- better internal statistics + memory performance optimization

src/CMakeLists.txt

@@ -360,13 +360,21 @@ endif()
 target_compile_definitions(rtphone PUBLIC ${DEFINES} )
 
 if (TARGET_LINUX)
-    target_link_options(rtphone PUBLIC -Wl,-Bstatic)
+    # PRIVATE, not PUBLIC: rtphone is a STATIC library, so these link options are
+    # never used to build rtphone itself and must not propagate to consumers.
+    # As PUBLIC they leaked into every consumer's LINK_FLAGS as an adjacent
+    # "-Wl,-Bstatic -Wl,-Bdynamic" pair (the wrapped libraries land in a separate
+    # LINK_LIBRARIES section, so nothing is actually wrapped). The trailing
+    # -Bdynamic forced the linker back into dynamic-search mode, which broke
+    # fully-static consumers (e.g. vq-core built with SERVER_STATIC_LINKING=ON:
+    # "attempted static link of dynamic object libz.so").
+    target_link_options(rtphone PRIVATE -Wl,-Bstatic)
     target_compile_options(rtphone PUBLIC -Wno-deprecated -Wno-deprecated-declarations)
 endif()
 target_link_libraries(rtphone PUBLIC ${LIBS_STATIC})
 
 if (TARGET_LINUX)
-    target_link_options(rtphone PUBLIC -Wl,-Bdynamic)
+    target_link_options(rtphone PRIVATE -Wl,-Bdynamic)
 endif()
 
 target_include_directories(rtphone

src/engine/agent/Agent_Impl.cpp

@@ -181,9 +181,6 @@ void AgentImpl::processConfig(JsonCpp::Value &d, JsonCpp::Value &answer)
     config()[CONFIG_IPV4] = d["ipv4"].asBool();
     config()[CONFIG_IPV6] = d["ipv6"].asBool();
 
-    if (transport == "tls")
-        config()[CONFIG_SIPS] = true;
-
     // Log file
     std::string logfile = d["logfile"].asString();
     ice::Logger& logger = ice::GLogger;
@@ -195,7 +192,7 @@ void AgentImpl::processConfig(JsonCpp::Value &d, JsonCpp::Value &answer)
 
     mUseNativeAudio = d["nativeaudio"].asBool();
     config()[CONFIG_OWN_DNS] = d["dns_servers"].asString();
-    config()[CONFIG_SIPS] = d["secure"].asBool();
+    config()[CONFIG_SIPS] = d["secure"].asBool() || transport == "tls";
     config()[CONFIG_STUNSERVER_IP] = d["stun_server"].asString();
 
     answer["status"] = Status_Ok;
@@ -483,18 +480,18 @@ void AgentImpl::processDestroySession(JsonCpp::Value& request, JsonCpp::Value& a
 
 void AgentImpl::processWaitForEvent(JsonCpp::Value &request, JsonCpp::Value &answer)
 {
-    std::unique_lock<std::recursive_mutex> l(mAgentMutex);
+    // Deliberately does NOT take mAgentMutex: events are produced by the worker
-
+    // thread inside process(), which needs mAgentMutex. Holding it here would
-    //int x = 0;
+    // stall all SIP/media processing for the whole timeout and guarantee that
-    //int y = 1/x;
+    // the awaited event can never arrive during the wait.
 
     int timeout = 0;
     if (request.isMember("timeout"))
         timeout = request["timeout"].asInt();
 
     std::unique_lock<std::mutex> eventLock(mEventListMutex);
-    if (mEventList.empty())
+    mEventListChangeCondVar.wait_for(eventLock, chrono::milliseconds(timeout),
-        mEventListChangeCondVar.wait_for(eventLock, chrono::milliseconds(timeout));
+                                     [this]() { return !mEventList.empty(); });
 
     if (!mEventList.empty())
     {
@@ -521,7 +518,7 @@ void AgentImpl::processGetMediaStats(JsonCpp::Value& request, JsonCpp::Value& an
             answer["codec"] = result[SessionInfo_AudioCodec].asStdString();
         if (result.exists(SessionInfo_NetworkMos))
             answer["network_mos"] = result[SessionInfo_NetworkMos].asFloat();
-        if (result.exists(SessionInfo_PacketLoss))
+        if (result.exists(SessionInfo_LostRtp))
             answer["rtp_lost"] = result[SessionInfo_LostRtp].asInt();
         if (result.exists(SessionInfo_DroppedRtp))
             answer["rtp_dropped"] = result[SessionInfo_DroppedRtp].asInt();
@@ -749,6 +746,7 @@ void AgentImpl::onSessionTerminated(PSession s, int responsecode, int reason)
   if (mOutgoingAudioDump)
     mOutgoingAudioDump->close();
   */
+    if (mAudioManager)
         mAudioManager->stop(mUseNativeAudio ? AudioManager::atReceiver : AudioManager::atNull);
     // Gather statistics before
     EVENT_WITH_NAME("session_terminated");

src/engine/agent/Agent_Impl.h

@@ -13,6 +13,7 @@
 #include "Agent_AudioManager.h"
 #include <mutex>
 #include <condition_variable>
+#include <atomic>
 
 
 class AgentImpl: public UserAgent, public MT::Stream::MediaObserver
@@ -32,7 +33,7 @@ protected:
 
 
     std::shared_ptr<std::thread> mThread;
-    volatile bool mShutdown;
+    std::atomic<bool> mShutdown;
     std::shared_ptr<MT::Terminal> mTerminal;
     std::shared_ptr<AudioManager> mAudioManager;
     Audio::DataConnection* mAudioMonitoring = nullptr;

src/engine/audio/Audio_DataWindow.cpp

@@ -9,11 +9,7 @@
 using namespace Audio;
 
 DataWindow::DataWindow()
-{
+{}
-    mFilled = 0;
-    mData = nullptr;
-    mCapacity = 0;
-}
 
 DataWindow::~DataWindow()
 {
@@ -24,10 +20,15 @@ DataWindow::~DataWindow()
     }
 }
 
-void DataWindow::setCapacity(int capacity)
+void DataWindow::setCapacity(size_t capacity)
 {
     Lock l(mMutex);
-    int tail  = capacity - mCapacity;
+
+    // The window only ever grows; a smaller request keeps the current buffer.
+    if (capacity <= mCapacity)
+        return;
+
+    size_t tail  = capacity - mCapacity;
     char* buffer = mData;
     mData = (char*)realloc(mData, capacity);
     if (!mData)
@@ -41,7 +42,7 @@ void DataWindow::setCapacity(int capacity)
     mCapacity = capacity;
 }
 
-void DataWindow::addZero(int length)
+void DataWindow::addZero(size_t length)
 {
     Lock l(mMutex);
 
@@ -60,7 +61,7 @@ void DataWindow::addZero(int length)
 }
 
 
-void DataWindow::add(const void* data, int length)
+void DataWindow::add(const void* data, size_t length)
 {
     Lock l(mMutex);
 
@@ -94,7 +95,7 @@ void DataWindow::add(short sample)
     add(&sample, sizeof sample);
 }
 
-void DataWindow::erase(int length)
+void DataWindow::erase(size_t length)
 {
     Lock l(mMutex);
     if (length > mFilled)
@@ -120,21 +121,21 @@ void DataWindow::clear()
     mFilled = 0;
 }
 
-short DataWindow::shortAt(int index) const
+short DataWindow::shortAt(size_t index) const
 {
     Lock l(mMutex);
     assert(index < mFilled / 2);
     return ((short*)mData)[index];
 }
 
-void DataWindow::setShortAt(short value, int index)
+void DataWindow::setShortAt(short value, size_t index)
 {
     Lock l(mMutex);
     assert(index < mFilled / 2);
     ((short*)mData)[index] = value;
 }
 
-int DataWindow::read(void* buffer, int length)
+size_t DataWindow::read(void* buffer, size_t length)
 {
     Lock l(mMutex);
     if (length > mFilled)
@@ -150,25 +151,27 @@ int DataWindow::read(void* buffer, int length)
     return length;
 }
 
-int DataWindow::filled() const
+size_t DataWindow::filled() const
 {
     Lock l(mMutex);
     return mFilled;
 }
 
-void DataWindow::setFilled(int filled)
+void DataWindow::setFilled(size_t filled)
 {
     Lock l(mMutex);
+    if (filled > mCapacity)
+        throw std::bad_alloc();
     mFilled = filled;
 }
 
-int DataWindow::capacity() const
+size_t DataWindow::capacity() const
 {
     Lock l(mMutex);
     return mCapacity;
 }
 
-void DataWindow::zero(int length)
+void DataWindow::zero(size_t length)
 {
     Lock l(mMutex);
     assert(length <= mCapacity);
@@ -189,10 +192,10 @@ size_t DataWindow::moveTo(DataWindow& dst, size_t size)
     return avail;
 }
 
-std::chrono::milliseconds DataWindow::getTimeLength(int samplerate, int channels) const
+std::chrono::milliseconds DataWindow::getTimeLength(const Audio::Format& fmt) const
 {
     Lock l(mMutex);
-    return std::chrono::milliseconds(mFilled / sizeof(short) / channels / (samplerate / 1000));
+    return std::chrono::milliseconds(mFilled / sizeof(short) / fmt.channels() / (fmt.rate()/ 1000));
 }
 
 void DataWindow::makeStereoFromMono(DataWindow& dst, DataWindow& src)

src/engine/audio/Audio_DataWindow.h

@@ -8,6 +8,7 @@
 
 #include "../helper/HL_ByteBuffer.h"
 #include "../helper/HL_Sync.h"
+#include "Audio_Interface.h"
 
 namespace Audio
 {
@@ -17,34 +18,34 @@ public:
     DataWindow();
     ~DataWindow();
 
-    void        setCapacity(int capacity);
+    void        setCapacity(size_t capacity);
-    int         capacity() const;
+    size_t      capacity() const;
 
-    void        addZero(int length);
+    void        addZero(size_t length);
-    void        add(const void* data, int length);
+    void        add(const void* data, size_t length);
     void        add(short sample);
-    int         read(void* buffer, int length);
+    size_t      read(void* buffer, size_t length);
-    void        erase(int length = -1);
+    void        erase(size_t length);
     const char* data() const;
     char*       mutableData();
-    int         filled() const;
+    size_t      filled() const;
-    void        setFilled(int filled);
+    void        setFilled(size_t filled);
     void        clear();
 
-    short       shortAt(int index) const;
+    short       shortAt(size_t index) const;
-    void        setShortAt(short value, int index);
+    void        setShortAt(short value, size_t index);
-    void        zero(int length);
+    void        zero(size_t length);
-    size_t      moveTo(DataWindow& dst, size_t size);
+    size_t      moveTo(DataWindow& dst, size_t size /* in bytes*/ );
 
-    std::chrono::milliseconds getTimeLength(int samplerate, int channels) const;
+    std::chrono::milliseconds getTimeLength(const Format& fmt) const;
 
     static void makeStereoFromMono(DataWindow& dst, DataWindow& src);
 
 protected:
     mutable Mutex mMutex;
-    char* mData;
+    char*   mData = nullptr;
-    int mFilled;
+    size_t  mFilled = 0;
-    int mCapacity;
+    size_t  mCapacity = 0;
 };
 }
 #endif

src/engine/audio/Audio_Mixer.cpp

@@ -94,6 +94,7 @@ Mixer::~Mixer()
 
 void Mixer::unregisterChannel(void* channel)
 {
+  Lock l(mMutex);
   for (int i=0; i<AUDIO_MIX_CHANNEL_COUNT; i++)
   {
     Stream& c = mChannelList[i];
@@ -108,6 +109,7 @@ void Mixer::unregisterChannel(void* channel)
 
 void Mixer::clear(void* context, unsigned ssrc)
 {
+  Lock l(mMutex);
   for (int i=0; i<AUDIO_MIX_CHANNEL_COUNT; i++)
   {
     Stream& c = mChannelList[i];
@@ -147,6 +149,7 @@ void Mixer::addPcm(void* context, unsigned ssrc,
 {
   assert(inputRate == 8000 || inputRate == 16000 || inputRate == 32000);
 
+  Lock l(mMutex);
   int i;
 
   // Locate a channel
@@ -172,6 +175,7 @@ void Mixer::addPcm(void* context, unsigned ssrc, Audio::DataWindow& w, int rate,
 {
   assert(rate == 8000 || rate == 16000 || rate == 32000 || rate == 48000);
 
+  Lock l(mMutex);
   int i;
 
   // Locate a channel
@@ -196,6 +200,8 @@ void Mixer::addPcm(void* context, unsigned ssrc, Audio::DataWindow& w, int rate,
 
 void Mixer::mix()
 {
+  Lock l(mMutex);
+
   // Current sample
   int sample = 0;
 
@@ -310,6 +316,8 @@ void Mixer::mix()
 
 int Mixer::getPcm(void* outputData, int outputLength)
 {
+  Lock l(mMutex);
+
   if (mOutput.filled() < outputLength)
     mix();
 
@@ -320,14 +328,26 @@ int Mixer::getPcm(void* outputData, int outputLength)
 
 int Mixer::mixAndGetPcm(Audio::DataWindow& output)
 {
+  Lock l(mMutex);
+
   // Mix
   mix();
 
-  // Set output space
+  size_t avail = mOutput.filled();
-  output.setCapacity(mOutput.filled());
+  if (!avail)
+  {
+    output.setFilled(0);
+    return 0;
+  }
 
-  // Read mixed data to output
+  // Make sure output has enough space (setCapacity only ever grows the window)
-  return mOutput.read(output.mutableData(), output.capacity());
+  if (output.capacity() < avail)
+    output.setCapacity(avail);
+
+  // Read mixed data to output and publish the real byte count
+  size_t got = mOutput.read(output.mutableData(), avail);
+  output.setFilled(got);
+  return static_cast<int>(got);
 }
 
 int Mixer::available()

src/engine/audio/Audio_Resampler.cpp

@@ -18,9 +18,7 @@ namespace Audio
 
 
 SpeexResampler::SpeexResampler()
-    :mContext(NULL), mErrorCode(0), mSourceRate(0), mDestRate(0), mLastSample(0), mChannels(0)
+{}
-{
-}
 
 void SpeexResampler::start(int channels, int sourceRate, int destRate)
 {
@@ -51,6 +49,11 @@ void SpeexResampler::stop()
     }
 }
 
+bool SpeexResampler::isOpened() const
+{
+    return mContext != nullptr;
+}
+
 SpeexResampler::~SpeexResampler()
 {
     stop();
@@ -113,22 +116,22 @@ size_t SpeexResampler::processBuffer(const void* src, size_t sourceLength, size_
     return outLen * sizeof(short) * mChannels;
 }
 
-int SpeexResampler::sourceRate()
+int SpeexResampler::sourceRate() const
 {
     return mSourceRate;
 }
 
