view src/audio/SDL_wave.c @ 968:4675910b0b7b

Date: Mon, 11 Oct 2004 15:17:27 +0300 (EEST) From: Hannu Savolainen Subject: Re: SDL uses obsolete OSS features I did some work on getting OSS to work better with SDL. There have been some problems with select which should be fixed now. I'm having some problems in understanding what is the purpose of the DSP_WaitAudio() routine. I added a return to the very beginning of this routine and commendted out the define for USE_BLOCKING_WRITES. At least lbreakout2 seems to work as well as earlier. The latencies are the same. An ordinary blocking write does exactly the same thing than DSP_WaitAudio does. So I would recommend using the USE_BLOCKING_WRITES approach and removing everything from the DSP_WaitAudio routine. Also enabling USE_BLOCKING_WRITES makes it possible to simplify DSP_PlayAudio() because you don't need to handle the partial writes (the do-while loop). Attached is a patch against SDL-1.2.7. After these changes SDL will use OSS as it's designed to be used (make it as simple as possible). This code should work with all OSS implementations because it uses only the very fundamental features that have been there since the jurassic times.
author Sam Lantinga <slouken@libsdl.org>
date Fri, 12 Nov 2004 21:39:04 +0000
parents b8d311d90021
children 80f8c94b5199
line wrap: on
line source

/*
    SDL - Simple DirectMedia Layer
    Copyright (C) 1997-2004 Sam Lantinga

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Library General Public
    License as published by the Free Software Foundation; either
    version 2 of the License, or (at your option) any later version.

    This library is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    Library General Public License for more details.

    You should have received a copy of the GNU Library General Public
    License along with this library; if not, write to the Free
    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA

    Sam Lantinga
    slouken@libsdl.org
*/

#ifdef SAVE_RCSID
static char rcsid =
 "@(#) $Id$";
#endif

#ifndef DISABLE_FILE

/* Microsoft WAVE file loading routines */

#include <stdlib.h>
#include <string.h>

#include "SDL_error.h"
#include "SDL_audio.h"
#include "SDL_wave.h"
#include "SDL_endian.h"

#ifndef NELEMS
#define NELEMS(array)	((sizeof array)/(sizeof array[0]))
#endif

static int ReadChunk(SDL_RWops *src, Chunk *chunk);

struct MS_ADPCM_decodestate {
	Uint8 hPredictor;
	Uint16 iDelta;
	Sint16 iSamp1;
	Sint16 iSamp2;
};
static struct MS_ADPCM_decoder {
	WaveFMT wavefmt;
	Uint16 wSamplesPerBlock;
	Uint16 wNumCoef;
	Sint16 aCoeff[7][2];
	/* * * */
	struct MS_ADPCM_decodestate state[2];
} MS_ADPCM_state;

static int InitMS_ADPCM(WaveFMT *format)
{
	Uint8 *rogue_feel;
	Uint16 extra_info;
	int i;

	/* Set the rogue pointer to the MS_ADPCM specific data */
	MS_ADPCM_state.wavefmt.encoding = SDL_SwapLE16(format->encoding);
	MS_ADPCM_state.wavefmt.channels = SDL_SwapLE16(format->channels);
	MS_ADPCM_state.wavefmt.frequency = SDL_SwapLE32(format->frequency);
	MS_ADPCM_state.wavefmt.byterate = SDL_SwapLE32(format->byterate);
	MS_ADPCM_state.wavefmt.blockalign = SDL_SwapLE16(format->blockalign);
	MS_ADPCM_state.wavefmt.bitspersample =
					 SDL_SwapLE16(format->bitspersample);
	rogue_feel = (Uint8 *)format+sizeof(*format);
	if ( sizeof(*format) == 16 ) {
		extra_info = ((rogue_feel[1]<<8)|rogue_feel[0]);
		rogue_feel += sizeof(Uint16);
	}
	MS_ADPCM_state.wSamplesPerBlock = ((rogue_feel[1]<<8)|rogue_feel[0]);
	rogue_feel += sizeof(Uint16);
	MS_ADPCM_state.wNumCoef = ((rogue_feel[1]<<8)|rogue_feel[0]);
	rogue_feel += sizeof(Uint16);
	if ( MS_ADPCM_state.wNumCoef != 7 ) {
		SDL_SetError("Unknown set of MS_ADPCM coefficients");
		return(-1);
	}
	for ( i=0; i<MS_ADPCM_state.wNumCoef; ++i ) {
		MS_ADPCM_state.aCoeff[i][0] = ((rogue_feel[1]<<8)|rogue_feel[0]);
		rogue_feel += sizeof(Uint16);
		MS_ADPCM_state.aCoeff[i][1] = ((rogue_feel[1]<<8)|rogue_feel[0]);
		rogue_feel += sizeof(Uint16);
	}
	return(0);
}

