Menu
I am making a sample application that utilizes AVFoundation to record video. The whole point is so I can have more control over how the video is recorded. In my sample project I have the video capturing but am struggling with handling orientation correctly.
I have done a lot of searching around the web and found that others are suggesting that I should NOT allow my capture view or capture session to rotate based on orientation, but rather set a transformation to rotate the video during playback. I have this working fine on iOS and Mac devices, but am wondering if I will have issues on other platforms such as Windows or Android.
Also, when I view the recorded video's metadata I see that the width and height are not set properly for the orientation. This makes sense as I am only transforming the presentation of the video and not it's actual resolution.
My question here is how do I correctly support portrait and landscape orientations and have it reflected correctly in the video file output? I need these videos to play on all platforms correctly so I am thinking the resolution is going to matter a great deal.
Below is the full source I have written thus far. I appreciate any advice you all can provide.
import UIKit
import AVFoundation
class ViewController: UIViewController, AVCaptureVideoDataOutputSampleBufferDelegate, AVCaptureAudioDataOutputSampleBufferDelegate {
//MARK: - Outlet
@IBOutlet weak var previewView: UIView!
@IBOutlet var playStopButton: UIButton!
//MARK: - Private Variables
let videoQueue = DispatchQueue(label: "VideoQueue", qos: .background, attributes: .concurrent, autoreleaseFrequency: .workItem, target: nil)
private let captureSession = AVCaptureSession()
var outputUrl: URL {
get {
if let url = _outputUrl {
return url
}
_outputUrl = outputDirectory.appendingPathComponent("video.mp4")
return _outputUrl!
}
}
private var _outputUrl: URL?
var outputDirectory: URL {
get {
if let url = _outputDirectory {
return url
}
_outputDirectory = getDocumentsDirectory().appendingPathComponent("recording")
return _outputDirectory!
}
}
private var _outputDirectory: URL?
private var assetWriter: AVAssetWriter?
private var videoInput: AVAssetWriterInput?
private var audioInput: AVAssetWriterInput?
private var videoOutput: AVCaptureVideoDataOutput?
private var audioOutput: AVCaptureAudioDataOutput?
private var isRecording = false
private var isWriting = false
private var videoSize = CGSize(width: 640, height: 480)
//MARK: - View Life-cycle
override func viewDidLoad() {
super.viewDidLoad()
videoQueue.async {
do {
try self.configureCaptureSession()
try self.configureAssetWriter()
DispatchQueue.main.async {
self.configurePreview()
}
} catch {
DispatchQueue.main.async {
self.showAlert("Unable to configure video output")
}
}
}
}
override var preferredInterfaceOrientationForPresentation: UIInterfaceOrientation {
return .portrait
}
override var supportedInterfaceOrientations: UIInterfaceOrientationMask {
return .portrait
}
//MARK: - Capture Session
private func configureCaptureSession() throws {
do {
// configure the session
if captureSession.canSetSessionPreset(AVCaptureSessionPreset640x480) {
captureSession.sessionPreset = AVCaptureSessionPreset640x480
}
// configure capture devices
let camDevice = AVCaptureDevice.defaultDevice(withMediaType: AVMediaTypeVideo)
let micDevice = AVCaptureDevice.defaultDevice(withMediaType: AVMediaTypeAudio)
let camInput = try AVCaptureDeviceInput(device: camDevice)
let micInput = try AVCaptureDeviceInput(device: micDevice)
if captureSession.canAddInput(camInput) {
captureSession.addInput(camInput)
}
if captureSession.canAddInput(micInput) {
captureSession.addInput(micInput)
}
// configure audio/video output
videoOutput = AVCaptureVideoDataOutput()
videoOutput?.alwaysDiscardsLateVideoFrames = false // TODO: is this necessary?
videoOutput?.setSampleBufferDelegate(self, queue: videoQueue)
if let v = videoOutput {
captureSession.addOutput(v)
}
audioOutput = AVCaptureAudioDataOutput()
audioOutput?.setSampleBufferDelegate(self, queue: videoQueue)
if let a = audioOutput {
captureSession.addOutput(a)
}
// configure audio session
let audioSession = AVAudioSession.sharedInstance()
try audioSession.setCategory(AVAudioSessionCategoryPlayAndRecord)
try audioSession.setActive(true)
var micPort: AVAudioSessionPortDescription?
