improve IPF writing

CMakeLists.txt

@@ -16,7 +16,7 @@ if("${CMAKE_CURRENT_SOURCE_DIR}" STREQUAL "${CMAKE_CURRENT_BINARY_DIR}")
 endif()
 
 
-project(exhale VERSION 1.1.2 LANGUAGES CXX)
+project(exhale VERSION 1.1.3 LANGUAGES CXX)
 
 if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES)
     set(CMAKE_BUILD_TYPE Release

README.md

@@ -33,7 +33,7 @@ exhale is being made available under an open-source license which is
 based on the 3-clause BSD license but modified to address particular
 aspects dictated by the nature and the output of this application.
 
-The license text and release notes for the current version 1.1.2 can
+The license text and release notes for the current version 1.1.3 can
 be found in the `include` subdirectory of the exhale distribution.
 
 

include/Release.htm

@@ -25,9 +25,14 @@
 <td valign="top">
 
 <h1><br><span class="pink">exhale</span> - <span class="pink">e</span>codis e<span class="pink">x</span>tended <span class="pink">h</span>igh-efficiency <span class="pink">a</span>nd <span class="pink">l</span>ow-complexity <span class="pink">e</span>ncoder<br><span class="gray"><sup><br>Software Release Notes, Version History, Known Issues, Upcoming Feature Roadmap</sup></span><br><br></h1>
-<h3>&nbsp; &nbsp;The version of this distribution of the &laquo;exhale&raquo; software release is <b>1.1.2</b> (official pub&shy;lic minor release) from January 2021. Please check <a href="http://www.ecodis.de/audio.htm#mpeg">www.ecodis.de</a> regularly for new versions of this software. A summary of each version up to this release, a list of known issues with this release, and a roadmap of additional functionality are provided below.</h3>
+<h3>&nbsp; &nbsp;The version of this distribution of the &laquo;exhale&raquo; software release is <b>1.1.3</b> (official pub&shy;lic minor release) from February 2021. Please check <a href="http://www.ecodis.de/audio.htm#mpeg">www.ecodis.de</a> regularly for new versions of this software. A summary of each version up to this release, a list of known issues with this release, and a roadmap of additional functionality are provided below.</h3>
 <h3><br><b>Chronological Version History</b></h3>
-<h3>&nbsp; &nbsp;Version <b>1.1.2 <span class="gray">&nbsp;Jan. 2021, this release</span></b></h3>
+<h3>&nbsp; &nbsp;Version <b>1.1.3 <span class="gray">&nbsp;Feb. 2021, this release</span></b></h3>
+<ul>
+ <li><h3>exhaleApp: allow specifying loudness (LUFS) and peak sample (dBFS) after preset</h3></li>
+ <li><h3>exhaleLib: write UsacConfig in immediate playout frames (increases compatibility)</h3></li>
+</ul>
+<h3>&nbsp; &nbsp;Version <b>1.1.2 <span class="gray">&nbsp;Jan. 2021</span></b></h3>
 <ul>
  <li><h3>further improved file interoperability and seekability with some playback software</h3></li>
  <li><h3>exhaleLib: write all frames in &laquo;stss&raquo; data as immediate playout frames (issue 15)</h3></li>

include/version.h

@@ -15,5 +15,5 @@
 # define EXHALELIB_VERSION_MINOR "1"
 #endif
 #ifndef EXHALELIB_VERSION_BUGFIX
-# define EXHALELIB_VERSION_BUGFIX ".2" // "RC" or ".0", ".1", ...
+# define EXHALELIB_VERSION_BUGFIX ".3" // "RC" or ".0", ".1", ...
 #endif

src/app/exhaleApp.cpp

@@ -230,6 +230,33 @@ static void eaApplyDownsampler (int32_t* const pcmBuffer, int32_t* const resampl
 }
 #endif // ENABLE_RESAMPLING
 
