Merge pull request #2017 from shoamano83/fix/sdl-0139_apt_clarification

Changes for "Clarification of audio format details of AudioPassThru" proposal

5 files changed, 82 insertions, 21 deletions

diff --git a/src/appMain/audio.8bit.wav b/src/appMain/audio.8bit.wav
index cebf6cc0e2..7a3f970a21 100644
--- a/src/appMain/audio.8bit.wav
+++ b/src/appMain/audio.8bit.wav
diff --git a/src/components/interfaces/HMI_API.xml b/src/components/interfaces/HMI_API.xml
index 18966ff00a..bdbb60f98a 100644
--- a/src/components/interfaces/HMI_API.xml
+++ b/src/components/interfaces/HMI_API.xml
@@ -1137,21 +1137,35 @@
 
 <enum name="SamplingRate">
   <description>Describes different sampling options for PerformAudioPassThru.</description>
-  <element name="8KHZ" internal_name="RATE_8KHZ"/>
-  <element name="16KHZ" internal_name="RATE_16KHZ"/>
-  <element name="22KHZ" internal_name="RATE_22KHZ"/>
-  <element name="44KHZ" internal_name="RATE_44KHZ"/>
+  <element name="8KHZ" internal_name="RATE_8KHZ">
+    <description>Sampling rate of 8000 Hz.</description>
+  </element>
+  <element name="16KHZ" internal_name="RATE_16KHZ">
+    <description>Sampling rate of 16000 Hz.</description>
+  </element>
+  <element name="22KHZ" internal_name="RATE_22KHZ">
+    <description>Sampling rate of 22050 Hz.</description>
+  </element>
+  <element name="44KHZ" internal_name="RATE_44KHZ">
+    <description>Sampling rate of 44100 Hz.</description>
+  </element>
 </enum>
 
 <enum name="BitsPerSample">
   <description>Describes different quality options for PerformAudioPassThru.</description>
-  <element name="8_BIT" internal_name="RATE_8_BIT"/>
-  <element name="16_BIT" internal_name="RATE_16_BIT"/>
+  <element name="8_BIT" internal_name="RATE_8_BIT">
+    <description>Audio sample is 8 bits wide, unsigned.</description>
+  </element>
+  <element name="16_BIT" internal_name="RATE_16_BIT">
+    <description>Audio sample is 16 bits wide, signed, and in little endian.</description>
+  </element>
 </enum>
 
 <enum name="AudioType">
   <description>Describes different audio type options for PerformAudioPassThru.</description>
-  <element name="PCM" />
+  <element name="PCM">
+    <description>Linear PCM.</description>
+  </element>
 </enum>
 
 <enum name="KeyboardLayout">
@@ -2396,6 +2410,7 @@
   <description>
     Describes different audio type configurations for PerformAudioPassThru.
     e.g. 8kHz,8-bit,PCM
+    The audio is recorded in monaural.
   </description>
   <param name="samplingRate" type="Common.SamplingRate" mandatory="true"/>
   <param name="bitsPerSample" type="Common.BitsPerSample" mandatory="true"/>
diff --git a/src/components/interfaces/MOBILE_API.xml b/src/components/interfaces/MOBILE_API.xml
index 11c8040722..97bc33ceae 100644
--- a/src/components/interfaces/MOBILE_API.xml
+++ b/src/components/interfaces/MOBILE_API.xml
@@ -480,27 +480,42 @@
     
     <enum name="SamplingRate">
         <description>Describes different sampling options for PerformAudioPassThru.</description>
-        <element name="8KHZ" internal_name="SamplingRate_8KHZ"/>
-        <element name="16KHZ" internal_name="SamplingRate_16KHZ"/>
-        <element name="22KHZ" internal_name="SamplingRate_22KHZ"/>
-        <element name="44KHZ" internal_name="SamplingRate_44KHZ"/>
+        <element name="8KHZ" internal_name="SamplingRate_8KHZ">
+            <description>Sampling rate of 8000 Hz.</description>
+        </element>
+        <element name="16KHZ" internal_name="SamplingRate_16KHZ">
+            <description>Sampling rate of 16000 Hz.</description>
+        </element>
+        <element name="22KHZ" internal_name="SamplingRate_22KHZ">
+            <description>Sampling rate of 22050 Hz.</description>
+        </element>
+        <element name="44KHZ" internal_name="SamplingRate_44KHZ">
+            <description>Sampling rate of 44100 Hz.</description>
+        </element>
     </enum>
     
     <enum name="BitsPerSample">
         <description>Describes different quality options for PerformAudioPassThru.</description>
-        <element name="8_BIT" internal_name="BitsPerSample_8_BIT"/>
-        <element name="16_BIT" internal_name="BitsPerSample_16_BIT"/>
+        <element name="8_BIT" internal_name="BitsPerSample_8_BIT">
+            <description>Audio sample is 8 bits wide, unsigned.</description>
+        </element>
+        <element name="16_BIT" internal_name="BitsPerSample_16_BIT">
+            <description>Audio sample is 16 bits wide, signed, and in little endian.</description>
+        </element>
     </enum>
     
     <enum name="AudioType">
         <description>Describes different audio type options for PerformAudioPassThru.</description>
-        <element name="PCM" />
+        <element name="PCM">
+            <description>Linear PCM.</description>
+        </element>
     </enum>
     
