| /* |
| * Copyright (c) 2013 The WebRTC project authors. All Rights Reserved. |
| * |
| * Use of this source code is governed by a BSD-style license |
| * that can be found in the LICENSE file in the root of the source |
| * tree. An additional intellectual property rights grant can be found |
| * in the file PATENTS. All contributing project authors may |
| * be found in the AUTHORS file in the root of the source tree. |
| */ |
| |
| #include "common_audio/resampler/push_sinc_resampler.h" |
| |
| #include <algorithm> |
| #include <cmath> |
| #include <cstring> |
| #include <memory> |
| |
| #include "common_audio/include/audio_util.h" |
| #include "common_audio/resampler/sinusoidal_linear_chirp_source.h" |
| #include "rtc_base/time_utils.h" |
| #include "test/gmock.h" |
| #include "test/gtest.h" |
| |
| namespace webrtc { |
| namespace { |
| |
| // Almost all conversions have an RMS error of around -14 dbFS. |
| const double kResamplingRMSError = -14.42; |
| |
| // Used to convert errors to dbFS. |
| template <typename T> |
| T DBFS(T x) { |
| return 20 * std::log10(x); |
| } |
| |
| } // namespace |
| |
| class PushSincResamplerTest : public ::testing::TestWithParam< |
| ::testing::tuple<int, int, double, double>> { |
| public: |
| PushSincResamplerTest() |
| : input_rate_(::testing::get<0>(GetParam())), |
| output_rate_(::testing::get<1>(GetParam())), |
| rms_error_(::testing::get<2>(GetParam())), |
| low_freq_error_(::testing::get<3>(GetParam())) {} |
| |
| ~PushSincResamplerTest() override {} |
| |
| protected: |
| void ResampleBenchmarkTest(bool int_format); |
| void ResampleTest(bool int_format); |
| |
| int input_rate_; |
| int output_rate_; |
| double rms_error_; |
| double low_freq_error_; |
| }; |
| |
| class ZeroSource : public SincResamplerCallback { |
| public: |
| void Run(size_t frames, float* destination) override { |
| std::memset(destination, 0, sizeof(float) * frames); |
| } |
| }; |
| |
| void PushSincResamplerTest::ResampleBenchmarkTest(bool int_format) { |
| const size_t input_samples = static_cast<size_t>(input_rate_ / 100); |
| const size_t output_samples = static_cast<size_t>(output_rate_ / 100); |
| const int kResampleIterations = 500000; |
| |
| // Source for data to be resampled. |
| ZeroSource resampler_source; |
| |
| std::unique_ptr<float[]> resampled_destination(new float[output_samples]); |
| std::unique_ptr<float[]> source(new float[input_samples]); |
| std::unique_ptr<int16_t[]> source_int(new int16_t[input_samples]); |
| std::unique_ptr<int16_t[]> destination_int(new int16_t[output_samples]); |
| |
| resampler_source.Run(input_samples, source.get()); |
| for (size_t i = 0; i < input_samples; ++i) { |
| source_int[i] = static_cast<int16_t>(floor(32767 * source[i] + 0.5)); |
| } |
| |
| printf("Benchmarking %d iterations of %d Hz -> %d Hz:\n", kResampleIterations, |
| input_rate_, output_rate_); |
| const double io_ratio = input_rate_ / static_cast<double>(output_rate_); |
| SincResampler sinc_resampler(io_ratio, SincResampler::kDefaultRequestSize, |
| &resampler_source); |
| int64_t start = rtc::TimeNanos(); |
| for (int i = 0; i < kResampleIterations; ++i) { |
| sinc_resampler.Resample(output_samples, resampled_destination.get()); |
| } |
| double total_time_sinc_us = |
| (rtc::TimeNanos() - start) / rtc::kNumNanosecsPerMicrosec; |
| printf("SincResampler took %.2f us per frame.\n", |
| total_time_sinc_us / kResampleIterations); |
| |
| PushSincResampler resampler(input_samples, output_samples); |
| start = rtc::TimeNanos(); |
| if (int_format) { |