if let inputs = audioSession.availableInputs {
for port in inputs {
if port.portType == AVAudioSessionPortBuiltInMic {
micPort = port
break;
}
}
}
if let port = micPort, let dataSources = port.dataSources {
for source in dataSources {
if source.orientation == AVAudioSessionOrientationFront {
try audioSession.setPreferredInput(port)
break
}
}
}
} catch {
print("Failed to configure audio/video capture session")
throw error
}
}
private func configureAssetWriter() throws {
prepareVideoFile()
do {
assetWriter = try AVAssetWriter(url: outputUrl, fileType: AVFileTypeMPEG4)
guard let writer = assetWriter else {
print("Asset writer not created")
return
}
let vidSize = videoSize
let videoSettings: [String: Any] = [AVVideoCodecKey: AVVideoCodecH264,
AVVideoWidthKey: NSNumber(value: Float(vidSize.width)),
AVVideoHeightKey: NSNumber(value: Float(vidSize.height))]
videoInput = AVAssetWriterInput(mediaType: AVMediaTypeVideo, outputSettings: videoSettings)
videoInput?.expectsMediaDataInRealTime = true
videoInput?.transform = getVideoTransform()
var channelLayout = AudioChannelLayout()
memset(&channelLayout, 0, MemoryLayout<AudioChannelLayout>.size);
channelLayout.mChannelLayoutTag = kAudioChannelLayoutTag_Stereo;
let audioSettings: [String: Any] = [AVFormatIDKey: kAudioFormatMPEG4AAC,
AVSampleRateKey: 44100,
AVNumberOfChannelsKey: 2]
audioInput = AVAssetWriterInput(mediaType: AVMediaTypeAudio, outputSettings: audioSettings)
audioInput?.expectsMediaDataInRealTime = true
guard let vi = videoInput else {
print("Video input not configured")
return
}
guard let ai = audioInput else {
print("Audio input not configured")
return
}
if writer.canAdd(vi) {
writer.add(vi)
}
if writer.canAdd(ai) {
writer.add(ai)
}
} catch {
print("Failed to configure asset writer")
throw error
}
}
private func prepareVideoFile() {
if FileManager.default.fileExists(atPath: outputUrl.path) {
do {
try FileManager.default.removeItem(at: outputUrl)
} catch {
print("Unable to remove file at URL \(outputUrl)")
}
}
if !FileManager.default.fileExists(atPath: outputDirectory.path) {
do {
try FileManager.default.createDirectory(at: outputDirectory, withIntermediateDirectories: true, attributes: nil)
} catch {
print("Unable to create directory at URL \(outputDirectory)")
}
}
}
private func configurePreview() {
if let previewLayer = AVCaptureVideoPreviewLayer(session: captureSession) {
previewLayer.frame = previewView.bounds
previewView.layer.addSublayer(previewLayer)
}
}
private func getVideoSize() -> CGSize {
if UIDevice.current.orientation == .landscapeLeft || UIDevice.current.orientation == .landscapeRight {
if videoSize.width > videoSize.height {
return videoSize
} else {
return CGSize(width: videoSize.height, height: videoSize.width)
}
} else {
if videoSize.width < videoSize.height {
return videoSize
} else {
return CGSize(width: videoSize.height, height: videoSize.width)
}
}
}
private func getVideoTransform() -> CGAffineTransform {
switch UIDevice.current.orientation {
case .portraitUpsideDown:
return CGAffineTransform(rotationAngle: CGFloat((M_PI * -90.0)) / 180.0)
case .landscapeLeft:
return CGAffineTransform(rotationAngle: CGFloat((M_PI * -180.0)) / 180.0) // TODO: Add support for front facing camera
// return CGAffineTransform(rotationAngle: CGFloat((M_PI * 0.0)) / 180.0) // TODO: For front facing camera
case .landscapeRight:
return CGAffineTransform(rotationAngle: CGFloat((M_PI * 0.0)) / 180.0) // TODO: Add support for front facing camera
// return CGAffineTransform(rotationAngle: CGFloat((M_PI * -180.0)) / 180.0) // TODO: For front facing camera
default:
return CGAffineTransform(rotationAngle: CGFloat((M_PI * 90.0)) / 180.0)
}
}
//MARK: - Controls
private func startRecording() {
videoQueue.async {
self.captureSession.startRunning()
}
isRecording = true
