feat(audio): real noise suppression via vendored RNNoise (send + receive)
The two-sided NR plumbing (RemoteStream::recv_ns + the per-listener vc_set_remote_stream noise_reduction toggle) was wired but inert: ApmProcessor::create() returned a no-op passthrough, because the originally-planned webrtc-audio-processing has no working Windows/macOS build. Drop in RNNoise as the real backend behind the same ApmProcessor interface, lighting up both NR paths. - Vendor RNNoise (BSD-3 + CC0) at third_party/rnnoise/ — the vcpkg port is !windows !arm, so it can't cover our primary targets. Shrunk int8 model (78MB -> 11.7MB via upstream scripts/shrink_model.sh), built as a standalone C static lib with no RTCD (portable scalar path on x86, auto-NEON on arm64) under -DDISABLE_DEBUG_FLOAT. Model is baked in (rnnoise_create(NULL)); no runtime file. - New RnnoiseProcessor (core/src/audio/apm_processor.cpp) selected by ApmProcessor::create() when VOICECAT_HAS_NS. Mono/48kHz/480-sample; our clock is fixed 48kHz and Opus frame sizes are multiples of 480, so no resampling. RT-safe: allocates at construction, lock-free in the capture/playback callbacks. - Receive-side: lit up via the factory; gated to mono streams (a stereo stream is a screen-audio share, not voice). - Send-side (new): vc_set_input_noise_reduction(client, enable) ABI + vc_client::mic_ns_, run before input gain/VAD in on_capture_frame. A stereo mic is downmixed to mono ONLY when NR is on — with NR off a stereo mic keeps full stereo (never collapse mic quality unasked). - Enable C as a project language for the vendored lib. - New noise_suppression test: white noise through ApmProcessor::create() drops ~99.9% RMS. ctest --preset dev green, 28/28. windows-client DLL builds clean with vc_set_input_noise_reduction exported, system-only deps. - Docs synced: voice.md §10, tech-stack.md §1/§5, third_party/README.md, vcpkg.json note, PROGRESS.md, CLAUDE.md. Client on/off UI toggles (Windows/macOS/iOS) are the remaining follow-up. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Claude Opus 4.8
+85
@@ -0,0 +1,85 @@
|
||||
/* Copyright (c) 2011-2019 Mozilla
|
||||
2023 Amazon */
|
||||
/*
|
||||
Redistribution and use in source and binary forms, with or without
|
||||
modification, are permitted provided that the following conditions
|
||||
are met:
|
||||
|
||||
- Redistributions of source code must retain the above copyright
|
||||
notice, this list of conditions and the following disclaimer.
|
||||
|
||||
- Redistributions in binary form must reproduce the above copyright
|
||||
notice, this list of conditions and the following disclaimer in the
|
||||
documentation and/or other materials provided with the distribution.
|
||||
|
||||
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
||||
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
||||
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
||||
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR
|
||||
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
||||
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
||||
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
||||
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
||||
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
||||
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
||||
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
||||
*/
|
||||
|
||||
#ifndef DNN_X86_H
|
||||
#define DNN_X86_H
|
||||
|
||||
#include "cpu_support.h"
|
||||
#include "opus_types.h"
|
||||
|
||||
void compute_linear_sse4_1(const LinearLayer *linear, float *out, const float *in);
|
||||
void compute_activation_sse4_1(float *output, const float *input, int N, int activation);
|
||||
void compute_conv2d_sse4_1(const Conv2dLayer *conv, float *out, float *mem, const float *in, int height, int hstride, int activation);
|
||||
|
||||
void compute_linear_avx2(const LinearLayer *linear, float *out, const float *in);
|
||||
void compute_activation_avx2(float *output, const float *input, int N, int activation);
|
||||
void compute_conv2d_avx2(const Conv2dLayer *conv, float *out, float *mem, const float *in, int height, int hstride, int activation);
|
||||
|
||||
|
||||
|
||||
#ifdef RNN_ENABLE_X86_RTCD
|
||||
|
||||
extern void (*const RNN_COMPUTE_LINEAR_IMPL[OPUS_ARCHMASK + 1])(
|
||||
const LinearLayer *linear,
|
||||
float *out,
|
||||
const float *in
|
||||
);
|
||||
#define OVERRIDE_COMPUTE_LINEAR
|
||||
#define compute_linear(linear, out, in, arch) \
|
||||
((*RNN_COMPUTE_LINEAR_IMPL[(arch) & OPUS_ARCHMASK])(linear, out, in))
|
||||
|
||||
|
||||
extern void (*const RNN_COMPUTE_ACTIVATION_IMPL[OPUS_ARCHMASK + 1])(
|
||||
float *output,
|
||||
const float *input,
|
||||
int N,
|
||||
int activation
|
||||
);
|
||||
#define OVERRIDE_COMPUTE_ACTIVATION
|
||||
#define compute_activation(output, input, N, activation, arch) \
|
||||
((*RNN_COMPUTE_ACTIVATION_IMPL[(arch) & OPUS_ARCHMASK])(output, input, N, activation))
|
||||
|
||||
|
||||
extern void (*const RNN_COMPUTE_CONV2D_IMPL[OPUS_ARCHMASK + 1])(
|
||||
const Conv2dLayer *conv,
|
||||
float *out,
|
||||
float *mem,
|
||||
const float *in,
|
||||
int height,
|
||||
int hstride,
|
||||
int activation
|
||||
);
|
||||
#define OVERRIDE_COMPUTE_CONV2D
|
||||
#define compute_conv2d(conv, out, mem, in, height, hstride, activation, arch) \
|
||||
((*RNN_COMPUTE_CONV2D_IMPL[(arch) & OPUS_ARCHMASK])(conv, out, mem, in, height, hstride, activation))
|
||||
|
||||
|
||||
#endif
|
||||
|
||||
|
||||
|
||||
#endif /* DNN_X86_H */
|
||||
+47
@@ -0,0 +1,47 @@
|
||||
/* Copyright (c) 2023 Amazon */
|
||||
/*
|
||||
Redistribution and use in source and binary forms, with or without
|
||||
modification, are permitted provided that the following conditions
|
||||
are met:
|
||||
|
||||
- Redistributions of source code must retain the above copyright
|
||||
notice, this list of conditions and the following disclaimer.