-int SpeexResampler::destRate()
+int SpeexResampler::destRate() const
 {
     return mDestRate;
 }
 
-size_t SpeexResampler::getDestLength(size_t sourceLen)
+size_t SpeexResampler::getDestLength(size_t sourceLen) const
 {
     return size_t(sourceLen * (float(mDestRate) / mSourceRate) + 0.5f);
 }
 
-size_t SpeexResampler::getSourceLength(size_t destLen)
+size_t SpeexResampler::getSourceLength(size_t destLen) const
 {
     // Here we want to get 'destLen' number of samples
     return size_t(destLen * (float(mSourceRate) / mDestRate) + 0.5f);

src/engine/audio/Audio_Resampler.h

@@ -24,23 +24,25 @@ namespace Audio
 
         void start(int channels, int sourceRate, int destRate);
         void stop();
+        bool isOpened() const;
+
         size_t processBuffer(const void* source, size_t sourceLength, size_t& sourceProcessed,
                              void* dest, size_t destCapacity);
-        int sourceRate();
+        int sourceRate() const;
-        int destRate();
+        int destRate() const;
-        size_t getDestLength(size_t sourceLen);
+        size_t getDestLength(size_t sourceLen) const;
-        size_t getSourceLength(size_t destLen);
+        size_t getSourceLength(size_t destLen) const;
 
         // Returns instance + speex encoder size in bytes
         size_t getSize() const;
 
     protected:
-        void* mContext;
+        void* mContext = nullptr;
-        int   mErrorCode;
+        int   mErrorCode = 0;
-        int   mSourceRate,
+        int   mSourceRate = 0,
-        mDestRate,
+              mDestRate = 0,
-        mChannels;
+              mChannels = 0;
-        short mLastSample;
+        short mLastSample = 0;
     };
 
     typedef SpeexResampler Resampler;

src/engine/audio/Audio_WavFile.cpp

@@ -60,17 +60,11 @@ std::string WavFileReader::readChunk()
     uint32_t size = 0;
     readBuffer(&size, 4);
 
-    if (result == "fact")
+    if (result == "data")
-    {
-        uint32_t dataLength = 0;
-        readBuffer(&dataLength, sizeof dataLength);
-        mDataLength = dataLength;
-    }
-    else
-    if (result != "data")
-        mInput->seekg(size, std::ios_base::beg);
-    else
         mDataLength = size;
+    else
+        // Skip the chunk body; RIFF chunks are word-aligned, so odd sizes carry a pad byte
+        mInput->seekg(std::streamoff(size + (size & 1)), std::ios_base::cur);
 
     return result;
 }
@@ -151,7 +145,9 @@ bool WavFileReader::open(const std::filesystem::path& p)
         mBits = 0;
         readBuffer(&mBits, sizeof(mBits));
 
-        if (mBits !=8 && mBits != 16)
+        // Only 16-bit PCM is supported: the read path feeds the data
+        // directly into a 16-bit resampler.
+        if (mBits != 16)
             THROW_READERROR;
 
         // Look for the chunk 'data'
@@ -222,7 +218,8 @@ size_t WavFileReader::read(short* buffer, size_t samples)
         auto filePosition = mInput->tellg();
 
         // Check how much data we can read
-        size_t fileAvailable = mDataLength + mDataOffset - filePosition;
+        std::streamoff dataEnd = std::streamoff(mDataLength) + mDataOffset;
+        size_t fileAvailable = filePosition < dataEnd ? size_t(dataEnd - filePosition) : 0;
         requiredBytes = fileAvailable < requiredBytes ? fileAvailable : requiredBytes;
     }
 
@@ -254,8 +251,9 @@ size_t WavFileReader::readRaw(short* buffer, size_t samples)
         auto filePosition = mInput->tellg();
 
         // Check how much data we can read
-        size_t fileAvailable = mDataLength + mDataOffset - filePosition;
+        std::streamoff dataEnd = std::streamoff(mDataLength) + mDataOffset;
-        requiredBytes = (int)fileAvailable < requiredBytes ? (int)fileAvailable : requiredBytes;
+        size_t fileAvailable = filePosition < dataEnd ? size_t(dataEnd - filePosition) : 0;
+        requiredBytes = fileAvailable < requiredBytes ? fileAvailable : requiredBytes;
     }
 
     size_t readBytes = tryReadBuffer(buffer, requiredBytes);
@@ -332,10 +330,11 @@ bool WavFileWriter::open(const std::filesystem::path& p, int samplerate, int cha
     mChannels = channels;
 
     mOutput = std::make_unique<std::ofstream>(p, std::ios::binary | std::ios::trunc);
-    if (!mOutput)
+    if (!mOutput->is_open())
     {
         int errorcode = errno;
         ICELogError(<< "Failed to create .wav file: filename = " << p << " , error = " << errorcode);
+        mOutput.reset();
         return false;
     }
 
@@ -420,7 +419,7 @@ size_t WavFileWriter::write(const void* buffer, size_t bytes)
 bool WavFileWriter::isOpened() const
 {
     LOCK;
-    return mOutput.get();
+    return mOutput && mOutput->is_open();
 }
 
 std::filesystem::path WavFileWriter::path() const

src/engine/endpoint/EP_AudioProvider.cpp

@@ -93,11 +93,13 @@ void AudioProvider::updateSdpOffer(resip::SdpContents::Session::Medium& sdp, Sdp
         // Check if SRTP suite is found already or not
         if (mSrtpSuite == SRTP_NONE)
         {
+            // RFC 4568 requires a unique tag per crypto attribute; use the suite id.
             for (int suite = SRTP_AES_128_AUTH_80; suite <= SRTP_LAST; suite++)
-                sdp.addAttribute("crypto", resip::Data(createCryptoAttribute((SrtpSuite)suite)));
+                sdp.addAttribute("crypto", resip::Data(createCryptoAttribute((SrtpSuite)suite, suite)));
         }
         else
-            sdp.addAttribute("crypto", resip::Data(createCryptoAttribute(mSrtpSuite)));
+            // Answer/re-offer: echo the tag of the negotiated attribute.
+            sdp.addAttribute("crypto", resip::Data(createCryptoAttribute(mSrtpSuite, mSrtpTag)));
     }
 
     // Use CodecListPriority mCodecPriority adapter to work with codec priorities
@@ -246,11 +248,13 @@ bool AudioProvider::processSdpOffer(const resip::SdpContents::Session::Medium& m
         {
             const resip::Data& attr = *attrIter;
             ByteBuffer tempkey;
-            SrtpSuite suite = processCryptoAttribute(attr, tempkey);
+            int tag = 1;
-            if (suite > ss)
+            SrtpSuite suite = processCryptoAttribute(attr, tempkey, &tag);
+            if (srtpSuiteStrength(suite) > srtpSuiteStrength(ss))
             {
                 ss = suite;
                 mSrtpSuite = suite;
+                mSrtpTag = tag;
                 key = tempkey;
             }
         }
@@ -295,26 +299,30 @@ MT::PStream AudioProvider::activeStream()
     return mActiveStream;
 }
 
-std::string AudioProvider::createCryptoAttribute(SrtpSuite suite)
+std::string AudioProvider::createCryptoAttribute(SrtpSuite suite, int tag)
 {
     if (!mActiveStream)
         return "";
 
     // Print key to base64 string
     PByteBuffer keyBuffer = mActiveStream->srtp().outgoingKey(suite).first;
+    if (!keyBuffer)
+        return "";
     resip::Data d(keyBuffer->data(), keyBuffer->size());
     resip::Data keyText = d.base64encode();
 
-    return std::format("{} {} inline:{}", 1, toString(suite), keyText.c_str());
+    return std::format("{} {} inline:{}", tag, toString(suite), keyText.c_str());
 }
 
-SrtpSuite AudioProvider::processCryptoAttribute(const resip::Data& value, ByteBuffer& key)
+SrtpSuite AudioProvider::processCryptoAttribute(const resip::Data& value, ByteBuffer& key, int* tag)
 {
     int srtpTag = 0;
     char suite[64], keyChunk[256];
     int components = sscanf(value.c_str(), "%d %63s inline: %255s", &srtpTag, suite, keyChunk);
     if (components != 3)
         return SRTP_NONE;
+    if (tag)
+        *tag = srtpTag;
 
     const char* delimiter = strchr(keyChunk, '|');
     resip::Data keyText;

src/engine/endpoint/EP_AudioProvider.h

@@ -74,7 +74,7 @@ public:
   void setupMirror(bool enable);
 
   void configureMediaObserver(MT::Stream::MediaObserver* observer, void* userTag);
-  static SrtpSuite processCryptoAttribute(const resip::Data& value, ByteBuffer& key);
+  static SrtpSuite processCryptoAttribute(const resip::Data& value, ByteBuffer& key, int* tag = nullptr);
 
 protected:
   // SDP's stream name
@@ -93,6 +93,7 @@ protected:
 
   unsigned mState;
   SrtpSuite mSrtpSuite;
+  int mSrtpTag = 1;                       // RFC 4568 tag of the negotiated crypto attribute
   struct RemoteCodec
   {
     RemoteCodec(MT::Codec::Factory* factory, int payloadType)
@@ -109,7 +110,7 @@ protected:
   MT::Stream::MediaObserver* mMediaObserver = nullptr;
   void* mMediaObserverTag = nullptr;
 
-  std::string createCryptoAttribute(SrtpSuite suite);
+  std::string createCryptoAttribute(SrtpSuite suite, int tag);
   void findRfc2833(const resip::SdpContents::Session::Medium::CodecContainer& codecs);
 
   // Implements setState() logic. This allows to be called from constructor (it is not virtual function)

src/engine/endpoint/EP_Engine.cpp

@@ -470,7 +470,9 @@ void UserAgent::process()
                         // Send generated packet via provider's method to allow custom scheme of encryption
                         ICELogDebug(<<"Sending ICE packet to " << buffer->remoteAddress().toStdString() << " with " << buffer->comment());
 
-                        PDatagramSocket s = iceComponentId == ICE_RTP_ID ? stream.socket4().mRtp : stream.socket4().mRtcp;
+                        RtpPair<PDatagramSocket>& pair = buffer->remoteAddress().family() == AF_INET6 ? stream.socket6() : stream.socket4();
+                        PDatagramSocket s = iceComponentId == ICE_RTP_ID ? pair.mRtp : pair.mRtcp;
+                        if (s)
                             stream.provider()->sendData(s, buffer->remoteAddress(), buffer->data(), buffer->size());
                         break;
                     }
@@ -805,7 +807,10 @@ void UserAgent::onEarlyMedia(resip::ClientInviteSessionHandle h, const resip::Si
 /// called when dialog enters the Early state - typically after getting 18x
 void UserAgent::onProvisional(resip::ClientInviteSessionHandle h, const resip::SipMessage& msg)
 {
-    PSession s = getUserSession(CAST2RESIPSESSION(h)->mSessionId);
+    ResipSession* rs = CAST2RESIPSESSION(h);
+    if (!rs)
+        return;
+    PSession s = getUserSession(rs->mSessionId);
     if (!s)
         return;
 
@@ -821,7 +826,10 @@ void UserAgent::onProvisional(resip::ClientInviteSessionHandle h, const resip::S
 /// called when a dialog initiated as a UAC enters the connected state
 void UserAgent::onConnected(resip::ClientInviteSessionHandle h, const resip::SipMessage& msg)
 {
-    PSession s = getUserSession(CAST2RESIPSESSION(h)->mSessionId);
+    ResipSession* rs = CAST2RESIPSESSION(h);
+    if (!rs)
+        return;
+    PSession s = getUserSession(rs->mSessionId);
     if (!s)
         return;
 
@@ -874,7 +882,10 @@ void UserAgent::onConnected(resip::InviteSessionHandle h, const resip::SipMessag
 
 void UserAgent::onTerminated(resip::InviteSessionHandle h, resip::InviteSessionHandler::TerminatedReason reason, const resip::SipMessage* related)
 {
-    PSession s = getUserSession(CAST2RESIPSESSION(h)->mSessionId);
+    ResipSession* rs = CAST2RESIPSESSION(h);
+    if (!rs)
+        return;
+    PSession s = getUserSession(rs->mSessionId);
     if (!s)
         return;
 
@@ -920,6 +931,8 @@ void UserAgent::onAnswer(resip::InviteSessionHandle h, const resip::SipMessage&
     if (!resipSession)
         return;
     Session* s = resipSession->session();
+    if (!s)
+        return;
 
     bool iceAvailable = true;
 
@@ -1069,7 +1082,8 @@ void UserAgent::onAnswer(resip::InviteSessionHandle h, const resip::SipMessage&
 /// Called when an SDP offer is received - must send an answer soon after this
 void UserAgent::onOffer(resip::InviteSessionHandle h, const resip::SipMessage& msg, const resip::SdpContents& sdp)
 {
-    PSession s = getUserSession(CAST2RESIPSESSION(h)->mSessionId);
+    ResipSession* rs = CAST2RESIPSESSION(h);
+    PSession s = rs ? getUserSession(rs->mSessionId) : PSession();
     if (!s)
     {
         h->reject(488);
@@ -1091,7 +1105,8 @@ void UserAgent::onOffer(resip::InviteSessionHandle h, const resip::SipMessage& m
 
     uint64_t version = sdp.session().origin().getVersion();
     std::string remoteIp = sdp.session().connection().getAddress().c_str();
-    int code;
+    // Default to 200: a retransmitted offer (same origin version) keeps the session.
+    int code = 200;
     if ((uint64_t)-1 == s->mRemoteOriginVersion)
     {
         code = s->processSdp(version, iceAvailable, icePwd, iceUfrag, remoteIp, sdp.session().media());
@@ -1299,6 +1314,8 @@ void UserAgent::onPresenceUpdate(PClientObserver observer, const std::string& pe
 void UserAgent::onNewSubscription(resip::ServerSubscriptionHandle h, const resip::SipMessage& sub)
 {
     ResipSession* s = CAST2RESIPSESSION(h);
+    if (!s)
+        return;
 
     // Get the event package name
     const char* event = sub.header(resip::h_Event).value().c_str();

src/engine/endpoint/EP_Session.cpp

@@ -346,6 +346,10 @@ void Session::stop()
             // Free socket
             SocketHeap::instance().freeSocketPair( dataStream.socket4() );
             SocketHeap::instance().freeSocketPair( dataStream.socket6() );
+
+            // Drop the references so the destructor's cleanup does not free them again
+            dataStream.setSocket4(RtpPair<PDatagramSocket>());
+            dataStream.setSocket6(RtpPair<PDatagramSocket>());
         }
     }
 
