AVMutableComposition -How将多个音频记录与1个视频记录合并

Question

我有几个音频剪辑，我用AVAudioRecorder录制了一段视频。使用AVMutableComposition，我希望在录制音频时将音频资产与视频合并。例如，视频是1分钟长，我录制了3个音频剪辑在5-10秒，20-25秒，和30-35秒。音频剪辑应该在这些特定的时间框架内与视频合并。当最后的视频播放时，音频将在这些时间帧上播放。

模型：

class AudioModel {

    var audioUrl: URL?
    var startTime: Double?
    var endTime: Double?
}

混合：

let mixComposition = AVMutableComposition()

guard let videoCompositionTrack = mixComposition.addMutableTrack(withMediaType: .video, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
guard let audioFromVideoCompositionTrack = mixComposition.addMutableTrack(withMediaType: .audio, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
guard let audioModelCompositionTrack = mixComposition.addMutableTrack(withMediaType: .audio, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }

let videoAsset = AVURLAsset(url: videoURL)
guard let videoTrack = videoAsset.tracks(withMediaType: .video).first else { return }

for audioModel in audioModels {

    let audioAsset = AVURLAsset(url: audioModel.url!)
    let startTime = CMTime(seconds: audioModel.startTime!, preferredTimescale: 1000)

    do {

        try videoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: videoAsset.duration), of: videoTrack, at: .zero)
            
        if let audioTrackFromAudioModel = audioAsset.tracks(withMediaType: .audio).first {
                
            try audioModelCompositionTrack.insertTimeRange(CMTimeRangeMake(start: startTime, duration: audioAsset.duration),
                                                               of: audioTrackFromAudioModel, at: .zero)
        }
            
        if let audioFromVideoTrack = videoAsset.tracks(withMediaType: .audio).first {
            try audioFromVideoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: CMTime.zero, duration: videoAsset.duration),
                                                                   of: audioFromVideoTrack, at: .zero)
        }

    } catch {
    }
}

let exporter = AVAssetExportSession(asset: mixComposition, presetName: AVAssetExportPresetHighestQuality)
// ... I know what to do from here

Answer 1

您的方法是正确的，但是您混淆了用于insertTimeRange的两个参数，并且多次从视频轨道中添加视频和音频。

insertTimeRange中的第一个参数是指原始音频资产中的timeRange，而不是组合；因此，假设对于要添加整个剪辑的每个音频剪辑，时间范围应该总是从.zero开始，而不是在startTime。at:参数不应该是.zero，而应该是"startTime“--在要添加音频的组合中的时间。

关于您的视频轨道和audioFromVideoTrack，我不会将它们作为循环的一部分添加，而是在循环之前添加它们。否则，您将多次添加它们(每个音频项一次)，而不仅仅是一次，这可能导致不必要的行为或导出会话完全失败。

我编辑了您的代码，但没有进行实际测试，所以请稍加考虑。

guard let videoCompositionTrack = mixComposition.addMutableTrack(withMediaType: .video, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
guard let audioFromVideoCompositionTrack = mixComposition.addMutableTrack(withMediaType: .audio, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
guard let audioModelCompositionTrack = mixComposition.addMutableTrack(withMediaType: .audio, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }

let videoAsset = AVURLAsset(url: videoURL)
guard let videoTrack = videoAsset.tracks(withMediaType: .video).first else { return }

do {
    try videoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: videoAsset.duration), of: videoTrack, at: .zero)
    if let audioFromVideoTrack = videoAsset.tracks(withMediaType: .audio).first {
        try audioFromVideoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: CMTime.zero, duration: videoAsset.duration), of: audioFromVideoTrack, at: .zero)
    }
} catch {
}

for audioModel in audioModels {
    let audioAsset = AVURLAsset(url: audioModel.url!)
    let startTime = CMTime(seconds: audioModel.startTime!, preferredTimescale: 1000)
    do {
        if let audioTrackFromAudioModel = audioAsset.tracks(withMediaType: .audio).first {
            try audioModelCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: audioAsset.duration), of: audioTrackFromAudioModel, at: startTime)
        }
    } catch {
    }
}

let exporter = AVAssetExportSession(asset: mixComposition, presetName: AVAssetExportPresetHighestQuality)
// ... I know what to do from here