+static uint32_t eaInitLoudnessInfo (const uint32_t defaultLoudStats, const bool loudnessInfoProvided,
+#ifdef EXHALE_APP_WCHAR
+                                    const wchar_t* const strings[])
+#else
+                                    const char* const strings[])
+#endif
+{
+  float loudnessInfo[2];
+  uint32_t qLoud, qPeak;
+
+  if (!loudnessInfoProvided) return defaultLoudStats;
+
+  for (int i = 0; i < 2; i++)
+  {
+    char cString[8];
+
+    for (int c = 0; c < 8; c++) cString[c] = char (strings[i][c]); // make sure string is of type char
+    cString[7] = '\0';
+    loudnessInfo[i] = (float) atof (cString);
+  }
+
+  qLoud = uint32_t (4.0f  * __max (0.0f, loudnessInfo[0] + 57.75f) + 0.5f); // quantize LUFS to 8 bits
+  qPeak = uint32_t (32.0f * __max (0.0f, 20.0f - loudnessInfo[1]) + 0.5f); // quantize peak to 12 bits
+
+  return EA_LOUD_INIT | (qPeak << 18) | (qLoud << 6) | 11u;  // measurementSystem = 2, reliability = 3
+}
+
 // main routine
 #ifdef EXHALE_APP_WCHAR
 # ifdef __MINGW32__
@@ -242,7 +269,7 @@ int main (const int argc, char* argv[])
 {
   if (argc <= 0) return argc; // for safety
 
-  const bool readStdin = (argc == 3);
+  const bool readStdin = (argc == 3 || argc == 5);
   BasicWavReader wavReader;
   int32_t* inPcmData = nullptr;  // 24-bit WAVE audio input buffer
 #if ENABLE_RESAMPLING
@@ -360,7 +387,7 @@ int main (const int argc, char* argv[])
   fprintf_s (stdout, "  ---------------------------------------------------------------------\n\n");
 
   // check arg. list, print usage if needed
-  if ((argc < 3) || (argc > 4) || (argc > 1 && argv[1][1] != 0))
+  if ((argc < 3) || (argc > 6) || (argc > 1 && argv[1][1] != 0))
   {
     fprintf_s (stdout, " Copyright 2018-2021 C.R.Helmrich, project ecodis. See License.htm for details.\n\n");
 
@@ -481,12 +508,12 @@ int main (const int argc, char* argv[])
     inFileHandle = fileno (stdin);
 #endif
   }
-  else // argc = 4, open input file
+  else // argc = 4/6, open WAV file
   {
 #ifdef EXHALE_APP_WCHAR
-    const wchar_t* inFileName = argv[2];
+    const wchar_t* inFileName = argv[argc - 2];
 #else
-    const char* inFileName = argv[2];
+    const char* inFileName = argv[argc - 2];
 #endif
     uint16_t    inPathEnd  = 0;
 
@@ -510,11 +537,11 @@ int main (const int argc, char* argv[])
 #ifdef EXHALE_APP_WCHAR
       inFileName = (const wchar_t*) malloc ((exePathEnd + i + 1) * sizeof (wchar_t));  // 0-terminated
       memcpy ((void*) inFileName, argv[0], exePathEnd * sizeof (wchar_t));  // prepend executable path
-      memcpy ((void*)(inFileName + exePathEnd), argv[2], (i + 1) * sizeof (wchar_t));  // to file name
+      memcpy ((void*)(inFileName + exePathEnd), argv[argc - 2], (i + 1) * sizeof (wchar_t));// to name
 #else
       inFileName = (const char*) malloc ((exePathEnd + i + 1) * sizeof (char)); // 0-terminated string
       memcpy ((void*) inFileName, argv[0], exePathEnd * sizeof (char)); // prepend executable path ...
-      memcpy ((void*)(inFileName + exePathEnd), argv[2], (i + 1) * sizeof (char)); // ... to file name
+      memcpy ((void*)(inFileName + exePathEnd), argv[argc - 2], (i + 1) * sizeof (char)); //...to name
 #endif
     }
 
@@ -716,7 +743,7 @@ int main (const int argc, char* argv[])
 #endif
       const unsigned indepPeriod = (sampleRate < 48000 ? (sampleRate - 320) / frameLength : 45 /*for 50-Hz video, use 50 for 60-Hz video*/);
       const unsigned mod3Percent = unsigned ((expectLength * (3 + (coreSbrFrameLengthIndex & 3))) >> 17);
-      uint32_t byteCount = 0, bw = (numChannels < 7 ? loudStats : 0);
+      uint32_t byteCount = 0, bw = (numChannels < 7 ? eaInitLoudnessInfo (loudStats, argc > 4, &argv[2]) : 0);
       uint32_t br, bwMax = 0; // br will be used to hold bytes read and/or bit-rate
       uint32_t headerRes = 0;
       // initialize LoudnessEstimator object
@@ -739,6 +766,7 @@ int main (const int argc, char* argv[])
       // init encoder, generate UsacConfig()
       memset (outAuData, 0, 108 * sizeof (uint8_t));  // max. allowed ASC + UC size
       i = exhaleEnc.initEncoder (outAuData, &bw); // bw stores actual ASC + UC size
+      if ((i == 256) && (argc > 4)) i = 0; // clear LUFS warning
 
       if ((i |= mp4Writer.open (outFileHandle, sampleRate, numChannels, inSampDepth, frameLength, startLength
 #if ENABLE_SIMPLE_SBR
@@ -984,7 +1012,7 @@ int main (const int argc, char* argv[])
         const uint32_t qPeak = uint32_t (32.0f * (20.0f - 20.0f * log10 (__max (EA_PEAK_MIN, float (loudStats & USHRT_MAX))) - EA_PEAK_NORM) + 0.5f);
 