| for (int i = 0; i < kResampleIterations; ++i) { |
| EXPECT_EQ(output_samples, |
| resampler.Resample(source_int.get(), input_samples, |
| destination_int.get(), output_samples)); |
| } |
| } else { |
| for (int i = 0; i < kResampleIterations; ++i) { |
| EXPECT_EQ(output_samples, resampler.Resample(source.get(), input_samples, |
| resampled_destination.get(), |
| output_samples)); |
| } |
| } |
| double total_time_us = |
| (rtc::TimeNanos() - start) / rtc::kNumNanosecsPerMicrosec; |
| printf( |
| "PushSincResampler took %.2f us per frame; which is a %.1f%% overhead " |
| "on SincResampler.\n\n", |
| total_time_us / kResampleIterations, |
| (total_time_us - total_time_sinc_us) / total_time_sinc_us * 100); |
| } |
| |
| // Disabled because it takes too long to run routinely. Use for performance |
| // benchmarking when needed. |
| TEST_P(PushSincResamplerTest, DISABLED_BenchmarkInt) { |
| ResampleBenchmarkTest(true); |
| } |
| |
| TEST_P(PushSincResamplerTest, DISABLED_BenchmarkFloat) { |
| ResampleBenchmarkTest(false); |
| } |
| |
| // Tests resampling using a given input and output sample rate. |
| void PushSincResamplerTest::ResampleTest(bool int_format) { |
| // Make comparisons using one second of data. |
| static const double kTestDurationSecs = 1; |
| // 10 ms blocks. |
| const size_t kNumBlocks = static_cast<size_t>(kTestDurationSecs * 100); |
| const size_t input_block_size = static_cast<size_t>(input_rate_ / 100); |
| const size_t output_block_size = static_cast<size_t>(output_rate_ / 100); |
| const size_t input_samples = |
| static_cast<size_t>(kTestDurationSecs * input_rate_); |
| const size_t output_samples = |
| static_cast<size_t>(kTestDurationSecs * output_rate_); |
| |
| // Nyquist frequency for the input sampling rate. |
| const double input_nyquist_freq = 0.5 * input_rate_; |
| |
| // Source for data to be resampled. |
| SinusoidalLinearChirpSource resampler_source(input_rate_, input_samples, |
| input_nyquist_freq, 0); |
| |
| PushSincResampler resampler(input_block_size, output_block_size); |
| |
| // TODO(dalecurtis): If we switch to AVX/SSE optimization, we'll need to |
| // allocate these on 32-byte boundaries and ensure they're sized % 32 bytes. |
| std::unique_ptr<float[]> resampled_destination(new float[output_samples]); |
| std::unique_ptr<float[]> pure_destination(new float[output_samples]); |
| std::unique_ptr<float[]> source(new float[input_samples]); |
| std::unique_ptr<int16_t[]> source_int(new int16_t[input_block_size]); |
| std::unique_ptr<int16_t[]> destination_int(new int16_t[output_block_size]); |
| |
| // The sinc resampler has an implicit delay of approximately half the kernel |
| // size at the input sample rate. By moving to a push model, this delay |
| // becomes explicit and is managed by zero-stuffing in PushSincResampler. We |
| // deal with it in the test by delaying the "pure" source to match. It must be |
| // checked before the first call to Resample(), because ChunkSize() will |
| // change afterwards. |
| const size_t output_delay_samples = |
| output_block_size - resampler.get_resampler_for_testing()->ChunkSize(); |
| |
| // Generate resampled signal. |
| // With the PushSincResampler, we produce the signal block-by-10ms-block |
| // rather than in a single pass, to exercise how it will be used in WebRTC. |
| resampler_source.Run(input_samples, source.get()); |
| if (int_format) { |
| for (size_t i = 0; i < kNumBlocks; ++i) { |
| FloatToS16(&source[i * input_block_size], input_block_size, |
| source_int.get()); |
| EXPECT_EQ(output_block_size, |
