From 9d09ada33b26704a7f5d285d2f0d115e41df41bf Mon Sep 17 00:00:00 2001 From: Joshua Reisenauer Date: Mon, 2 May 2016 21:59:55 -0700 Subject: [PATCH] new boolean floatingPoint option Now floating point is either on or off. This simplifies the use of 16bit vs float. --- src/audio.c | 132 ++++++++++++++++++++++++++++++++------------------- src/audio.h | 7 ++- src/raylib.h | 7 ++- 3 files changed, 88 insertions(+), 58 deletions(-) diff --git a/src/audio.c b/src/audio.c index ff4f2858e..dc0637960 100644 --- a/src/audio.c +++ b/src/audio.c @@ -59,15 +59,17 @@ //---------------------------------------------------------------------------------- // Defines and Macros //---------------------------------------------------------------------------------- -#define MUSIC_STREAM_BUFFERS 2 -#define MAX_AUDIO_CONTEXTS 4 +#define MAX_STREAM_BUFFERS 2 +#define MAX_AUDIO_CONTEXTS 4 // Number of open AL sources #if defined(PLATFORM_RPI) || defined(PLATFORM_ANDROID) // NOTE: On RPI and Android should be lower to avoid frame-stalls - #define MUSIC_BUFFER_SIZE 4096*2 // PCM data buffer (short) - 16Kb (RPI) + #define MUSIC_BUFFER_SIZE_SHORT 4096*2 // PCM data buffer (short) - 16Kb (RPI) + #define MUSIC_BUFFER_SIZE_FLOAT 4096 // PCM data buffer (float) - 16Kb (RPI) #else // NOTE: On HTML5 (emscripten) this is allocated on heap, by default it's only 16MB!...just take care... - #define MUSIC_BUFFER_SIZE 4096*8 // PCM data buffer (short) - 64Kb + #define MUSIC_BUFFER_SIZE_SHORT 4096*8 // PCM data buffer (short) - 64Kb + #define MUSIC_BUFFER_SIZE_FLOAT 4096*4 // PCM data buffer (float) - 64Kb #endif //---------------------------------------------------------------------------------- @@ -80,7 +82,7 @@ typedef struct Music { stb_vorbis *stream; jar_xm_context_t *chipctx; // Stores jar_xm context - ALuint buffers[MUSIC_STREAM_BUFFERS]; + ALuint buffers[MAX_STREAM_BUFFERS]; ALuint source; ALenum format; @@ -96,12 +98,13 @@ typedef struct Music { // no more than 4 concurrent audio contexts in use. This is due to each active context being tied to // a dedicated mix channel. All audio is 32bit floating point in stereo. typedef struct AudioContext_t { - unsigned short sampleRate; // default is 48000 - unsigned char channels; // 1=mono,2=stereo - unsigned char mixChannel; // 0-3 or mixA-mixD, each mix channel can receive up to one dedicated audio stream - ALenum alFormat; // openAL format specifier - ALuint alSource; // openAL source - ALuint alBuffer[2]; // openAL sample buffer + unsigned short sampleRate; // default is 48000 + unsigned char channels; // 1=mono,2=stereo + unsigned char mixChannel; // 0-3 or mixA-mixD, each mix channel can receive up to one dedicated audio stream + bool floatingPoint; // if false then the short datatype is used instead + ALenum alFormat; // openAL format specifier + ALuint alSource; // openAL source + ALuint alBuffer[MAX_STREAM_BUFFERS]; // openAL sample buffer } AudioContext_t; #if defined(AUDIO_STANDALONE) @@ -126,7 +129,7 @@ static void UnloadWave(Wave wave); // Unload wave data static bool BufferMusicStream(ALuint buffer); // Fill music buffers with data static void EmptyMusicStream(void); // Empty music buffers -static void FillAlBufferWithSilence(AudioContext_t *ac, ALuint buffer);// fill buffer with zeros +static unsigned short FillAlBufferWithSilence(AudioContext_t *context, ALuint buffer);// fill buffer with zeros, returns number processed static void ResampleShortToFloat(short *shorts, float *floats, unsigned short len); // pass two arrays of the same legnth in static void ResampleByteToFloat(char *chars, float *floats, unsigned