@@ -475,7 +479,7 @@ void Session::getSessionInfo(Session::InfoOptions options, VariantMap& info)
     if (stat.mReceivedRtp)
         info[SessionInfo_PacketLoss] = static_cast<int>((stat.mPacketLoss * 1000) / stat.mReceivedRtp);
 
-    if (media)
+    if (media && mIceStack)
         info[SessionInfo_AudioPeer] = mIceStack->remoteAddress(media->iceInfo().mStreamId, media->iceInfo().mComponentId.mRtp).toStdString();
 
     info[SessionInfo_Jitter] = stat.mJitter;
@@ -485,7 +489,8 @@ void Session::getSessionInfo(Session::InfoOptions options, VariantMap& info)
     info[SessionInfo_BitrateSwitchCounter] = stat.mBitrateSwitchCounter;
     info[SessionInfo_CngCounter] = stat.mCng;
 #endif
-    info[SessionInfo_SSRC] = stat.mSsrc;
+    // Variant stores VTYPE_INT here; keep the 32 bits (consumers read it back with asInt()).
+    info[SessionInfo_SSRC] = static_cast<int>(stat.mSsrc);
     info[SessionInfo_RemotePeer] = stat.mRemotePeer.toStdString();
 }
 
@@ -741,9 +746,12 @@ PDataProvider Session::findProviderByPort(int family, unsigned short port)
     {
         Stream& s = mStreamList[i];
 
-        if ((s.socket4().mRtp->localport() == port || s.socket4().mRtcp->localport() == port) && family == AF_INET)
+        // Sockets may not be allocated yet (stream created from SDP, sockets follow later)
+        if (family == AF_INET && s.socket4().mRtp && s.socket4().mRtcp &&
+            (s.socket4().mRtp->localport() == port || s.socket4().mRtcp->localport() == port))
             return s.provider();
-        if ((s.socket6().mRtp->localport() == port || s.socket6().mRtcp->localport() == port) && family == AF_INET6)
+        if (family == AF_INET6 && s.socket6().mRtp && s.socket6().mRtcp &&
+            (s.socket6().mRtp->localport() == port || s.socket6().mRtcp->localport() == port))
             return s.provider();
     }
 

src/engine/helper/HL_IuUP.cpp

@@ -5,6 +5,10 @@ bool IuUP::TwoBytePseudoheader = false;
 
 bool IuUP::parse(const uint8_t *packet, int size, IuUP::Frame &result)
 {
+    // Data-with-CRC frames carry a 4 byte header
+    if (size < 4)
+        return false;
+
     // Wrap incoming packet in byte buffer
     BitReader reader(packet, size);
 
@@ -45,6 +49,10 @@ bool IuUP::parse2(const uint8_t* packet, int size, Frame& result)
         size -= 2;
     }
 
+    // Frame header is 3 bytes (no CRC) or 4 bytes (with CRC)
+    if (size < 3)
+        return false;
+
     BitReader reader(packet, size);
     result.mPduType = (PduType)reader.readBits(4);
 
@@ -52,6 +60,9 @@ bool IuUP::parse2(const uint8_t* packet, int size, Frame& result)
     if (result.mPduType != PduType::DataNoCrc && result.mPduType != PduType::DataWithCrc)
         return false;
 
+    if (result.mPduType == PduType::DataWithCrc && size < 4)
+        return false;
+
     result.mFrameNumber = reader.readBits(4);
     result.mFqc = reader.readBits(2);
     result.mRfci = reader.readBits(6);

src/engine/helper/HL_Rtp.cpp

@@ -93,7 +93,8 @@ bool RtpHelper::isRtp(const void* buffer, size_t length)
 
 bool RtpHelper::isRtpOrRtcp(const void* buffer, size_t length)
 {
-    if (length < 12)
+    // A minimal RTCP packet (e.g. an empty receiver report) is 8 bytes
+    if (length < 8)
         return false;
     const RtcpHeader* h = reinterpret_cast<const RtcpHeader*>(buffer);
     return h->version == 2;
@@ -390,7 +391,8 @@ void RtpDump::add(const void* buffer, size_t len, uint32_t offsetMs)
     if (!buffer || len == 0)
         return;
 
-    if (len > MAX_RTP_PACKET_SIZE)
+    // The record length field is 16-bit and covers payload + 8 byte header
+    if (len > MAX_RTP_PACKET_SIZE - 8)
         throw std::runtime_error("Packet too large: " + std::to_string(len));
 
     RtpData entry;

src/engine/helper/HL_SocketHeap.cpp

@@ -97,7 +97,20 @@ RtpPair<PDatagramSocket> SocketHeap::allocSocketPair(int family, SocketSink *sin
                 rtcp = allocSocket(family, sink, rtp->localport() + 1);
         }
         catch(...)
-        {}
+        {
+            // Release a partially allocated pair before retrying - otherwise
+            // the RTP socket from this attempt leaks into the socket map.
+            if (rtp)
+            {
+                freeSocket(rtp);
+                rtp.reset();
+            }
+            if (rtcp)
+            {
+                freeSocket(rtcp);
+                rtcp.reset();
+            }
+        }
     }
 
     if (!rtp || !rtcp)
@@ -139,6 +152,9 @@ PDatagramSocket SocketHeap::allocSocket(int family, SocketSink* sink, int port)
     sockaddr_in6 addr6;
     int result = 0;
     int testport;
+    // A fixed port cannot be retried (it would loop forever if the port is
+    // owned by another process); random ports get a bounded number of attempts.
+    int attemptsLeft = port ? 1 : 100;
     do
     {
         testport = port ? port : rand() % ((mFinish - mStart) / 2) * 2 + mStart;
@@ -164,7 +180,7 @@ PDatagramSocket SocketHeap::allocSocket(int family, SocketSink* sink, int port)
             break;
         }
 
-    } while (result == WSAEADDRINUSE);
+    } while (result == WSAEADDRINUSE && --attemptsLeft > 0);
 
     if (result)
     {

src/engine/helper/HL_String.cpp

@@ -48,9 +48,12 @@ std::string strx::appendPath(const std::string& s1, const std::string& s2)
 std::string strx::makeUtf8(const std::tstring &arg)
 {
 #if defined(TARGET_WIN)
-    size_t required = WideCharToMultiByte(CP_UTF8, 0, arg.c_str(), -1, NULL, 0, NULL, NULL);
+    int required = WideCharToMultiByte(CP_UTF8, 0, arg.c_str(), -1, NULL, 0, NULL, NULL);
-    char *result = (char*)_alloca(required + 1);
+    if (required <= 0)
-    WideCharToMultiByte(CP_UTF8, 0, arg.c_str(), -1, result, required+1, NULL, NULL);
+        return std::string();
+    std::string result(static_cast<size_t>(required), '\0');
+    WideCharToMultiByte(CP_UTF8, 0, arg.c_str(), -1, &result[0], required, NULL, NULL);
+    result.resize(strlen(result.c_str())); // strip the trailing NUL written by the API
     return result;
 #else
     return arg;
@@ -65,9 +68,12 @@ std::string strx::toUtf8(const std::tstring &arg)
 std::tstring strx::makeTstring(const std::string& arg)
 {
 #if defined(TARGET_WIN)
-    size_t count = MultiByteToWideChar(CP_UTF8, 0, arg.c_str(), -1, NULL, 0);
+    int count = MultiByteToWideChar(CP_UTF8, 0, arg.c_str(), -1, NULL, 0);
-    wchar_t* result = (wchar_t*)_alloca(count * 2);
+    if (count <= 0)
-    MultiByteToWideChar(CP_UTF8, 0, arg.c_str(), -1, result, count);
+        return std::tstring();
+    std::wstring result(static_cast<size_t>(count), L'\0');
+    MultiByteToWideChar(CP_UTF8, 0, arg.c_str(), -1, &result[0], count);
+    result.resize(wcslen(result.c_str())); // strip the trailing NUL written by the API
     return result;
 #else
     return arg;
@@ -93,11 +99,7 @@ int strx::toInt(const char *s, int defaultValue, bool* isOk)
 uint64_t strx::toUint64(const char* s, uint64_t def, bool *isOk)
 {
     uint64_t result = def;
-#if defined(TARGET_WIN)
+    if (sscanf(s, "%" SCNu64, &result) != 1)
-    if (sscanf(s, "%I64d", &result) != 1)
-#else
-    if (sscanf(s, "%llu", &result) != 1)
-#endif
     {
         if (isOk)
             *isOk = false;
@@ -143,7 +145,6 @@ std::string strx::toHex(const uint8_t* input, size_t inputLength)
         *r++ = hexmap[hi];
         *r++ = hexmap[low];
     }
-    *r = 0;
 
     return result;
 }
@@ -171,11 +172,9 @@ std::string strx::doubleToString(double value, int precision)
 
 const char* strx::findSubstring(const char* buffer, const char* substring, size_t bufferLength)
 {
-#if defined(TARGET_WIN)
+    // The buffer is not necessarily NUL-terminated, so a bounded search is
-    return (const char*)strstr(buffer, substring);
+    // required on every platform (a memmem replacement for MSVC is provided below).
-#else
     return (const char*)memmem(buffer, bufferLength, substring, strlen(substring));
-#endif
 }
 
 
@@ -332,7 +331,7 @@ std::string strx::fromHex2String(const std::string& s)
     std::string result; result.resize(s.size() / 2);
     const char* t = s.c_str();
     for (size_t i = 0; i < result.size(); i++)
-        result[i] = hex2code(t[i*2]);
+        result[i] = static_cast<char>((hex2code(t[i*2]) << 4) | hex2code(t[i*2+1]));
 
     return result;
 }
@@ -367,12 +366,13 @@ std::string strx::decodeUri(const std::string& s)
 
     char ch;
 
-    int i, ii;
+    int i, ii = 0;
     for (i=0; i<(int)s.length(); i++)
     {
-        if (s[i] == 37)
+        if (s[i] == '%' && i + 2 < (int)s.length())
+        {
+            if (sscanf(s.substr(i+1,2).c_str(), "%x", &ii) == 1)
             {
-            sscanf(s.substr(i+1,2).c_str(), "%x", &ii);
                 ch = static_cast<char>(ii);
                 ret += ch;
                 i += 2;
@@ -380,19 +380,24 @@ std::string strx::decodeUri(const std::string& s)
             else
                 ret += s[i];
         }
+        else
+            ret += s[i];
+    }
     return ret;
 }
 
 bool strx::startsWith(const std::string& s, const std::string& prefix)
 {
-    std::string::size_type p = s.find(prefix);
+    if (prefix.size() > s.size())
-    return p == 0;
+        return false;
+    return s.compare(0, prefix.size(), prefix) == 0;
 }
 
 bool strx::endsWith(const std::string& s, const std::string& suffix)
 {
-    std::string::size_type p = s.rfind(suffix);
+    if (suffix.size() > s.size())
-    return (p == s.size() - suffix.size());
+        return false;
+    return s.compare(s.size() - suffix.size(), suffix.size(), suffix) == 0;
 }
 
 int strx::stringToDuration(const std::string& s)

src/engine/helper/HL_Sync.cpp

@@ -20,7 +20,8 @@
 void SyncHelper::delay(unsigned int microseconds)
 {
 #ifdef TARGET_WIN
-    ::Sleep(microseconds/1000);
+    // Round up so sub-millisecond delays do not become Sleep(0)
+    ::Sleep((microseconds + 999) / 1000);
 #endif
 #if defined(TARGET_OSX) || defined(TARGET_LINUX)
     timespec requested, remaining;
@@ -93,8 +94,9 @@ uint32_t chronox::getDelta(uint32_t later, uint32_t earlier)
     if (later > earlier)
         return later - earlier;
 
+    // Counter wrapped: unsigned subtraction yields the correct modulo-2^32 delta
     if (later < earlier && later < 0x7FFFFFFF && earlier >= 0x7FFFFFFF)
-        return 0xFFFFFFFF - earlier + later;
+        return later - earlier;
 
     return 0;
 }
@@ -115,8 +117,8 @@ uint64_t chronox::toTimestamp(const timeval& ts)
 
 int64_t chronox::getDelta(const timespec& a, const timespec& b)
 {
-    uint64_t ms_a = (uint64_t)a.tv_sec * 1000 + a.tv_nsec / 10000000;
+    int64_t ms_a = (int64_t)a.tv_sec * 1000 + a.tv_nsec / 1000000;
-    uint64_t ms_b = (uint64_t)b.tv_sec * 1000 + b.tv_nsec / 10000000;
+    int64_t ms_b = (int64_t)b.tv_sec * 1000 + b.tv_nsec / 1000000;
     return ms_a - ms_b;
 }
 
@@ -162,13 +164,11 @@ void BufferQueue::push(const void* data, int bytes)
 BufferQueue::PBlock BufferQueue::pull(int milliseconds)
 {
     std::unique_lock<std::mutex> l(mMutex);
-    std::cv_status status = mBlockList.empty() ? std::cv_status::timeout : std::cv_status::no_timeout;
+    mSignal.wait_for(l, std::chrono::milliseconds(milliseconds),
-
+                     [this]() { return !mBlockList.empty(); });
-    if (mBlockList.empty())
-        status = mSignal.wait_for(l, std::chrono::milliseconds(milliseconds));
 
     PBlock r;
-    if (status == std::cv_status::no_timeout && !mBlockList.empty())
+    if (!mBlockList.empty())
     {
         r = mBlockList.front();
         mBlockList.pop_front();

src/engine/media/MT_AmrCodec.cpp

@@ -265,7 +265,19 @@ PCodec AmrNbCodec::CodecFactory::create()
 AmrNbCodec::AmrNbCodec(const AmrCodecConfig& config)
     :mConfig(config)
 {
+    // Contexts are created lazily (see ensureEncoder/ensureDecoder) - a codec
+    // resolved only for network-MOS metadata never allocates them.
+}
+
+void AmrNbCodec::ensureEncoder()
+{
+    if (!mEncoderCtx)
         mEncoderCtx = Encoder_Interface_init(1);
+}
+
+void AmrNbCodec::ensureDecoder()
+{
+    if (!mDecoderCtx)
         mDecoderCtx = Decoder_Interface_init();
 }
 
@@ -298,6 +310,8 @@ Codec::Info AmrNbCodec::info()
 
 Codec::EncodeResult AmrNbCodec::encode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
+    ensureEncoder();
+
     if (input.size_bytes() % pcmLength())
         return {.mEncoded = 0};
 
@@ -324,6 +338,8 @@ Codec::EncodeResult AmrNbCodec::encode(std::span<const uint8_t> input, std::span
 #define AMR_BITRATE_DTX 15
 Codec::DecodeResult AmrNbCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
+    ensureDecoder();
+
     if (mConfig.mOctetAligned)
         return {.mDecoded = 0};
 
