将 Metal MTKView 实时捕获为电影？
Posted
技术标签:
【中文标题】将 Metal MTKView 实时捕获为电影？【英文标题】：Capture Metal MTKView as Movie in realtime? 【发布时间】：2017-10-05 21:24:22 【问题描述】：从MTKView 捕获帧的最有效方法是什么？如果可能的话，我想实时保存帧中的 .mov 文件。是否可以渲染成 AVPlayer 框架之类的？
目前正在使用此代码进行绘制（基于@warrenm PerformanceShaders project）：
func draw(in view: MTKView)
_ = inflightSemaphore.wait(timeout: DispatchTime.distantFuture)
updateBuffers()
let commandBuffer = commandQueue.makeCommandBuffer()
commandBuffer.addCompletedHandler [weak self] commandBuffer in
if let strongSelf = self
strongSelf.inflightSemaphore.signal()
// Dispatch the current kernel to perform the selected image filter
selectedKernel.encode(commandBuffer: commandBuffer,
sourceTexture: kernelSourceTexture!,
destinationTexture: kernelDestTexture!)
if let renderPassDescriptor = view.currentRenderPassDescriptor, let currentDrawable = view.currentDrawable
let clearColor = MTLClearColor(red: 0, green: 0, blue: 0, alpha: 1)
renderPassDescriptor.colorAttachments[0].clearColor = clearColor
let renderEncoder = commandBuffer.makeRenderCommandEncoder(descriptor: renderPassDescriptor)
renderEncoder.label = "Main pass"
renderEncoder.pushDebugGroup("Draw textured square")
renderEncoder.setFrontFacing(.counterClockwise)
renderEncoder.setCullMode(.back)
renderEncoder.setRenderPipelineState(pipelineState)
renderEncoder.setVertexBuffer(vertexBuffer, offset: MBEVertexDataSize * bufferIndex, at: 0)
renderEncoder.setVertexBuffer(uniformBuffer, offset: MBEUniformDataSize * bufferIndex , at: 1)
renderEncoder.setFragmentTexture(kernelDestTexture, at: 0)
renderEncoder.setFragmentSamplerState(sampler, at: 0)
renderEncoder.drawPrimitives(type: .triangleStrip, vertexStart: 0, vertexCount: 4)
renderEncoder.popDebugGroup()
renderEncoder.endEncoding()
commandBuffer.present(currentDrawable)
bufferIndex = (bufferIndex + 1) % MBEMaxInflightBuffers
commandBuffer.commit()
【问题讨论】：
这当然是可能的。由于在 ios 上从IOSurface 创建 CVPixelBuffer 不是受支持的操作，因此最快的方法可能是使用 AVAssetWriterInputPixelBufferAdaptor（包装像素缓冲池），您可以在其中读取内容每一帧的drawable的纹理。有一些示例说明如何使用浮动的 OpenGL ES 执行此操作。我最终会尝试发布一个包含所有细节的更完整的答案，但也许这会对你有所帮助。另一个要考虑的选择是 ReplayKit，如果它足够灵活，可以做你想做的事。
【参考方案1】：
这是一个小类，它执行编写捕获 Metal 视图内容的电影文件的基本功能：
class MetalVideoRecorder
var isRecording = false
var recordingStartTime = TimeInterval(0)
private var assetWriter: AVAssetWriter
private var assetWriterVideoInput: AVAssetWriterInput
private var assetWriterPixelBufferInput: AVAssetWriterInputPixelBufferAdaptor
init?(outputURL url: URL, size: CGSize)
do
assetWriter = try AVAssetWriter(outputURL: url, fileType: .m4v)
catch
return nil
let outputSettings: [String: Any] = [ AVVideoCodecKey : AVVideoCodecType.h264,
AVVideoWidthKey : size.width,
AVVideoHeightKey : size.height ]
assetWriterVideoInput = AVAssetWriterInput(mediaType: .video, outputSettings: outputSettings)
assetWriterVideoInput.expectsMediaDataInRealTime = true
let sourcePixelBufferAttributes: [String: Any] = [
kCVPixelBufferPixelFormatTypeKey as String : kCVPixelFormatType_32BGRA,
kCVPixelBufferWidthKey as String : size.width,
kCVPixelBufferHeightKey as String : size.height ]
assetWriterPixelBufferInput = AVAssetWriterInputPixelBufferAdaptor(assetWriterInput: assetWriterVideoInput,
sourcePixelBufferAttributes: sourcePixelBufferAttributes)
assetWriter.add(assetWriterVideoInput)
func startRecording()
assetWriter.startWriting()
assetWriter.startSession(atSourceTime: .zero)
recordingStartTime = CACurrentMediaTime()
isRecording = true
func endRecording(_ completionHandler: @escaping () -> ())
isRecording = false
assetWriterVideoInput.markAsFinished()
assetWriter.finishWriting(completionHandler: completionHandler)
func writeFrame(forTexture texture: MTLTexture)
if !isRecording
return
while !assetWriterVideoInput.isReadyForMoreMediaData
guard let pixelBufferPool = assetWriterPixelBufferInput.pixelBufferPool else
print("Pixel buffer asset writer input did not have a pixel buffer pool available; cannot retrieve frame")
return
var maybePixelBuffer: CVPixelBuffer? = nil
let status = CVPixelBufferPoolCreatePixelBuffer(nil, pixelBufferPool, &maybePixelBuffer)
if status != kCVReturnSuccess
print("Could not get pixel buffer from asset writer input; dropping frame...")
