avfilter/af_sofalizer: remove reorder hack

Signed-off-by: Paul B Mahol <onemda@gmail.com>
This commit is contained in:
Paul B Mahol 2015-12-13 12:56:30 +01:00
parent 49d97d9bca
commit 780c4223a5

View File

@ -53,11 +53,11 @@ typedef struct SOFAlizerContext {
char *filename; /* name of SOFA file */
NCSofa sofa; /* contains data of the SOFA file */
const int8_t *reorder; /* reorder in SOFA channel order */
int sample_rate; /* sample rate from SOFA file */
float *speaker_azim; /* azimuth of the virtual loudspeakers */
float *speaker_elev; /* elevation of the virtual loudspeakers */
float gain_lfe; /* gain applied to LFE channel */
int lfe_channel; /* LFE channel position in channel layout */
int n_conv; /* number of channels to convolute */
@ -81,8 +81,6 @@ typedef struct SOFAlizerContext {
float elevation; /* elevation of virtual loudspeakers (in deg.) */
float radius; /* distance virtual loudspeakers to listener (in metres) */
int lfe; /* whether or not the LFE channel is used */
AVFloatDSPContext *fdsp;
} SOFAlizerContext;
@ -353,54 +351,31 @@ error:
return ret;
}
static const int8_t reorder[18][9] = {
{ 0, -1, -1, -1, -1, -1, -1, -1, -1 },
{ 0, 1, -1, -1, -1, -1, -1, -1, -1 },
{ 0, 1, 2, -1, -1, -1, -1, -1, -1 },
{ 0, 1, 2, -1, -1, -1, -1, -1, -1 },
{ 0, 1, 2, 3, -1, -1, -1, -1, -1 },
{ 0, 1, 2, 3, -1, -1, -1, -1, -1 },
{ 0, 1, 2, 3, -1, -1, -1, -1, -1 },
{ 0, 1, 3, 4, 2, -1, -1, -1, -1 },
{ 0, 1, 3, 4, 2, -1, -1, -1, -1 },
{ 0, 1, 4, 5, 2, 3, -1, -1, -1 },
{ 0, 1, 4, 5, 2, 3, -1, -1, -1 },
{ 0, 1, 5, 6, 4, 2, 3, -1, -1 },
{ 0, 1, 5, 6, 3, 4, 2, -1, -1 },
{ 0, 1, 6, 7, 4, 5, 2, 3, -1 },
{ 0, 1, 2, 3, 4, 5, 6, 7, 8 },
{ 0, 1, 2, 3, 4, 5, 6, 7, -1 },
{ 0, 1, 3, 4, 2, 5, -1, -1, -1 },
{ 0, 1, 4, 5, 2, 6, 3, -1, -1 },
};
static int get_speaker_pos(AVFilterContext *ctx,
float *speaker_azim, float *speaker_elev)
{
struct SOFAlizerContext *s = ctx->priv;
uint64_t channels_layout = ctx->inputs[0]->channel_layout;
float azim[9] = { 0 };
float elev[9] = { 0 };
int nb_input_channels = ctx->inputs[0]->channels; /* get no. input channels */
int n_conv = nb_input_channels;
float azim[10] = { 0 };
float elev[10] = { 0 };
int n_conv = ctx->inputs[0]->channels; /* get no. input channels */
if (channels_layout & AV_CH_LOW_FREQUENCY) { /* if LFE is used */
/* decrease number of channels to be convolved: */
n_conv = nb_input_channels - 1;
}
s->lfe_channel = -1;
/* set speaker positions according to input channel configuration: */
switch (channels_layout) {
case AV_CH_LAYOUT_MONO:
azim[0] = 0;
break;
case AV_CH_LAYOUT_STEREO:
case AV_CH_LAYOUT_2POINT1:
s->lfe_channel = 2;
case AV_CH_LAYOUT_STEREO:
azim[0] = 30;
azim[1] = 330;
break;
case AV_CH_LAYOUT_SURROUND:
case AV_CH_LAYOUT_3POINT1:
s->lfe_channel = 3;
case AV_CH_LAYOUT_SURROUND:
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
@ -422,63 +397,101 @@ static int get_speaker_pos(AVFilterContext *ctx,
azim[2] = 120;
azim[3] = 240;
break;
case AV_CH_LAYOUT_4POINT0:
case AV_CH_LAYOUT_4POINT1:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[4] = 180;
break;
case AV_CH_LAYOUT_4POINT0:
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[3] = 180;
break;
case AV_CH_LAYOUT_5POINT0:
case AV_CH_LAYOUT_5POINT1:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 90;
azim[3] = 270;
azim[4] = 0;
azim[2] = 0;
azim[4] = 90;
azim[5] = 270;
break;
