C# FFmpeg合併多個音頻文件，並指定每段音頻的插入時間。

合成音頻的FFmpeg命令爲：

ffmpeg -i 11.wav -i 13.wav -filter_complex "[0]adelay=1s:all=1[0a];[1]adelay=26s:all=1[1a];[0a][1a]amix=inputs=2[a]" -map "[a]" output.wav

不知道具體啥意思，大概猜測一下，
-i <fileName> 輸入文件的路徑，有幾個文件就幾個-i命令
[0]adelay=1s:all=1[0a] [0]和[0a]指的是第幾個文件，索引從0開始，adelay=1s 從音頻的第一秒開始插入（4.2版本以下不支持1s這種寫法），all：Use last set delay for all remaining channels. By default is disabled. This option if enabled changes how option delays is interpreted.
amix=inputs=2 2指的是一共有幾個文件
output.wav 輸出文件的路徑或者文件名
文件名都支持全路徑的寫法。可以是 C:\Data\CHN\Other\FFmpegMergeMp3\11.wav 也可以是 11.wav

ffmpeg版本：https://www.gyan.dev/ffmpeg/builds/ffmpeg-git-essentials.7z
latest git master branch build version: 2021-11-25-git-522f577d7e
使用 <package id="Xabe.FFmpeg" version="5.0.2" targetFramework="net48" /> 包來執行命令

public static string FFMPEGPath { get; set; } = @"C:\Data\PATHVariables";

static async Task Main(string[] args)
{
    string inputFile1 = @"C:\Data\CHN\Other\FFmpegMergeMp3\11.wav";
    string inputFile3 = @"C:\Data\CHN\Other\FFmpegMergeMp3\13.wav";
    string outputPath = $@"C:\Data\CHN\Other\FFmpegMergeMp3\{Guid.NewGuid()}.wav";

    var listFile = new List<TextToSppechConversionModel>();
    listFile.Add(new TextToSppechConversionModel() { Title = "Section 1", AudioFilePath = inputFile1, StartTime = "00:0:05" });
    listFile.Add(new TextToSppechConversionModel() { Title = "Section 3", AudioFilePath = inputFile3, StartTime = "00:0:15" });

    await new Program().CombineAudioToOneFile(listFile, outputPath);
}

public async Task CombineAudioToOneFile(List<TextToSppechConversionModel> listSourceFile, string outputPath)
{
    FFmpeg.SetExecutablesPath(FFMPEGPath, "ffmpeg", "ffprobe");

    //https://stackoverflow.com/questions/60027460/how-to-add-multiple-audio-files-at-specific-times-on-a-silence-audio-file-using
    //https://www.ffmpeg.org/ffmpeg-all.html#adelay
    //https://ffmpeg.org/ffmpeg-filters.html#amix
    //string inputFile1 = @"C:\Data\CHN\Other\FFmpegMergeMp3\11.wav";
    //string inputFile3 = @"C:\Data\CHN\Other\FFmpegMergeMp3\13.wav";
    //outputPath = $@"C:\Data\CHN\Other\FFmpegMergeMp3\{Guid.NewGuid()}.wav";

    //var mediaFile1 = await FFmpeg.GetMediaInfo(inputFile1);
    //var mediaFile3 = await FFmpeg.GetMediaInfo(inputFile3);

    //var audioStream1 = mediaFile1.AudioStreams.FirstOrDefault();
    //var audioStream3 = mediaFile3.AudioStreams.FirstOrDefault();
    //audioStream1 = audioStream1.SetCodec(AudioCodec.aac);
    //audioStream3 = audioStream3.SetCodec(AudioCodec.aac);

    //var conversion = FFmpeg.Conversions.New().AddStream(new List<IStream>() { audioStream1, audioStream3 }).SetOutput(outputPath);
    /*
        * ffmpeg -i 11.wav -i 13.wav -filter_complex \"[0]adelay=1s:all=1[0a];[1]adelay=26s:all=1[1a];[0a][1a]amix=inputs=2[a]\" 
        * -map \"[a]\" output11132.wav
        */
    //string param = $" -i {inputFile1} -i {inputFile3} -filter_complex \"[0]adelay=1s:all=1[0a];[1]adelay=26s:all=1[1a];[0a][1a]amix=inputs=2[a]\" -map \"[a]\" {outputPath}";
    //var conversionResult = await conversion.Start(param);

    StringBuilder sbInputFile = new StringBuilder();
    StringBuilder sbAdelay = new StringBuilder();
    var listStream = new List<IStream>();
    var strA = "";
    for (int i = 0; i < listSourceFile.Count; i++)
    {
        var sourceModel = listSourceFile[i];
        if (!File.Exists(sourceModel.AudioFilePath))
            throw new Exception($"the file {sourceModel.AudioFilePath} not found.");

        var mediaFile1 = await FFmpeg.GetMediaInfo(sourceModel.AudioFilePath);
        var audioStream1 = mediaFile1.AudioStreams.FirstOrDefault().SetCodec(AudioCodec.ac3);
        listStream.Add(audioStream1);
        //sourceModel.StartTime;//"00:01:02"
        var timeSpan = DateTime.Parse(sourceModel.StartTime) - DateTime.Parse("00:00:00");
        var startSeconds = timeSpan.TotalSeconds; //seconds
        //check current length
        if (i + 1 < listSourceFile.Count)
        {
            //get next start time
            var nextStartTime = listSourceFile[i + 1].StartTime;
            var timeSpanNext = DateTime.Parse(nextStartTime) - DateTime.Parse("00:00:00");
            var currEndTime = mediaFile1.Duration.TotalSeconds + timeSpan.TotalSeconds;
            if (currEndTime > timeSpanNext.TotalSeconds)
            {
                var endTimeStr = DateTime.Parse("00:00:00").AddSeconds(currEndTime).ToString("HH:mm:ss");
                throw new Exception($"the section [{sourceModel.Title}] media end time({endTimeStr}) is greater than next section[{listSourceFile[i + 1].Title}] start time({nextStartTime})");
            }
        }
        sbInputFile.Append($" -i {sourceModel.AudioFilePath}");
        //[0]adelay=1s:all=1[0a];
        //[1]adelay=26s:all=1[1a];
        //[0a][1a]
        sbAdelay.Append($"[{i}]adelay={startSeconds}s:all=1[{i}a];");
        strA += $"[{i}a]";
    }
    //param
    /*
        * ffmpeg -i 11.wav -i 13.wav -filter_complex \"[0]adelay=1s:all=1[0a];[1]adelay=26s:all=1[1a];[0a][1a]amix=inputs=2[a]\" 
        * -map \"[a]\" output11132.wav
        */
    string param = $"{sbInputFile} -filter_complex \"{sbAdelay}{strA}amix=inputs={listSourceFile.Count}[a]\" -map \"[a]\" {outputPath}";
    var conversion = FFmpeg.Conversions.New().AddStream(listStream).SetOutput(outputPath);
    var conversionResult = await conversion.Start(param);
    //return conversionResult.Duration;
}

internal class TextToSppechConversionModel
{
    public string Title { get; set; }
    public string AudioFilePath { get; set; }
    public string StartTime { get; set; }
}

參考：
//https://stackoverflow.com/questions/60027460/how-to-add-multiple-audio-files-at-specific-times-on-a-silence-audio-file-using
//https://www.ffmpeg.org/ffmpeg-all.html#adelay
//https://ffmpeg.org/ffmpeg-filters.html#amix
//https://ffmpeg.xabe.net/docs.html
//https://ffmpeg.xabe.net/tutorial.html