short len); // pass two arrays of same length in @@ -201,11 +204,10 @@ bool IsAudioDeviceReady(void) // Audio contexts are for outputing custom audio waveforms, This will shut down any other sound sources currently playing // The mixChannel is what mix channel you want to operate on, 0-3 are the ones available. Each mix channel can only be used one at a time. -// exmple usage is InitAudioContext(48000, 0, 2); // mixchannel 1, 48khz, stereo -// all samples are floating point by default -AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels) +// exmple usage is InitAudioContext(48000, 0, 2, true); // mixchannel 1, 48khz, stereo, floating point +AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels, bool floatingPoint) { - if(mixChannel > MAX_AUDIO_CONTEXTS) return NULL; + if(mixChannel >= MAX_AUDIO_CONTEXTS) return NULL; if(!IsAudioDeviceReady()) InitAudioDevice(); else StopMusicStream(); @@ -214,13 +216,24 @@ AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChanne ac->sampleRate = sampleRate; ac->channels = channels; ac->mixChannel = mixChannel; + ac->floatingPoint = floatingPoint; mixChannelsActive_g[mixChannel] = ac; // setup openAL format if(channels == 1) - ac->alFormat = AL_FORMAT_MONO_FLOAT32; - else - ac->alFormat = AL_FORMAT_STEREO_FLOAT32; + { + if(floatingPoint) + ac->alFormat = AL_FORMAT_MONO_FLOAT32; + else + ac->alFormat = AL_FORMAT_MONO16; + } + else if(channels == 2) + { + if(floatingPoint) + ac->alFormat = AL_FORMAT_STEREO_FLOAT32; + else + ac->alFormat = AL_FORMAT_STEREO16; + } // Create an audio source alGenSources(1, &ac->alSource); @@ -230,16 +243,17 @@ AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChanne alSource3f(ac->alSource, AL_VELOCITY, 0, 0, 0); // Create Buffer - alGenBuffers(2, ac->alBuffer); + alGenBuffers(MAX_STREAM_BUFFERS, ac->alBuffer); //fill buffers - FillAlBufferWithSilence(ac, ac->alBuffer[0]); - FillAlBufferWithSilence(ac, ac->alBuffer[1]); - alSourceQueueBuffers(ac->alSource, 2, ac->alBuffer); + int x; + for(x=0;xalBuffer[x]); + + alSourceQueueBuffers(ac->alSource, MAX_STREAM_BUFFERS, ac->alBuffer); alSourcei(ac->alSource, AL_LOOPING, AL_FALSE); // this could cause errors alSourcePlay(ac->alSource); - return ac; } return NULL; @@ -264,20 +278,22 @@ void CloseAudioContext(AudioContext ctx) //delete source and buffers alDeleteSources(1, &context->alSource); - alDeleteBuffers(2, context->alBuffer); + alDeleteBuffers(MAX_STREAM_BUFFERS, context->alBuffer); mixChannelsActive_g[context->mixChannel] = NULL; free(context); ctx = NULL; } } -// Pushes more audio data into context mix channel, if none are ever pushed then zeros are fed in -// Call "UpdateAudioContext(ctx, NULL, 0)" every game tick if you want to pause the audio -// Returns number of floats that where processed -unsigned short UpdateAudioContext(AudioContext ctx, float *data, unsigned short dataLength) +// Pushes more audio data into context mix channel, if none are ever pushed then zeros are fed in. +// Call "UpdateAudioContext(ctx, NULL, 0)" every game tick if you want to pause the audio. +// @Returns number of samples that where processed. +// All data streams should be of a length that is evenly divisible by MUSIC_BUFFER_SIZE, +// otherwise the remaining data will not be pushed. +unsigned short UpdateAudioContext(AudioContext ctx, void *data, unsigned short numberElements) { unsigned short numberProcessed = 0; - unsigned short numberRemaining = dataLength; + unsigned short numberRemaining = numberElements; AudioContext_t *context = (AudioContext_t*)ctx; if (context && mixChannelsActive_g[context->mixChannel] == context) @@ -288,44 +304,60 @@ unsigned short UpdateAudioContext(AudioContext ctx, float *data, unsigned short if(!processed) return 0;//nothing to process, queue is still full - if (!data || !dataLength)// play silence + if (!data || !numberElements)// play silence + { while (processed > 0) { alSourceUnqueueBuffers(context->alSource, 1, &buffer); - FillAlBufferWithSilence(context, buffer); + numberProcessed += FillAlBufferWithSilence(context, buffer); alSourceQueueBuffers(context->alSource, 1, &buffer); processed--; - numberProcessed+=MUSIC_BUFFER_SIZE; } + } if(numberRemaining)// buffer data stream in increments of MUSIC_BUFFER_SIZE + { while (processed > 0) { - alSourceUnqueueBuffers(context->alSource, 1, &buffer); - if(numberRemaining >= MUSIC_BUFFER_SIZE) + if(context->floatingPoint && numberRemaining >= MUSIC_BUFFER_SIZE_FLOAT) // process float buffers { - float pcm[MUSIC_BUFFER_SIZE]; - memcpy(pcm, &data[numberProcessed], MUSIC_BUFFER_SIZE); - alBufferData(buffer, context->alFormat, pcm, MUSIC_BUFFER_SIZE*sizeof(float), context->sampleRate); + float *ptr = (float*)data; + alSourceUnqueueBuffers(context->alSource, 1, &buffer); + alBufferData(buffer, context->alFormat, &ptr[numberProcessed], MUSIC_BUFFER_SIZE_FLOAT*sizeof(float), context->sampleRate); alSourceQueueBuffers(context->alSource, 1, &buffer); - numberProcessed+=MUSIC_BUFFER_SIZE; - numberRemaining-=MUSIC_BUFFER_SIZE; + numberProcessed+=MUSIC_BUFFER_SIZE_FLOAT; + numberRemaining-=MUSIC_BUFFER_SIZE_FLOAT; } - else // less than MUSIC_BUFFER_SIZE number of samples left to buffer, the remaining data is padded with zeros + else if(!context->floatingPoint && numberRemaining >= MUSIC_BUFFER_SIZE_SHORT) // process short buffers { - float pcm[MUSIC_BUFFER_SIZE] = {0.f}; // pad with zeros + short *ptr = (short*)data; + alSourceUnqueueBuffers(context->alSource, 1, &buffer); + alBufferData(buffer, context->alFormat, &ptr[numberProcessed], MUSIC_BUFFER_SIZE_SHORT*sizeof(short), context->sampleRate); + alSourceQueueBuffers(context->alSource, 1, &buffer); + numberProcessed+=MUSIC_BUFFER_SIZE_SHORT; + numberRemaining-=MUSIC_BUFFER_SIZE_SHORT; } processed--; } + } } return numberProcessed; } -// fill buffer with zeros -static void FillAlBufferWithSilence(AudioContext_t *context, ALuint buffer) +// fill buffer with zeros, returns number processed +static unsigned short FillAlBufferWithSilence(AudioContext_t *context, ALuint buffer) { - float pcm[MUSIC_BUFFER_SIZE] = {0.f}; - alBufferData(buffer, context->alFormat, pcm, MUSIC_BUFFER_SIZE*sizeof(float), context->sampleRate); + if(context->floatingPoint){ + float pcm[MUSIC_BUFFER_SIZE_FLOAT] = {0.f}; + alBufferData(buffer, context->alFormat, pcm, MUSIC_BUFFER_SIZE_FLOAT*sizeof(float), context->sampleRate); + return MUSIC_BUFFER_SIZE_FLOAT; + } + else + { + short pcm[MUSIC_BUFFER_SIZE_SHORT] = {0}; + alBufferData(buffer, context->alFormat, pcm, MUSIC_BUFFER_SIZE_SHORT*sizeof(short), context->sampleRate); + return MUSIC_BUFFER_SIZE_SHORT; + } } // example usage: @@ -920,7 +952,7 @@ float GetMusicTimePlayed(void) // Fill music buffers with new data from music stream static bool BufferMusicStream(ALuint buffer) { - short pcm[MUSIC_BUFFER_SIZE]; + short pcm[MUSIC_BUFFER_SIZE_SHORT]; int size = 0; // Total size of data steamed (in bytes) int streamedBytes = 0; // samples of data obtained, channels are not included in calculation @@ -930,15 +962,15 @@ static bool BufferMusicStream(ALuint buffer) { if (currentMusic.chipTune) // There is no end of stream for xmfiles, once the end is reached zeros are generated for non looped chiptunes. { - int readlen = MUSIC_BUFFER_SIZE / 2; + int readlen = MUSIC_BUFFER_SIZE_SHORT / 2; jar_xm_generate_samples_16bit(currentMusic.chipctx, pcm, readlen); // reads 2*readlen shorts and moves them to buffer+size memory location size += readlen * currentMusic.channels; // Not sure if this is what it needs } else { - while (size < MUSIC_BUFFER_SIZE) + while (size < MUSIC_BUFFER_SIZE_SHORT) { - streamedBytes = stb_vorbis_get_samples_short_interleaved(currentMusic.stream, currentMusic.channels, pcm + size, MUSIC_BUFFER_SIZE - size); + streamedBytes = stb_vorbis_get_samples_short_interleaved(currentMusic.stream, currentMusic.channels, pcm + size, MUSIC_BUFFER_SIZE_SHORT - size); if (streamedBytes > 0) size += (streamedBytes*currentMusic.channels); else break; } diff --git a/src/audio.h b/src/audio.h index 2b36b3f04..9d6fa5d09 100644 --- a/src/audio.h +++ b/src/audio.h @@ -84,11 +84,10 @@ bool IsAudioDeviceReady(void); // True if call // Audio contexts are for outputing custom audio waveforms, This will shut down any other sound sources currently playing // The mixChannel is what mix channel you want to operate on, 0-3 are the ones available. Each mix channel can only be used one at a time. -// exmple usage is InitAudioContext(48000, 0, 2); // mixchannel 1, 48khz, stereo -// all samples are floating point by default -AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels); +// exmple usage is InitAudioContext(48000, 0, 2, true); // mixchannel 1, 48khz, stereo, floating point +AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels, bool floatingPoint); void CloseAudioContext(AudioContext ctx); // Frees audio context -unsigned short UpdateAudioContext(AudioContext ctx, float *data, unsigned short dataLength); // Pushes more audio data into context mix channel, if NULL is passed to data then zeros are played +unsigned short UpdateAudioContext(AudioContext ctx, void *data, unsigned short numberElements); // Pushes more audio data into context mix channel, if NULL is passed to data then zeros are played Sound LoadSound(char *fileName); // Load sound to memory Sound LoadSoundFromWave(Wave wave); // Load sound to memory from wave data diff --git a/src/raylib.h b/src/raylib.h index 5c727b615..c9ba5b384 100644 --- a/src/raylib.h +++ b/src/raylib.h @@ -870,11 +870,10 @@ bool IsAudioDeviceReady(void); // True if call // Audio contexts are for outputing custom audio waveforms, This will shut down any other sound sources currently playing // The mixChannel is what mix channel you want to operate on, 0-3 are the ones available. Each mix channel can only be used one at a time. -// exmple usage is InitAudioContext(48000, 0, 2); // mixchannel 1, 48khz, stereo -// all samples are floating point by default -AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels); +// exmple usage is InitAudioContext(48000, 0, 2, true); // mixchannel 1, 48khz, stereo, floating point +AudioContext InitAudioContext(unsigned short sampleRate, unsigned char mixChannel, unsigned char channels, bool floatingPoint); void CloseAudioContext(AudioContext ctx); // Frees audio context -unsigned short UpdateAudioContext(AudioContext ctx, float *data, unsigned short dataLength); // Pushes more audio data into context mix channel, if NULL is passed to data then zeros are played +unsigned short UpdateAudioContext(AudioContext ctx, void *data, unsigned short numberElements); // Pushes more audio data into context mix channel, if NULL is passed to data then zeros are played Sound LoadSound(char *fileName); // Load sound to memory Sound LoadSoundFromWave(Wave wave); // Load sound to memory from wave data