d734880b07a48b73d6d5b3128d70697b9cbaf348 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/d734880b07a48b73d6d5b3128d70697b9cbaf348
Authored: 2025-01-09 17:30:45 -0500
Committed: 2025-10-20 15:34:46 -0400

Making the normalized shift changeable

---

e0c4c71150022b6b74639121a9be6f96d36cecfc by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e0c4c71150022b6b74639121a9be6f96d36cecfc
Authored: 2024-12-19 13:58:58 -0500
Committed: 2025-10-20 15:34:46 -0400

Fixed-point conversion of cubic_synthesis()

---

30c4936cba3bbca6af03a62590934b45ab3d64a6 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/30c4936cba3bbca6af03a62590934b45ab3d64a6
Authored: 2024-12-17 15:00:06 -0500
Committed: 2025-10-20 15:34:46 -0400

Using celt_rsqrt_norm32() in normalization

---

2973cdd0f3f08ed4d485f424764e5590dc1089e3 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2973cdd0f3f08ed4d485f424764e5590dc1089e3
Authored: 2024-12-16 15:02:48 -0500
Committed: 2025-10-20 15:34:45 -0400

Allow for more than 1275 bytes for qext packets

---

29bad3b7c3dfd79d99a24abed883cfd7c85b4e56 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/29bad3b7c3dfd79d99a24abed883cfd7c85b4e56
Authored: 2024-12-16 13:40:13 -0500
Committed: 2025-10-20 15:34:45 -0400

Code qext high bands after low bands

---

019f2b6d2e647d00985b607aa05c4732eb75aa07 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/019f2b6d2e647d00985b607aa05c4732eb75aa07
Authored: 2024-12-16 11:33:20 -0500
Committed: 2025-10-20 15:34:44 -0400

More conservative qext allocation at lower bitrate

---

10691b246a5a1bd48464ec8b68f7c4b17d7d3928 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/10691b246a5a1bd48464ec8b68f7c4b17d7d3928
Authored: 2024-12-13 13:34:55 -0500
Committed: 2025-10-20 15:34:44 -0400

Reallocate mid/side bits over cap as extra

---

844919e46fcaecf17d870ac6e2045eb28d3a127d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/844919e46fcaecf17d870ac6e2045eb28d3a127d
Authored: 2024-12-13 11:13:30 -0500
Committed: 2025-10-20 15:34:43 -0400

No need for EPSILON in stereo_itheta()

---

3210cdff17c22c44f856ada647370978ce67a37d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/3210cdff17c22c44f856ada647370978ce67a37d
Authored: 2024-12-12 13:37:10 -0500
Committed: 2025-10-20 15:34:43 -0400

Adding QEXT flags for intensity and dual stereo

---

5b5218ebd0511629caab1544a2fe67bfa3d15410 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5b5218ebd0511629caab1544a2fe67bfa3d15410
Authored: 2024-11-29 12:01:32 -0500
Committed: 2025-10-20 15:34:43 -0400

Add entropy coding of qext bit allocation

---

dd38542a908acdda23660d9cebb216606ad9c9f2 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/dd38542a908acdda23660d9cebb216606ad9c9f2
Authored: 2024-11-28 16:36:03 -0500
Committed: 2025-10-20 15:34:42 -0400

Add some entropy coding to PVQ refinement

---

2c7e434f1a4be98794ad5e90a653a11537c64330 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2c7e434f1a4be98794ad5e90a653a11537c64330
Authored: 2024-11-27 16:50:09 -0500
Committed: 2025-10-20 15:34:42 -0400

Try allocating at least 1 bit depth to high bands

---

0c9fc09fcca018497f43fd6cd2e039a9f2886104 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0c9fc09fcca018497f43fd6cd2e039a9f2886104
Authored: 2024-12-11 23:28:12 +0000
Committed: 2025-10-20 15:34:41 -0400

Add the fixed-point cos(PI0.5x) function (cos_norm2).

