quita el glitch al resproducir después de cargar un fichero.

varias optimizaciones y comprobaciones

docs/changelog.txt

@@ -1,10 +1,10 @@
 *******************************************************************************
-Libre Media Server Audio - An Open source Media Server for arts and performing.
+LibreMediaServer Audio - An open source media server for arts and performing.
 (c) Criptomart - Santiago Noreña 2012-2024 <lms@criptomart.net>
 https://git.criptomart.net/libremediaserver
 *******************************************************************************
 
-Libre Media Server ChangeLog
+LibreMediaServer Changelog
 
 v 0.2.0 Antígona (26/05/2024)
 + Change audio engine to miniaudio because is imposible pan in SFML and it has not access to low API and audio processing.
@@ -20,9 +20,7 @@ v 0.2.0 Antígona (26/05/2024)
 + OlaThread send double channels (volume, entry point, load media) only once for each dmx frame buffer.
 + Terminal mode without graphical interface. All audio methods has been refactorized out of QWidget world.
 + Compilation without GUI (-DNOGUI).
-+ New Status "Iddle" in playbacks if is not loaded.
 + New DMX personality version, better sort for audio needs (first load media, set vol, pan, etc, last playback order);
-+ Refresh layer values when it loads a new sound file.
 + No QtSignals for sending data, better performance about 20% in my machine. Now, libremediaserver only updates values in AudioWidget, ui refresh is doing with a timer in audiowidget, so there is not problems between graphical and ola thread (the callback). Signals are used only from Ui to libreMediaServer to notify user interactions and Qtimers.
 + Load media files from ui clicking in the media labels.
 + New Play Modes:
@@ -31,6 +29,7 @@ v 0.2.0 Antígona (26/05/2024)
     - Play all medias in one folder randomly.
 + Multi audio devices output.
 + Vumeter for each layer
++ Show device name on Ui and ouput bus slider.
 
 v 0.1.3 Leúcade (19/04/2024)
 + Ubuntu 22.04 jammy.

docs/roadmap.txt

@@ -6,38 +6,32 @@ https://git.criptomart.net/libremediaserver
 
 Libre Media Server Roadmap
 
-v  0.2.x
+v  0.3.0
-- skin, UI/UX
+- Ui/Ux: skin, style.
+- Ui/Ux; Keyboards strokes.
+- Ui/Ux: Dar la opción clickeando en el widget de tiempo de poner una cuenta atrás en vez de hacia delante.
+- Ui/Ux: seek cursor playback
 - live input.
-- insertar/bypass/eliminar audio procesadores sin reiniciar por capa y master. (compresor, equs).
+- remove ola and use sACN directly.
-- FX en capas master para que se puedan usar como envíos de auxiliar.
-- Enroutado de masters en otros masters (retorno de efectos).
-
-v 0.2.2
-- Use sACN directly.
    + la instalación de OLA es mediante compilación, el repo de paquetes no está actualizado, nada user-friendly.
    + hay que empaquetar OLA, incluirlo en el binario, o implementar sACN y linkarlo estáticamente.
    + https://github.com/ETCLabs/sACN
 - Qt6.
-- Audio processing (eq, rev, compresor, ...) by master and layer.
+- CIPT/MSex.
-- CIPT/MSex, send icons play/pause/stop.
 - Rasp build.
 - Octopus Sound Card support (6 outputs - 8 inputs).
-
-v 0.2.1
-- mute/panic on layer.
 - Master Bus Layer:
-    - each layer will have one "Gain" prefader that acts in source, "Vol" in v 1.3.
-    - each layer will have one volume dmx channel for each bus layer. One aux "Send" prefader.
     - mute/panic.
     - fader + value
-    - pan.
+    - pan
-    - magicq .hed
+    - magicq personality .hed
     - audio device linked, outputs will be redirected there.
     - dmx address + universe settings.
-- Rose noise and sine generator in menu to test system.
+    - compresor/limiter.
-- Ui/Ux; Keyboards strokes.
+- Layer:
-- Ui/Ux: Dar la opción clickeando en el widget de tiempo de poner una cuenta atrás en vez de hacia delante.
+    - audio procesadores (compresor, reveb, delay).
+    - mute/panic.
+- Rose noise and sine generator.
 - Logs, verbosity, timestamp.
 - New play mode without pitch control, it saves resources. MA_SOUND_FLAG_NO_PITCH
 - SettingsDialog.
@@ -45,8 +39,4 @@ v 0.2.1
 - ¿Exit Point? is it needed?
 - Hardening: check return errors, try/catch exceptions, i'm too happy....
 - Tests: errors on wrong conf file.
-- Ui/Ux: seek cursor playback
 - ampliar writer para recibir un número n de entradas y escribirlas cada una en un buffer
-
-v0.2.0:
-- mostrad información de envíos y dispositivos en ui