@@ -427,6 +443,8 @@ Codec::DecodeResult AmrNbCodec::decode(std::span<const uint8_t> input, std::span
 
 size_t AmrNbCodec::plc(int lostFrames, std::span<uint8_t> output)
 {
+    ensureDecoder();
+
     if (output.size_bytes() < lostFrames * pcmLength())
         return 0;
 
@@ -496,6 +514,13 @@ AmrWbStatistics MT::GAmrWbStatistics;
 AmrWbCodec::AmrWbCodec(const AmrCodecConfig& config)
     :mConfig(config)
 {
+    // Decoder context is created lazily (see ensureDecoder) - a codec resolved
+    // only for network-MOS metadata never allocates the AMR-WB decoder state.
+}
+
+void AmrWbCodec::ensureDecoder()
+{
+    if (!mDecoderCtx)
         mDecoderCtx = D_IF_init();
 }
 
@@ -630,6 +655,8 @@ Codec::DecodeResult AmrWbCodec::decodePlain(std::span<const uint8_t> input, std:
 
 Codec::DecodeResult AmrWbCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
+    ensureDecoder();
+
     if (mConfig.mIuUP)
         return decodeIuup(input, output);
     else

src/engine/media/MT_AmrCodec.h

@@ -33,6 +33,13 @@ protected:
     int mPreviousPacketLength = 0;
     size_t mCngCounter = 0;
     size_t mSwitchCounter = 0;
+
+    // opencore-amr encoder/decoder state is allocated lazily on first encode/decode.
+    // Network-MOS-only streams resolve codec metadata (name/samplerate/frame timing)
+    // but never decode, so they must not pay for a context they never use - at scale
+    // this is ~a decoder state (several KB) saved per network-only stream.
+    void ensureEncoder();
+    void ensureDecoder();
 public:
     class CodecFactory: public Factory
     {
@@ -85,6 +92,10 @@ protected:
 
     int mPreviousPacketLength;
 
+    // Decoder state is allocated lazily on first decode/plc (see AmrNbCodec) so
+    // network-MOS-only streams never instantiate the AMR-WB decoder.
+    void ensureDecoder();
+
     DecodeResult decodeIuup(std::span<const uint8_t> input, std::span<uint8_t> output);
     DecodeResult decodePlain(std::span<const uint8_t> input, std::span<uint8_t> output);
 

src/engine/media/MT_AudioCodec.cpp

@@ -24,6 +24,7 @@
 #include <memory.h>
 #include <string.h>
 #include <algorithm>
+#include <vector>
 
 #define LOG_SUBSYSTEM "media"
 
@@ -434,9 +435,10 @@ Codec::Info OpusCodec::info() {
 
 Codec::EncodeResult OpusCodec::encode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
-    // Send number of samples for input and number of bytes for output
+    // opus_encode() takes the frame size in samples per channel and the output
+    // capacity in bytes.
     int written = opus_encode(mEncoderCtx, (const opus_int16*)input.data(), input.size_bytes() / (sizeof(short) * channels()),
-                              output.data(), output.size_bytes() / (sizeof(short) * channels()));
+                              output.data(), output.size_bytes());
     if (written < 0)
         return {.mEncoded = 0};
     else
@@ -445,10 +447,13 @@ Codec::EncodeResult OpusCodec::encode(std::span<const uint8_t> input, std::span<
 
 Codec::DecodeResult OpusCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
-    int result = 0;
+    if (input.empty())
+        return {0};
 
     // Examine the number of channels available in incoming packet
     int nr_of_channels = opus_packet_get_nb_channels(input.data());
+    if (nr_of_channels != 1 && nr_of_channels != 2)
+        return {0};
 
     // Recreate decoder if needed
     if (mDecoderChannels != nr_of_channels)
@@ -473,80 +478,97 @@ Codec::DecodeResult OpusCodec::decode(std::span<const uint8_t> input, std::span<
     if (nr_of_frames <= 0)
         return {0};
 
-    // We support stereo and mono here.
+    // Output must match channels() - that is what info() promises downstream.
-    int buffer_capacity = nr_of_frames * sizeof(opus_int16) * nr_of_channels;
+    size_t needed = (size_t)nr_of_frames * sizeof(opus_int16) * channels();
-    opus_int16 *buffer_decode = (opus_int16 *)alloca(buffer_capacity);
+    if (needed > output.size_bytes())
+        return {0};
+
+    if (nr_of_channels == channels())
+    {
         int decoded = opus_decode(mDecoderCtx, input.data(), input.size_bytes(),
-                              buffer_decode, nr_of_frames, 0);
+                                  (opus_int16*)output.data(), nr_of_frames, 0);
+        if (decoded < 0)
+        {
+            ICELogCritical(<< "opus_decode() returned " << decoded);
+            return {0};
+        }
+        return {.mDecoded = (size_t)decoded * sizeof(opus_int16) * nr_of_channels};
+    }
+
+    // Channel count differs from the negotiated one - decode to a temporary
+    // buffer and convert.
+    std::vector<opus_int16> temp((size_t)nr_of_frames * nr_of_channels);
+    int decoded = opus_decode(mDecoderCtx, input.data(), input.size_bytes(),
+                              temp.data(), nr_of_frames, 0);
     if (decoded < 0)
     {
         ICELogCritical(<< "opus_decode() returned " << decoded);
         return {0};
     }
 
-    opus_int16 *buffer_stereo = nullptr;
+    opus_int16* out = (opus_int16*)output.data();
-    int buffer_stereo_capacity = buffer_capacity * 2;
+    if (channels() == 2 && nr_of_channels == 1)
-
+    {
-    switch (nr_of_channels) {
+        for (int i = 0; i < decoded; i++)
-        case 1:
+            out[i * 2] = out[i * 2 + 1] = temp[i];
-            // Convert to stereo before
+        return {.mDecoded = (size_t)decoded * sizeof(opus_int16) * 2};
-            buffer_stereo = (opus_int16 *) alloca(buffer_stereo_capacity);
-            for (int i = 0; i < nr_of_frames; i++) {
-                buffer_stereo[i * 2 + 1] = buffer_decode[i];
-                buffer_stereo[i * 2] = buffer_decode[i];
     }
-            assert(buffer_stereo_capacity <= output.size_bytes());
+    else // mono negotiated, stereo packet
-            memcpy(output.data(), buffer_stereo, buffer_stereo_capacity);
+    {
-            result = buffer_stereo_capacity;
+        for (int i = 0; i < decoded; i++)
-            break;
+            out[i] = (opus_int16)((int(temp[i * 2]) + temp[i * 2 + 1]) / 2);
-
+        return {.mDecoded = (size_t)decoded * sizeof(opus_int16)};
-        case 2:
-            assert(buffer_capacity <= output.size_bytes());
-            memcpy(output.data(), buffer_decode, buffer_capacity);
-            result = buffer_capacity;
-            break;
-
-        default:
-            assert(0);
     }
-
-    return {.mDecoded = (size_t)result};
 }
 
 size_t OpusCodec::plc(int lostPackets, std::span<uint8_t> output)
 {
-    // Find how much frames do we need to produce and prefill it with silence
+    if (lostPackets <= 0 || output.empty())
-    int frames_per_packet = (int)pcmLength() / (sizeof(opus_int16) * channels());
+        return 0;
-    memset(output.data(), 0, output.size_bytes());
 
-    // Use this pointer as output
+    // Total bytes we are asked to conceal, clamped to the output capacity.
-    opus_int16* data_output = reinterpret_cast<opus_int16*>(output.data());
+    size_t packet_bytes = (size_t)pcmLength();
+    size_t total = std::min(output.size_bytes(), packet_bytes * (size_t)lostPackets);
+    memset(output.data(), 0, total);
 
-    int nr_of_decoded_frames = 0;
+    // No decoder yet (PLC before the first decoded packet) - leave silence.
+    if (!mDecoderCtx || (mDecoderChannels != 1 && mDecoderChannels != 2))
+        return total;
 
-    // Buffer for single lost frame
+    int samples_per_packet = (int)(packet_bytes / (sizeof(opus_int16) * channels()));
-    opus_int16* buffer_plc = (opus_int16*)alloca(frames_per_packet * mDecoderChannels * sizeof(opus_int16));
+    if (samples_per_packet <= 0)
-    for (int i=0; i<lostPackets; i++)
+        return total;
+
+    opus_int16* out = reinterpret_cast<opus_int16*>(output.data());
+    std::vector<opus_int16> temp((size_t)samples_per_packet * mDecoderChannels);
+
+    for (int packet = 0; packet < lostPackets; packet++)
     {
-        nr_of_decoded_frames = opus_decode(mDecoderCtx, nullptr, 0, buffer_plc, frames_per_packet, 0);
+        size_t offset_bytes = (size_t)packet * packet_bytes;
-        assert(nr_of_decoded_frames == frames_per_packet);
+        if (offset_bytes + packet_bytes > total)
-        switch (mDecoderChannels)
-        {
-            case 1:
-                // Convert mono to stereo
-                for (int i=0; i < nr_of_decoded_frames; i++)
-                    data_output[i * 2] = data_output[i * 2 + 1] = buffer_plc[i];
-                data_output += frames_per_packet * mChannels;
             break;
 
-            case 2:
+        int decoded = opus_decode(mDecoderCtx, nullptr, 0, temp.data(), samples_per_packet, 0);
-                // Just copy data
+        if (decoded <= 0)
-                memcpy(data_output, buffer_plc, frames_per_packet * sizeof(opus_int16) * mDecoderChannels);
+            break; // keep the pre-filled silence
-                data_output += frames_per_packet * mChannels;
+
-                break;
+        opus_int16* dst = out + offset_bytes / sizeof(opus_int16);
+        if (mDecoderChannels == channels())
+        {
+            memcpy(dst, temp.data(), (size_t)decoded * sizeof(opus_int16) * mDecoderChannels);
+        }
+        else if (channels() == 2 && mDecoderChannels == 1)
+        {
+            for (int i = 0; i < decoded; i++)
+                dst[i * 2] = dst[i * 2 + 1] = temp[i];
+        }
+        else // mono negotiated, stereo decoder
+        {
+            for (int i = 0; i < decoded; i++)
+                dst[i] = (opus_int16)((int(temp[i * 2]) + temp[i * 2 + 1]) / 2);
         }
     }
-    return ((uint8_t*)data_output - output.data());
+    return total;
 }
 
 size_t OpusCodec::getNumberOfSamples(std::span<const uint8_t> payload)
@@ -1021,14 +1043,29 @@ Codec::EncodeResult GsmCodec::encode(std::span<const uint8_t> input, std::span<u
 
 Codec::DecodeResult GsmCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
-    if (input.size_bytes() % rtpLength() != 0)
+    const size_t frameSize = (size_t)rtpLength();
+    if (!frameSize || input.size_bytes() % frameSize != 0)
         return {.mDecoded = 0};
 
-    int i=0;
+    // Bytes_65 carries a WAV49 frame pair (33 + 32 bytes) and produces 320 samples
-    for (i = 0; i < input.size_bytes() / rtpLength(); i++)
+    const size_t pcmPerFrame = (mCodecType == Type::Bytes_65) ? 640 : 320;
-        gsm_decode(mGSM, (gsm_byte *)input.data() + 33 * i, (gsm_signal *)output.data() + 160 * i);
+    size_t frames = input.size_bytes() / frameSize;
 
-    return {.mDecoded = (size_t)i * 320};
+    size_t i;
+    for (i = 0; i < frames; i++)
+    {
+        if ((i + 1) * pcmPerFrame > output.size_bytes())
+            break;
+
+        const gsm_byte* in = (const gsm_byte*)input.data() + frameSize * i;
+        gsm_signal* out = (gsm_signal*)output.data() + (pcmPerFrame / 2) * i;
+
+        gsm_decode(mGSM, (gsm_byte*)in, out);
+        if (mCodecType == Type::Bytes_65)
+            gsm_decode(mGSM, (gsm_byte*)(in + 33), out + 160);
+    }
+
+    return {.mDecoded = i * pcmPerFrame};
 }
 
 size_t GsmCodec::plc(int lostFrames, std::span<uint8_t> output)
@@ -1327,8 +1364,11 @@ hr_ref_from_canon(uint16_t *hr_ref, const uint8_t *canon)
 */
 Codec::DecodeResult GsmHrCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
-    ByteBuffer bb(input, ByteBuffer::CopyBehavior::UseExternal);
+    // hr_ref_from_canon() reads 112 bits (14 bytes) starting at offset 1,
-    BitReader br(bb);
+    // and the decoder produces 160 samples (320 bytes).
+    if (input.size_bytes() < 15 || output.size_bytes() < 320)
+        return {.mDecoded = 0};
+
     uint16_t hr_ref[22];
 
     hr_ref_from_canon(hr_ref, input.data() + 1);

src/engine/media/MT_AudioReceiver.cpp

@@ -28,8 +28,7 @@ using namespace MT;
 // ----------------- RtpBuffer::Packet --------------
 RtpBuffer::Packet::Packet(const std::shared_ptr<RTPPacket>& packet, std::chrono::milliseconds timelength, int samplerate)
     :mRtp(packet), mTimelength(timelength), mSamplerate(samplerate)
-{
+{}
-}
 
 std::shared_ptr<RTPPacket> RtpBuffer::Packet::rtp() const
 {
@@ -60,12 +59,11 @@ std::vector<short>& RtpBuffer::Packet::pcm()
 RtpBuffer::RtpBuffer(Statistics& stat)
     :mStat(stat)
 {
-    if (mStat.mPacketLoss)
-        std::cout << "Warning: packet loss is not zero" << std::endl;
 }
 
 RtpBuffer::~RtpBuffer()
 {
+    if (mAddCounter)
         ICELogDebug(<< "Number of add packets: " << mAddCounter << ", number of retrieved packets " << mReturnedCounter);
 }
 
@@ -126,19 +124,19 @@ std::shared_ptr<RtpBuffer::Packet> RtpBuffer::add(const std::shared_ptr<jrtplib:
         mStat.mPacketInterval.process(t - mLastAddTime);
         mLastAddTime = t;
     }
-    mStat.mSsrc = static_cast<uint16_t>(packet->GetSSRC());
+    mStat.mSsrc = packet->GetSSRC();
 
     // Update jitter
-    ICELogMedia(<< "Adding new packet into jitter buffer");
+    ICELogMedia(<< "Adding new packet seqno " << packet->GetSequenceNumber() << " into jitter buffer");
     mAddCounter++;
 
     // Look for maximum&minimal sequence number; check for dublicates
-    unsigned maxno = 0xFFFFFFFF, minno = 0;
+    unsigned maxno = 0, minno = 0xFFFFFFFF;
 
     // New sequence number
     unsigned newSeqno = packet->GetExtendedSequenceNumber();
 
-    for (std::shared_ptr<Packet>& p: mPacketList)
+    for (auto& p: mPacketList)
     {
         unsigned seqno = p->rtp()->GetExtendedSequenceNumber();
 