| resampler.Resample(source_int.get(), input_block_size, |
| destination_int.get(), output_block_size)); |
| S16ToFloat(destination_int.get(), output_block_size, |
| &resampled_destination[i * output_block_size]); |
| } |
| } else { |
| for (size_t i = 0; i < kNumBlocks; ++i) { |
| EXPECT_EQ( |
| output_block_size, |
| resampler.Resample(&source[i * input_block_size], input_block_size, |
| &resampled_destination[i * output_block_size], |
| output_block_size)); |
| } |
| } |
| |
| // Generate pure signal. |
| SinusoidalLinearChirpSource pure_source( |
| output_rate_, output_samples, input_nyquist_freq, output_delay_samples); |
| pure_source.Run(output_samples, pure_destination.get()); |
| |
| // Range of the Nyquist frequency (0.5 * min(input rate, output_rate)) which |
| // we refer to as low and high. |
| static const double kLowFrequencyNyquistRange = 0.7; |
| static const double kHighFrequencyNyquistRange = 0.9; |
| |
| // Calculate Root-Mean-Square-Error and maximum error for the resampling. |
| double sum_of_squares = 0; |
| double low_freq_max_error = 0; |
| double high_freq_max_error = 0; |
| int minimum_rate = std::min(input_rate_, output_rate_); |
| double low_frequency_range = kLowFrequencyNyquistRange * 0.5 * minimum_rate; |
| double high_frequency_range = kHighFrequencyNyquistRange * 0.5 * minimum_rate; |
| |
| for (size_t i = 0; i < output_samples; ++i) { |
| double error = fabs(resampled_destination[i] - pure_destination[i]); |
| |
| if (pure_source.Frequency(i) < low_frequency_range) { |
| if (error > low_freq_max_error) |
| low_freq_max_error = error; |
| } else if (pure_source.Frequency(i) < high_frequency_range) { |
| if (error > high_freq_max_error) |
| high_freq_max_error = error; |
| } |
| // TODO(dalecurtis): Sanity check frequencies > kHighFrequencyNyquistRange. |
| |
| sum_of_squares += error * error; |
| } |
| |
| double rms_error = sqrt(sum_of_squares / output_samples); |
| |
| rms_error = DBFS(rms_error); |
| // In order to keep the thresholds in this test identical to SincResamplerTest |
| // we must account for the quantization error introduced by truncating from |
| // float to int. This happens twice (once at input and once at output) and we |
| // allow for the maximum possible error (1 / 32767) for each step. |
| // |
| // The quantization error is insignificant in the RMS calculation so does not |
| // need to be accounted for there. |
| low_freq_max_error = DBFS(low_freq_max_error - 2.0 / 32767); |
| high_freq_max_error = DBFS(high_freq_max_error - 2.0 / 32767); |
| |
| EXPECT_LE(rms_error, rms_error_); |
| EXPECT_LE(low_freq_max_error, low_freq_error_); |
| |
| // All conversions currently have a high frequency error around -6 dbFS. |
| static const double kHighFrequencyMaxError = -6.02; |
| EXPECT_LE(high_freq_max_error, kHighFrequencyMaxError); |
| } |
| |
| TEST_P(PushSincResamplerTest, ResampleInt) { |
| ResampleTest(true); |
| } |
| |
| TEST_P(PushSincResamplerTest, ResampleFloat) { |
| ResampleTest(false); |
| } |
| |
| // Thresholds chosen arbitrarily based on what each resampling reported during |
| // testing. All thresholds are in dbFS, http://en.wikipedia.org/wiki/DBFS. |
| INSTANTIATE_TEST_SUITE_P( |
| PushSincResamplerTest, |
| PushSincResamplerTest, |
| ::testing::Values( |
| // First run through the rates tested in SincResamplerTest. The |
| // thresholds are identical. |
| // |
| // We don't test rates which fail to provide an integer number of |
| // samples in a 10 ms block (22050 and 11025 Hz). WebRTC doesn't support |
