diff --git a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.cc b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.cc index 8f0e7bf6b9..38a7ea32cf 100644 --- a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.cc +++ b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.cc @@ -30,7 +30,7 @@ const int kChunkSizeMs = 10; // Size provided by APM. const float kClipFreqKhz = 0.2f; const float kKbdAlpha = 1.5f; const float kLambdaBot = -1.0f; // Extreme values in bisection -const float kLambdaTop = -10e-18f; // search for lamda. +const float kLambdaTop = -1e-5f; // search for lamda. const float kVoiceProbabilityThreshold = 0.02f; // Number of chunks after voice activity which is still considered speech. const size_t kSpeechOffsetDelay = 80; @@ -164,15 +164,13 @@ void IntelligibilityEnhancer::ProcessClearBlock( const float power_bot = DotProduct(gains_eq_.get(), filtered_clear_pow_.get(), bank_size_); if (power_target >= power_bot && power_target <= power_top) { - SolveForLambda(power_target, power_bot, power_top); + SolveForLambda(power_target); UpdateErbGains(); } // Else experiencing power underflow, so do nothing. gain_applier_.Apply(in_block, out_block); } -void IntelligibilityEnhancer::SolveForLambda(float power_target, - float power_bot, - float power_top) { +void IntelligibilityEnhancer::SolveForLambda(float power_target) { const float kConvergeThresh = 0.001f; // TODO(ekmeyerson): Find best values const int kMaxIters = 100; // for these, based on experiments. @@ -183,7 +181,7 @@ void IntelligibilityEnhancer::SolveForLambda(float power_target, float power_ratio = 2.f; // Ratio of achieved power to target power. int iters = 0; while (std::fabs(power_ratio - 1.f) > kConvergeThresh && iters <= kMaxIters) { - const float lambda = lambda_bot + (lambda_top - lambda_bot) / 2.f; + const float lambda = (lambda_bot + lambda_top) / 2.f; SolveForGainsGivenLambda(lambda, start_freq_, gains_eq_.get()); const float power = DotProduct(gains_eq_.get(), filtered_clear_pow_.get(), bank_size_); @@ -286,7 +284,8 @@ std::vector> IntelligibilityEnhancer::CreateErbBank( void IntelligibilityEnhancer::SolveForGainsGivenLambda(float lambda, size_t start_freq, float* sols) { - bool quadratic = (kRho < 1.f); + const float kMinPower = 1e-5f; + const float* pow_x0 = filtered_clear_pow_.get(); const float* pow_n0 = filtered_noise_pow_.get(); @@ -295,20 +294,24 @@ void IntelligibilityEnhancer::SolveForGainsGivenLambda(float lambda, } // Analytic solution for optimal gains. See paper for derivation. - for (size_t n = start_freq - 1; n < bank_size_; ++n) { - float alpha0, beta0, gamma0; - gamma0 = 0.5f * kRho * pow_x0[n] * pow_n0[n] + - lambda * pow_x0[n] * pow_n0[n] * pow_n0[n]; - beta0 = lambda * pow_x0[n] * (2 - kRho) * pow_x0[n] * pow_n0[n]; - if (quadratic) { - alpha0 = lambda * pow_x0[n] * (1 - kRho) * pow_x0[n] * pow_x0[n]; - sols[n] = - (-beta0 - sqrtf(beta0 * beta0 - 4 * alpha0 * gamma0)) / - (2 * alpha0 + std::numeric_limits::epsilon()); + for (size_t n = start_freq; n < bank_size_; ++n) { + if (pow_x0[n] < kMinPower || pow_n0[n] < kMinPower) { + sols[n] = 1.f; } else { - sols[n] = -gamma0 / beta0; + const float gamma0 = 0.5f * kRho * pow_x0[n] * pow_n0[n] + + lambda * pow_x0[n] * pow_n0[n] * pow_n0[n]; + const float beta0 = + lambda * pow_x0[n] * (2.f - kRho) * pow_x0[n] * pow_n0[n]; + const float alpha0 = + lambda * pow_x0[n] * (1.f - kRho) * pow_x0[n] * pow_x0[n]; + RTC_DCHECK_LT(alpha0, 0.f); + // The quadratic equation should always have real roots, but to guard + // against numerical errors we limit it to a minimum of zero. + sols[n] = std::max( + 0.f, (-beta0 - std::sqrt(std::max( + 0.f, beta0 * beta0 - 4.f * alpha0 * gamma0))) / + (2.f * alpha0)); } - sols[n] = fmax(0, sols[n]); } } diff --git a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.h b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.h index c18bac0d85..22a3eab697 100644 --- a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.h +++ b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer.h @@ -67,7 +67,7 @@ class IntelligibilityEnhancer { std::complex* out_block); // Bisection search for optimal |lambda|. - void SolveForLambda(float power_target, float power_bot, float power_top); + void SolveForLambda(float power_target); // Transforms freq gains to ERB gains. void UpdateErbGains(); diff --git a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer_unittest.cc b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer_unittest.cc index b59ae36d8b..ebfb67a90d 100644 --- a/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer_unittest.cc +++ b/webrtc/modules/audio_processing/intelligibility/intelligibility_enhancer_unittest.cc @@ -186,16 +186,11 @@ static_assert(arraysize(kTestCenterFreqs) == arraysize(kTestFilterBank), // Target output for gain solving test. Generated with matlab. const size_t kTestStartFreq = 12; // Lowest integral frequency for ERBs. -const float kTestZeroVar[] = { - 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 0.f, 0.f, 0.f, - 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, - 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0}; -static_assert(arraysize(kTestCenterFreqs) == arraysize(kTestZeroVar), - "Power test data badly initialized."); +const float kTestZeroVar = 1.f; const float kTestNonZeroVarLambdaTop[] = { - 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 0.f, 0.f, 0.f, + 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 1.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, - 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0}; + 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f, 0.f}; static_assert(arraysize(kTestCenterFreqs) == arraysize(kTestNonZeroVarLambdaTop), "Power test data badly initialized."); @@ -280,7 +275,7 @@ TEST_F(IntelligibilityEnhancerTest, TestSolveForGains) { } enh_->SolveForGainsGivenLambda(lambda, enh_->start_freq_, &sols[0]); for (size_t i = 0; i < enh_->bank_size_; i++) { - EXPECT_NEAR(kTestZeroVar[i], sols[i], kMaxTestError); + EXPECT_NEAR(kTestZeroVar, sols[i], kMaxTestError); } for (size_t i = 0; i < enh_->bank_size_; i++) { enh_->filtered_clear_pow_[i] = static_cast(i + 1); @@ -293,7 +288,7 @@ TEST_F(IntelligibilityEnhancerTest, TestSolveForGains) { lambda = -1.f; enh_->SolveForGainsGivenLambda(lambda, enh_->start_freq_, &sols[0]); for (size_t i = 0; i < enh_->bank_size_; i++) { - EXPECT_NEAR(kTestZeroVar[i], sols[i], kMaxTestError); + EXPECT_NEAR(kTestNonZeroVarLambdaTop[i], sols[i], kMaxTestError); } }