@@ -171,7 +169,7 @@ std::shared_ptr<RtpBuffer::Packet> RtpBuffer::add(const std::shared_ptr<jrtplib:
         available = findTimelength();
 
         if (available > mHigh)
-            ICELogMedia(<< "Available " << available << "ms with limit " << mHigh << "ms");
+            ICELogMedia(<< "Available " << available << " with limit " << mHigh);
 
         return p;
     }
@@ -186,16 +184,18 @@ std::shared_ptr<RtpBuffer::Packet> RtpBuffer::add(const std::shared_ptr<jrtplib:
     return std::shared_ptr<Packet>();
 }
 
-RtpBuffer::FetchResult RtpBuffer::fetch()
+void RtpBuffer::trimToHighWater(size_t maxPackets)
 {
     Lock l(mGuard);
 
-    FetchResult result;
-
-    // See if there is enough information in buffer
     auto total = findTimelength();
 
-    while (total > mHigh && mPacketList.size() > 1 && 0ms != mHigh)
+    // Drop the oldest packet while either bound is exceeded: the time-based
+    // high-water mark (mHigh, when set) or, if maxPackets != 0, the packet-count
+    // cap. Always keep at least one packet so loss/gap accounting has a reference.
+    while (mPacketList.size() > 1 &&
+           ((0ms != mHigh && total > mHigh) ||
+            (maxPackets != 0 && mPacketList.size() > maxPackets)))
     {
         ICELogMedia( << "Dropping RTP packets from jitter buffer");
         total -= mPacketList.front()->timelength();
@@ -235,6 +235,19 @@ RtpBuffer::FetchResult RtpBuffer::fetch()
         // Increase number in statistics
         mStat.mPacketDropped++;
     }
+}
+
+RtpBuffer::FetchResult RtpBuffer::fetch()
+{
+    Lock l(mGuard);
+
+    FetchResult result;
+
+    // Bound the buffer to the high-water mark before fetching.
+    trimToHighWater();
+
+    // See how much audio is buffered now.
+    auto total = findTimelength();
 
     if (total < mLow || total == 0ms)
     {
@@ -346,33 +359,32 @@ int RtpBuffer::getNumberOfAddPackets() const
 //-------------- Receiver ---------------
 Receiver::Receiver(Statistics& stat)
     :mStat(stat)
-{
+{}
-}
 
 Receiver::~Receiver()
-{
+{}
-}
 
 //-------------- AudioReceiver ----------------
 AudioReceiver::AudioReceiver(const CodecList::Settings& settings, MT::Statistics &stat)
-    :Receiver(stat), mBuffer(stat), mDtmfBuffer(stat), mCodecSettings(settings), mCodecList(settings), mDtmfReceiver(stat)
+    :Receiver(stat), mRtpBuffer(stat), mDtmfBuffer(stat), mCodecSettings(settings), mCodecList(settings), mDtmfReceiver(stat)
 {
-    // Init resamplers
-    mResampler8.start(AUDIO_CHANNELS, 8000, AUDIO_SAMPLERATE);
-    mResampler16.start(AUDIO_CHANNELS, 16000, AUDIO_SAMPLERATE);
-    mResampler32.start(AUDIO_CHANNELS, 32000, AUDIO_SAMPLERATE);
-    mResampler48.start(AUDIO_CHANNELS, 48000, AUDIO_SAMPLERATE);
-
     // Init codecs
     mCodecList.setSettings(settings);
     mCodecList.fillCodecMap(mCodecMap);
 
-    mAvailable.setCapacity(AUDIO_SAMPLERATE * sizeof(short));
-
     mDtmfBuffer.setPrebuffer(0ms);
     mDtmfBuffer.setLow(0ms);
     mDtmfBuffer.setHigh(1ms);
 
+    // Avoid collecting too much data
+    mRtpBuffer.setHigh(240ms);
+
+    // Resamplers are lazy inside; there is no actual memory allocation
+    mResampler8.start(AUDIO_CHANNELS, 8000, AUDIO_SAMPLERATE);
+    mResampler16.start(AUDIO_CHANNELS, 16000, AUDIO_SAMPLERATE);
+    mResampler32.start(AUDIO_CHANNELS, 32000, AUDIO_SAMPLERATE);
+    mResampler48.start(AUDIO_CHANNELS, 48000, AUDIO_SAMPLERATE);
+
 #if defined(DUMP_DECODED)
     mDecodedDump = std::make_shared<Audio::WavFileWriter>();
     mDecodedDump->open("decoded.wav", 8000 /*G711*/, AUDIO_CHANNELS);
@@ -386,6 +398,11 @@ AudioReceiver::~AudioReceiver()
     mResampler32.stop();
     mResampler48.stop();
     mDecodedDump.reset();
+
+    if (mRequestedAudio != 0ms)
+        ICELogDebug(<< "Requested " << mRequestedAudio << ", produced " << mProducedAudio);
+    if (mDecodeCount)
+        ICELogDebug(<< "Average interval between packet decoding " << mIntervalBetweenDecode / mDecodeCount);
 }
 
 // Update codec settings
@@ -406,40 +423,6 @@ CodecList::Settings& AudioReceiver::getCodecSettings()
     return mCodecSettings;
 }
 
-size_t decode_packet(Codec& codec, RTPPacket& p, void* output_buffer, size_t output_capacity)
-{
-    // How much data was produced
-    size_t result = 0;
-
-    // Handle here regular RTP packets
-    // Check if payload length is ok
-    int tail = codec.rtpLength() ? p.GetPayloadLength() % codec.rtpLength() : 0;
-
-    if (!tail)
-    {
-        // Find number of frames
-        int frame_count = codec.rtpLength() ? p.GetPayloadLength() / codec.rtpLength() : 1;
-        int frame_length = codec.rtpLength() ? codec.rtpLength() : (int)p.GetPayloadLength();
-
-        // Save last packet time length
-        // mLastPacketTimeLength = mFrameCount * mCodec->frameTime();
-
-        // Decode
-
-        for (int i=0; i < frame_count; i++)
-        {
-            auto r = codec.decode({p.GetPayloadData() + i * codec.rtpLength(), (size_t)frame_length},
-                                  {(uint8_t*)output_buffer, output_capacity});
-
-            result += r.mDecoded;
-        }
-    }
-    else
-        ICELogMedia(<< "RTP packet with tail.");
-
-    return result;
-}
-
 Codec* AudioReceiver::add(const std::shared_ptr<jrtplib::RTPPacket>& p)
 {
     Codec* codec = nullptr;
@@ -450,7 +433,7 @@ Codec* AudioReceiver::add(const std::shared_ptr<jrtplib::RTPPacket>& p)
         payloadLength = p->GetPayloadLength(),
         ptype = p->GetPayloadType();
 
-    ICELogMedia(<< "Adding packet No " << p->GetSequenceNumber());
+    // ICELogMedia(<< "Adding packet No " << p->GetSequenceNumber());
 
     // Increase codec counter
     mStat.mCodecCount[ptype]++;
@@ -508,12 +491,12 @@ Codec* AudioReceiver::add(const std::shared_ptr<jrtplib::RTPPacket>& p)
         {
             // It will cause statistics to report about bad RTP packet
             // I have to replay last packet payload here to avoid report about lost packet
-            mBuffer.add(p, std::chrono::milliseconds(time_length), samplerate);
+            mRtpBuffer.add(p, std::chrono::milliseconds(time_length), samplerate);
             return nullptr;
         }
 
         // Queue packet to buffer
-        mBuffer.add(p, std::chrono::milliseconds(time_length), samplerate).get();
+        mRtpBuffer.add(p, std::chrono::milliseconds(time_length), samplerate).get();
     }
     return codec;
 }
@@ -522,32 +505,36 @@ void AudioReceiver::processDecoded(Audio::DataWindow& output, DecodeOptions opti
 {
     // Write to audio dump if requested
     if (mDecodedDump && mDecodedLength)
-        mDecodedDump->write(mDecodedFrame, mDecodedLength);
+        mDecodedDump->write(mDecodedFrame.data(), mDecodedLength);
 
     // Resample to target rate
     makeMonoAndResample(options.mResampleToMainRate ? mCodec->samplerate() : 0, mCodec->channels());
 
     // Send to output
-    output.add(mResampledFrame, mResampledLength);
+    output.add(mResampledFrame.data(), mResampledLength);
 }
 
 void AudioReceiver::produceSilence(std::chrono::milliseconds length, Audio::DataWindow& output, DecodeOptions options)
 {
+    if (!mCodec)
+        return;
+
     // Fill mDecodeBuffer as much as needed and call processDecoded()
     // Depending on used codec mono or stereo silence should be produced
+
     size_t chunks = length.count() / 10;
     size_t tail = length.count() % 10;
     size_t chunk_size = 10 * sizeof(int16_t) * mCodec->samplerate() / 1000 * mCodec->channels();
     size_t tail_size = tail * sizeof(int16_t) * mCodec->samplerate() / 1000 * mCodec->channels();
     for (size_t i = 0; i < chunks; i++)
     {
-        memset(mDecodedFrame, 0, chunk_size);
+        memset(mDecodedFrame.data(), 0, chunk_size);
         mDecodedLength = chunk_size;
         processDecoded(output, options);
     }
     if (tail)
     {
-        memset(mDecodedFrame, 0, tail_size);
+        memset(mDecodedFrame.data(), 0, tail_size);
         mDecodedLength = tail_size;
         processDecoded(output, options);
     }
@@ -561,7 +548,7 @@ void AudioReceiver::produceCNG(std::chrono::milliseconds length, Audio::DataWind
         if (options.mSkipDecode)
             mDecodedLength = 0;
         else
-            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), 100, mDecodedFrame, false);
+            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), 100, mDecodedFrame.data(), false);
 
         if (mDecodedLength)
             processDecoded(output, options);
@@ -574,7 +561,7 @@ void AudioReceiver::produceCNG(std::chrono::milliseconds length, Audio::DataWind
         if (options.mSkipDecode)
             mDecodedLength = 0;
         else
-            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), tail, reinterpret_cast<short*>(mDecodedFrame), false);
+            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), tail, reinterpret_cast<short*>(mDecodedFrame.data()), false);
 
         if (mDecodedLength)
             processDecoded(output, options);
@@ -592,7 +579,7 @@ AudioReceiver::DecodeResult AudioReceiver::decodeGapTo(Audio::DataWindow& output
         {
             // Synthesize comfort noise. It will be done on AUDIO_SAMPLERATE rate directly to mResampledFrame buffer.
             // Do not forget to send this noise to analysis
-            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), mLastPacketTimeLength, reinterpret_cast<short*>(mDecodedFrame), false);
+            mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), mLastPacketTimeLength, reinterpret_cast<short*>(mDecodedFrame.data()), false);
         }
         else
             decodePacketTo(output, options, mCngPacket);
@@ -605,14 +592,14 @@ AudioReceiver::DecodeResult AudioReceiver::decodeGapTo(Audio::DataWindow& output
             mDecodedLength = 0;
         else
         {
-            mDecodedLength = mCodec->plc(mFrameCount, {(uint8_t*)mDecodedFrame, sizeof mDecodedFrame});
+            mDecodedLength = mCodec->plc(mFrameCount, {(uint8_t*)mDecodedFrame.data(), mDecodedFrame.size() * sizeof(int16_t)});
             if (!mDecodedLength)
             {
                 // PLC is not support or failed
                 // So substitute the silence
                 size_t nr_of_samples = mCodec->frameTime() * mCodec->samplerate() / 1000 * sizeof(short);
                 mDecodedLength = nr_of_samples * sizeof(short);
-                memset(mDecodedFrame, 0, mDecodedLength);
+                memset(mDecodedFrame.data(), 0, mDecodedLength);
             }
         }
     }
@@ -635,7 +622,8 @@ AudioReceiver::DecodeResult AudioReceiver::decodePacketTo(Audio::DataWindow& out
     auto& rtp = *packet->rtp(); // Syntax sugar
 
     mFailedCount = 0;
-    // Check if we need to emit silence or CNG - previously CNG packet was detected. Emit CNG audio here if needed.
+
+    // Check if we need to emit silence - it may happen in the case if next packet has RTP timestamp much beyond the previous one; maybe DTX was active.
     if (mLastPacketTimestamp && mLastPacketTimeLength && mCodec)
     {
          int units = rtp.GetTimestamp() - *mLastPacketTimestamp;
@@ -643,7 +631,8 @@ AudioReceiver::DecodeResult AudioReceiver::decodePacketTo(Audio::DataWindow& out
          if (milliseconds > mLastPacketTimeLength)
          {
              auto silenceLength = std::chrono::milliseconds(milliseconds - mLastPacketTimeLength);
-
+             ICELogDebug(<< "Emit " << silenceLength << " silence while requested " << options.mElapsed);
+             silenceLength = std::min(silenceLength, options.mElapsed);
              if (mCngPacket && options.mFillGapByCNG)
                  produceCNG(silenceLength, output, options);
              else
@@ -677,11 +666,12 @@ AudioReceiver::DecodeResult AudioReceiver::decodePacketTo(Audio::DataWindow& out
                 mDecodedLength = 0;
             else
             {
+                ICELogDebug(<< "Decoding CNG");
                 mCngPacket = packet;
                 mCngDecoder.decode3389(rtp.GetPayloadData(), rtp.GetPayloadLength());
 