         // recreate ASC + UC + loudness data
-        bw = EA_LOUD_INIT | (qPeak << 18) | (qLoud << 6) | 11; // measurementSystem
+        bw = EA_LOUD_INIT | (qPeak << 18) | (qLoud << 6) | 11u;
         memset (outAuData, 0, 108 * sizeof (uint8_t)); // max allowed ASC + UC size
         i = exhaleEnc.initEncoder (outAuData, &bw); // with finished loudnessInfo()
       }
@@ -1049,12 +1077,12 @@ mainFinish:
       {
         fprintf_s (stderr, " ERROR while trying to close stdin stream! Has it already been closed?\n\n");
       }
-      else  // argc = 4, file
+      else // argc = 4/6, WAV
       {
 #ifdef EXHALE_APP_WCHAR
-        fwprintf_s (stderr, L" ERROR while trying to close input file %s! Does it still exist?\n\n", argv[2]);
+        fwprintf_s (stderr, L" ERROR while trying to close input file %s! Does it still exist?\n\n", argv[argc - 2]);
 #else
-        fprintf_s (stderr, " ERROR while trying to close input file %s! Does it still exist?\n\n", argv[2]);
+        fprintf_s (stderr, " ERROR while trying to close input file %s! Does it still exist?\n\n", argv[argc - 2]);
 #endif
       }
     }

src/app/exhaleApp.rc

@@ -13,7 +13,7 @@
 
 0 ICON "exhaleApp.ico"
 VS_VERSION_INFO VERSIONINFO
-FILEVERSION 1,1,2
+FILEVERSION 1,1,3
 BEGIN
   BLOCK "StringFileInfo"
   BEGIN

src/lib/bitStreamWriter.cpp

@@ -715,7 +715,10 @@ unsigned BitStreamWriter::createAudioConfig (const char samplingFrequencyIndex,
 {
   const uint8_t fli = (sbrRatioShiftValue == 0 ? 1 /*no SBR*/ : __min (2, sbrRatioShiftValue) + 2);
   const int8_t usfi = __max (0, samplingFrequencyIndex - 3 * sbrRatioShiftValue); // TODO: non-standard sampling rates
-  unsigned bitCount = 37;
+  unsigned bitCount = 37, auLen;
+#ifndef NO_PREROLL_DATA
+  unsigned ucOffset = (samplingFrequencyIndex < AAC_NUM_SAMPLE_RATES ? 2 : 5);
+#endif
 
   if ((elementType == nullptr) || (audioConfig == nullptr) || (chConfigurationIndex >= USAC_MAX_NUM_ELCONFIGS) ||
 #if !RESTRICT_TO_AAC
@@ -816,8 +819,12 @@ unsigned BitStreamWriter::createAudioConfig (const char samplingFrequencyIndex,
 
   bitCount += (8 - m_auBitStream.heldBitCount) & 7;
   writeByteAlignment ();  // flush bytes
-
+  auLen = __min (18u + fli, bitCount >> 3);
-  memcpy (audioConfig, &m_auBitStream.stream.front (), __min (17u + fli, bitCount >> 3));
+#ifndef NO_PREROLL_DATA
+  m_usacConfigLen = uint16_t (__max (9, auLen - ucOffset)); // excl. ASC payload
+  memcpy (m_usacConfig, &m_auBitStream.stream.at (ucOffset), auLen - ucOffset);
+#endif
+  memcpy (audioConfig,  &m_auBitStream.stream.front (), auLen);
 