     <struct name="AudioPassThruCapabilities">
         <description>
             Describes different audio type configurations for PerformAudioPassThru.
             e.g. {8kHz,8-bit,PCM}
+            The audio is recorded in monaural.
         </description>
         <param name="samplingRate" type="SamplingRate" mandatory="true"/>
         <param name="bitsPerSample" type="BitsPerSample" mandatory="true"/>
diff --git a/src/components/media_manager/src/audio/audio_stream_sender_thread.cc b/src/components/media_manager/src/audio/audio_stream_sender_thread.cc
index e1b932b611..23181e7431 100644
--- a/src/components/media_manager/src/audio/audio_stream_sender_thread.cc
+++ b/src/components/media_manager/src/audio/audio_stream_sender_thread.cc
@@ -61,6 +61,15 @@ const int32_t AudioStreamSenderThread::kAudioPassThruTimeout = 1000;
 #endif
 const uint32_t kMqueueMessageSize = 4095;
 
+// Size of RIFF header contained in a .wav file: 12 bytes for 'RIFF' chunk
+// descriptor, 24 bytes for 'fmt ' sub-chunk and 8 bytes for 'data' sub-chunk
+// header.
+// The correct format of audio stream for AudioPassThru feature is to include
+// only audio sample data. Since both pre-recorded file (audio.8bit.wav) and
+// GStreamer-generated file contain RIFF header, we will skip it when reading
+// the files.
+static const uint32_t kRIFFHeaderSize = 44;
+
 CREATE_LOGGERPTR_GLOBAL(logger_, "MediaManager")
 
 AudioStreamSenderThread::AudioStreamSenderThread(
@@ -69,7 +78,7 @@ AudioStreamSenderThread::AudioStreamSenderThread(
     application_manager::ApplicationManager& app_mngr)
     : session_key_(session_key)
     , fileName_(fileName)
-    , offset_(0)
+    , offset_(kRIFFHeaderSize)
     , shouldBeStoped_(false)
     , shouldBeStoped_lock_()
     , shouldBeStoped_cv_()
@@ -82,7 +91,7 @@ AudioStreamSenderThread::~AudioStreamSenderThread() {}
 void AudioStreamSenderThread::threadMain() {
   LOG4CXX_AUTO_TRACE(logger_);
 
-  offset_ = 0;
+  offset_ = kRIFFHeaderSize;
 
   while (false == getShouldBeStopped()) {
     AutoLock auto_lock(shouldBeStoped_lock_);
@@ -125,7 +134,7 @@ void AudioStreamSenderThread::sendAudioChunkToMobile() {
   }
 #if !defined(EXTENDED_MEDIA_MODE)
   // without recording stream restart reading 1-sec file
-  offset_ = 0;
+  offset_ = kRIFFHeaderSize;
 #endif
 }
 
diff --git a/src/components/media_manager/src/audio/from_mic_to_file_recorder_thread.cc b/src/components/media_manager/src/audio/from_mic_to_file_recorder_thread.cc
index 99548e71bd..5c12614662 100644
--- a/src/components/media_manager/src/audio/from_mic_to_file_recorder_thread.cc
+++ b/src/components/media_manager/src/audio/from_mic_to_file_recorder_thread.cc
@@ -42,6 +42,9 @@ CREATE_LOGGERPTR_GLOBAL(logger_, "MediaManager")
 
 GMainLoop* FromMicToFileRecorderThread::loop = NULL;
 
+// As per spec, AudioPassThru recording is in monaural
+static const int kNumAudioChannels = 1;
+
 FromMicToFileRecorderThread::FromMicToFileRecorderThread(
     const std::string& output_file, int32_t duration)
     : threads::ThreadDelegate()
@@ -119,7 +122,8 @@ void FromMicToFileRecorderThread::threadMain() {
   initArgs();
 
   GstElement* pipeline;
-  GstElement* alsasrc, *wavenc, *filesink;
+  GstElement* alsasrc, *audioconvert, *capsfilter, *wavenc, *filesink;
+  GstCaps* audiocaps;
   GstBus* bus;
 
   const gchar* device = "hw:0,0";
@@ -196,11 +200,18 @@ void FromMicToFileRecorderThread::threadMain() {
 
   // Create all of the elements to be added to the pipeline
   alsasrc = gst_element_factory_make("alsasrc", "alsasrc0");
+  audioconvert = gst_element_factory_make("audioconvert", "audioconvert0");
+  capsfilter = gst_element_factory_make("capsfilter", "filter0");
   wavenc = gst_element_factory_make("wavenc", "wavenc0");
   filesink = gst_element_factory_make("filesink", "filesink0");
 
+  // create a capability to downmix the recorded audio to monaural
+  audiocaps = gst_caps_new_simple(
+      "audio/x-raw", "channels", G_TYPE_INT, kNumAudioChannels, NULL);
+
   // Assert that all the elements were created
-  if (!alsasrc || !wavenc || !filesink) {
+  if (!alsasrc || !audioconvert || !capsfilter || !wavenc || !filesink ||
+      !audiocaps) {
     g_error("Failed creating one or more of the pipeline elements.\n");
   }
 
@@ -209,10 +220,21 @@ void FromMicToFileRecorderThread::threadMain() {
   g_object_set(G_OBJECT(filesink), "location", outfile, NULL);
 
   // Add the elements to the pipeline
-  gst_bin_add_many(GST_BIN(pipeline), alsasrc, wavenc, filesink, NULL);
+  gst_bin_add_many(GST_BIN(pipeline),
+                   alsasrc,
+                   audioconvert,
+                   capsfilter,
+                   wavenc,
+                   filesink,
+                   NULL);
 
   // Link the elements
-  gst_element_link_many(alsasrc, wavenc, filesink, NULL);
+  gst_element_link_many(
+      alsasrc, audioconvert, capsfilter, wavenc, filesink, NULL);
+
+  // set the capability
+  g_object_set(G_OBJECT(capsfilter), "caps", audiocaps, NULL);
+  gst_caps_unref(audiocaps);
 
   gst_element_set_state(pipeline, GST_STATE_PLAYING);