Merge pull request #1385 from stream-labs/bugfix/audioapp-timestamp

Implement timestamp recalculation for audio buffers
shogo4405 · Mar 11, 2024 · 17d67f5 · 17d67f5
2 parents 9b96908 + e840bdb
commit 17d67f5
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 7 deletions.
diff --git a/Sources/IO/IOAudioResampler.swift b/Sources/IO/IOAudioResampler.swift
@@ -77,10 +77,6 @@ struct IOAudioResamplerSettings {
 final class IOAudioResampler<T: IOAudioResamplerDelegate> {
     private static func makeAudioFormat(_ inSourceFormat: inout AudioStreamBasicDescription) -> AVAudioFormat? {
         if inSourceFormat.mFormatID == kAudioFormatLinearPCM && kLinearPCMFormatFlagIsBigEndian == (inSourceFormat.mFormatFlags & kLinearPCMFormatFlagIsBigEndian) {
-            // ReplayKit audioApp.
-            guard inSourceFormat.mBitsPerChannel == 16 else {
-                return nil
-            }
             let interleaved = !((inSourceFormat.mFormatFlags & kLinearPCMFormatFlagIsNonInterleaved) == kLinearPCMFormatFlagIsNonInterleaved)
             if let channelLayout = Self.makeChannelLayout(inSourceFormat.mChannelsPerFrame) {
                 return .init(
@@ -168,8 +164,17 @@ final class IOAudioResampler<T: IOAudioResamplerDelegate> {
 
     func append(_ sampleBuffer: CMSampleBuffer) {
         inSourceFormat = sampleBuffer.formatDescription?.audioStreamBasicDescription
+        guard let inSourceFormat else {
+            return
+        }
         if sampleTime == kIOAudioResampler_sampleTime {
-            sampleTime = sampleBuffer.presentationTimeStamp.value
+            let targetSampleTime: CMTimeValue
+            if sampleBuffer.presentationTimeStamp.timescale == Int32(inSourceFormat.mSampleRate) {
+                targetSampleTime = sampleBuffer.presentationTimeStamp.value
+            } else {
+                targetSampleTime = Int64(Double(sampleBuffer.presentationTimeStamp.value) * inSourceFormat.mSampleRate / Double(sampleBuffer.presentationTimeStamp.timescale))
+            }
+            sampleTime = AVAudioFramePosition(targetSampleTime)
             if let outputFormat {
                 anchor = .init(hostTime: AVAudioTime.hostTime(forSeconds: sampleBuffer.presentationTimeStamp.seconds), sampleTime: sampleTime, atRate: outputFormat.sampleRate)
             }

diff --git a/Sources/IO/IOAudioRingBuffer.swift b/Sources/IO/IOAudioRingBuffer.swift
@@ -40,8 +40,14 @@ final class IOAudioRingBuffer {
         guard CMSampleBufferDataIsReady(sampleBuffer) else {
             return
         }
+        let targetSampleTime: CMTimeValue
+        if sampleBuffer.presentationTimeStamp.timescale == Int32(inputBuffer.format.sampleRate) {
+            targetSampleTime = sampleBuffer.presentationTimeStamp.value
+        } else {
+            targetSampleTime = Int64(Double(sampleBuffer.presentationTimeStamp.value) * inputBuffer.format.sampleRate / Double(sampleBuffer.presentationTimeStamp.timescale))
+        }
         if sampleTime == 0 {
-            sampleTime = sampleBuffer.presentationTimeStamp.value
+            sampleTime = targetSampleTime
         }
         if inputBuffer.frameLength < sampleBuffer.numSamples {
             if let buffer = AVAudioPCMBuffer(pcmFormat: inputFormat, frameCapacity: AVAudioFrameCount(sampleBuffer.numSamples)) {
@@ -67,7 +73,7 @@ final class IOAudioRingBuffer {
                 }
             }
         }
-        skip = max(Int(sampleBuffer.presentationTimeStamp.value - sampleTime), 0)
+        skip = max(Int(targetSampleTime - sampleTime), 0)
         sampleTime += Int64(skip)
         append(inputBuffer)
     }