case AV_CH_LAYOUT_5POINT0:
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[3] = 90;
azim[4] = 270;
break;
case AV_CH_LAYOUT_5POINT1_BACK:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[4] = 120;
azim[5] = 240;
break;
case AV_CH_LAYOUT_5POINT0_BACK:
case AV_CH_LAYOUT_5POINT1_BACK:
azim[0] = 30;
azim[1] = 330;
azim[2] = 120;
azim[3] = 240;
azim[4] = 0;
azim[2] = 0;
azim[3] = 120;
azim[4] = 240;
break;
case AV_CH_LAYOUT_6POINT1:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[4] = 180;
azim[5] = 90;
azim[6] = 270;
break;
case AV_CH_LAYOUT_6POINT0:
case AV_CH_LAYOUT_6POINT1:
azim[0] = 30;
azim[1] = 330;
azim[2] = 90;
azim[3] = 270;
azim[4] = 0;
azim[5] = 180;
azim[2] = 0;
azim[3] = 180;
azim[4] = 90;
azim[5] = 270;
break;
case AV_CH_LAYOUT_6POINT1_BACK:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 120;
azim[3] = 240;
azim[4] = 0;
azim[4] = 180;
azim[2] = 0;
azim[4] = 120;
azim[5] = 240;
azim[6] = 180;
break;
case AV_CH_LAYOUT_HEXAGONAL:
azim[0] = 30;
azim[1] = 330;
azim[2] = 120;
azim[3] = 240;
azim[4] = 0;
azim[2] = 0;
azim[3] = 120;
azim[4] = 240;
azim[5] = 180;
break;
case AV_CH_LAYOUT_7POINT0:
case AV_CH_LAYOUT_7POINT1:
s->lfe_channel = 3;
azim[0] = 30;
azim[1] = 330;
azim[2] = 90;
azim[3] = 270;
azim[2] = 0;
azim[4] = 150;
azim[5] = 210;
azim[6] = 0;
azim[6] = 90;
azim[7] = 270;
break;
case AV_CH_LAYOUT_7POINT0:
azim[0] = 30;
azim[1] = 330;
azim[2] = 0;
azim[3] = 150;
azim[4] = 210;
azim[5] = 90;
azim[6] = 270;
break;
case AV_CH_LAYOUT_OCTAGONAL:
azim[0] = 30;
@ -494,66 +507,6 @@ static int get_speaker_pos(AVFilterContext *ctx,
return -1;
}
switch (channels_layout) {
case AV_CH_LAYOUT_MONO:
s->reorder = reorder[0];
break;
case AV_CH_LAYOUT_STEREO:
s->reorder = reorder[1];
break;
case AV_CH_LAYOUT_2_1:
case AV_CH_LAYOUT_2POINT1:
s->reorder = reorder[2];
break;
case AV_CH_LAYOUT_SURROUND:
s->reorder = reorder[3];
break;
case AV_CH_LAYOUT_3POINT1:
case AV_CH_LAYOUT_2_2:
s->reorder = reorder[4];
break;
case AV_CH_LAYOUT_QUAD:
s->reorder = reorder[5];
break;
case AV_CH_LAYOUT_4POINT0:
s->reorder = reorder[6];
break;
case AV_CH_LAYOUT_4POINT1:
s->reorder = reorder[7];
break;
case AV_CH_LAYOUT_5POINT0:
case AV_CH_LAYOUT_5POINT0_BACK:
s->reorder = reorder[8];
break;
case AV_CH_LAYOUT_5POINT1:
case AV_CH_LAYOUT_5POINT1_BACK:
s->reorder = reorder[9];
break;
case AV_CH_LAYOUT_6POINT0:
s->reorder = reorder[10];
break;
case AV_CH_LAYOUT_HEXAGONAL:
s->reorder = reorder[16];
break;
case AV_CH_LAYOUT_6POINT1:
s->reorder = reorder[11];
break;
case AV_CH_LAYOUT_6POINT1_BACK:
s->reorder = reorder[17];
break;
case AV_CH_LAYOUT_7POINT0:
s->reorder = reorder[12];
break;
case AV_CH_LAYOUT_7POINT1:
s->reorder = reorder[13];
break;
case AV_CH_LAYOUT_OCTAGONAL:
s->reorder = reorder[15];
break;
default:
return -1;
}
memcpy(speaker_azim, azim, n_conv * sizeof(float));
memcpy(speaker_elev, elev, n_conv * sizeof(float));
@ -654,7 +607,7 @@ static int sofalizer_convolute(AVFilterContext *ctx, void *arg, int jobnr, int n
const int n_samples = s->sofa.n_samples; /* length of one IR */
const float *src = (const float *)in->data[0]; /* get pointer to audio input buffer */
float *dst = (float *)out->data[0]; /* get pointer to audio output buffer */
const int in_channels = in->channels; /* number of input channels */
const int in_channels = s->n_conv; /* number of input channels */