static Sint32 MS_ADPCM_nibble(struct MS_ADPCM_decodestate *state,
					Uint8 nybble, Sint16 *coeff)
{
	const Sint32 max_audioval = ((1<<(16-1))-1);
	const Sint32 min_audioval = -(1<<(16-1));
	const Sint32 adaptive[] = {
		230, 230, 230, 230, 307, 409, 512, 614,
		768, 614, 512, 409, 307, 230, 230, 230
	};
	Sint32 new_sample, delta;

	new_sample = ((state->iSamp1 * coeff[0]) +
		      (state->iSamp2 * coeff[1]))/256;
	if ( nybble & 0x08 ) {
		new_sample += state->iDelta * (nybble-0x10);
	} else {
		new_sample += state->iDelta * nybble;
	}
	if ( new_sample < min_audioval ) {
		new_sample = min_audioval;
	} else
	if ( new_sample > max_audioval ) {
		new_sample = max_audioval;
	}
	delta = ((Sint32)state->iDelta * adaptive[nybble])/256;
	if ( delta < 16 ) {
		delta = 16;
	}
	state->iDelta = delta;
	state->iSamp2 = state->iSamp1;
	state->iSamp1 = new_sample;
	return(new_sample);
}

static int MS_ADPCM_decode(Uint8 **audio_buf, Uint32 *audio_len)
{
	struct MS_ADPCM_decodestate *state[2];
	Uint8 *freeable, *encoded, *decoded;
	Sint32 encoded_len, samplesleft;
	Sint8 nybble, stereo;
	Sint16 *coeff[2];
	Sint32 new_sample;

	/* Allocate the proper sized output buffer */
	encoded_len = *audio_len;
	encoded = *audio_buf;
	freeable = *audio_buf;
	*audio_len = (encoded_len/MS_ADPCM_state.wavefmt.blockalign) * 
				MS_ADPCM_state.wSamplesPerBlock*
				MS_ADPCM_state.wavefmt.channels*sizeof(Sint16);
	*audio_buf = (Uint8 *)malloc(*audio_len);
	if ( *audio_buf == NULL ) {
		SDL_Error(SDL_ENOMEM);
		return(-1);
	}
	decoded = *audio_buf;