src/audiolayerwidget.cpp

@@ -198,6 +198,7 @@ void AudioLayerWidget::openMediaDialog()
     fileNames = dialog.selectedFiles();
     emit uiLoadMedia(m_layer, fileNames.at(0));
     this->setMediaFile(fileNames.at(0));
+    this->setPlaybackStatus(Status::Stopped);
 }
 
 // from DMX signals
@@ -267,18 +268,18 @@ void AudioLayerWidget::setCurrentTime(float progress)
 
 void AudioLayerWidget::setFilterParam(int channel, int value)
 {
-    if (channel <= FILTER_BANK_GAIN - FILTER_CHANNELS){
+    if (channel <= FILTER_BANK_GAIN - HP_FREQ){
         m_filterBank->blockSignals(true);
         m_filterBank->setValue(channel, value);
         m_filterBank->blockSignals(false);
     } else if (channel == SEND1 - HP_FREQ) {
-        m_bus1->blockSignals(false);
-        m_bus1->setValue((value * 256) + 255);
         m_bus1->blockSignals(true);
+        m_bus1->setValue((value * 256) + 255);
+        m_bus1->blockSignals(false);
     } else if (channel == SEND2 - HP_FREQ) {
-        m_bus2->blockSignals(false);
-        m_bus2->setValue(value * 256 + 255);
         m_bus2->blockSignals(true);
+        m_bus2->setValue(value * 256 + 255);
+        m_bus2->blockSignals(false);
     }
 }
 

src/libremediaserver-audio.cpp

@@ -70,8 +70,7 @@ void libreMediaServerAudio::loadMedia(int layer, int folder, int file)
     if (strcmp(mediaFile.toLatin1().constData(), m_currentMedia[layer].toLatin1().constData()) == 0)
         return;
     if (QFile::exists(mediaFile)){
-        m_mae.loadMedia(layer, mediaFile.toLatin1().data(),\
+        m_mae.loadMedia(layer, mediaFile.toLatin1().data());
-                        m_dmxSettings.at(layer).audioDevice);
         m_currentMedia[layer] = mediaFile;
 #ifndef NOGUI
     if (m_ui)
@@ -135,7 +134,7 @@ void libreMediaServerAudio::dmxInput(int layer, int channel, int value)
         }
 #endif
     } else if (channel >= HP_FREQ) {
-        m_mae.filterParamChanged(layer, m_dmxSettings.at(layer).audioDevice, channel, value);
+        m_mae.filterParamChanged(layer, channel, value);
 #ifndef NOGUI
         if (m_ui) {
             m_lmsUi->m_aw->filterParamChanged(layer, channel, value);
@@ -252,10 +251,10 @@ void libreMediaServerAudio::uiSliderChanged(int layer, Slider s, int value)
         m_mae.setBypass(m_dmxSettings.at(layer).audioDevice, layer, value);
         break;
     case Slider::Bus1:
-        m_mae.filterParamChanged(layer, m_dmxSettings.at(layer).audioDevice, SEND1, value / 255);
+        m_mae.filterParamChanged(layer, SEND1, value / 255.0f);
         break;
     case Slider::Bus2:
-        m_mae.filterParamChanged(layer, m_dmxSettings.at(layer).audioDevice, SEND2, value / 255);
+        m_mae.filterParamChanged(layer, SEND2, value / 255.0f);
         break;
     }
 }
@@ -278,7 +277,7 @@ void libreMediaServerAudio::uiLoadMedia(int layer, QString mediaFile)
 
