annotate libmpcodecs/ae_twolame.c @ 17588:79081ba52e00

Move the v{Y,C}CoeffsBank vectors into the SwsContext, filling them in just once when the scaler is initialized, instead of building them and freeing them over and over. This gives massive performance improvements. patch by Alan Curry, pacman*at*TheWorld*dot*com

#include <stdio.h>
#include <stdlib.h>
#include <inttypes.h>
#include <unistd.h>
#include <string.h>
#include <sys/types.h>
#include "m_option.h"
#include "mp_msg.h"
#include "aviheader.h"
#include "libaf/af_format.h"
#include "ms_hdr.h"
#include "muxer.h"
#include "ae_twolame.h"
#include "libmpdemux/mp3_hdr.h"


static int 
    param_bitrate = 192,
    param_psy = 3,
    param_maxvbr = 0,
    param_errprot = 0,
    param_debug = 0;
    
static float param_vbr = 0;
static char *param_mode = "stereo";
    
m_option_t twolameopts_conf[] = {
	{"br", &param_bitrate, CONF_TYPE_INT, 0, 0, 0, NULL},
	{"mode", &param_mode, CONF_TYPE_STRING, 0, 0, 0, NULL},
	{"psy", &param_psy, CONF_TYPE_INT, CONF_RANGE, -1, 4, NULL},
	{"vbr", &param_vbr, CONF_TYPE_FLOAT, CONF_RANGE, -50, 50, NULL},
	{"maxvbr", &param_maxvbr, CONF_TYPE_INT, 0, 0, 0, NULL},
	{"errprot", &param_errprot, CONF_TYPE_INT, CONF_RANGE, 0, 1, NULL},
	{"debug", &param_debug, CONF_TYPE_INT, CONF_RANGE, 0, 100000000, NULL},
	{NULL, NULL, 0, 0, 0, 0, NULL}
};


static int bind_twolame(audio_encoder_t *encoder, muxer_stream_t *mux_a)
{
	mpae_twolame_ctx *ctx = (mpae_twolame_ctx *) encoder->priv;
	
	mux_a->wf = malloc(sizeof(WAVEFORMATEX)+256);
	mux_a->wf->wFormatTag = 0x50;
	mux_a->wf->nChannels = encoder->params.channels;
	mux_a->wf->nSamplesPerSec = encoder->params.sample_rate;
	mux_a->wf->nAvgBytesPerSec = encoder->params.bitrate / 8;
	
	if(ctx->vbr || ((mux_a->wf->nAvgBytesPerSec * encoder->params.samples_per_frame) % mux_a->wf->nSamplesPerSec))
	{
		mux_a->h.dwScale = encoder->params.samples_per_frame;
		mux_a->h.dwRate = encoder->params.sample_rate;
		mux_a->h.dwSampleSize = 0; // Blocksize not constant
	}
	else
	{
		mux_a->h.dwScale = (mux_a->wf->nAvgBytesPerSec * encoder->params.samples_per_frame)/ mux_a->wf->nSamplesPerSec; /* for cbr */
		mux_a->h.dwRate = mux_a->wf->nAvgBytesPerSec;
		mux_a->h.dwSampleSize = mux_a->h.dwScale;
	}
	mux_a->wf->nBlockAlign = mux_a->h.dwScale;
	mux_a->h.dwSuggestedBufferSize = (encoder->params.audio_preload*mux_a->wf->nAvgBytesPerSec)/1000;
	mux_a->h.dwSuggestedBufferSize -= mux_a->h.dwSuggestedBufferSize % mux_a->wf->nBlockAlign;
	
	mux_a->wf->cbSize = 0; //12;
	mux_a->wf->wBitsPerSample = 0; /* does not apply */
	((MPEGLAYER3WAVEFORMAT *) (mux_a->wf))->wID = 1;
	((MPEGLAYER3WAVEFORMAT *) (mux_a->wf))->fdwFlags = 2;
	((MPEGLAYER3WAVEFORMAT *) (mux_a->wf))->nBlockSize = mux_a->wf->nBlockAlign;
	((MPEGLAYER3WAVEFORMAT *) (mux_a->wf))->nFramesPerBlock = 1;
	((MPEGLAYER3WAVEFORMAT *) (mux_a->wf))->nCodecDelay = 0;
	