return
guard let pixelBuffer = maybePixelBuffer else return
CVPixelBufferLockBaseAddress(pixelBuffer, [])
let pixelBufferBytes = CVPixelBufferGetBaseAddress(pixelBuffer)!
// Use the bytes per row value from the pixel buffer since its stride may be rounded up to be 16-byte aligned
let bytesPerRow = CVPixelBufferGetBytesPerRow(pixelBuffer)
let region = MTLRegionMake2D(0, 0, texture.width, texture.height)
texture.getBytes(pixelBufferBytes, bytesPerRow: bytesPerRow, from: region, mipmapLevel: 0)
let frameTime = CACurrentMediaTime() - recordingStartTime
let presentationTime = CMTimeMakeWithSeconds(frameTime, preferredTimescale: 240)
assetWriterPixelBufferInput.append(pixelBuffer, withPresentationTime: presentationTime)
CVPixelBufferUnlockBaseAddress(pixelBuffer, [])
在初始化其中一个并调用 startRecording() 后，您可以将调度的处理程序添加到包含渲染命令的命令缓冲区并调用 writeFrame（在结束编码之后，但在呈现可绘制对象或提交缓冲区之前）：
let texture = currentDrawable.texture
commandBuffer.addCompletedHandler commandBuffer in
self.recorder.writeFrame(forTexture: texture)
录制完成后，只需拨打endRecording，视频文件将完成并关闭。
注意事项：
此类假定源纹理为默认格式.bgra8Unorm。如果不是，您将遇到崩溃或损坏。如有必要，使用计算或片段着色器转换纹理，或使用 Accelerate。
该类还假设纹理与视频帧的大小相同。如果不是这种情况（如果可绘制大小发生变化，或者您的屏幕自动旋转），输出将被破坏并且您可能会看到崩溃。根据您的应用程序需要，通过缩放或裁剪源纹理来缓解此问题。
【讨论】：
你好沃伦。我有一个问题可能是因为pixelFormat。我在这里有一个问题：***.com/questions/46146593/… 和 github 上的一个问题：github.com/svtek/SceneKitVideoRecorder/issues/3 你为什么用'addScheduledHandler'而不是'addCompletedHandler'？ 我认为使用调度处理程序是不正确的，因为该块将在 GPU 执行命令缓冲区之前运行。您需要使用完成处理程序。 还应该注意MTKView的属性framebufferOnly应该设置为false。由于这个原因，我最近遇到了问题，并且在视频中出现了伪影/崩溃 我在texture.getBytes(...) 收到Thread 1: EXC_BAD_ACCESS。有任何想法吗？花了几天时间尝试调试。【参考方案2】：
升级到 Swift 5
import AVFoundation
class MetalVideoRecorder
var isRecording = false
var recordingStartTime = TimeInterval(0)
private var assetWriter: AVAssetWriter
private var assetWriterVideoInput: AVAssetWriterInput
private var assetWriterPixelBufferInput: AVAssetWriterInputPixelBufferAdaptor
init?(outputURL url: URL, size: CGSize)
do
assetWriter = try AVAssetWriter(outputURL: url, fileType: AVFileType.m4v)
catch
return nil
let outputSettings: [String: Any] = [ AVVideoCodecKey : AVVideoCodecType.h264,
AVVideoWidthKey : size.width,
AVVideoHeightKey : size.height ]
assetWriterVideoInput = AVAssetWriterInput(mediaType: AVMediaType.video, outputSettings: outputSettings)
assetWriterVideoInput.expectsMediaDataInRealTime = true
let sourcePixelBufferAttributes: [String: Any] = [
kCVPixelBufferPixelFormatTypeKey as String : kCVPixelFormatType_32BGRA,