     if (strcmp(mediaFile.toLatin1().constData(), m_currentMedia[layer].toLatin1().constData()) == 0)
         return;
-    result = m_mae.loadMedia(layer, mediaFile.toLatin1().data(), m_dmxSettings[layer].audioDevice);
+    result = m_mae.loadMedia(layer, mediaFile.toLatin1().data());
     if (result == MA_SUCCESS) {
         m_currentMedia[layer] = mediaFile;
         m_lmsUi->m_aw->mediaLoaded(layer, mediaFile, m_mae.getDuration(layer));

src/ma_writer_node.c

@@ -38,8 +38,6 @@ static ma_node_vtable g_ma_writer_node_vtable =
     2,
     1,
     0
-//    MA_NODE_FLAG_CONTINUOUS_PROCESSING
-//    MA_NODE_FLAG_SILENT_OUTPUT
 };
 
 MA_API ma_result ma_writer_node_init(ma_node_graph* pNodeGraph, const ma_writer_node_config* pConfig, const ma_allocation_callbacks* pAllocationCallbacks, ma_writer_node* pWriteNode)
@@ -178,7 +176,7 @@ void ma_data_source_rb_uninit(ma_data_source_rb* pMyDataSource)
  * vumeter
  */
 
-MA_API ma_vumeter_node_config ma_vumeter_node_config_init(ma_uint32 channels, ma_uint32 sampleRate)
+MA_API ma_vumeter_node_config ma_vumeter_node_config_init(ma_uint32 channels, ma_uint32 format, ma_uint32 sampleRate)
 {
     ma_vumeter_node_config config;
 
@@ -186,6 +184,7 @@ MA_API ma_vumeter_node_config ma_vumeter_node_config_init(ma_uint32 channels, ma
     config.nodeConfig = ma_node_config_init();
     config.channels   = channels;
     config.sampleRate = sampleRate;
+    config.format     = format;
 
     return config;
 }
@@ -201,7 +200,7 @@ static void ma_vumeter_node_process_pcm_frames(ma_node* pNode, const float** ppF
         float input = fabsf(ppFramesIn[0][i]);
         pVumeterNode->level += pVumeterNode->alpha * (input - pVumeterNode->level);
     }
-    ma_copy_pcm_frames(ppFramesOut[0], ppFramesIn[0], *pFrameCountOut, ma_format_f32, pVumeterNode->channels);
+    ma_copy_pcm_frames(ppFramesOut[0], ppFramesIn[0], *pFrameCountOut, pVumeterNode->format, pVumeterNode->channels);
 }
 
 static ma_node_vtable g_ma_vumeter_node_vtable =
@@ -238,8 +237,9 @@ MA_API ma_result ma_vumeter_node_init(ma_node_graph* pNodeGraph, const ma_vumete
     }
     pVumeterNode->sampleRate = pConfig->sampleRate;
     pVumeterNode->channels = pConfig->channels;
+    pVumeterNode->format = pConfig->format;
     pVumeterNode->level = 0;
-    pVumeterNode->TC = 10000.0f;
+    pVumeterNode->TC = 0.250f;
     pVumeterNode->alpha = 1.0 - expf( (-2.0 * M_PI) / (pVumeterNode->TC * pConfig->sampleRate));
     return MA_SUCCESS;
 }

src/ma_writer_node.h

@@ -59,15 +59,17 @@ typedef struct
     ma_node_config nodeConfig;
     ma_uint32 channels;
     ma_uint32 sampleRate;
+    ma_uint32 format;
 } ma_vumeter_node_config;
 