| // these rates in any case (for the same reason). |
| |
| // To 44.1kHz |
| ::testing::make_tuple(8000, 44100, kResamplingRMSError, -62.73), |
| ::testing::make_tuple(16000, 44100, kResamplingRMSError, -62.54), |
| ::testing::make_tuple(32000, 44100, kResamplingRMSError, -63.32), |
| ::testing::make_tuple(44100, 44100, kResamplingRMSError, -73.53), |
| ::testing::make_tuple(48000, 44100, -15.01, -64.04), |
| ::testing::make_tuple(96000, 44100, -18.49, -25.51), |
| ::testing::make_tuple(192000, 44100, -20.50, -13.31), |
| |
| // To 48kHz |
| ::testing::make_tuple(8000, 48000, kResamplingRMSError, -63.43), |
| ::testing::make_tuple(16000, 48000, kResamplingRMSError, -63.96), |
| ::testing::make_tuple(32000, 48000, kResamplingRMSError, -64.04), |
| ::testing::make_tuple(44100, 48000, kResamplingRMSError, -62.63), |
| ::testing::make_tuple(48000, 48000, kResamplingRMSError, -73.52), |
| ::testing::make_tuple(96000, 48000, -18.40, -28.44), |
| ::testing::make_tuple(192000, 48000, -20.43, -14.11), |
| |
| // To 96kHz |
| ::testing::make_tuple(8000, 96000, kResamplingRMSError, -63.19), |
| ::testing::make_tuple(16000, 96000, kResamplingRMSError, -63.39), |
| ::testing::make_tuple(32000, 96000, kResamplingRMSError, -63.95), |
| ::testing::make_tuple(44100, 96000, kResamplingRMSError, -62.63), |
| ::testing::make_tuple(48000, 96000, kResamplingRMSError, -73.52), |
| ::testing::make_tuple(96000, 96000, kResamplingRMSError, -73.52), |
| ::testing::make_tuple(192000, 96000, kResamplingRMSError, -28.41), |
| |
| // To 192kHz |
| ::testing::make_tuple(8000, 192000, kResamplingRMSError, -63.10), |
| ::testing::make_tuple(16000, 192000, kResamplingRMSError, -63.14), |
| ::testing::make_tuple(32000, 192000, kResamplingRMSError, -63.38), |
| ::testing::make_tuple(44100, 192000, kResamplingRMSError, -62.63), |
| ::testing::make_tuple(48000, 192000, kResamplingRMSError, -73.44), |
| ::testing::make_tuple(96000, 192000, kResamplingRMSError, -73.52), |
| ::testing::make_tuple(192000, 192000, kResamplingRMSError, -73.52), |
| |
| // Next run through some additional cases interesting for WebRTC. |
| // We skip some extreme downsampled cases (192 -> {8, 16}, 96 -> 8) |
| // because they violate |kHighFrequencyMaxError|, which is not |
| // unexpected. It's very unlikely that we'll see these conversions in |
| // practice anyway. |
| |
| // To 8 kHz |
| ::testing::make_tuple(8000, 8000, kResamplingRMSError, -75.50), |
| ::testing::make_tuple(16000, 8000, -18.56, -28.79), |
| ::testing::make_tuple(32000, 8000, -20.36, -14.13), |
| ::testing::make_tuple(44100, 8000, -21.00, -11.39), |
| ::testing::make_tuple(48000, 8000, -20.96, -11.04), |
| |
| // To 16 kHz |
| ::testing::make_tuple(8000, 16000, kResamplingRMSError, -70.30), |
| ::testing::make_tuple(16000, 16000, kResamplingRMSError, -75.51), |
| ::testing::make_tuple(32000, 16000, -18.48, -28.59), |
| ::testing::make_tuple(44100, 16000, -19.30, -19.67), |
| ::testing::make_tuple(48000, 16000, -19.81, -18.11), |
| ::testing::make_tuple(96000, 16000, -20.95, -10.9596), |
| |
| // To 32 kHz |
| ::testing::make_tuple(8000, 32000, kResamplingRMSError, -70.30), |
| ::testing::make_tuple(16000, 32000, kResamplingRMSError, -75.51), |
| ::testing::make_tuple(32000, 32000, kResamplingRMSError, -75.51), |
| ::testing::make_tuple(44100, 32000, -16.44, -51.0349), |
| ::testing::make_tuple(48000, 32000, -16.90, -43.9967), |
| ::testing::make_tuple(96000, 32000, -19.61, -18.04), |
| ::testing::make_tuple(192000, 32000, -21.02, -10.94))); |
| |
| } // namespace webrtc |