	/* Get ready... Go! */
	stereo = (MS_ADPCM_state.wavefmt.channels == 2);
	state[0] = &MS_ADPCM_state.state[0];
	state[1] = &MS_ADPCM_state.state[stereo];
	while ( encoded_len >= MS_ADPCM_state.wavefmt.blockalign ) {
		/* Grab the initial information for this block */
		state[0]->hPredictor = *encoded++;
		if ( stereo ) {
			state[1]->hPredictor = *encoded++;
		}
		state[0]->iDelta = ((encoded[1]<<8)|encoded[0]);
		encoded += sizeof(Sint16);
		if ( stereo ) {
			state[1]->iDelta = ((encoded[1]<<8)|encoded[0]);
			encoded += sizeof(Sint16);
		}
		state[0]->iSamp1 = ((encoded[1]<<8)|encoded[0]);
		encoded += sizeof(Sint16);
		if ( stereo ) {
			state[1]->iSamp1 = ((encoded[1]<<8)|encoded[0]);
			encoded += sizeof(Sint16);
		}
		state[0]->iSamp2 = ((encoded[1]<<8)|encoded[0]);
		encoded += sizeof(Sint16);
		if ( stereo ) {
			state[1]->iSamp2 = ((encoded[1]<<8)|encoded[0]);
			encoded += sizeof(Sint16);
		}
		coeff[0] = MS_ADPCM_state.aCoeff[state[0]->hPredictor];
		coeff[1] = MS_ADPCM_state.aCoeff[state[1]->hPredictor];

		/* Store the two initial samples we start with */
		decoded[0] = state[0]->iSamp2&0xFF;
		decoded[1] = state[0]->iSamp2>>8;
		decoded += 2;
		if ( stereo ) {
			decoded[0] = state[1]->iSamp2&0xFF;
			decoded[1] = state[1]->iSamp2>>8;
			decoded += 2;
		}
		decoded[0] = state[0]->iSamp1&0xFF;
		decoded[1] = state[0]->iSamp1>>8;
		decoded += 2;
		if ( stereo ) {
			decoded[0] = state[1]->iSamp1&0xFF;
			decoded[1] = state[1]->iSamp1>>8;
			decoded += 2;
		}

		/* Decode and store the other samples in this block */
		samplesleft = (MS_ADPCM_state.wSamplesPerBlock-2)*
					MS_ADPCM_state.wavefmt.channels;
		while ( samplesleft > 0 ) {
			nybble = (*encoded)>>4;
			new_sample = MS_ADPCM_nibble(state[0],nybble,coeff[0]);
			decoded[0] = new_sample&0xFF;
			new_sample >>= 8;
			decoded[1] = new_sample&0xFF;
			decoded += 2;

			nybble = (*encoded)&0x0F;
			new_sample = MS_ADPCM_nibble(state[1],nybble,coeff[1]);
			decoded[0] = new_sample&0xFF;
			new_sample >>= 8;
			decoded[1] = new_sample&0xFF;
			decoded += 2;

			++encoded;
			samplesleft -= 2;
		}
		encoded_len -= MS_ADPCM_state.wavefmt.blockalign;
	}
	free(freeable);
	return(0);
}

struct IMA_ADPCM_decodestate {
	Sint32 sample;
	Sint8 index;
};
static struct IMA_ADPCM_decoder {
	WaveFMT wavefmt;
	Uint16 wSamplesPerBlock;
	/* * * */
	struct IMA_ADPCM_decodestate state[2];
} IMA_ADPCM_state;

static int InitIMA_ADPCM(WaveFMT *format)
{
	Uint8 *rogue_feel;
	Uint16 extra_info;

	/* Set the rogue pointer to the IMA_ADPCM specific data */
	IMA_ADPCM_state.wavefmt.encoding = SDL_SwapLE16(format->encoding);
	IMA_ADPCM_state.wavefmt.channels = SDL_SwapLE16(format->channels);
	IMA_ADPCM_state.wavefmt.frequency = SDL_SwapLE32(format->frequency);
	IMA_ADPCM_state.wavefmt.byterate = SDL_SwapLE32(format->byterate);
	IMA_ADPCM_state.wavefmt.blockalign = SDL_SwapLE16(format->blockalign);
	IMA_ADPCM_state.wavefmt.bitspersample =
					 SDL_SwapLE16(format->bitspersample);
	rogue_feel = (Uint8 *)format+sizeof(*format);
	if ( sizeof(*format) == 16 ) {
		extra_info = ((rogue_feel[1]<<8)|rogue_feel[0]);
		rogue_feel += sizeof(Uint16);
	}
	IMA_ADPCM_state.wSamplesPerBlock = ((rogue_feel[1]<<8)|rogue_feel[0]);
	return(0);
}