-MA_API ma_vumeter_node_config ma_vumeter_node_config_init(ma_uint32 channels, ma_uint32 sampleRate);
+MA_API ma_vumeter_node_config ma_vumeter_node_config_init(ma_uint32 channels, ma_uint32 format, ma_uint32 sampleRate);
 
 typedef struct
 {
     ma_node_base baseNode;
     ma_uint32 channels;
     ma_uint32 sampleRate;
+    ma_uint32 format;
     float level;
     float TC;
     float alpha;

src/miniaudioengine.cpp

@@ -132,7 +132,7 @@ ma_result MiniAudioEngine::createFilterBank(uint layer)
         cout << "ERROR " << result << ": Failed to init hi shelf filter node." << endl;
         return result;
     }
-    ma_vumeter_node_config vuc = ma_vumeter_node_config_init(CHANNELS, FORMAT);
+    ma_vumeter_node_config vuc = ma_vumeter_node_config_init(CHANNELS, FORMAT, SAMPLE_RATE);
     ma_vumeter_node_init(ng, &vuc, NULL, &fb->vumeter);
     if (result != MA_SUCCESS) {
         cout << "ERROR " << result << ": Failed to init vumeter node." << endl;
@@ -278,7 +278,7 @@ ma_result MiniAudioEngine::startDevices()
     deviceConfig.dataCallback       = audioDataCallback;
     engineConfig                  = ma_engine_config_init();
     engineConfig.pResourceManager = &m_mae.resourceManager;
-    engineConfig.gainSmoothTimeInMilliseconds = SAMPLE_RATE / 25;
+    engineConfig.defaultVolumeSmoothTimeInPCMFrames = SAMPLE_RATE / 20;
     engineConfig.noAutoStart      = MA_TRUE;
 
     for (uint internalId = 0; internalId < m_mae.audioDevicesQty; internalId++) {
@@ -346,27 +346,30 @@ char* MiniAudioEngine::getDeviceName(uint id)
 
 }
 
-ma_result MiniAudioEngine::loadMedia(int layer, char *file, uint audioDevice)
+ma_result MiniAudioEngine::loadMedia(int layer, char *file)
 {
     ma_result result;
 
     if (m_mae.mediaLoaded[layer] == true)
     {
+        ma_sound_set_volume(&m_mae.sounds[layer], 0.0f);
+        ma_sound_stop(&m_mae.sounds[layer]);
         ma_sound_uninit(&m_mae.sounds[layer]);
         m_mae.mediaLoaded[layer] = false;
     }
-    result = ma_sound_init_from_file(&m_mae.engines[0], file, \
+    ma_sound_config soundConfig = ma_sound_config_init();
-                MA_SOUND_FLAG_NO_SPATIALIZATION \
+    soundConfig = ma_sound_config_init();
-                , NULL, NULL, &m_mae.sounds[layer]);
+    soundConfig.pFilePath = file;
+    soundConfig.pInitialAttachment = &m_mae.filters[layer].input;
+    soundConfig.initialAttachmentInputBusIndex = 0;
+    soundConfig.channelsIn = 0;
+    soundConfig.channelsOut = CHANNELS;
+    soundConfig.flags = MA_SOUND_FLAG_NO_SPATIALIZATION | MA_SOUND_FLAG_NO_DEFAULT_ATTACHMENT | MA_SOUND_FLAG_STREAM; //| MA_SOUND_FLAG_NO_PITCH
+    result = ma_sound_init_ex(&m_mae.engines[0], &soundConfig, &m_mae.sounds[layer]);
     if (result != MA_SUCCESS) {
         cout << "Error" << result << ": Failed to load file " << file << endl;
         return result;
     }
-    result = ma_node_attach_output_bus(&m_mae.sounds[layer], 0, &m_mae.filters[layer].input, 0);
-    if (result != MA_SUCCESS) {
-        cout << "Error " << result << ": Failed to attach sound output bus " << audioDevice << endl;
-        return result;
-    }
     m_mae.mediaLoaded[layer] = true;
     this->refreshValues(layer);
     m_mae.currentStatus[layer].media = file;
@@ -435,7 +438,7 @@ void MiniAudioEngine::volChanged(int layer, int vol)
         db = 0;
     } else
         db = ma_volume_db_to_linear(db);
-    ma_sound_group_set_fade_in_milliseconds(&m_mae.sounds[layer], -1, db, FADE_TIME);
+    ma_sound_set_fade_in_milliseconds(&m_mae.sounds[layer], -1, db, FADE_TIME);
     m_mae.currentStatus[layer].vol = vol;
 }
 