                 // Emit CNG mLastPacketLength milliseconds
-                mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), mLastPacketTimeLength, (short*)mDecodedFrame, true);
+                mDecodedLength = mCngDecoder.produce(mCodec->samplerate(), mLastPacketTimeLength, (short*)mDecodedFrame.data(), true);
                 if (mDecodedLength)
                     processDecoded(output, options);
             }
@@ -717,7 +707,7 @@ AudioReceiver::DecodeResult AudioReceiver::decodePacketTo(Audio::DataWindow& out
                     {
                         // Decode frame by frame
                         auto codecInput = std::span{rtp.GetPayloadData() + i * mCodec->rtpLength(), (size_t)frameLength};
-                        auto codecOutput = std::span{(uint8_t*)mDecodedFrame, sizeof mDecodedFrame};
+                        auto codecOutput = std::span{(uint8_t*)mDecodedFrame.data(), mDecodedFrame.size() * sizeof(int16_t)};
                         auto r = mCodec->decode(codecInput, codecOutput);
                         mDecodedLength = r.mDecoded;
                         if (mDecodedLength > 0)
@@ -760,8 +750,15 @@ AudioReceiver::DecodeResult AudioReceiver::decodeEmptyTo(Audio::DataWindow& outp
             // Try to decode it - replay previous audio decoded or use CNG decoder (if payload type is 13)
             if (mCngPacket->rtp()->GetPayloadType() == 13)
             {
-                // Using latest CNG packet to produce comfort noise
+                // Using latest CNG packet to produce comfort noise.
-                auto produced = mCngDecoder.produce(fmt.rate(), options.mElapsed.count(), (short*)(output.data() + output.filled()), false);
+                // Clamp the produced amount to the remaining capacity of the output window -
+                // the CNG decoder writes straight into its buffer.
+                size_t bytesPerMs = (size_t)fmt.rate() / 1000 * sizeof(short) * fmt.channels();
+                size_t room = output.capacity() - output.filled();
+                int ms = bytesPerMs ? (int)std::min<int64_t>(options.mElapsed.count(), (int64_t)(room / bytesPerMs)) : 0;
+                if (ms <= 0)
+                    return {.mStatus = DecodeResult::Status::Skip};
+                auto produced = mCngDecoder.produce(fmt.rate(), ms, (short*)(output.mutableData() + output.filled()), false);
                 output.setFilled(output.filled() + produced);
                 return {.mStatus = DecodeResult::Status::Ok, .mSamplerate = fmt.rate(), .mChannels = fmt.channels()};
             }
@@ -775,7 +772,7 @@ AudioReceiver::DecodeResult AudioReceiver::decodeEmptyTo(Audio::DataWindow& outp
         else
         {
             // Emit silence if codec information is available - it is to properly handle the gaps
-            auto avail = output.getTimeLength(fmt.rate(), fmt.channels());
+            auto avail = output.getTimeLength(fmt);
             if (options.mElapsed > avail)
                 output.addZero(fmt.sizeFromTime(options.mElapsed - avail));
         }
@@ -785,86 +782,18 @@ AudioReceiver::DecodeResult AudioReceiver::decodeEmptyTo(Audio::DataWindow& outp
     return {.mStatus = DecodeResult::Status::Skip};
 }
 
-AudioReceiver::DecodeResult AudioReceiver::getAudioTo(Audio::DataWindow& output, DecodeOptions options)
+void MT::AudioReceiver::processDtmf()
+{
+    if (mDtmfBuffer.getCount())
     {
-    DecodeResult result = {.mStatus = DecodeResult::Status::Skip};
-
-    // Process RFC2833 here; it doesn't result in any audio - only callbacks and statistics
         auto fr = mDtmfBuffer.fetch();
         if (fr.mPacket && fr.mStatus == RtpBuffer::FetchResult::Status::RegularPacket)
             mDtmfReceiver.add(fr.mPacket->rtp());
-
-
-    auto produced = 0ms;
-    if (mAvailable.filled() && mCodec && options.mElapsed != 0ms)
-    {
-        Audio::Format fmt = options.mResampleToMainRate ? Audio::Format(AUDIO_SAMPLERATE, 1) : mCodec->getAudioFormat();
-        auto initiallyAvailable = mCodec ? mAvailable.getTimeLength(fmt.rate(), fmt.channels()) : 0ms;
-        if (initiallyAvailable != 0ms)
-        {
-            std::chrono::milliseconds resultTime = std::min(initiallyAvailable, options.mElapsed);
-            auto resultLen = fmt.sizeFromTime(resultTime);
-            mAvailable.moveTo(output, resultLen);
-            produced += resultTime;
-
-            // Maybe request is satisfied ?
-            if (produced >= options.mElapsed)
-                return {.mStatus = DecodeResult::Status::Ok, .mSamplerate = fmt.rate(), .mChannels = fmt.channels()};
     }
 }
 
-    std::chrono::milliseconds decoded = 0ms;
+void MT::AudioReceiver::updateDecodingTimeStatistics()
-    do
 {
-        // Get next packet from buffer
-        RtpBuffer::ResultList rl;
-        RtpBuffer::FetchResult fr = mBuffer.fetch();
-        // ICELogDebug(<< fr.toString() << " " << mBuffer.findTimelength());
-
-        switch (fr.mStatus)
-        {
-        case RtpBuffer::FetchResult::Status::Gap:           result = decodeGapTo(mAvailable, options);                  break;
-        case RtpBuffer::FetchResult::Status::NoPacket:      result = decodeEmptyTo(mAvailable, options);                break;
-        case RtpBuffer::FetchResult::Status::RegularPacket: result = decodePacketTo(mAvailable, options, fr.mPacket);   break;
-        default:
-            assert(0);
-        }
-
-        // Was there decoding at all ?
-        if (!mCodec)
-            break; // No sense to continue - we have no information at all
-
-        Audio::Format fmt = options.mResampleToMainRate ? Audio::Format(AUDIO_SAMPLERATE, 1) : mCodec->getAudioFormat();
-        result.mSamplerate = fmt.rate();
-        result.mChannels = fmt.channels();
-
-        // Have we anything interesting in the buffer ?
-        auto bufferAvailable = mAvailable.getTimeLength(fmt.rate(), fmt.channels());
-        if (bufferAvailable == 0ms)
-            break; // No sense to continue - decoding / CNG / PLC stopped totally
-
-        // How much data should be moved to result buffer ?
-        if (options.mElapsed != 0ms)
-        {
-            std::chrono::milliseconds resultTime = std::min(bufferAvailable, options.mElapsed - produced);
-            auto resultLen = fmt.sizeFromTime(resultTime);
-            mAvailable.moveTo(output, resultLen);
-            produced += resultTime;
-        }
-        else
-            mAvailable.moveTo(output, mAvailable.filled());
-
-        decoded += bufferAvailable;
-    }
-    while (produced < options.mElapsed);
-
-    if (produced != 0ms)
-        result.mStatus = DecodeResult::Status::Ok;
-
-    // Time statistics
-    if (result.mStatus == DecodeResult::Status::Ok)
-    {
-        // Decode statistics
     if (!mDecodeTimestamp)
         mDecodeTimestamp = std::chrono::steady_clock::now();
     else
@@ -874,9 +803,114 @@ AudioReceiver::DecodeResult AudioReceiver::getAudioTo(Audio::DataWindow& output,
         mDecodeTimestamp = t;
     }
 }
+
+AudioReceiver::DecodeResult AudioReceiver::getAudioTo(Audio::DataWindow& output, DecodeOptions options)
+{
+    // ICELogDebug(<< "getAudioTo() for " << options.mElapsed);
+    assert (options.mElapsed != 0ms);
+
+    // First decode on this receiver: allocate the scratch buffers. Network-MOS-only
+    // streams never reach this point, so they never pay for them.
+    ensureDecodeBuffers();
+
+    // Increase counter of requested audio
+    mRequestedAudio += options.mElapsed;
+
+    DecodeResult result = {.mStatus = DecodeResult::Status::Skip};
+
+    // Process RFC2833 here; it doesn't result in any audio - only callbacks and statistics
+    processDtmf();
+
+    // How much time length audio we produced here
+    auto produced = 0ms;
+    Audio::Format fmt;
+
+    // Have we anything from the previous decode attempts ?
+    if (mAvailable.filled())
+    {
+        // Find what audio format is used in mAvailable data
+        fmt = options.mResampleToMainRate ? Audio::Format(AUDIO_SAMPLERATE, 1) : mCodec->getAudioFormat();
+
+        // How much milliseconds are available ?
+        auto availTime = mAvailable.getTimeLength(fmt);
+        if (availTime != 0ms)
+        {
+            // How much we can consume from the mAvailable buffer ?
+            std::chrono::milliseconds resultTime = std::min(availTime, options.mElapsed);
+
+            // Number of bytes
+            mAvailable.moveTo(output, fmt.sizeFromTime(resultTime));
+
+            // Increase the counter of produced milliseconds
+            produced += resultTime;
+        }
+    }
+
+    while (produced < options.mElapsed)
+    {
+        // Get next packet from buffer
+        RtpBuffer::FetchResult fr = mRtpBuffer.fetch();
+
+        // Decode to mAvailable buffer
+        switch (fr.mStatus)
+        {
+        case RtpBuffer::FetchResult::Status::Gap:           result = decodeGapTo(mAvailable, options.decreaseElapsedBy(produced));                                                      break;
+        case RtpBuffer::FetchResult::Status::NoPacket:      result = decodeEmptyTo(mAvailable, options.decreaseElapsedBy(produced));                                                    break;
+        case RtpBuffer::FetchResult::Status::RegularPacket: result = decodePacketTo(mAvailable, options.decreaseElapsedBy(produced), fr.mPacket);   updateDecodeIntervalStatistics();   break;
+        default:
+            assert(0);
+        }
+
+        // Was there decoding at all ?
+        if (!mCodec)
+            break; // No sense to continue - we have no information at all
+
+        fmt = options.mResampleToMainRate ? Audio::Format(AUDIO_SAMPLERATE, 1) : mCodec->getAudioFormat();
+        result.mSamplerate  = fmt.rate();
+        result.mChannels    = fmt.channels();
+
+        // How much milliseconds we have in audio buffer ?
+        auto bufferAvailable = mAvailable.getTimeLength(fmt);
+        if (bufferAvailable == 0ms)
+            break; // No sense to continue - decoding / CNG / PLC stopped totally
+
+        // How much data should be moved to result buffer ?
+        std::chrono::milliseconds resultTime = std::min(bufferAvailable, options.mElapsed - produced);
+        mAvailable.moveTo(output, fmt.sizeFromTime(resultTime));
+        produced += resultTime;
+    }
+
+    if (produced != 0ms)
+    {
+        result.mStatus = DecodeResult::Status::Ok;
+        updateDecodingTimeStatistics();
+    }
+
+    mProducedAudio += produced;
+    // ICELogDebug(<< "Requested " << options.mElapsed << ", produced " << produced << ", remains " << mAvailable.getTimeLength(fmt) << ", packets " << getRtpBuffer().getCount());
     return result;
 }
 
+void AudioReceiver::ensureDecodeBuffers()
+{
+    // Allocate the decode/convert/resample scratch buffers to full capacity on the
+    // first decode. mDecodedFrame being empty means none are allocated yet; they
+    // are always allocated together, so checking one is enough.
+    if (mDecodedFrame.empty())
+    {
+        mDecodedFrame.resize(MT_MAX_DECODEBUFFER);
+        mConvertedFrame.resize(MT_MAX_DECODEBUFFER * 2);
+        mResampledFrame.resize(MT_MAX_DECODEBUFFER);
+    }
+
+    if (!mAvailable.capacity())
+    {
+        // 10 seconds is the maximum length of decoded audio in single step
+        // It is important - DTX may produce silence up to few seconds easily
+        mAvailable.setCapacity(AUDIO_SAMPLERATE * 10 * sizeof(short));
+    }
+}
+
 void AudioReceiver::makeMonoAndResample(int rate, int channels)
 {
     // Make mono from stereo - engine works with mono only for now
@@ -884,12 +918,12 @@ void AudioReceiver::makeMonoAndResample(int rate, int channels)
     if (channels != AUDIO_CHANNELS)
     {
         if (channels == 1)
-            mConvertedLength = Audio::ChannelConverter::monoToStereo(mDecodedFrame, mDecodedLength, mConvertedFrame, mDecodedLength * 2);
+            mConvertedLength = Audio::ChannelConverter::monoToStereo(mDecodedFrame.data(), mDecodedLength, mConvertedFrame.data(), mDecodedLength * 2);
         else
-            mDecodedLength = Audio::ChannelConverter::stereoToMono(mDecodedFrame, mDecodedLength, mDecodedFrame, mDecodedLength / 2);
+            mDecodedLength = Audio::ChannelConverter::stereoToMono(mDecodedFrame.data(), mDecodedLength, mDecodedFrame.data(), mDecodedLength / 2);
     }
 
-    void* frames = mConvertedLength ? mConvertedFrame : mDecodedFrame;
+    void* frames = mConvertedLength ? (void*)mConvertedFrame.data() : (void*)mDecodedFrame.data();
     unsigned length = mConvertedLength ? mConvertedLength : mDecodedLength;
 
     Audio::Resampler* r = nullptr;
@@ -900,13 +934,13 @@ void AudioReceiver::makeMonoAndResample(int rate, int channels)
     case 32000:    r = &mResampler32; break;
     case 48000:    r = &mResampler48; break;
     default:
-        memcpy(mResampledFrame, frames, length);
+        memcpy(mResampledFrame.data(), frames, length);
         mResampledLength = length;
         return;
     }
 
     size_t processedInput = 0;
-    mResampledLength = r->processBuffer(frames, length, processedInput, mResampledFrame, r->getDestLength(length));
+    mResampledLength = r->processBuffer(frames, length, processedInput, mResampledFrame.data(), r->getDestLength(length));
     // processedInput result value is ignored - it is always equal to length as internal sample rate is 8/16/32/48K
 }
 
@@ -987,43 +1021,16 @@ AudioReceiver::MediaInfo AudioReceiver::infoFor(jrtplib::RTPPacket& p)
     return {packetTime, codec->samplerate()};
 }
 
-// int AudioReceiver::timelengthFor(jrtplib::RTPPacket& p)
+void AudioReceiver::updateDecodeIntervalStatistics()
-// {
+{
-//     CodecMap::iterator codecIter = mCodecMap.find(p.GetPayloadType());
+    auto now = std::chrono::steady_clock::now();
-//     if (codecIter == mCodecMap.end())
+    if (mLastDecodeTimestamp)
-//         return 0;
+    {
-
+        mIntervalBetweenDecode += std::chrono::duration_cast<std::chrono::microseconds>(now - *mLastDecodeTimestamp);
-//     PCodec codec = codecIter->second;
+        mDecodeCount ++;
-//     if (codec)
+    }
-//     {
+    mLastDecodeTimestamp = now;
-//         int frame_count = 0;
+}
-//         if (codec->rtpLength() != 0)
-//         {
-//             frame_count = static_cast<int>(p.GetPayloadLength() / codec->rtpLength());
-//             if (p.GetPayloadType() == 9/*G729A silence*/ && p.GetPayloadLength() % codec->rtpLength())
-//                 frame_count++;
-//         }
-//         else
-//             frame_count = 1;
-
-//         return frame_count * codec->frameTime();
-//     }
-//     else
-//         return 0;
-// }
-
-// int AudioReceiver::samplerateFor(jrtplib::RTPPacket& p)
-// {
-//     CodecMap::iterator codecIter = mCodecMap.find(p.GetPayloadType());
-//     if (codecIter != mCodecMap.end())
-//     {
-//         PCodec codec = codecIter->second;
-//         if (codec)
-//             return codec->samplerate();
-//     }
-
-//     return 8000;
-// }
 
