server1/ThreeLib/signalstats/sample/demo.cpp

#include <iostream>
#include <vector>
#include <string>
#include <cmath>
#include <sndfile.h>
#include <nlohmann/json.hpp>
#include "signalstats_wrapper.h"

using json = nlohmann::json;

// 将音频数据按指定时长分段
std::vector<std::vector<double>> split_audio(const std::vector<double>& audio, int sample_rate, double segment_duration) {
    int samples_per_segment = static_cast<int>(sample_rate * segment_duration);
    int num_segments = static_cast<int>(std::ceil(static_cast<double>(audio.size()) / samples_per_segment));
    std::vector<std::vector<double>> segments;

    for (int i = 0; i < num_segments; ++i) {
        int start = i * samples_per_segment;
        int end = std::min(start + samples_per_segment, static_cast<int>(audio.size()));
        segments.push_back(std::vector<double>(audio.begin() + start, audio.begin() + end));
    }

    return segments;
}

int main(int argc, char* argv[]) {
    if (argc != 2) {
        std::cerr << "Usage: " << argv[0] << " <audio_file>" << std::endl;
        return 1;
    }

    // 打开音频文件
    SF_INFO sf_info;
    SNDFILE* file = sf_open(argv[1], SFM_READ, &sf_info);
    if (!file) {
        std::cerr << "Error: Could not open audio file: " << argv[1] << std::endl;
        std::cerr << "Error message: " << sf_strerror(nullptr) << std::endl;
        return 1;
    }

    // 读取音频数据
    std::vector<double> buffer(sf_info.frames * sf_info.channels);
    sf_count_t count = sf_readf_double(file, buffer.data(), sf_info.frames);
    sf_close(file);

    if (count != sf_info.frames) {
        std::cerr << "Error: Could not read all frames from audio file" << std::endl;
        return 1;
    }

    // 分离声道
    std::vector<std::vector<double>> channels(sf_info.channels);
    for (int ch = 0; ch < sf_info.channels; ++ch) {
        channels[ch].resize(sf_info.frames);
        for (sf_count_t i = 0; i < sf_info.frames; ++i) {
            channels[ch][i] = buffer[i * sf_info.channels + ch];
        }
    }

    // 初始化Python解释器
    signalstats_wrapper::initialize();

    // 按0.5秒分段检测每个声道
    const double segment_duration = 0.5; // 分段时长（秒）
    json result;
    result["channels"] = json::array();

    for (int ch = 0; ch < sf_info.channels; ++ch) {
        json channel_result;
        channel_result["channel"] = ch;
        channel_result["segments"] = json::array();

        // 将声道数据分段
        auto segments = split_audio(channels[ch], sf_info.samplerate, segment_duration);

        // 检测每个分段
        for (size_t i = 0; i < segments.size(); ++i) {
            json segment_result;
            json output;
            signalstats_wrapper::detect_signal_wrapper(
                output,
                segments[i],
                sf_info.samplerate,
                3e-3,  // silence_threshold
                -70.0, // db_threshold
                -70.0, // cv_threshold
                {"tukey", "0.25"}, // window_params
                256,   // nperseg
                32,    // noverlap
                256,   // nfft
                false  // debug
            );

            segment_result["start_time"] = i * segment_duration;
            segment_result["end_time"] = std::min((i + 1) * segment_duration, 
                static_cast<double>(sf_info.frames) / sf_info.samplerate);
            segment_result["detection_result"] = output;

            channel_result["segments"].push_back(segment_result);
        }

        result["channels"].push_back(channel_result);
    }

    // 输出检测结果
    std::cout << result.dump(2) << std::endl;
    
    signalstats_wrapper::finalize();

    return 0;
}