static Sint32 IMA_ADPCM_nibble(struct IMA_ADPCM_decodestate *state,Uint8 nybble)
{
	const Sint32 max_audioval = ((1<<(16-1))-1);
	const Sint32 min_audioval = -(1<<(16-1));
	const int index_table[16] = {
		-1, -1, -1, -1,
		 2,  4,  6,  8,
		-1, -1, -1, -1,
		 2,  4,  6,  8
	};
	const Sint32 step_table[89] = {
		7, 8, 9, 10, 11, 12, 13, 14, 16, 17, 19, 21, 23, 25, 28, 31,
		34, 37, 41, 45, 50, 55, 60, 66, 73, 80, 88, 97, 107, 118, 130,
		143, 157, 173, 190, 209, 230, 253, 279, 307, 337, 371, 408,
		449, 494, 544, 598, 658, 724, 796, 876, 963, 1060, 1166, 1282,
		1411, 1552, 1707, 1878, 2066, 2272, 2499, 2749, 3024, 3327,
		3660, 4026, 4428, 4871, 5358, 5894, 6484, 7132, 7845, 8630,
		9493, 10442, 11487, 12635, 13899, 15289, 16818, 18500, 20350,
		22385, 24623, 27086, 29794, 32767
	};
	Sint32 delta, step;

	/* Compute difference and new sample value */
	step = step_table[state->index];
	delta = step >> 3;
	if ( nybble & 0x04 ) delta += step;
	if ( nybble & 0x02 ) delta += (step >> 1);
	if ( nybble & 0x01 ) delta += (step >> 2);
	if ( nybble & 0x08 ) delta = -delta;
	state->sample += delta;

	/* Update index value */
	state->index += index_table[nybble];
	if ( state->index > 88 ) {
		state->index = 88;
	} else
	if ( state->index < 0 ) {
		state->index = 0;
	}

	/* Clamp output sample */
	if ( state->sample > max_audioval ) {
		state->sample = max_audioval;
	} else
	if ( state->sample < min_audioval ) {
		state->sample = min_audioval;
	}
	return(state->sample);
}

/* Fill the decode buffer with a channel block of data (8 samples) */
static void Fill_IMA_ADPCM_block(Uint8 *decoded, Uint8 *encoded,
	int channel, int numchannels, struct IMA_ADPCM_decodestate *state)
{
	int i;
	Sint8 nybble;
	Sint32 new_sample;

	decoded += (channel * 2);
	for ( i=0; i<4; ++i ) {
		nybble = (*encoded)&0x0F;
		new_sample = IMA_ADPCM_nibble(state, nybble);
		decoded[0] = new_sample&0xFF;
		new_sample >>= 8;
		decoded[1] = new_sample&0xFF;
		decoded += 2 * numchannels;

		nybble = (*encoded)>>4;
		new_sample = IMA_ADPCM_nibble(state, nybble);
		decoded[0] = new_sample&0xFF;
		new_sample >>= 8;
		decoded[1] = new_sample&0xFF;
		decoded += 2 * numchannels;

		++encoded;
	}
}

static int IMA_ADPCM_decode(Uint8 **audio_buf, Uint32 *audio_len)
{
	struct IMA_ADPCM_decodestate *state;
	Uint8 *freeable, *encoded, *decoded;
	Sint32 encoded_len, samplesleft;
	int c, channels;

	/* Check to make sure we have enough variables in the state array */
	channels = IMA_ADPCM_state.wavefmt.channels;
	if ( channels > NELEMS(IMA_ADPCM_state.state) ) {
		SDL_SetError("IMA ADPCM decoder can only handle %d channels",
						NELEMS(IMA_ADPCM_state.state));
		return(-1);
	}
	state = IMA_ADPCM_state.state;