---

eb75f10521eb402928a372e1996ab73922d6f595 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/eb75f10521eb402928a372e1996ab73922d6f595
Authored: 2025-01-30 09:57:00 -0500
Committed: 2025-10-20 15:34:41 -0400

Using odd K on transients to avoid reduce pre-echo

---

0ffac18fe2a879e5ea1bb19800eff8a61e4dcf4c by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0ffac18fe2a879e5ea1bb19800eff8a61e4dcf4c
Authored: 2025-01-30 00:30:36 -0500
Committed: 2025-10-20 15:34:40 -0400

Use raw bits for cubic codebook

---

184005fbce989118c76fc94df91befb7849f12e4 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/184005fbce989118c76fc94df91befb7849f12e4
Authored: 2024-11-27 11:29:30 -0500
Committed: 2025-10-20 15:34:40 -0400

Add band splitting to cubic quantizer

---

5b8c9faecacda9258467c7cb8059d6e17199542e by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5b8c9faecacda9258467c7cb8059d6e17199542e
Authored: 2024-11-21 18:24:29 -0500
Committed: 2025-10-20 15:34:39 -0400

Increase allocation of high bands for HF tones

---

cb383ca68d448b82ea44dbd0617065dd12255939 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/cb383ca68d448b82ea44dbd0617065dd12255939
Authored: 2024-11-21 17:26:35 -0500
Committed: 2025-10-20 15:34:39 -0400

Add slope to HF bit allocation

---

332154ec8306b8720bf178020884dfd731d41d3d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/332154ec8306b8720bf178020884dfd731d41d3d
Authored: 2024-11-19 16:44:22 -0500
Committed: 2025-10-20 15:34:39 -0400

Don't force post-filter multiples to powers of two

---

a2ffbfd41fb7f4cfd91ee3765b93ff878cf40dbe by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/a2ffbfd41fb7f4cfd91ee3765b93ff878cf40dbe
Authored: 2025-06-13 10:50:10 -0400
Committed: 2025-10-20 15:34:38 -0400

Add bit allocation fuzzing to QEXT

---

0dba819301ae1244be54b2102e558af890bf9a72 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0dba819301ae1244be54b2102e558af890bf9a72
Authored: 2024-11-12 10:41:39 -0500
Committed: 2025-10-20 15:34:38 -0400

Adds cubic codebook

---

e3a1c20dbcf449213ccc4f75ce19df75bb29b97e by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e3a1c20dbcf449213ccc4f75ce19df75bb29b97e
Authored: 2025-06-13 16:06:54 -0400
Committed: 2025-10-20 15:34:37 -0400

Fix maxDepth VBR cap for QEXT

---

7426f9a64365db278a5c07204953f14ba3fbb015 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7426f9a64365db278a5c07204953f14ba3fbb015
Authored: 2024-09-20 12:06:56 -0400
Committed: 2025-10-20 15:34:37 -0400

Coding fine energy and PVQ up to 48 kHz

---

c8e10bc0a30b63322d19f4fce06f9725842c2c8d by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/c8e10bc0a30b63322d19f4fce06f9725842c2c8d
Authored: 2024-08-16 11:24:40 -0400
Committed: 2025-10-20 15:34:35 -0400

Add coarse energy quantization for 20-48 kHz

---

f4b25cc25a7e6834ce097da1063743e6419b855f by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/f4b25cc25a7e6834ce097da1063743e6419b855f
Authored: 2024-05-27 12:37:21 -0400
Committed: 2025-10-20 15:31:02 -0400

Make 96 kHz interoperable with 48 kHz

---

7dee79e7ef86fdf1642dbf6ffb5c167eda1dc371 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/7dee79e7ef86fdf1642dbf6ffb5c167eda1dc371
Authored: 2024-07-20 02:10:20 -0400
Committed: 2025-10-20 15:31:01 -0400

Extra bits allocator based on band energy

---

570583e46514d022f60f2dd1890ba0a11033a8f7 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/570583e46514d022f60f2dd1890ba0a11033a8f7
Authored: 2024-07-09 18:14:19 -0400
Committed: 2025-10-20 15:31:01 -0400

Code extra bits even for split theta of 0 or pi/2

---

a6410625151564759d1a37ca86a72fbcec3ba9c3 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/a6410625151564759d1a37ca86a72fbcec3ba9c3
Authored: 2024-07-19 17:32:16 -0400
Committed: 2025-10-20 15:31:00 -0400