   return (bitCount >> 3);  // byte count
 }
@@ -854,7 +861,7 @@ unsigned BitStreamWriter::createAudioFrame (CoreCoderData** const elementData,
   if ((ipf == 2) || (ipf == 1 && (numElements > 1 || !noiseFilling[0])))
   {
     bitCount = __min (nSamplesInFrame << 2, (uint32_t) m_auBitStream.stream.size ());
-    memcpy (tempBuffer, &m_auBitStream.stream.front (), bitCount);
+    memcpy (tempBuffer, &m_auBitStream.stream.front (), bitCount); // prev fr AU
   }
 #endif
   m_auBitStream.reset ();
@@ -869,19 +876,28 @@ unsigned BitStreamWriter::createAudioFrame (CoreCoderData** const elementData,
   {
     const uint16_t idxPreRollExt = elementData[0]->elementType & 1;
     const bool lowRatePreRollExt = (ipf == 1 && numElements == 1 && noiseFilling[0]);
-    const unsigned payloadLength = (lowRatePreRollExt ? (8 + idxPreRollExt * 6) >> (sbrRatioShiftValue > 0 ? 0 : 1) : bitCount) + 3; // in bytes!
+    const unsigned   extraLength = (m_usacConfigLen > 14 ? 4 : 3) + m_usacConfigLen;
+    const unsigned payloadLength = (lowRatePreRollExt ? (8 + idxPreRollExt * 6) >> (sbrRatioShiftValue > 0 ? 0 : 1) : bitCount) + extraLength; // in bytes
 
     m_auBitStream.write (0, 1); // usacExtElementUseDefaultLength = 0 (variable)
     m_auBitStream.write (CLIP_UCHAR (payloadLength), 8);
     if (payloadLength > 254) m_auBitStream.write (payloadLength - 253, 16);
 
-    m_auBitStream.write (0, 6); // start AudioPreRoll - configLen = reserved = 0
+    m_auBitStream.write (__min (15, m_usacConfigLen), 4); // configLen (part #1)
+    if (m_usacConfigLen > 14) m_auBitStream.write (m_usacConfigLen - 15, 4);
+
+    m_auBitStream.write (m_usacConfig[ci++] & 31, 5); // 1st 3 bits are from ASC
+    while (ci < m_usacConfigLen) m_auBitStream.write (m_usacConfig[ci++], 8);
+    ci = 0;
+    m_auBitStream.write (0, 8 - 5); // pad end of UsacConfig() data
+
+    m_auBitStream.write (0, 2); // applyCrossfade = 0 and reserved = 0 (part #2)
     m_auBitStream.write (1, 2); // numPreRollFrames, only one supported for now!
-    m_auBitStream.write (payloadLength - 3, 16); // auLen
+    m_auBitStream.write (payloadLength - extraLength, 16); // auLen
 
     if (lowRatePreRollExt)
     {
-      bitCount = payloadLength - 3;
+      bitCount = payloadLength - extraLength;
       memcpy (tempBuffer, zeroAu[idxPreRollExt], bitCount);
       if (elementData[0]->elementType < ID_USAC_LFE)  // correct window_sequence
       {
@@ -893,6 +909,8 @@ unsigned BitStreamWriter::createAudioFrame (CoreCoderData** const elementData,
     }
     while (ci < bitCount) m_auBitStream.write (tempBuffer[ci++], 8); // write AU
     ci = 0;
+    if (m_usacConfigLen > 14) m_auBitStream.write (0, 4); // pad end of ext data
+
     bitCount = (payloadLength > 254 ? 26 : 10) + (payloadLength << 3); // for PR
   }
   bitCount++; // for ElementPresent flag

src/lib/bitStreamWriter.h

@@ -16,8 +16,8 @@
 
 // constants, experimental macros
 #define CORE_MODE_FD            0
-#define ID_EXT_LOUDNESS_INFO    2
 #define ID_EXT_ELE_FILL         0
+#define ID_EXT_LOUDNESS_INFO    2
 #define SFB_PER_PRED_BAND       2
 
 // output bit-stream writer class
@@ -30,6 +30,10 @@ private:
   uint32_t     m_frameLength; // number of samples in full frame
   uint8_t      m_numSwbShort; // max. SFB count in short windows
   uint8_t*     m_uCharBuffer; // temporary buffer for ungrouping
+#ifndef NO_PREROLL_DATA
+  uint8_t      m_usacConfig[20]; // buffer for UsacConfig in IPF
+  uint16_t     m_usacConfigLen; // length of UsacConfig in bytes
+#endif
 
   // helper functions
   void     writeByteAlignment (); // write 0s for byte alignment
@@ -52,7 +56,8 @@ private:
 public:
 
   // constructor
-  BitStreamWriter () { m_auBitStream.reset (); m_frameLength = 0; m_numSwbShort = 0; m_uCharBuffer = nullptr; }
+  BitStreamWriter () { m_auBitStream.reset ();  m_frameLength = 0;  m_numSwbShort = 0;  m_uCharBuffer = nullptr;
+                       memset (m_usacConfig, 0, 20u);  m_usacConfigLen = 0; }
   // destructor
   ~BitStreamWriter() { m_auBitStream.reset (); }
   // public functions