playStopButton.setTitle("Stop Recording", for: .normal)
print("Recording did start")
}
private func stopRecording() {
if !isRecording {
return
}
videoQueue.async {
self.assetWriter?.finishWriting {
print("Asset writer did finish writing")
self.isWriting = false
}
self.captureSession.stopRunning()
}
isRecording = false
playStopButton.setTitle("Start Recording", for: .normal)
print("Recording did stop")
}
//MARK: - AVCaptureVideoDataOutputSampleBufferDelegate
func captureOutput(_ captureOutput: AVCaptureOutput!, didOutputSampleBuffer sampleBuffer: CMSampleBuffer!, from connection: AVCaptureConnection!) {
guard let w = assetWriter else {
print("Asset writer not configured")
return
}
guard let vo = videoOutput else {
print("Video output not configured")
return
}
guard let ao = audioOutput else {
print("Audio output not configured")
return
}
guard let vi = videoInput else {
print("Video input not configured")
return
}
guard let ai = audioInput else {
print("Audio input not configured")
return
}
let st = CMSampleBufferGetPresentationTimeStamp(sampleBuffer)
print("Writer status \(w.status.rawValue)")
if let e = w.error {
print("Writer error \(e)")
stopRecording()
return
}
switch w.status {
case .unknown:
if !isWriting {
isWriting = true
w.startWriting()
w.startSession(atSourceTime: st)
}
return
case .completed:
print("Video writing completed")
return
case .cancelled:
print("Video writing cancelled")
return
case .failed:
print("Video writing failed")
return
default:
print("Video is writing")
}
if vo == captureOutput {
if !vi.append(sampleBuffer) {
print("Unable to write to video buffer")
}
} else if ao == captureOutput {
if !ai.append(sampleBuffer) {
print("Unable to write to audio buffer")
}
}
}
//MARK: Helpers
private func getDocumentsDirectory() -> URL {
let paths = FileManager.default.urls(for: .documentDirectory, in: .userDomainMask)
let documentsDirectory = paths[0]
return documentsDirectory
}
//MARK: Actions
@IBAction func startStopTapped(sender: AnyObject) {
if isRecording {
stopRecording()
} else {
startRecording()
}
}
}
259
Video orientation is handled by the AVAssetWriterInput.transform, looks like the getVideoTransform() implementation is not correct - CGAffineTransform expects the rotation angle to be in radians, so need to change to something like this:
private func getVideoTransform() -> CGAffineTransform {
switch UIDevice.current.orientation {
case .portrait:
return .identity
case .portraitUpsideDown:
return CGAffineTransform(rotationAngle: .pi)
case .landscapeLeft:
return CGAffineTransform(rotationAngle: .pi/2)
case .landscapeRight:
return CGAffineTransform(rotationAngle: -.pi/2)
default:
return .identity
}
}
From Apple Technical Q&A: https://developer.apple.com/library/archive/qa/qa1744/_index.html
If you are using an
AVAssetWriterobject to write a movie file, you can use thetransformproperty of the associatedAVAssetWriterInputto specify the output file orientation. This will write a display transform property into the output file as the preferred transformation of the visual media data for display purposes. See theAVAssetWriterInput.hinterface file for the details.
155
I have found a solution to my problem. The solution is to export the video using AVAssetExportSession to handle setting the video size and then handling the rotation at the time of export and not during recording. I still have an issue were I need to fix the scale factor to go from my original video size to a smaller 640x480 resolution, but at least I solved my rotation issues. Please see updated code below.
import UIKit
import AVFoundation
class ViewController: UIViewController, AVCaptureVideoDataOutputSampleBufferDelegate, AVCaptureAudioDataOutputSampleBufferDelegate {
//MARK: - Outlet
@IBOutlet weak var previewView: UIView!
@IBOutlet var playStopButton: UIButton!