 // ----------------------- DtmfReceiver -------------------
 DtmfReceiver::DtmfReceiver(Statistics& stat)
@@ -1036,21 +1043,25 @@ DtmfReceiver::~DtmfReceiver()
 void DtmfReceiver::add(const std::shared_ptr<RTPPacket>& p)
 {
     auto ev = DtmfBuilder::parseRfc2833({p->GetPayloadData(), p->GetPayloadLength()});
-    if (ev.mTone != mEvent || ev.mEnd != mEventEnded)
+    if (!ev.mTone)
+        return; // Malformed or unknown event payload
+
+    // A new digit begins when the tone changes, or when the same tone starts
+    // again after the previous occurrence ended. Retransmitted start/end
+    // packets keep both fields unchanged and are ignored. The end packet of
+    // the current tone only updates state - the digit was already reported.
+    bool newEvent = (ev.mTone != mEvent) || (mEventEnded && !ev.mEnd);
+
+    if (newEvent)
     {
-        if (!(mEvent == ev.mTone && !mEventEnded && ev.mEnd))
-        {
-            // New tone is here
         if (mCallback)
             mCallback(ev.mTone);
 
         // Queue statistics item
         mStat.mDtmf2833Timeline.emplace_back(Dtmf2833Event{.mTone = ev.mTone,
                                                            .mTimestamp = RtpHelper::toMicroseconds(p->GetReceiveTime())});
+    }
 
-            // Store to avoid triggering on the packet
     mEvent = ev.mTone;
     mEventEnded = ev.mEnd;
 }
-    }
-}

src/engine/media/MT_AudioReceiver.h

@@ -20,6 +20,8 @@
 
 #include <optional>
 #include <chrono>
+#include <vector>
+#include <cstdint>
 using namespace std::chrono_literals;
 
 namespace MT
@@ -104,6 +106,18 @@ public:
 
     FetchResult fetch();
 
+    // Drop oldest packets so buffered audio stays within the high-water mark,
+    // recording packet-loss events for any sequence gaps crossed (the same
+    // accounting fetch() performs). Used to bound memory on streams that never
+    // call fetch() - i.e. network-MOS-only streams with audio decode disabled,
+    // which would otherwise retain every packet for the whole call.
+    //
+    // maxPackets, when non-zero, additionally caps the buffer to that many packets
+    // regardless of buffered time. The decode path (fetch()) leaves it 0 so jitter
+    // tolerance stays governed by the time-based high-water mark; the network-only
+    // path passes a small cap since those packets are never decoded.
+    void trimToHighWater(size_t maxPackets = 0);
+
 protected:
     unsigned    mSsrc = 0;
     std::chrono::milliseconds   mHigh = std::chrono::milliseconds(RTP_BUFFER_HIGH),
@@ -122,6 +136,7 @@ protected:
     std::optional<uint32_t> mLastSeqno;
     std::optional<jrtplib::RTPTime> mLastReceiveTime;
 
+
     // To calculate average interval between packet add. It is close to jitter but more useful in debugging.
     float mLastAddTime = 0.0f;
 };
@@ -169,10 +184,22 @@ public:
 
     struct DecodeOptions
     {
+        bool mRealtimeProcessing = false;               // Target PCAP parsing by default
         bool mResampleToMainRate = true;                // Resample all decoded audio to AUDIO_SAMPLERATE
         bool mFillGapByCNG = false;                     // Use CNG information if available
         bool mSkipDecode = false;                       // Don't do decode, just dry run - fetch packets, remove them from the jitter buffer
         std::chrono::milliseconds mElapsed = 0ms;       // How much milliseconds should be decoded; zero value means "decode just next packet from the buffer"
+        DecodeOptions decreaseElapsedBy(std::chrono::milliseconds delta)
+        {
+            return
+            {
+                .mRealtimeProcessing = mRealtimeProcessing,
+                .mResampleToMainRate = mResampleToMainRate,
+                .mFillGapByCNG = mFillGapByCNG,
+                .mSkipDecode = mSkipDecode,
+                .mElapsed = std::max(mElapsed - delta, 0ms)
+            };
+        }
     };
 
     struct DecodeResult
@@ -193,7 +220,7 @@ public:
 
     // Looks for codec by payload type
     Codec*      findCodec(int payloadType);
-    RtpBuffer&  getRtpBuffer() { return mBuffer; }
+    RtpBuffer&  getRtpBuffer() { return mRtpBuffer; }
 
     // Returns size of AudioReceiver's instance in bytes (including size of all data + codecs + etc.)
     int getSize() const;
@@ -205,14 +232,12 @@ public:
     };
     MediaInfo infoFor(jrtplib::RTPPacket& p);
 
-    // // Returns timelength for given packet
+    void processDtmf();
-    // int timelengthFor(jrtplib::RTPPacket& p);
 
-    // // Return samplerate for given packet
+    void updateDecodingTimeStatistics();
-    // int samplerateFor(jrtplib::RTPPacket& p);
 
 protected:
-    RtpBuffer                           mBuffer;                // Jitter buffer itself
+    RtpBuffer                           mRtpBuffer;             // RTP jitter buffer itself; here are audio packets
     RtpBuffer                           mDtmfBuffer;            // These two (mDtmfBuffer / mDtmfReceiver) are for our analyzer stack only; in normal softphone logic DTMF packets goes via SingleAudioStream::mDtmfReceiver
     DtmfReceiver                        mDtmfReceiver;
 
@@ -229,16 +254,22 @@ protected:
     // Already decoded data that can be retrieved without actual decoding - it may happen because of getAudioTo() may be limited by time interval
     Audio::DataWindow mAvailable;
 
-    // Temporary buffer to hold decoded data (it is better than allocate data on stack)
+    // Decode/convert/resample scratch buffers. These were inline arrays
-    int16_t mDecodedFrame[MT_MAX_DECODEBUFFER];
+    // (MT_MAX_DECODEBUFFER * {1,2,1} * int16_t = 256 KB total) carried by every
+    // AudioReceiver, hence by every StreamDecoder - including network-MOS-only
+    // streams that never decode. They are now allocated lazily on the first
+    // getAudioTo() call via ensureDecodeBuffers(); non-decoding streams keep them
+    // empty. Once allocated they are sized to full capacity and reused, so decode
+    // behaviour is unchanged.
+    std::vector<int16_t> mDecodedFrame;     // sized to MT_MAX_DECODEBUFFER
     size_t mDecodedLength = 0;
 
     // Buffer to hold data converted to stereo/mono; there is multiplier 2 as it can be stereo audio
-    int16_t mConvertedFrame[MT_MAX_DECODEBUFFER * 2];
+    std::vector<int16_t> mConvertedFrame;   // sized to MT_MAX_DECODEBUFFER * 2
     size_t mConvertedLength = 0;
 
     // Buffer to hold data resampled to AUDIO_SAMPLERATE
-    int16_t mResampledFrame[MT_MAX_DECODEBUFFER];
+    std::vector<int16_t> mResampledFrame;   // sized to MT_MAX_DECODEBUFFER
     size_t mResampledLength = 0;
 
     // Last packet time length
@@ -258,6 +289,15 @@ protected:
     float mIntervalSum = 0.0f;
     int mIntervalCount = 0;
 
+    std::chrono::milliseconds mRequestedAudio = 0ms;
+    std::chrono::milliseconds mProducedAudio = 0ms;
+
+    // Lazily allocate the decode/convert/resample scratch buffers (mDecodedFrame,
+    // mConvertedFrame, mResampledFrame) to full capacity on the first decode. A
+    // no-op once allocated. Called at the top of getAudioTo(); network-MOS-only
+    // streams never reach it, so they never pay the 256 KB.
+    void ensureDecodeBuffers();
+
     // Zero rate will make audio mono but resampling will be skipped
     void makeMonoAndResample(int rate, int channels);
 
@@ -272,6 +312,12 @@ protected:
     DecodeResult decodeGapTo(Audio::DataWindow& output, DecodeOptions options);
     DecodeResult decodePacketTo(Audio::DataWindow& output, DecodeOptions options, const std::shared_ptr<RtpBuffer::Packet>& p);
     DecodeResult decodeEmptyTo(Audio::DataWindow& output, DecodeOptions options);
+
+    std::optional<std::chrono::steady_clock::time_point> mLastDecodeTimestamp;
+    std::chrono::microseconds mIntervalBetweenDecode = 0us;
+    size_t mDecodeCount = 0;
+    void updateDecodeIntervalStatistics();
+
 };
 
 }

src/engine/media/MT_AudioStream.cpp

@@ -238,6 +238,9 @@ void AudioStream::addData(const void* buffer, int bytes)
 
 void AudioStream::copyDataTo(Audio::Mixer& mixer, int needed)
 {
+    // mStreamMap is also mutated from the network thread (dataArrived)
+    Lock l(mMutex);
+
     // Local audio mixer - used to send audio to media observer
     Audio::Mixer localMixer;
     Audio::DataWindow forObserver;
@@ -282,22 +285,27 @@ void AudioStream::copyDataTo(Audio::Mixer& mixer, int needed)
 
     if (mMediaObserver)
     {
-        localMixer.mixAndGetPcm(forObserver);
+        int mixedBytes = localMixer.mixAndGetPcm(forObserver);
-        mMediaObserver->onMedia(forObserver.data(), forObserver.capacity(), MT::Stream::MediaDirection::Incoming, this, mMediaObserverTag);
+        if (mixedBytes > 0)
+            mMediaObserver->onMedia(forObserver.data(), mixedBytes, MT::Stream::MediaDirection::Incoming, this, mMediaObserverTag);
     }
 }
 
 void AudioStream::dataArrived(PDatagramSocket s, const void* buffer, int length, InternetAddress& source)
 {
+    // Protects mStreamMap (also iterated by copyDataTo on the audio thread)
+    // and the receive/decrypt buffers.
+    Lock l(mMutex);
+
     jrtplib::RTPIPv6Address addr6;
     jrtplib::RTPIPv4Address addr4;
     jrtplib::RTPExternalTransmissionInfo* info = dynamic_cast<jrtplib::RTPExternalTransmissionInfo*>(mRtpSession.GetTransmissionInfo());
     assert(info);
 
     // Drop RTP packets if stream is not receiving now; let RTCP go
-    if (!(state() & (int)StreamState::Receiving) && RtpHelper::isRtpOrRtcp(buffer, length))
+    if (!(state() & (int)StreamState::Receiving) && RtpHelper::isRtp(buffer, length))
     {
-        ICELogMedia(<< "Stream is not allowed to receive RTP stream. Ignore the RT(C)P packet");
+        ICELogMedia(<< "Stream is not allowed to receive RTP stream. Ignore the RTP packet");
         return;
     }
 

src/engine/media/MT_CngHelper.cpp

@@ -135,7 +135,7 @@ namespace MT
 
     // Get noise level
     unsigned char noiseLevel = *dataIn;
-    float linear = float(1.0 / noiseLevel ? noiseLevel : 1);
+    float linear = 1.0f / float(noiseLevel ? noiseLevel : 1);
 
     // Generate white noise for 16KHz sample rate
     LPFilter lpf; HPFilter hpf;

src/engine/media/MT_CodecList.cpp

@@ -137,7 +137,6 @@ std::string CodecList::Settings::toString() const
         oss << "OPUS ptype: " << spec.mPayloadType << ", rate: " << spec.mRate << ", channels: " << spec.mChannels << std::endl;
     }
 
-
     return oss.str();
 }
 
@@ -234,7 +233,7 @@ static int findOctetMode(const char* line)
     p += strlen(param_name);
     char int_buf[8] = {0};
     size_t int_buf_offset = 0;
-    while (*p && isdigit(*p) && int_buf_offset < sizeof(int_buf))
+    while (*p && isdigit(*p) && int_buf_offset < sizeof(int_buf) - 1)
         int_buf[int_buf_offset++] = *p++;
     return atoi(int_buf);
 }

src/engine/media/MT_Dtmf.cpp

@@ -37,12 +37,11 @@ void DtmfBuilder::buildRfc2833(const Rfc2833Event& ev, void* output)
 
     char* packet = (char*)output;
 
+    // RFC 4733: byte 1 is E(1) R(1) volume(6)
     packet[0] = toneValue;
-    packet[1] = 1 | (ev.mVolume << 2);
+    packet[1] = ev.mVolume & 0x3F;
     if (ev.mEnd)
-        packet[1] |= 128;
+        packet[1] |= 0x80;
-    else
-        packet[1] &= 127;
 
     unsigned short durationValue = htons(ev.mDuration);
     memcpy(packet + 2, &durationValue, 2);
@@ -58,11 +57,11 @@ DtmfBuilder::Rfc2833Event DtmfBuilder::parseRfc2833(std::span<uint8_t> payload)
     uint8_t b0 = payload[0];
     uint8_t b1 = payload[1];
 
-    if (b0 >=0 && b0 <= 9)
+    if (b0 <= 9)
         r.mTone = '0' + b0;
     else
-    if (b0 >= 12 && b0 <= 17)
+    if (b0 >= 12 && b0 <= 15)
-        r.mTone = 'A' + b0;
+        r.mTone = 'A' + b0 - 12;
     else
     if (b0 == 10)
         r.mTone = '*';
@@ -70,9 +69,10 @@ DtmfBuilder::Rfc2833Event DtmfBuilder::parseRfc2833(std::span<uint8_t> payload)
     if (b0 == 11)
         r.mTone = '#';
 
-    r.mEnd = (b1 & 128);
+    // RFC 4733: byte 1 is E(1) R(1) volume(6); duration is bytes 2-3, network order
-    r.mVolume = (b1 & 127) >> 2;
+    r.mEnd = (b1 & 0x80) != 0;
-    r.mDuration = ntohs(*(uint16_t*)payload.data()+2);
+    r.mVolume = b1 & 0x3F;
+    r.mDuration = (uint16_t(payload[2]) << 8) | payload[3];
 
     return r;
 }
@@ -202,7 +202,7 @@ void PDTMFEncoder_AddTone(double f1, double f2, unsigned ms1, unsigned ms2, unsi
         int ival = ifix(val);
         if (ival < -32768)
             ival = -32768;
-        else if (val > 32767)
+        else if (ival > 32767)
             ival = 32767;
 
         result[dataPtr++] = ival / 2;
@@ -280,8 +280,9 @@ void DtmfContext::stopTone()
         switch (mType)
         {
         case Dtmf_Rfc2833:
+            // Mark stopped but keep the entry: getRfc2833() emits the end
+            // packet(s) for a stopped tone and erases it afterwards.
             mQueue.front().mStopped = true;
-            mQueue.erase(mQueue.begin());
             break;
 
         case Dtmf_Inband:
@@ -769,7 +770,7 @@ int zap_dtmf_detect (dtmf_detect_state_t *s,
             s->fax_tone.v2 = s->fax_tone.v3;
             s->fax_tone.v3 = s->fax_tone.fac*s->fax_tone.v2 - v1 + famp;
 