Allowing extra PVQ bits for any N

---

ea35d9350820a21ee8f9927cc355387fed0c4ed0 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ea35d9350820a21ee8f9927cc355387fed0c4ed0
Authored: 2024-07-19 17:31:31 -0400
Committed: 2025-10-20 15:31:00 -0400

Supporting more than 8 extra PVQ bits

---

4df4a2055f89fe4a6b1dff9722e96e6fc5ec7f4e by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/4df4a2055f89fe4a6b1dff9722e96e6fc5ec7f4e
Authored: 2024-12-11 23:17:43 +0000
Committed: 2025-10-20 15:30:59 -0400

Implement fixed-point normalized atan and atan2p functions.

---

9cf42995add32a86935b12e89ee4ce6e930f1f56 by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9cf42995add32a86935b12e89ee4ce6e930f1f56
Authored: 2024-12-10 22:01:16 +0000
Committed: 2025-10-20 15:30:59 -0400

Improve rsqrt for better accuracy.

---

4dac7a4a71874ffc09ae1a0b69d3ced7a483260c by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/4dac7a4a71874ffc09ae1a0b69d3ced7a483260c
Authored: 2024-11-03 08:55:41 -0500
Committed: 2025-10-20 15:30:59 -0400

Add a new cos with precision of 20-24 bits.

---

fb66bb567315d52ae4ec90861ed2cc107ad4105f by Yunho Huh <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/fb66bb567315d52ae4ec90861ed2cc107ad4105f
Authored: 2024-10-31 00:56:01 +0000
Committed: 2025-10-20 15:30:58 -0400

Improve atan2's precision to 20-24 bits.

---

0126b89bc9fba5bae631d74a41ec8f38453529e6 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/0126b89bc9fba5bae631d74a41ec8f38453529e6
Authored: 2024-06-28 19:51:36 -0400
Committed: 2025-10-20 15:30:58 -0400

Use float atan temporarily for QEXT

---

d54f2bc8c9378864c51a3da3a79515f97cec33ea by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/d54f2bc8c9378864c51a3da3a79515f97cec33ea
Authored: 2024-07-18 11:08:26 -0400
Committed: 2025-10-20 15:30:57 -0400

Extend resolution for theta

---

ba4e5f2070a7c3a9a12357e459609da08be26e89 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ba4e5f2070a7c3a9a12357e459609da08be26e89
Authored: 2024-06-21 17:20:12 -0400
Committed: 2025-10-20 15:30:57 -0400

Extend PVQ resolution for N=4

---

499b65c41d6145424c4d4821b1c52e5d1132d81e by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/499b65c41d6145424c4d4821b1c52e5d1132d81e
Authored: 2024-07-18 10:49:22 -0400
Committed: 2025-10-20 15:30:56 -0400

Extend PVQ resolution for N=2

---

bc41149a7b2c93d70ae0d9884af03d3d52be2ac9 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/bc41149a7b2c93d70ae0d9884af03d3d52be2ac9
Authored: 2024-06-21 17:40:51 -0400
Committed: 2025-10-20 15:30:56 -0400

Add extra energy bits

---

ded5c88cbf45a2390b288b30433c31760828bbc8 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ded5c88cbf45a2390b288b30433c31760828bbc8
Authored: 2024-06-21 17:35:52 -0400
Committed: 2025-10-20 15:30:56 -0400

Defining basic code for CELT extension

---

075ed58a8e63e5cccfd00ca8e383187766f7ae16 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/075ed58a8e63e5cccfd00ca8e383187766f7ae16
Authored: 2025-10-07 15:18:21 -0700
Committed: 2025-10-20 15:30:51 -0400

Fix noise PLC: add post-filtering

---

34bba701ae97c913de719b1f7c10686f62cddb15 by Jan Buethe <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/34bba701ae97c913de719b1f7c10686f62cddb15
Authored: 2025-09-27 22:50:03 -0700
Committed: 2025-09-27 22:50:03 -0700

switching to quantized BBWENet model (v2)

---

b5dc74f26ff9a9ef1f1af9a473b9ae07520eed7d by Alfred E. Heggestad <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/b5dc74f26ff9a9ef1f1af9a473b9ae07520eed7d
Authored: 2025-09-13 18:24:42 +0200
Committed: 2025-09-14 02:01:05 -0400

celt: fix a small typo in comment in interp_bits2pulses() (rate.c)