//MARK: - Private Variables
let videoQueue = DispatchQueue(label: "VideoQueue", qos: .background, attributes: .concurrent, autoreleaseFrequency: .workItem, target: nil)
private let captureSession = AVCaptureSession()
var outputUrl: URL {
get {
if let url = _outputUrl {
return url
}
_outputUrl = outputDirectory.appendingPathComponent("video.mp4")
return _outputUrl!
}
}
private var _outputUrl: URL?
var exportUrl: URL {
get {
if let url = _exportUrl {
return url
}
_exportUrl = outputDirectory.appendingPathComponent("video_encoded.mp4")
return _exportUrl!
}
}
private var _exportUrl: URL?
var outputDirectory: URL {
get {
if let url = _outputDirectory {
return url
}
_outputDirectory = getDocumentsDirectory().appendingPathComponent("recording")
return _outputDirectory!
}
}
private var _outputDirectory: URL?
private var assetWriter: AVAssetWriter?
private var videoInput: AVAssetWriterInput?
private var audioInput: AVAssetWriterInput?
private var videoOutput: AVCaptureVideoDataOutput?
private var audioOutput: AVCaptureAudioDataOutput?
private var isRecording = false
private var isWriting = false
private var videoSize = CGSize(width: 640, height: 480)
private var exportPreset = AVAssetExportPreset640x480
//MARK: - View Life-cycle
override func viewDidLoad() {
super.viewDidLoad()
videoQueue.async {
do {
try self.configureCaptureSession()
DispatchQueue.main.sync {
self.configurePreview()
}
} catch {
DispatchQueue.main.async {
self.showAlert("Unable to configure capture session")
}
}
}
}
override var preferredInterfaceOrientationForPresentation: UIInterfaceOrientation {
return .portrait
}
override var supportedInterfaceOrientations: UIInterfaceOrientationMask {
return .portrait
}
//MARK: - Capture Session
private func configureCaptureSession() throws {
do {
// configure capture devices
let camDevice = AVCaptureDevice.defaultDevice(withMediaType: AVMediaTypeVideo)
let micDevice = AVCaptureDevice.defaultDevice(withMediaType: AVMediaTypeAudio)
let camInput = try AVCaptureDeviceInput(device: camDevice)
let micInput = try AVCaptureDeviceInput(device: micDevice)
if captureSession.canAddInput(camInput) {
captureSession.addInput(camInput)
}
if captureSession.canAddInput(micInput) {
captureSession.addInput(micInput)
}
// configure audio/video output
videoOutput = AVCaptureVideoDataOutput()
videoOutput?.alwaysDiscardsLateVideoFrames = false // TODO: is this necessary?
videoOutput?.setSampleBufferDelegate(self, queue: videoQueue)
if let v = videoOutput {
captureSession.addOutput(v)
}
audioOutput = AVCaptureAudioDataOutput()
audioOutput?.setSampleBufferDelegate(self, queue: videoQueue)
if let a = audioOutput {
captureSession.addOutput(a)
}
// configure audio session
let audioSession = AVAudioSession.sharedInstance()
try audioSession.setCategory(AVAudioSessionCategoryPlayAndRecord)
try audioSession.setActive(true)
var micPort: AVAudioSessionPortDescription?