/* ring buffer length is: longest IR plus max. delay -> next power of 2 */
const int buffer_length = s->buffer_length;
/* -1 for AND instead of MODULO (applied to powers of 2): */
@ -676,13 +629,21 @@ static int sofalizer_convolute(AVFilterContext *ctx, void *arg, int jobnr, int n
*dst = 0;
for (l = 0; l < in_channels; l++) {
/* write current input sample to ringbuffer (for each channel) */
*(buffer[l] + wr) = src[s->reorder[l]];
*(buffer[l] + wr) = src[l];
}
/* loop goes through all channels to be convolved (excl. LFE): */
for (l = 0; l < s->n_conv; l++) {
/* loop goes through all channels to be convolved */
for (l = 0; l < in_channels; l++) {
const float *const bptr = buffer[l];
if (l == s->lfe_channel) {
/* LFE is an input channel but requires no convolution */
/* apply gain to LFE signal and add to output buffer */
*dst += *(buffer[s->lfe_channel] + wr) * s->gain_lfe;
temp_ir += n_samples;
continue;
}
/* current read position in ringbuffer: input sample write position
* - delay for l-th ch. + diff. betw. IR length and buffer length
* (mod buffer length) */
@ -700,11 +661,6 @@ static int sofalizer_convolute(AVFilterContext *ctx, void *arg, int jobnr, int n
temp_ir += n_samples;
}
if (s->lfe) { /* LFE */
/* apply gain to LFE signal and add to output buffer */
*dst += *(buffer[s->n_conv] + wr) * s->gain_lfe;
}
/* clippings counter */
if (fabs(*dst) > 1)
*n_clippings += 1;
@ -817,7 +773,7 @@ static int load_data(AVFilterContext *ctx, int azim, int elev, float radius)
{
struct SOFAlizerContext *s = ctx->priv;
const int n_samples = s->sofa.n_samples;
int n_conv = s->n_conv; /* no. channels to convolve (excl. LFE) */
int n_conv = s->n_conv; /* no. channels to convolve */
int delay_l[10]; /* broadband delay for each IR */
int delay_r[10];
int nb_input_channels = ctx->inputs[0]->channels; /* no. input channels */
@ -825,7 +781,7 @@ static int load_data(AVFilterContext *ctx, int azim, int elev, float radius)
float *data_ir_l = NULL;
float *data_ir_r = NULL;
int offset = 0; /* used for faster pointer arithmetics in for-loop */
int m[s->n_conv]; /* measurement index m of IR closest to required source positions */
int m[10]; /* measurement index m of IR closest to required source positions */
int i, j, azim_orig = azim, elev_orig = elev;
if (!s->sofa.ncid) { /* if an invalid SOFA file has been selected */
@ -935,9 +891,7 @@ static int config_input(AVFilterLink *inlink)
/* gain -3 dB per channel, -6 dB to get LFE on a similar level */
s->gain_lfe = expf((s->gain - 3 * inlink->channels - 6) / 20 * M_LN10);
s->lfe = !!(inlink->channel_layout & AV_CH_LOW_FREQUENCY);
/* LFE is an input channel but requires no convolution */
s->n_conv = nb_input_channels - s->lfe;
s->n_conv = nb_input_channels;
/* get size of ringbuffer (longest IR plus max. delay) */
/* then choose next power of 2 for performance optimization */
@ -953,7 +907,7 @@ static int config_input(AVFilterLink *inlink)
s->buffer_length = exp2(32 - clz((uint32_t)n_max));
/* Allocate memory for the impulse responses, delays and the ringbuffers */
/* size: (longest IR) * (number of channels to convolute), without LFE */
/* size: (longest IR) * (number of channels to convolute) */
s->data_ir[0] = av_malloc_array(n_max_ir, sizeof(float) * s->n_conv);
s->data_ir[1] = av_malloc_array(n_max_ir, sizeof(float) * s->n_conv);
/* length: number of channels to convolute */