	/* Allocate the proper sized output buffer */
	encoded_len = *audio_len;
	encoded = *audio_buf;
	freeable = *audio_buf;
	*audio_len = (encoded_len/IMA_ADPCM_state.wavefmt.blockalign) * 
				IMA_ADPCM_state.wSamplesPerBlock*
				IMA_ADPCM_state.wavefmt.channels*sizeof(Sint16);
	*audio_buf = (Uint8 *)malloc(*audio_len);
	if ( *audio_buf == NULL ) {
		SDL_Error(SDL_ENOMEM);
		return(-1);
	}
	decoded = *audio_buf;

	/* Get ready... Go! */
	while ( encoded_len >= IMA_ADPCM_state.wavefmt.blockalign ) {
		/* Grab the initial information for this block */
		for ( c=0; c<channels; ++c ) {
			/* Fill the state information for this block */
			state[c].sample = ((encoded[1]<<8)|encoded[0]);
			encoded += 2;
			if ( state[c].sample & 0x8000 ) {
				state[c].sample -= 0x10000;
			}
			state[c].index = *encoded++;
			/* Reserved byte in buffer header, should be 0 */
			if ( *encoded++ != 0 ) {
				/* Uh oh, corrupt data?  Buggy code? */;
			}

			/* Store the initial sample we start with */
			decoded[0] = state[c].sample&0xFF;
			decoded[1] = state[c].sample>>8;
			decoded += 2;
		}

		/* Decode and store the other samples in this block */
		samplesleft = (IMA_ADPCM_state.wSamplesPerBlock-1)*channels;
		while ( samplesleft > 0 ) {
			for ( c=0; c<channels; ++c ) {
				Fill_IMA_ADPCM_block(decoded, encoded,
						c, channels, &state[c]);
				encoded += 4;
				samplesleft -= 8;
			}
			decoded += (channels * 8 * 2);
		}
		encoded_len -= IMA_ADPCM_state.wavefmt.blockalign;
	}
	free(freeable);
	return(0);
}

SDL_AudioSpec * SDL_LoadWAV_RW (SDL_RWops *src, int freesrc,
		SDL_AudioSpec *spec, Uint8 **audio_buf, Uint32 *audio_len)
{
	int was_error;
	Chunk chunk;
	int lenread;
	int MS_ADPCM_encoded, IMA_ADPCM_encoded;
	int samplesize;

	/* WAV magic header */
	Uint32 RIFFchunk;
	Uint32 wavelen;
	Uint32 WAVEmagic;

	/* FMT chunk */
	WaveFMT *format = NULL;

	/* Make sure we are passed a valid data source */
	was_error = 0;
	if ( src == NULL ) {
		was_error = 1;
		goto done;
	}
		
	/* Check the magic header */
	RIFFchunk	= SDL_ReadLE32(src);
	wavelen		= SDL_ReadLE32(src);
	if ( wavelen == WAVE ) { /* The RIFFchunk has already been read */
		WAVEmagic = wavelen;
		wavelen   = RIFFchunk;
		RIFFchunk = RIFF;
	} else {
		WAVEmagic = SDL_ReadLE32(src);
	}
	if ( (RIFFchunk != RIFF) || (WAVEmagic != WAVE) ) {
		SDL_SetError("Unrecognized file type (not WAVE)");
		was_error = 1;
		goto done;
	}

	/* Read the audio data format chunk */
	chunk.data = NULL;
	do {
		if ( chunk.data != NULL ) {
			free(chunk.data);
		}
		lenread = ReadChunk(src, &chunk);
		if ( lenread < 0 ) {
			was_error = 1;
			goto done;
		}
	} while ( (chunk.magic == FACT) || (chunk.magic == LIST) );