if let inputs = audioSession.availableInputs {
for port in inputs {
if port.portType == AVAudioSessionPortBuiltInMic {
micPort = port
break;
}
}
}
if let port = micPort, let dataSources = port.dataSources {
for source in dataSources {
if source.orientation == AVAudioSessionOrientationFront {
try audioSession.setPreferredInput(port)
break
}
}
}
} catch {
print("Failed to configure audio/video capture session")
throw error
}
}
private func configureAssetWriter() throws {
prepareVideoFile()
do {
if assetWriter != nil {
assetWriter = nil
videoInput = nil
audioInput = nil
}
assetWriter = try AVAssetWriter(url: outputUrl, fileType: AVFileTypeMPEG4)
guard let writer = assetWriter else {
print("Asset writer not created")
return
}
let videoSettings: [String: Any] = [AVVideoCodecKey: AVVideoCodecH264,
AVVideoWidthKey: NSNumber(value: Float(videoSize.width)),
AVVideoHeightKey: NSNumber(value: Float(videoSize.height))]
videoInput = AVAssetWriterInput(mediaType: AVMediaTypeVideo, outputSettings: videoSettings)
videoInput?.expectsMediaDataInRealTime = true
var channelLayout = AudioChannelLayout()
memset(&channelLayout, 0, MemoryLayout<AudioChannelLayout>.size);
channelLayout.mChannelLayoutTag = kAudioChannelLayoutTag_Stereo;
let audioSettings: [String: Any] = [AVFormatIDKey: kAudioFormatMPEG4AAC,
AVSampleRateKey: 44100,
AVNumberOfChannelsKey: 2]
audioInput = AVAssetWriterInput(mediaType: AVMediaTypeAudio, outputSettings: audioSettings)
audioInput?.expectsMediaDataInRealTime = true
guard let vi = videoInput else {
print("Video input not configured")
return
}
guard let ai = audioInput else {
print("Audio input not configured")
return
}
if writer.canAdd(vi) {
writer.add(vi)
}
if writer.canAdd(ai) {
writer.add(ai)
}
} catch {
print("Failed to configure asset writer")
throw error
}
}
private func prepareVideoFile() {
if FileManager.default.fileExists(atPath: outputUrl.path) {
do {
try FileManager.default.removeItem(at: outputUrl)
} catch {
print("Unable to remove file at URL \(outputUrl)")
}
}
if !FileManager.default.fileExists(atPath: outputDirectory.path) {
do {
try FileManager.default.createDirectory(at: outputDirectory, withIntermediateDirectories: true, attributes: nil)
} catch {
print("Unable to create directory at URL \(outputDirectory)")
}
}
}
private func configurePreview() {
if let previewLayer = AVCaptureVideoPreviewLayer(session: captureSession) {
previewLayer.frame = previewView.bounds
previewView.layer.addSublayer(previewLayer)
}
}
private func getVideoSize() -> CGSize {
if UIDevice.current.orientation == .landscapeLeft || UIDevice.current.orientation == .landscapeRight {
if videoSize.width > videoSize.height {
return videoSize
} else {
return CGSize(width: videoSize.height, height: videoSize.width)
}
} else {
if videoSize.width < videoSize.height {
return videoSize
} else {
return CGSize(width: videoSize.height, height: videoSize.width)
}
}
}
//MARK: - Controls
private func startRecording() {
videoQueue.async {
do {
try self.configureAssetWriter()
self.captureSession.startRunning()
} catch {
print("Unable to start recording")
DispatchQueue.main.async { self.showAlert("Unable to start recording") }
}
}
isRecording = true
playStopButton.setTitle("Stop Recording", for: .normal)
print("Recording did start")
}
private func stopRecording() {
if !isRecording {
return
}
videoQueue.async {
self.assetWriter?.finishWriting {
print("Asset writer did finish writing")
self.isWriting = false
}
self.captureSession.stopRunning()
do {
try self.export()
} catch {
print("Export failed")
DispatchQueue.main.async { self.showAlert("Unable to export video") }
}
}
isRecording = false
playStopButton.setTitle("Start Recording", for: .normal)
print("Recording did stop")
}
//MARK: - AVCaptureVideoDataOutputSampleBufferDelegate
func captureOutput(_ captureOutput: AVCaptureOutput!, didOutputSampleBuffer sampleBuffer: CMSampleBuffer!, from connection: AVCaptureConnection!) {
guard let w = assetWriter else {
print("Asset writer not configured")
return
}
guard let vo = videoOutput else {
print("Video output not configured")
return
}
guard let ao = audioOutput else {
print("Audio output not configured")
return
}
guard let vi = videoInput else {
print("Video input not configured")
return
}
guard let ai = audioInput else {
print("Audio input not configured")
return
}
let st = CMSampleBufferGetPresentationTimeStamp(sampleBuffer)