kCVPixelBufferWidthKey as String : size.width,
kCVPixelBufferHeightKey as String : size.height ]
assetWriterPixelBufferInput = AVAssetWriterInputPixelBufferAdaptor(assetWriterInput: assetWriterVideoInput,
sourcePixelBufferAttributes: sourcePixelBufferAttributes)
assetWriter.add(assetWriterVideoInput)
func startRecording()
assetWriter.startWriting()
assetWriter.startSession(atSourceTime: CMTime.zero)
recordingStartTime = CACurrentMediaTime()
isRecording = true
func endRecording(_ completionHandler: @escaping () -> ())
isRecording = false
assetWriterVideoInput.markAsFinished()
assetWriter.finishWriting(completionHandler: completionHandler)
func writeFrame(forTexture texture: MTLTexture)
if !isRecording
return
while !assetWriterVideoInput.isReadyForMoreMediaData
guard let pixelBufferPool = assetWriterPixelBufferInput.pixelBufferPool else
print("Pixel buffer asset writer input did not have a pixel buffer pool available; cannot retrieve frame")
return
var maybePixelBuffer: CVPixelBuffer? = nil
let status = CVPixelBufferPoolCreatePixelBuffer(nil, pixelBufferPool, &maybePixelBuffer)
if status != kCVReturnSuccess
print("Could not get pixel buffer from asset writer input; dropping frame...")
return
guard let pixelBuffer = maybePixelBuffer else return
CVPixelBufferLockBaseAddress(pixelBuffer, [])
let pixelBufferBytes = CVPixelBufferGetBaseAddress(pixelBuffer)!
// Use the bytes per row value from the pixel buffer since its stride may be rounded up to be 16-byte aligned
let bytesPerRow = CVPixelBufferGetBytesPerRow(pixelBuffer)
let region = MTLRegionMake2D(0, 0, texture.width, texture.height)
texture.getBytes(pixelBufferBytes, bytesPerRow: bytesPerRow, from: region, mipmapLevel: 0)
let frameTime = CACurrentMediaTime() - recordingStartTime
let presentationTime = CMTimeMakeWithSeconds(frameTime, preferredTimescale: 240)
assetWriterPixelBufferInput.append(pixelBuffer, withPresentationTime: presentationTime)
CVPixelBufferUnlockBaseAddress(pixelBuffer, [])
【讨论】：
【参考方案3】：我正在使用这篇文章来记录自定义金属视图，但我遇到了一些问题。当我开始录制时，我在 iPhone 12 Pro Max 上从 60fps 提高到 ~20fps。在 Profiling 之后，减慢一切的函数是 texture.getBytes，因为它正在将缓冲区从 GPU 抓取到 CPU。
另一个问题是视频和音频不同步，不确定是否会导致这种情况。我不确定我是否应该进入信号量路线来解决这个问题，或者还有其他潜在的解决方法。
在我的例子中，纹理大小与屏幕大小一样大，因为我从相机流创建它，然后通过几个 CIFilter 对其进行处理。我不确定问题是否在于它太大，所以 getBytes 无法实时支持这种大小的纹理。
如果我需要定义优先级，我的第一要务是解决音频和视频之间的不同步问题。任何想法都会非常有帮助。
【讨论】：
@warrenm 也许你对此有一些想法 您似乎发布了一个问题作为答案；您可能应该发布自己的问题。以上是关于将 Metal MTKView 实时捕获为电影？的主要内容，如果未能解决你的问题，请参考以下文章
如何将 Vignette CIFilter 应用于 iOS 中的实时摄像机源？