@@ -485,12 +488,13 @@ ma_result MiniAudioEngine::playbackChanged(int layer, Status status)
             ma_sound_set_stop_time_in_milliseconds(&m_mae.sounds[layer], ~(ma_uint64)0);
             ma_sound_set_looping(&m_mae.sounds[layer], loop);
             result = ma_sound_start(&m_mae.sounds[layer]);
-            if (m_mae.currentStatus[layer].cursor != 0) {
+            //this->volChanged(layer, m_mae.currentStatus[layer].vol);
-                usleep(1000 * 50); // Avoid small glitch at start, how to flush the cached buffers in audio pipe line?
+            float db = (m_mae.currentStatus[layer].vol / 771.0f) - 85.0f;
-            }
+            if (db <= -85.0f) {
-            this->volChanged(layer, m_mae.currentStatus[layer].vol);
+                db = 0;
-        default:
+            } else
-            break;
+                db = ma_volume_db_to_linear(db);
+            ma_sound_set_fade_in_milliseconds(&m_mae.sounds[layer], -1, db, 0);
     }
     if (result == MA_SUCCESS)
         m_mae.currentStatus[layer].status = status;
@@ -506,7 +510,7 @@ ma_result MiniAudioEngine::seekToCursor(int layer, int cursor)
         return MA_DOES_NOT_EXIST;
     result = ma_sound_get_length_in_pcm_frames(&m_mae.sounds[layer], &end);
     if (result != MA_SUCCESS) { return result; }
-    start = (cursor * end) / 65025;
+    start = (cursor * end) / 65535;
     Status oldStatus = m_mae.currentStatus[layer].status;
     result = ma_sound_seek_to_pcm_frame(&m_mae.sounds[layer], start);
     //result = ma_data_source_set_loop_point_in_pcm_frames(&m_mae.sounds[layer], start, end); // this do nothing here, it must be done after set_looping or start?
@@ -531,14 +535,14 @@ void MiniAudioEngine::refreshValues(int layer)
 {
     this->seekToCursor(layer, m_mae.currentStatus[layer].cursor);
     this->panChanged(layer, m_mae.currentStatus[layer].pan);
-    this->volChanged(layer, m_mae.currentStatus[layer].vol);
     this->pitchChanged(layer, m_mae.currentStatus[layer].pitch);
+    ma_sound_group_set_fade_in_milliseconds(&m_mae.sounds[layer], -1, 0, 0.0f);
     this->playbackChanged(layer, m_mae.currentStatus[layer].status);
+    this->volChanged(layer, m_mae.currentStatus[layer].vol);
 }
 
-ma_result MiniAudioEngine::filterParamChanged(int layer, int audioDevice, int channel, int value)
+ma_result MiniAudioEngine::filterParamChanged(int layer, int channel, int value)
 {
-    (void)audioDevice;
     ma_result result = MA_SUCCESS;
     filterBank *fb = &m_mae.filters[layer];
 

src/miniaudioengine.h

@@ -65,7 +65,7 @@ protected:
     MiniAudioEngine();
     void stopEngine();
     bool startEngine(uint layersQty, uint* audioDevicesID, uint audioDevicesQty);
-    ma_result loadMedia(int layer, char *media, uint audioDevice);
+    ma_result loadMedia(int layer, char *media);
     void volChanged(int layer, int vol);
     void panChanged(int layer, float pan);
     void pitchChanged(int layer, float pitch);
@@ -79,11 +79,11 @@ protected:
         return ma_sound_get_volume(&m_mae.sounds[layer]);
     };
     inline bool getAtEnd(int layer) { return m_mae.sounds[layer].atEnd; }
-    ma_result filterParamChanged(int layer, int audioDevice, int channel, int value);
+    ma_result filterParamChanged(int layer, int channel, int value);
     bool setBypass(int audioDevice, int layer, bool bypass);
     inline float getLevel(int layer) {
         float level = ma_vumeter_node_get_level(&m_mae.filters[layer].vumeter);
-        return ma_volume_linear_to_db(level);
+        return ma_volume_linear_to_db(level) - 4.0f;
     };
     char* getDeviceName(uint id);