同时使用AVComposition和AVVideoComposition时转换错误

Question

我正在创建一个AVMutableComposition。我需要将资产水平翻转，所以我正在设置如下组合轨道的转换：

compositionTrack.preferredTransform = assetTrack.preferredTransform.scaledBy(x: -1, y: 1)

如果我导出这个(我使用AVAssetPreset960x640作为我的预置)，它就会像预期的那样工作。

不过，我还需要添加一个AVMutableVideoComposition覆盖，以便与此复制呈现。这个覆盖层不应该水平翻转。我是这样说的：

// Create video composition
let videoComposition = AVMutableVideoComposition()
videoComposition.renderSize = videoSize
videoComposition.frameDuration = CMTime(value: 1, timescale: 30)
videoComposition.animationTool = AVVideoCompositionCoreAnimationTool(
    postProcessingAsVideoLayer: videoLayer,
    in: outputLayer
)

let instruction = AVMutableVideoCompositionInstruction()
instruction.timeRange = CMTimeRange(
  start: .zero,
  duration: composition.duration
)
let layerInstruction = AVMutableVideoCompositionLayerInstruction(assetTrack: compositionTrack)
layerInstruction.setTransform(assetTrack.preferredTransform, at: .zero)
    
instruction.layerInstructions = [layerInstruction]
videoComposition.instructions = [instruction]

当我用这个导出视频时，它不是水平翻转的。它阻止应用于组合的preferredTransform被执行。大概是这条线造成了这种情况：

layerInstruction.setTransform(assetTrack.preferredTransform, at: .zero)

如果我将图层转换设置为assetTrack.preferredTransform.scaledBy(x: -1, y: 1)或compositionTrack.preferredTransform，则会在导出时显示一个黑色屏幕。

苹果让这些医生在AVVideoComposition上解释转换。如果我理解正确的话，他们说我应该设置图层指令的转换。如果我这样做--对图层指令而不是构图进行转换--我仍然会看到一个黑色屏幕。

这是为什么，我怎么能解决这个问题？

Answer 1

编辑:成功的转换是

CGAffineTransform(a: -1.0, b: 0.0, c: 0.0, d: 1.0, tx: videoSize.width, ty: 0.0)

因此，我在前一段时间遇到了这个问题，遵循这个RayWenderlich 教程 (您的代码看起来也非常熟悉！)。问题是assetTrack.preferredTransform不可靠，在一些视频中它可以工作，而在另一些视频上，它将是一个黑色的屏幕。

而不是

let layerInstruction = AVMutableVideoCompositionLayerInstruction(assetTrack: compositionTrack)
layerInstruction.setTransform(assetTrack.preferredTransform, at: .zero)

试试这个：

func compositionLayerInstruction(for track: AVCompositionTrack, assetTrack: AVAssetTrack, orientation: UIImage.Orientation) -> AVMutableVideoCompositionLayerInstruction {

    let instruction = AVMutableVideoCompositionLayerInstruction(assetTrack: track)
    
    var transform = CGAffineTransform.identity
    let assetSize = assetTrack.naturalSize
    
    /// you should be able to play with these values to make a horizontal flip (try changing `a` and `d`)
    switch orientation {
    case .up:
        transform = CGAffineTransform(a: 1, b: 0, c: 0, d: 1, tx: 0, ty: 0)
    case .down:
        transform = CGAffineTransform(a: -1, b: 0, c: 0, d: -1, tx: assetSize.width, ty: assetSize.height)
    case .left:
        transform = CGAffineTransform(a: 0, b: -1, c: 1, d: 0, tx: 0, ty: assetSize.width)
    case .right:
        transform = CGAffineTransform(a: 0, b: 1, c: -1, d: 0, tx: assetSize.height, ty: 0)
     default:
        print("Unsupported orientation")
    }
    
    instruction.setTransform(transform, at: .zero)
    
    return instruction
}

let layerInstruction = compositionLayerInstruction(
    for: compositionTrack, assetTrack: assetTrack, orientation: videoInfo.orientation)
/// assetTrack is the original video

我是根据另一个答案得出的，所以，我记不起在哪里了。

下面是一个帮助函数，用于获取orientation

/// adjust the video orientation is the source has a different orientation
private func orientation(from transform: CGAffineTransform) -> (orientation: UIImage.Orientation, isPortrait: Bool) {
    var assetOrientation = UIImage.Orientation.up
    var isPortrait = false
    if transform.a == 0 && transform.b == 1.0 && transform.c == -1.0 && transform.d == 0 {
        assetOrientation = .right
        isPortrait = true
    } else if transform.a == 0 && transform.b == -1.0 && transform.c == 1.0 && transform.d == 0 {
        assetOrientation = .left
        isPortrait = true
    } else if transform.a == 1.0 && transform.b == 0 && transform.c == 0 && transform.d == 1.0 {
        assetOrientation = .up
    } else if transform.a == -1.0 && transform.b == 0 && transform.c == 0 && transform.d == -1.0 {
        assetOrientation = .down
    }
    return (assetOrientation, isPortrait)
}

充分使用：

/// get adjusted orientation of the video
let videoInfo = orientation(from: assetTrack.preferredTransform)
let videoSize: CGSize

if videoInfo.isPortrait {
    videoSize = CGSize(
    width: assetTrack.naturalSize.height,
    height: assetTrack.naturalSize.width)
} else {
    videoSize = assetTrack.naturalSize
}


 /// the video
let videoLayer = CALayer()
videoLayer.frame = CGRect(origin: .zero, size: videoSize)
        
/// overlay where you add graphics and other stuff
let overlayLayer = CALayer()
overlayLayer.frame = CGRect(origin: .zero, size: videoSize)

let outputLayer = CALayer()
outputLayer.frame = CGRect(origin: .zero, size: videoSize)
outputLayer.addSublayer(videoLayer)
outputLayer.addSublayer(overlayLayer)

let videoComposition = AVMutableVideoComposition()
videoComposition.renderSize = videoSize
videoComposition.frameDuration = CMTime(value: 1, timescale: 30)
videoComposition.animationTool = AVVideoCompositionCoreAnimationTool(
postProcessingAsVideoLayer: videoLayer,
in: outputLayer)

// MARK: - previous `let layerInstruction` code goes here...

instruction.layerInstructions = [layerInstruction]

/// export session code here...