|
||||
|
||||
- Redistributions in binary form must reproduce the above copyright
|
||||
notice, this list of conditions and the following disclaimer in the
|
||||
documentation and/or other materials provided with the distribution.
|
||||
|
||||
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
||||
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
||||
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
||||
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR
|
||||
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
||||
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
||||
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
||||
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
||||
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
||||
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
||||
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
||||
*/
|
||||
|
||||
#ifdef _MSC_VER
|
||||
|
||||
# ifdef OPUS_X86_MAY_HAVE_SSE
|
||||
# ifndef __SSE__
|
||||
# define __SSE__
|
||||
# endif
|
||||
# endif
|
||||
|
||||
# ifdef OPUS_X86_MAY_HAVE_SSE2
|
||||
# ifndef __SSE2__
|
||||
# define __SSE2__
|
||||
# endif
|
||||
# endif
|
||||
|
||||
# ifdef OPUS_X86_MAY_HAVE_SSE4_1
|
||||
# ifndef __SSE4_1__
|
||||
# define __SSE4_1__
|
||||
# endif
|
||||
# endif
|
||||
|
||||
#endif
|
||||
Vendored
+88
@@ -0,0 +1,88 @@
|
||||
/* Copyright (c) 2014, Cisco Systems, INC
|
||||
Written by XiangMingZhu WeiZhou MinPeng YanWang
|
||||
|
||||
Redistribution and use in source and binary forms, with or without
|
||||
modification, are permitted provided that the following conditions
|
||||
are met:
|
||||
|
||||
- Redistributions of source code must retain the above copyright
|
||||
notice, this list of conditions and the following disclaimer.
|
||||
|
||||
- Redistributions in binary form must reproduce the above copyright
|
||||
notice, this list of conditions and the following disclaimer in the
|
||||
documentation and/or other materials provided with the distribution.
|
||||
|
||||
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
||||
``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
||||
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
||||
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
|
||||
OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
||||
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
||||
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
||||
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
||||
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
||||
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
||||
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
||||
*/
|
||||
|
||||
#if !defined(X86CPU_H)
|
||||
# define X86CPU_H
|
||||
|
||||
# define MAY_HAVE_SSE4_1(name) name ## _sse4_1
|
||||
|
||||
# define MAY_HAVE_AVX2(name) name ## _avx2
|
||||
|
||||
# ifdef RNN_ENABLE_X86_RTCD
|
||||
int opus_select_arch(void);
|
||||
# endif
|
||||
|
||||
# if defined(__SSE2__)
|
||||
# include "common.h"
|
||||
|
||||
/*MOVD should not impose any alignment restrictions, but the C standard does,
|
||||
and UBSan will report errors if we actually make unaligned accesses.
|
||||
Use this to work around those restrictions (which should hopefully all get
|
||||
optimized to a single MOVD instruction).
|
||||
GCC implemented _mm_loadu_si32() since GCC 11; HOWEVER, there is a bug!
|
||||
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99754
|
||||
LLVM implemented _mm_loadu_si32() since Clang 8.0, however the
|
||||
__clang_major__ version number macro is unreliable, as vendors
|
||||
(specifically, Apple) will use different numbering schemes than upstream.
|
||||
Clang's advice is "use feature detection", but they do not provide feature
|
||||
detection support for specific SIMD functions.
|
||||
We follow the approach from the SIMDe project and instead detect unrelated
|
||||
features that should be available in the version we want (see
|
||||
<https://github.com/simd-everywhere/simde/blob/master/simde/simde-detect-clang.h>).*/
|
||||
# if defined(__clang__)
|
||||
# if __has_warning("-Wextra-semi-stmt") || \
|
||||
__has_builtin(__builtin_rotateleft32)
|
||||
# define OPUS_CLANG_8 (1)
|
||||
# endif
|
||||
# endif
|
||||
# if !defined(_MSC_VER) && !OPUS_GNUC_PREREQ(11,3) && !defined(OPUS_CLANG_8)
|
||||
# include <string.h>
|
||||
# include <emmintrin.h>
|
||||
|
||||
# ifdef _mm_loadu_si32
|
||||
# undef _mm_loadu_si32
|
||||
# endif
|
||||
# define _mm_loadu_si32 WORKAROUND_mm_loadu_si32
|
||||
static inline __m128i WORKAROUND_mm_loadu_si32(void const* mem_addr) {
|
||||
int val;
|
||||
memcpy(&val, mem_addr, sizeof(val));
|
||||
return _mm_cvtsi32_si128(val);
|
||||
}
|
||||
# elif defined(_MSC_VER)
|
||||
/* MSVC needs this for _mm_loadu_si32 */
|
||||
# include <immintrin.h>
|
||||
# endif
|
||||
|
||||
# define OP_CVTEPI8_EPI32_M32(x) \
|
||||
(_mm_cvtepi8_epi32(_mm_loadu_si32(x)))
|
||||
|
||||
# define OP_CVTEPI16_EPI32_M64(x) \
|
||||
(_mm_cvtepi16_epi32(_mm_loadl_epi64((__m128i *)(void*)(x))))
|
||||
|
||||
# endif
|
||||
|
||||
#endif
|
||||
Reference in New Issue
Block a user