---

213230eee4e2d93f4db6eddaeec07484b769ea29 by Alfred E. Heggestad <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/213230eee4e2d93f4db6eddaeec07484b769ea29
Authored: 2025-09-13 10:43:52 +0200
Committed: 2025-09-14 02:00:22 -0400

celt: fix DB_SHIFT typo in GCONST macro comment

---

15a34cba8f6a201aff53f976fd2c788660fcbffd by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/15a34cba8f6a201aff53f976fd2c788660fcbffd
Authored: 2025-08-27 12:20:27 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: silk: optimize silk_SMLAWB for MIPS32+

---

cc6750e43d2ea3fdbadef39bb26d2fa2f2ff3095 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/cc6750e43d2ea3fdbadef39bb26d2fa2f2ff3095
Authored: 2025-08-24 12:58:38 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: delete unused header noise_shape_analysis_FIX_mipsr1.h

---

e9661ef9e78d27f348e6ebd6be917a54f7689250 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/e9661ef9e78d27f348e6ebd6be917a54f7689250
Authored: 2025-08-24 12:24:42 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: delete unused header prefilter_FIX_mipsr1.h

---

9022688874ab31f9d75f5fc9763da62b762bf7b2 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/9022688874ab31f9d75f5fc9763da62b762bf7b2
Authored: 2025-08-24 12:20:45 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: silk: optimize silk_warped_autocorrelation_FIX for MIPS32+

---

b08c8757a1150c6c63a82cecce20c2b128769408 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/b08c8757a1150c6c63a82cecce20c2b128769408
Authored: 2025-08-24 10:52:05 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: allow __builtin_clz for MIPS32+

---

247aa220f71b2c830ce5aa3d0a14588c250b360f by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/247aa220f71b2c830ce5aa3d0a14588c250b360f
Authored: 2025-08-24 10:45:26 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: silk: optimize silk_SMULWB for MIPS32+

---

260679b16ef5aacdfe494de9cd2ef4327fed23ca by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/260679b16ef5aacdfe494de9cd2ef4327fed23ca
Authored: 2025-08-23 23:40:46 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: tune NSQ for MIPS

---

51644d58847510189688dea64903c9e1e965fc27 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/51644d58847510189688dea64903c9e1e965fc27
Authored: 2025-08-23 17:40:32 +0300
Committed: 2025-09-14 01:25:36 -0400

refactor: _celt_lpc performance improvement

---

ab9082c00bf02763d47bab09698b37e4ef32d12f by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ab9082c00bf02763d47bab09698b37e4ef32d12f
Authored: 2025-08-23 17:26:29 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: unroll fft_downshift loops for performance

---

ea8f175fa4fda3ae8f6e1c0d07d4a831b288d16b by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ea8f175fa4fda3ae8f6e1c0d07d4a831b288d16b
Authored: 2025-08-23 17:22:40 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: optimize comb_filter for MIPS32r1-r5

---

cb20faaaacef7e597d38329b703bb2a7a188fdc1 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/cb20faaaacef7e597d38329b703bb2a7a188fdc1
Authored: 2025-08-23 17:18:28 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS DSP: utilize dpaq_s.w.ph for celt_inner_prod

---

6ecee3715bdc51454e923ad9ba046b02836b4406 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/6ecee3715bdc51454e923ad9ba046b02836b4406
Authored: 2025-08-23 17:17:01 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS DSP: utilize vector insns for celt_maxabs16

---

325585954d0875afe6c0cbb6d9cf2c95db3d22cb by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/325585954d0875afe6c0cbb6d9cf2c95db3d22cb
Authored: 2025-08-22 14:00:01 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: tune vq for MIPS

---

81fc1497b1e5092e2a3df5945705c252b66bd74d by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/81fc1497b1e5092e2a3df5945705c252b66bd74d
Authored: 2025-08-22 09:37:34 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: tune celt pitch for MIPS

---

ee90c140359eacbc950cc1ee4b140ecf9efcefea by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ee90c140359eacbc950cc1ee4b140ecf9efcefea
Authored: 2025-08-22 07:42:07 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: optimize MULT16_32_Q16 for MIPS without DSP

