WI:P

Author: NicolasHug

src/torchcodec/decoders/_core/VideoDecoder.cpp

@@ -1365,10 +1365,20 @@ void VideoDecoder::convertAudioAVFrameToFrameOutputOnCPU(
       static_cast<AVSampleFormat>(avFrameStream.avFrame->format);
   AVSampleFormat desiredSampleFormat = AV_SAMPLE_FMT_FLTP;
 
+  int sourceSampleRate = avFrameStream.avFrame->sample_rate;
+  int desiredSampleRate =
+      streamInfos_[activeStreamIndex_].audioStreamOptions.sampleRate.value_or(
+          sourceSampleRate);
+
   UniqueAVFrame convertedAVFrame;
-  if (sourceSampleFormat != desiredSampleFormat) {
-    convertedAVFrame = convertAudioAVFrameSampleFormat(
-        avFrameStream.avFrame, sourceSampleFormat, desiredSampleFormat);
+  if (sourceSampleFormat != desiredSampleFormat ||
+      sourceSampleRate != desiredSampleRate) {
+    convertedAVFrame = convertAudioAVFrameSampleFormatAndSampleRate(
+        avFrameStream.avFrame,
+        sourceSampleFormat,
+        desiredSampleFormat,
+        sourceSampleRate,
+        desiredSampleRate);
   }
   const UniqueAVFrame& avFrame = (sourceSampleFormat != desiredSampleFormat)
       ? convertedAVFrame
@@ -1398,12 +1408,12 @@ void VideoDecoder::convertAudioAVFrameToFrameOutputOnCPU(
   frameOutput.data = outputData;
 }
 
-UniqueAVFrame VideoDecoder::convertAudioAVFrameSampleFormat(
+UniqueAVFrame VideoDecoder::convertAudioAVFrameSampleFormatAndSampleRate(
     const UniqueAVFrame& avFrame,
     AVSampleFormat sourceSampleFormat,
-    AVSampleFormat desiredSampleFormat
-
-) {
+    AVSampleFormat desiredSampleFormat,
+    int sourceSampleRate,
+    int desiredSampleRate) {
   auto& streamInfo = streamInfos_[activeStreamIndex_];
   const auto& streamMetadata =
       containerMetadata_.allStreamMetadata[activeStreamIndex_];

src/torchcodec/decoders/_core/VideoDecoder.h

@@ -142,7 +142,6 @@ class VideoDecoder {
   struct AudioStreamOptions {
     AudioStreamOptions() {}
 
-    // explicit AudioStreamOptions(const std::string& optionsString);
     std::optional<int> sampleRate;
   };
 
@@ -413,10 +412,12 @@ class VideoDecoder {
       const AVFrame* avFrame,
       torch::Tensor& outputTensor);
 
-  UniqueAVFrame convertAudioAVFrameSampleFormat(
+  UniqueAVFrame convertAudioAVFrameSampleFormatAndSampleRate(
       const UniqueAVFrame& avFrame,
       AVSampleFormat sourceSampleFormat,
-      AVSampleFormat desiredSampleFormat);
+      AVSampleFormat desiredSampleFormat,
+      int sourceSampleRate,
+      int desiredSampleRate);
 
   // --------------------------------------------------------------------------
   // COLOR CONVERSION LIBRARIES HANDLERS CREATION