	/* Decode the audio data format */
	format = (WaveFMT *)chunk.data;
	if ( chunk.magic != FMT ) {
		SDL_SetError("Complex WAVE files not supported");
		was_error = 1;
		goto done;
	}
	MS_ADPCM_encoded = IMA_ADPCM_encoded = 0;
	switch (SDL_SwapLE16(format->encoding)) {
		case PCM_CODE:
			/* We can understand this */
			break;
		case MS_ADPCM_CODE:
			/* Try to understand this */
			if ( InitMS_ADPCM(format) < 0 ) {
				was_error = 1;
				goto done;
			}
			MS_ADPCM_encoded = 1;
			break;
		case IMA_ADPCM_CODE:
			/* Try to understand this */
			if ( InitIMA_ADPCM(format) < 0 ) {
				was_error = 1;
				goto done;
			}
			IMA_ADPCM_encoded = 1;
			break;
		default:
			SDL_SetError("Unknown WAVE data format: 0x%.4x",
					SDL_SwapLE16(format->encoding));
			was_error = 1;
			goto done;
	}
	memset(spec, 0, (sizeof *spec));
	spec->freq = SDL_SwapLE32(format->frequency);
	switch (SDL_SwapLE16(format->bitspersample)) {
		case 4:
			if ( MS_ADPCM_encoded || IMA_ADPCM_encoded ) {
				spec->format = AUDIO_S16;
			} else {
				was_error = 1;
			}
			break;
		case 8:
			spec->format = AUDIO_U8;
			break;
		case 16:
			spec->format = AUDIO_S16;
			break;
		default:
			was_error = 1;
			break;
	}
	if ( was_error ) {
		SDL_SetError("Unknown %d-bit PCM data format",
			SDL_SwapLE16(format->bitspersample));
		goto done;
	}
	spec->channels = (Uint8)SDL_SwapLE16(format->channels);
	spec->samples = 4096;		/* Good default buffer size */

	/* Read the audio data chunk */
	*audio_buf = NULL;
	do {
		if ( *audio_buf != NULL ) {
			free(*audio_buf);
		}
		lenread = ReadChunk(src, &chunk);
		if ( lenread < 0 ) {
			was_error = 1;
			goto done;
		}
		*audio_len = lenread;
		*audio_buf = chunk.data;
	} while ( chunk.magic != DATA );

	if ( MS_ADPCM_encoded ) {
		if ( MS_ADPCM_decode(audio_buf, audio_len) < 0 ) {
			was_error = 1;
			goto done;
		}
	}
	if ( IMA_ADPCM_encoded ) {
		if ( IMA_ADPCM_decode(audio_buf, audio_len) < 0 ) {
			was_error = 1;
			goto done;
		}
	}

	/* Don't return a buffer that isn't a multiple of samplesize */
	samplesize = ((spec->format & 0xFF)/8)*spec->channels;
	*audio_len &= ~(samplesize-1);

done:
	if ( format != NULL ) {
		free(format);
	}
	if ( freesrc && src ) {
		SDL_RWclose(src);
	}
	if ( was_error ) {
		spec = NULL;
	}
	return(spec);
}

/* Since the WAV memory is allocated in the shared library, it must also
   be freed here.  (Necessary under Win32, VC++)
 */
void SDL_FreeWAV(Uint8 *audio_buf)
{
	if ( audio_buf != NULL ) {
		free(audio_buf);
	}
}

static int ReadChunk(SDL_RWops *src, Chunk *chunk)
{
	chunk->magic	= SDL_ReadLE32(src);
	chunk->length	= SDL_ReadLE32(src);
	chunk->data = (Uint8 *)malloc(chunk->length);
	if ( chunk->data == NULL ) {
		SDL_Error(SDL_ENOMEM);
		return(-1);
	}
	if ( SDL_RWread(src, chunk->data, chunk->length, 1) != 1 ) {
		SDL_Error(SDL_EFREAD);
		free(chunk->data);
		return(-1);
	}
	return(chunk->length);
}

#endif /* ENABLE_FILE */