I have a link I can DM for a minimum working example!
Recording Videos
For recording, the AVCaptureConnection for an AVCaptureSession, I set isVideoMirrored to true when using the front camera and false when using the back camera. All in portrait orientation.
Saving Videos
When I save videos, I perform an AVAssetExportSession. If I used the front camera, I want to maintain the isVideoMirrored = true, so I create an AVMutableComposition to set the AVAsset video track's preferredTransform to CGAffineTransform(scaleX: -1.0, y: 1.0).rotated(by: CGFloat(Double.pi/2)). For the back camera, I export the AVAsset as outputted.
Part of my saving code:
if didCaptureWithFrontCamera {
let composition = AVMutableComposition()
let assetVideoTrack = asset.tracks(withMediaType: .video).last!
let assetAudioTrack = asset.tracks(withMediaType: .audio).last!
let compositionVideoTrack = composition.addMutableTrack(withMediaType: AVMediaType.video, preferredTrackID: CMPersistentTrackID(kCMPersistentTrackID_Invalid))
let compositionAudioTrack = composition.addMutableTrack(withMediaType: AVMediaType.audio, preferredTrackID: CMPersistentTrackID(kCMPersistentTrackID_Invalid))
try? compositionVideoTrack?.insertTimeRange(CMTimeRangeMake(start: CMTime.zero, duration: asset.duration), of: assetVideoTrack, at: CMTime.zero)
try? compositionAudioTrack?.insertTimeRange(CMTimeRangeMake(start: CMTime.zero, duration: asset.duration), of: assetAudioTrack, at: CMTime.zero)
compositionVideoTrack?.preferredTransform = CGAffineTransform(scaleX: -1.0, y: 1.0).rotated(by: CGFloat(Double.pi/2))
guard let exportSession = AVAssetExportSession(asset: composition, presetName: AVAssetExportPreset1280x720) else {
handler(nil)
return
}
exportSession.outputURL = outputURL
exportSession.outputFileType = .mp4
exportSession.shouldOptimizeForNetworkUse = true
exportSession.exportAsynchronously { handler(exportSession) }
} else {
guard let exportSession = AVAssetExportSession(asset: asset, presetName: AVAssetExportPreset1280x720) else {
handler(nil)
return
}
exportSession.outputURL = outputURL
exportSession.outputFileType = .mp4
exportSession.shouldOptimizeForNetworkUse = true
exportSession.exportAsynchronously { handler(exportSession) }
}
Merging Videos
Later, to view the saved videos, I want to merge them together as a single video and maintain each by their original orientation via AVMutableComposition.
What partially has worked is setting the video track of AVMutableComposition to the preferredTransform property of the video track of an individual AVAsset video. The only problem is that a single orientation is applied to all the videos (i.e. mirroring isn't applied in a back camera recorded video and the same is applied to the front camera video too).
From solutions I've come across it appears I need to apply AVMutableVideoCompositionInstruction, but in trying to do so, the AVAssetExportSession doesn't seem to factor in the videoComposition instructions at all.
Any guidance would be extremely appreciated as I haven't been able to solve it for the life of me...
My attempted merge code:
func merge(videos: [AVURLAsset], for date: Date, completion: @escaping (_ url: URL, _ asset: AVAssetExportSession)->()) {
let videoComposition = AVMutableComposition()
var lastTime: CMTime = .zero
var count = 0
var instructions = [AVMutableVideoCompositionInstruction]()
let renderSize = CGSize(width: 720, height: 1280)
guard let videoCompositionTrack = videoComposition.addMutableTrack(withMediaType: .video, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
guard let audioCompositionTrack = videoComposition.addMutableTrack(withMediaType: .audio, preferredTrackID: Int32(kCMPersistentTrackID_Invalid)) else { return }
for video in videos {
if let videoTrack = video.tracks(withMediaType: .video)[safe: 0] {
//this is the only thing that seems to work, but work not in the way i'd hope where each video keeps its original orientation
//videoCompositionTrack.preferredTransform = videoTrack.preferredTransform
if let audioTrack = video.tracks(withMediaType: .audio)[safe: 0] {
do {
try videoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: video.duration), of: videoTrack, at: lastTime)
try audioCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: video.duration), of: audioTrack, at: lastTime)
let layerInstruction = videoCompositionInstruction(videoTrack, asset: video, count: count)
let videoCompositionInstruction = AVMutableVideoCompositionInstruction()