print("Writer status \(w.status.rawValue)")
if let e = w.error {
print("Writer error \(e)")
stopRecording()
return
}
switch w.status {
case .unknown:
if !isWriting {
isWriting = true
w.startWriting()
w.startSession(atSourceTime: st)
}
return
case .completed:
print("Video writing completed")
return
case .cancelled:
print("Video writing cancelled")
return
case .failed:
print("Video writing failed")
return
default:
print("Video is writing")
}
if vo == captureOutput {
if !vi.append(sampleBuffer) {
print("Unable to write to video buffer")
}
} else if ao == captureOutput {
if !ai.append(sampleBuffer) {
print("Unable to write to audio buffer")
}
}
}
//MARK: - Export
private func getVideoComposition(asset: AVAsset, videoSize: CGSize) -> AVMutableVideoComposition? {
guard let videoTrack = asset.tracks(withMediaType: AVMediaTypeVideo).first else {
print("Unable to get video tracks")
return nil
}
let videoComposition = AVMutableVideoComposition()
videoComposition.renderSize = videoSize
let seconds: Float64 = Float64(1.0 / videoTrack.nominalFrameRate)
videoComposition.frameDuration = CMTimeMakeWithSeconds(seconds, 600);
let layerInst = AVMutableVideoCompositionLayerInstruction(assetTrack: videoTrack)
var transforms = asset.preferredTransform
var isPortrait = true;
if (transforms.a == 0.0 && transforms.b == 1.0 && transforms.c == -1.0 && transforms.d == 0)
|| (transforms.a == 0.0 && transforms.b == -1.0 && transforms.c == 1.0 && transforms.d == 0) {
isPortrait = false;
}
if isPortrait {
transforms = transforms.concatenating(CGAffineTransform(rotationAngle: CGFloat(90.0.degreesToRadians)))
transforms = transforms.concatenating(CGAffineTransform(translationX: videoSize.width, y: 0))
}
layerInst.setTransform(transforms, at: kCMTimeZero)
let inst = AVMutableVideoCompositionInstruction()
inst.backgroundColor = UIColor.black.cgColor
inst.layerInstructions = [layerInst]
inst.timeRange = CMTimeRange(start: kCMTimeZero, duration: asset.duration)
videoComposition.instructions = [inst]
return videoComposition
}
private func export() throws {
let videoAsset = AVURLAsset(url: outputUrl)
if FileManager.default.fileExists(atPath: exportUrl.path) {
try FileManager.default.removeItem(at: exportUrl)
}
let videoSize = getVideoSize()
guard let encoder = AVAssetExportSession(asset: videoAsset, presetName: exportPreset) else {
print("Unable to create encoder")
return
}
guard let vidcomp = getVideoComposition(asset: videoAsset, videoSize: videoSize) else {
print("Unable to create video composition")
return
}
encoder.videoComposition = vidcomp
encoder.outputFileType = AVFileTypeMPEG4 // MP4 format
encoder.outputURL = exportUrl
encoder.shouldOptimizeForNetworkUse = true
encoder.exportAsynchronously(completionHandler: {
print("Video exported successfully")
})
}
//MARK: Helpers
private func getDocumentsDirectory() -> URL {
let paths = FileManager.default.urls(for: .documentDirectory, in: .userDomainMask)
let documentsDirectory = paths[0]
return documentsDirectory
}
//MARK: Actions
@IBAction func startStopTapped(sender: AnyObject) {
if isRecording {
stopRecording()
} else {
startRecording()
}
}
}
I found that the easiest is to set the preferredTransform on the video composition track depending on the required orientation.
Solution
// Devices orientation
var orientation = UIDevice.current.orientation
// The composition
let audioVideoComposition = AVMutableComposition()
// The video track of the composition
let videoCompositionTrack = audioVideoComposition
.addMutableTrack(withMediaType: .video, preferredTrackID: .init())!
// Set preferred transform
videoCompositionTrack.preferredTransform = getVideoTransform()
Helper function and extension
func getVideoTransform() -> CGAffineTransform {
switch orientation {
case .portrait:
return CGAffineTransform(rotationAngle: 90.degreesToRadians)
case .portraitUpsideDown:
return CGAffineTransform(rotationAngle: 180)
case .landscapeLeft:
return CGAffineTransform(rotationAngle: 0.degreesToRadians)
case .landscapeRight:
return CGAffineTransform(rotationAngle: 180.degreesToRadians)
default:
return CGAffineTransform(rotationAngle: 90.degreesToRadians)
}
}
extension BinaryInteger {
var degreesToRadians: CGFloat { CGFloat(self) * .pi / 180 }
}
extension FloatingPoint {
var degreesToRadians: Self { self * .pi / 180 }
var radiansToDegrees: Self { self * 180 / .pi }
}
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