---

03ef24ebe086c65ca3fcdf7a5808fc3bb702f2b6 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/03ef24ebe086c65ca3fcdf7a5808fc3bb702f2b6
Authored: 2025-08-22 07:08:13 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: fixed_generic_mipsr1.h remove unused functions

---

2cd012096907fa4ac03b2e1665af95d02ee0d579 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/2cd012096907fa4ac03b2e1665af95d02ee0d579
Authored: 2025-08-22 00:02:58 +0300
Committed: 2025-09-14 01:25:36 -0400

MIPS: optimize fft and mdct for MIPS32 without DSP

---

d9df94947de8a754af8ab35b9ca1808d43a26f9e by Alfred E. Heggestad <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/d9df94947de8a754af8ab35b9ca1808d43a26f9e
Authored: 2025-09-11 12:27:00 +0200
Committed: 2025-09-12 14:37:50 -0400

celt: correct naming of LPC include guard

---

10a8990bf2b0c7e22e186876f7622204bad14ed2 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/10a8990bf2b0c7e22e186876f7622204bad14ed2
Authored: 2025-09-12 14:22:26 -0400
Committed: 2025-09-12 14:22:26 -0400

Fix LBRR detection for ToC-only single-byte packet

---

8b2ec5760776ef44ca772c3f25a46cc417bfb217 by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/8b2ec5760776ef44ca772c3f25a46cc417bfb217
Authored: 2025-09-02 22:54:46 +0300
Committed: 2025-09-04 11:49:08 -0400

prevent double precision multiplication for RES2FLOAT and INT24TORES

---

32f66877d438744c08476c1e7d04623cb5cffcea by Siarhei Volkau <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/32f66877d438744c08476c1e7d04623cb5cffcea
Authored: 2025-09-02 22:46:04 +0300
Committed: 2025-09-04 11:49:08 -0400

opus_decode24: prevent aliasing pcm and st->channels

---

247ea8367395a95e5be2a7a9c5dfb8025a3f367e by Jan Buethe <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/247ea8367395a95e5be2a7a9c5dfb8025a3f367e
Authored: 2025-09-02 16:47:46 -0700
Committed: 2025-09-02 20:09:08 -0700

added --initial-epoch option to adv_train_bwe_model.py

---

ca1f8f2ff946d178a1785b4fc408075b717b5450 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/ca1f8f2ff946d178a1785b4fc408075b717b5450
Authored: 2025-08-26 15:35:06 -0400
Committed: 2025-08-28 10:04:45 -0400

Enable res24 by default

---

5d3f93fb5fc92cb765eeba1cd9b229498e22b285 by Jean-Marc Valin <[e-mail]>

https://gitlab.xiph.org/xiph/opus/commit/5d3f93fb5fc92cb765eeba1cd9b229498e22b285
Authored: 2025-08-26 15:32:36 -0400
Committed: 2025-08-28 10:04:42 -0400

Update random build test

---

Comment 2

[Update Bot]

I've submitted a try run for this commit: https://treeherder.mozilla.org/jobs?repo=try&revision=03b5e76ccbe3d80766a452244e36fa99a9353130

Comment 3

[Update Bot]

Attachment: Bug 2003771 - Update opus to d38db32cffb00460dc52543ba3e07ab16c2bf380

Comment 4

[Update Bot]

All jobs completed, we found the following issues.

Known Issues:

• test-windows11-64-24h2/debug-mochitest-browser-chrome-spi-nw-4 - 1 of 4 failed on the same (retriggered) task (failed: Lfs81J6ETnyw-DHeb6wgZg)

These failures may mean that the library update succeeded; you'll need to review
them yourself and decide. If there are lint failures, you will need to fix them in
a follow-up patch. (Or ignore the patch I made, and recreate it yourself with
./mach vendor media/libopus/moz.yaml.)

In either event, I have done all I can, so you will need to take it from here.
When reviewing, please note that this is external code, which needs a full and
careful inspection - not a rubberstamp.

Comment 5

[Update Bot]

Updatebot encountered an error while trying to set you as a reviewer in phabricator.
Updatebot will be unable to do anything more for this library version.