videoCompositionInstruction.timeRange = CMTimeRangeMake(start: lastTime, duration: video.duration)
videoCompositionInstruction.layerInstructions = [layerInstruction]
instructions.append(videoCompositionInstruction)
} catch {
return
}
lastTime = CMTimeAdd(lastTime, video.duration)
count += 1
} else {
do {
try videoCompositionTrack.insertTimeRange(CMTimeRangeMake(start: .zero, duration: video.duration), of: videoTrack, at: lastTime)
let layerInstruction = videoCompositionInstruction(videoTrack, asset: video, count: count)
let videoCompositionInstruction = AVMutableVideoCompositionInstruction()
videoCompositionInstruction.timeRange = CMTimeRangeMake(start: lastTime, duration: video.duration)
videoCompositionInstruction.layerInstructions = [layerInstruction]
instructions.append(videoCompositionInstruction)
} catch {
return
}
lastTime = CMTimeAdd(lastTime, video.duration)
count += 1
}
}
}
let mutableVideoComposition = AVMutableVideoComposition()
mutableVideoComposition.instructions = instructions
mutableVideoComposition.frameDuration = CMTimeMake(value: 1, timescale: 30)
mutableVideoComposition.renderSize = renderSize
dateFormatter.dateStyle = .long
dateFormatter.timeStyle = .short
let date = dateFormatter.string(from: date)
let mergedURL = NSURL.fileURL(withPath: NSTemporaryDirectory() + "merged-\(date)" + ".mp4")
guard let exporter = AVAssetExportSession(asset: videoComposition, presetName: AVAssetExportPresetHighestQuality) else { return }
exporter.outputURL = mergedURL
exporter.outputFileType = .mp4
exporter.videoComposition = mutableVideoComposition
exporter.shouldOptimizeForNetworkUse = true
completion(mergedURL, exporter)
}
func videoCompositionInstruction(_ firstTrack: AVAssetTrack, asset: AVAsset, count: Int) -> AVMutableVideoCompositionLayerInstruction {
let renderSize = CGSize(width: 720, height: 1280)
let instruction = AVMutableVideoCompositionLayerInstruction(assetTrack: firstTrack)
let assetTrack = asset.tracks(withMediaType: .video)[0]
let t = assetTrack.fixedPreferredTransform // new transform fix
let assetInfo = orientationFromTransform(t)
if assetInfo.isPortrait {
let scaleToFitRatio = renderSize.width / assetTrack.naturalSize.height
let scaleFactor = CGAffineTransform(scaleX: scaleToFitRatio, y: scaleToFitRatio)
var finalTransform = assetTrack.fixedPreferredTransform.concatenating(scaleFactor)
if assetInfo.orientation == .rightMirrored || assetInfo.orientation == .leftMirrored {
finalTransform = finalTransform.translatedBy(x: -t.ty, y: 0)
}
instruction.setTransform(t, at: CMTime.zero)
} else {
let renderRect = CGRect(x: 0, y: 0, width: renderSize.width, height: renderSize.height)
let videoRect = CGRect(origin: .zero, size: assetTrack.naturalSize).applying(assetTrack.fixedPreferredTransform)
let scale = renderRect.width / videoRect.width
let transform = CGAffineTransform(scaleX: renderRect.width / videoRect.width, y: (videoRect.height * scale) / assetTrack.naturalSize.height)
let translate = CGAffineTransform(translationX: .zero, y: ((renderSize.height - (videoRect.height * scale))) / 2)
instruction.setTransform(assetTrack.fixedPreferredTransform.concatenating(transform).concatenating(translate), at: .zero)
}
if count == 0 {
instruction.setOpacity(0.0, at: asset.duration)
}
return instruction
}
func orientationFromTransform(_ 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 = .rightMirrored
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 == 0 && transform.b == -1.0 && transform.c == -1.0 && transform.d == 0 {
assetOrientation = .leftMirrored
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)
}
extension AVAssetTrack {
var fixedPreferredTransform: CGAffineTransform {
var t = preferredTransform
switch(t.a, t.b, t.c, t.d) {
case (1, 0, 0, 1):
t.tx = 0
t.ty = 0
case (1, 0, 0, -1):
t.tx = 0
t.ty = naturalSize.height
case (-1, 0, 0, 1):
t.tx = naturalSize.width
t.ty = 0
case (-1, 0, 0, -1):
t.tx = naturalSize.width
t.ty = naturalSize.height
case (0, -1, 1, 0):
t.tx = 0
t.ty = naturalSize.width
case (0, 1, -1, 0):
t.tx = naturalSize.height
t.ty = 0
case (0, 1, 1, 0):
t.tx = 0
t.ty = 0
case (0, -1, -1, 0):
t.tx = naturalSize.height
t.ty = naturalSize.width
default:
break
}
return t
}
}
Assuming your transformations are correct, I updated your
mergefunction.The main change is using a single
AVMutableVideoCompositionInstructionwith multipleAVMutableVideoCompositionLayerInstruction, and passing the correctCMTimevalue to for the layer instruction to be executed at.