-            v1 = s->fax_tone.v2;
+            v1 = s->fax_tone2nd.v2;
             s->fax_tone2nd.v2 = s->fax_tone2nd.v3;
             s->fax_tone2nd.v3 = s->fax_tone2nd.fac*s->fax_tone2nd.v2 - v1 + famp;
         }
@@ -865,7 +866,7 @@ printf("Fax energy/Second Harmonic: %f/%f\n", fax_energy, fax_energy_2nd);
                 s->detected_digits++;
                 if (s->current_digits < MAX_DTMF_DIGITS)
                 {
-                    s->digits[s->current_digits++] = hit;
+                    s->digits[s->current_digits++] = 'f';
                     s->digits[s->current_digits] = '\0';
                 }
                 else

src/engine/media/MT_EvsCodec.cpp

@@ -152,6 +152,27 @@ EVSCodec::EVSCodec(const StreamParameters &sp)
 {
 	EVSCodec::sp = sp;
 
+    // Metadata only - the heavy decoder state is created lazily (ensureDecoder()).
+    mOutputFs = outputFsFromBw(sp.bw);
+}
+
+int EVSCodec::outputFsFromBw(int bw)
+{
+    switch (bw)
+    {
+    case NB:  return 8000;
+    case WB:  return 16000;
+    case SWB: return 32000;
+    case FB:  return 48000;
+    }
+    return 0;
+}
+
+void EVSCodec::ensureDecoder()
+{
+    if (st_dec)
+        return;
+
     if ((st_dec = reinterpret_cast<evs::Decoder_State*>(malloc(sizeof(evs::Decoder_State)))) == nullptr)
         throw std::bad_alloc();
 
@@ -170,9 +191,9 @@ EVSCodec::~EVSCodec()
 Codec::Info EVSCodec::info() {
     return {
         .mName = MT_EVS_CODECNAME,
-        .mSamplerate = st_dec->output_Fs,
+        .mSamplerate = mOutputFs,
         .mChannels = 1,
-        .mPcmLength = st_dec->output_Fs / 1000 * sp.ptime * 2,
+        .mPcmLength = mOutputFs / 1000 * sp.ptime * 2,
         .mFrameTime = sp.ptime,
         .mRtpLength = 0
     };
@@ -187,6 +208,8 @@ Codec::EncodeResult EVSCodec::encode(std::span<const uint8_t> input, std::span<u
 
 Codec::DecodeResult EVSCodec::decode(std::span<const uint8_t> input, std::span<uint8_t> output)
 {
+    ensureDecoder();
+
     if (output.size_bytes() < pcmLength())
         return {.mDecoded = 0};
 

src/engine/media/MT_EvsCodec.h

@@ -57,7 +57,21 @@ public:
 private:
     evs::Decoder_State* st_dec = nullptr;
     StreamParameters sp;
+
+    // Output sample rate, derived from the negotiated bandwidth (sp.bw) at
+    // construction. Cached so info()/samplerate()/pcmLength() work for network-MOS
+    // metadata without allocating the (large) EVS decoder state - see ensureDecoder.
+    int mOutputFs = 0;
+
     void initDecoder(const StreamParameters& sp);
+
+    // Allocate + initialize the EVS decoder state lazily on first decode().
+    // Network-MOS-only streams resolve metadata but never decode, so they never
+    // pay for the EVS decoder (Decoder_State + CLDFB/FD-CNG sub-allocations).
+    void ensureDecoder();
+
+    // Maps an EVS bandwidth (NB/WB/SWB/FB) to its output sample rate in Hz.
+    static int outputFsFromBw(int bw);
 };
 
 } // End of namespace

src/engine/media/MT_SingleAudioStream.cpp

@@ -34,7 +34,21 @@ void SingleAudioStream::copyPcmTo(Audio::DataWindow& output, int needed)
     // Packet by packet
     while (output.filled() < needed)
     {
-        if (mReceiver.getAudioTo(output, {}).mStatus != AudioReceiver::DecodeResult::Status::Ok)
+        // Number of bytes to fill on this step
+        auto requested = needed - output.filled();
+
+        auto options = AudioReceiver::DecodeOptions{
+            .mRealtimeProcessing = true,
+            .mResampleToMainRate = true,
+            .mSkipDecode = false,
+            .mElapsed  = std::chrono::milliseconds(requested / (AUDIO_SAMPLERATE / 1000))
+        };
+
+        // Try to get the data from receiver / decoder
+        if (options.mElapsed != 0ms) {
+            if (mReceiver.getAudioTo(output, options).mStatus != AudioReceiver::DecodeResult::Status::Ok)
+                break;
+        } else
             break;
     }
 

src/engine/media/MT_SrtpHelper.cpp

@@ -48,6 +48,24 @@ extern std::string_view toString(SrtpSuite suite)
     return {};
 }
 
+extern int srtpSuiteStrength(SrtpSuite suite)
+{
+    switch (suite)
+    {
+    case SRTP_NONE:                 return 0;
+    case SRTP_AES_128_AUTH_NULL:    return 1; // no authentication - weakest
+    case SRTP_AES_128_AUTH_32:      return 2;
+    case SRTP_AES_192_AUTH_32:      return 3;
+    case SRTP_AES_256_AUTH_32:      return 4;
+    case SRTP_AES_128_AUTH_80:      return 5;
+    case SRTP_AES_192_AUTH_80:      return 6;
+    case SRTP_AES_256_AUTH_80:      return 7;
+    case SRTP_AED_AES_128_GCM:      return 8;
+    case SRTP_AED_AES_256_GCM:      return 9;
+    }
+    return 0;
+}
+
 typedef void (*set_srtp_policy_function) (srtp_crypto_policy_t*);
 
 set_srtp_policy_function findPolicyFunction(SrtpSuite suite)
@@ -95,6 +113,7 @@ SrtpSession::SrtpSession()
 
     // Generate outgoing keys for all ciphers
     auto putKey = [this](SrtpSuite suite, size_t length){
+        assert(suite > SRTP_NONE && suite <= SRTP_LAST);
         auto key = std::make_shared<ByteBuffer>();
         key->resize(length);
         RAND_bytes(key->mutableData(), key->size());
@@ -103,9 +122,9 @@ SrtpSession::SrtpSession()
     putKey(SRTP_AES_128_AUTH_80, 30); putKey(SRTP_AES_128_AUTH_32, 30);
     putKey(SRTP_AES_192_AUTH_80, 38); putKey(SRTP_AES_192_AUTH_32, 38);
     putKey(SRTP_AES_256_AUTH_80, 46); putKey(SRTP_AES_256_AUTH_32, 46);
+    putKey(SRTP_AES_128_AUTH_NULL, 30); // NULL auth still encrypts - it needs a key+salt
     putKey(SRTP_AED_AES_128_GCM, 28);
     putKey(SRTP_AED_AES_256_GCM, 44);
-
 }
 
 SrtpSession::~SrtpSession()
@@ -214,7 +233,9 @@ SrtpKeySalt& SrtpSession::outgoingKey(SrtpSuite suite)
 {
     assert(suite > SRTP_NONE && suite <= SRTP_LAST);
     Lock l(mGuard);
-    return mOutgoingKey[int(suite)-1];  // The automated review sometimes give the hints about the possible underflow array index access
+    // Must use the same indexing as the constructor and open(): the SDP
+    // crypto attribute has to advertise the key the session encrypts with.
+    return mOutgoingKey[int(suite)];
 }
 
 bool SrtpSession::protectRtp(void* buffer, int* length)

src/engine/media/MT_SrtpHelper.h

@@ -36,6 +36,10 @@ enum SrtpSuite
 extern SrtpSuite        toSrtpSuite(const std::string_view& s);
 extern std::string_view toString(SrtpSuite suite);
 
+// Relative cryptographic strength used to pick a suite from an SDP offer.
+// Bigger is stronger. The raw enum values do NOT follow strength order.
+extern int              srtpSuiteStrength(SrtpSuite suite);
+
 typedef std::pair<PByteBuffer, PByteBuffer> SrtpKeySalt;
 typedef std::pair<unsigned, srtp_policy_t> SrtpStream;
 
@@ -68,8 +72,10 @@ protected:
     srtp_t          mInboundSession,
     mOutboundSession;
 
+    // Outgoing keys are indexed by the SrtpSuite enum value directly;
+    // index 0 (SRTP_NONE) is unused.
     SrtpKeySalt     mIncomingKey,
-                    mOutgoingKey[SRTP_LAST];
+                    mOutgoingKey[SRTP_LAST + 1];
     srtp_policy_t   mInboundPolicy;
     srtp_policy_t   mOutboundPolicy;
     SrtpSuite       mSuite;

src/engine/media/MT_Statistics.h

@@ -112,7 +112,7 @@ public:
     std::map<int,int>               mLoss;                  // Every item is number of loss of corresping length
     std::chrono::milliseconds       mAudioTime = 0ms;       // Decoded/found time in milliseconds
     size_t                          mDecodedSize = 0;       // Number of decoded bytes
-    uint16_t                        mSsrc = 0;              // Last known SSRC ID in a RTP stream
+    uint32_t                        mSsrc = 0;              // Last known SSRC ID in a RTP stream
     ice::NetworkAddress             mRemotePeer;            // Last known remote RTP address
 
     // AMR codec bitrate switch counter

src/libs/ice/ICELog.cpp

@@ -342,8 +342,6 @@ void Logger::beginLine(LogLevel level, const char* filename, int linenumber, con
     mFilename = filenamestart;
     mLine = linenumber;
     mSubsystem = subsystem;
-
-    // mStream << std::setw(8) << ICETimeHelper::timestamp() << " | " << std::setw(8) << ThreadInfo::currentThread() << " | " << std::setw(30) << filenamestart << " | " << std::setw(4) << linenumber << " | " << std::setw(12) << subsystem << " | ";
 }
 
 void

src/libs/ice/ICENetworkHelper.cpp

@@ -163,7 +163,7 @@ void NetworkHelper::reload(int networkType)
   fillUwpInterfaceList(AF_INET6, networkType, mIPList);
 #else
   // https://github.com/golang/go/issues/40569
-  struct ifaddrs* il = NULL;
+  struct ifaddrs* il = nullptr;
   if (getifaddrs(&il))
     throw Exception(GETIFADDRS_FAILED, errno);
   if (il)
@@ -171,6 +171,15 @@ void NetworkHelper::reload(int networkType)
     struct ifaddrs* current = il;
     while (current)
     {
+      // getifaddrs() may return entries with a null ifa_addr (interfaces with
+      // no assigned address, point-to-point/tunnel interfaces, link-layer
+      // entries, ...). Dereferencing it would crash, so skip such entries.
+      if (current->ifa_addr == nullptr)
+      {
+        current = current->ifa_next;
+        continue;
+      }
+
       //char ipbuffer[64];
       NetworkAddress addr;
       addr.setPort(1000); // Set fake address to keep NetworkAddress initialized

src/libs/jrtplib/src/rtpsources.h

@@ -44,7 +44,19 @@
 #include "rtptypes.h"
 #include "rtpmemoryobject.h"
 
-#define RTPSOURCES_HASHSIZE							8317
+// Number of buckets in the per-RTPSession SSRC->source hash table. This is an
+// inline array of pointers in every RTPSources instance (sizeof == hashsize *
+// sizeof(void*)), so it is paid by every RTPSession object regardless of how many
+// sources it actually tracks. The original jrtplib default (8317) targets RTP
+// mixers/conferences that demultiplex thousands of distinct SSRCs on one session;
+// it costs ~65 KB per session. Sevana's per-stream capture sessions carry ~1 SSRC,
+// so a far smaller table is ample - collisions are resolved by linked lists, so a
+// small size only affects lookup cost (negligible at our source counts), never
+// correctness. Overridable at build time for products that genuinely need many
+// sources per session.
+#ifndef RTPSOURCES_HASHSIZE
+#define RTPSOURCES_HASHSIZE							251
+#endif
 
 namespace jrtplib
 {

src/libs/resiprocate/resip/stack/Message.cxx

@@ -9,8 +9,12 @@
 using namespace resip;
 
 
+std::atomic<long> Message::sInstanceCount{0};
+
 Message::Message() : mTu(0)
-{}
+{
+   ++sInstanceCount;
+}
 
 Message::Brief
 Message::brief() const

src/libs/resiprocate/resip/stack/Message.hxx

@@ -7,6 +7,7 @@
 
 #include "rutil/Data.hxx"
 #include <iosfwd>
+#include <atomic>
 #include "rutil/resipfaststreams.hxx"
 
 namespace resip
@@ -21,7 +22,11 @@ class Message
 {
    public:
       Message();
-      virtual ~Message() {}
+      virtual ~Message() { --sInstanceCount; }
+
+      /// Live instance count of all Message-derived objects (leak indicator).
+      static std::atomic<long> sInstanceCount;
+      static long getInstanceCount() { return sInstanceCount.load(std::memory_order_relaxed); }
 
       /// facet for brief output to streams
       class Brief

src/libs/resiprocate/resip/stack/SipMessage.cxx

@@ -28,6 +28,8 @@ using namespace std;
 
 bool SipMessage::checkContentLength=true;
 
+std::atomic<long> SipMessage::sInstanceCount{0};
+
 SipMessage::SipMessage(const Tuple *receivedTransportTuple)
    : mIsDecorated(false),
      mIsBadAck200(false),     
@@ -51,6 +53,7 @@ SipMessage::SipMessage(const Tuple *receivedTransportTuple)
    // !bwc! TODO make this tunable
    mHeaders.reserve(16);
    clear();
+   ++sInstanceCount;
 }
 
 SipMessage::SipMessage(const SipMessage& from)
@@ -63,6 +66,7 @@ SipMessage::SipMessage(const SipMessage& from)
      mCreatedTime(Timer::getTimeMicroSec())
 {
    init(from);
+   ++sInstanceCount;
 }
 
 Message*
@@ -98,6 +102,7 @@ SipMessage::~SipMessage()
    }
 #endif
    freeMem();
+   --sInstanceCount;
 }
 
 void

src/libs/resiprocate/resip/stack/SipMessage.hxx

@@ -7,6 +7,7 @@
 #include <vector>
 #include <utility>
 #include <memory>
+#include <atomic>
 
 #include "resip/stack/Contents.hxx"
 #include "resip/stack/Headers.hxx"
@@ -154,6 +155,10 @@ class SipMessage : public TransactionMessage
 {
    public:
       RESIP_HeapCount(SipMessage);
+
+      /// Live instance count of SipMessage objects (leak indicator).
+      static std::atomic<long> sInstanceCount;
+      static long getInstanceCount() { return sInstanceCount.load(std::memory_order_relaxed); }
 #ifndef __SUNPRO_CC
       typedef std::list< std::pair<Data, HeaderFieldValueList*>, StlPoolAllocator<std::pair<Data, HeaderFieldValueList*>, PoolBase > > UnknownHeaders;
 #else