	// Fix allocation    
	mux_a->wf = realloc(mux_a->wf, sizeof(WAVEFORMATEX)+mux_a->wf->cbSize);
	
	encoder->input_format = AF_FORMAT_S16_NE;
	encoder->min_buffer_size = mux_a->h.dwSuggestedBufferSize;
	encoder->max_buffer_size = mux_a->h.dwSuggestedBufferSize*2;

	return 1;
}

static int encode_twolame(audio_encoder_t *encoder, uint8_t *dest, void *src, int len, int max_size)
{
	mpae_twolame_ctx *ctx = (mpae_twolame_ctx *)encoder->priv;
	int ret_size = 0, r2;
	
	len /= (2*encoder->params.channels);
	ret_size = twolame_encode_buffer_interleaved(ctx->twolame_ctx, src, len, dest, max_size);
	r2 = mp_decode_mp3_header(dest);	
	mp_msg(MSGT_MENCODER, MSGL_DBG2, "\nSIZE: %d, max: %d, r2: %d\n", ret_size, max_size, r2);
	if(r2 > 0)
		ret_size = r2;
	return ret_size;
}

int close_twolame(audio_encoder_t *encoder)
{
	free(encoder->priv);
	return 1;
}

static int get_frame_size(audio_encoder_t *encoder)
{
	int sz;
	if(encoder->stream->buffer_len < 4)
		return 0;
	sz = mp_decode_mp3_header(encoder->stream->buffer);
	if(sz <= 0)
		return 0;
	return sz;
}


int mpae_init_twolame(audio_encoder_t *encoder)
{
	int mode;
	mpae_twolame_ctx *ctx = NULL;
	
	if(encoder->params.channels == 1)
	{
		mp_msg(MSGT_MENCODER, MSGL_INFO, "ae_twolame, 1 audio channel, forcing mono mode\n");
		mode = TWOLAME_MONO;
	}
	else if(encoder->params.channels == 2)
	{
		if(! strcasecmp(param_mode, "dual"))
			mode = TWOLAME_DUAL_CHANNEL;
		else if(! strcasecmp(param_mode, "jstereo"))
			mode = TWOLAME_JOINT_STEREO;
		else if(! strcasecmp(param_mode, "stereo"))
			mode = TWOLAME_STEREO;
		else
		{
			mp_msg(MSGT_MENCODER, MSGL_ERR, "ae_twolame, unknown mode %s, exiting\n", param_mode);
		}
	}
	else
		mp_msg(MSGT_MENCODER, MSGL_ERR, "ae_twolame, Twolame can't encode > 2 channels, exiting\n");
	
	ctx = (mpae_twolame_ctx *) calloc(1, sizeof(mpae_twolame_ctx));
	if(ctx == NULL)
	{
		mp_msg(MSGT_MENCODER, MSGL_ERR, "ae_twolame, couldn't alloc a %d bytes context, exiting\n", sizeof(mpae_twolame_ctx));
		return 0;
	}
	
	ctx->twolame_ctx = twolame_init();
	if(ctx->twolame_ctx == NULL)
	{
		mp_msg(MSGT_MENCODER, MSGL_ERR, "ae_twolame, couldn't initial parameters from libtwolame, exiting\n");
		free(ctx);
		return 0;
	}
	ctx->vbr = 0;

	if(twolame_set_num_channels(ctx->twolame_ctx, encoder->params.channels) != 0)
		return 0;
	if(twolame_set_mode(ctx->twolame_ctx, mode) != 0)
		return 0;
		
	if(twolame_set_in_samplerate(ctx->twolame_ctx, encoder->params.sample_rate) != 0)
		return 0;
		
	if(twolame_set_out_samplerate(ctx->twolame_ctx, encoder->params.sample_rate) != 0)
		return 0;
	
	if(encoder->params.sample_rate < 32000)
		twolame_set_version(ctx->twolame_ctx, TWOLAME_MPEG2);
	else
		twolame_set_version(ctx->twolame_ctx, TWOLAME_MPEG1);
	
	if(twolame_set_psymodel(ctx->twolame_ctx, param_psy) != 0)
		return 0;
	
	if(twolame_set_bitrate(ctx->twolame_ctx, param_bitrate) != 0)
		return 0;
	
	if(param_errprot)
		if(twolame_set_error_protection(ctx->twolame_ctx, TRUE) != 0)
			return 0;
	
	if(param_vbr != 0)
	{
		if(twolame_set_VBR(ctx->twolame_ctx, TRUE) != 0)
			return 0;
		if(twolame_set_VBR_q(ctx->twolame_ctx, param_vbr) != 0)
			return 0;
		if(twolame_set_padding(ctx->twolame_ctx, FALSE) != 0)
			return 0;
		if(param_maxvbr)
		{
			if(twolame_set_VBR_max_bitrate_kbps(ctx->twolame_ctx, param_maxvbr) != 0)
				return 0;
		}
		ctx->vbr = 1;
	}
	
	if(twolame_set_verbosity(ctx->twolame_ctx, param_debug) != 0)
		return 0;
	
	if(twolame_init_params(ctx->twolame_ctx) != 0)
		return 0;
	
	encoder->params.bitrate = param_bitrate * 1000;
	encoder->params.samples_per_frame = 1152;
	encoder->priv = ctx;
	encoder->decode_buffer_size = 1152 * 2 * encoder->params.channels;
	
	encoder->bind = bind_twolame;
	encoder->get_frame_size = get_frame_size;
	encoder->encode = encode_twolame;
	encoder->close = close_twolame;
	
	return 1;
}