ffmpeg/libswresample/rematrix.c

/*
 * Copyright (C) 2011-2012 Michael Niedermayer (michaelni@gmx.at)
 *
 * This file is part of libswresample
 *
 * libswresample is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * libswresample 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
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with libswresample; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "swresample_internal.h"
#include "libavutil/avassert.h"
#include "libavutil/channel_layout.h"
#include "libavutil/mem.h"

#define TEMPLATE_REMATRIX_FLT
#include "rematrix_template.c"
#undef TEMPLATE_REMATRIX_FLT

#define TEMPLATE_REMATRIX_DBL
#include "rematrix_template.c"
#undef TEMPLATE_REMATRIX_DBL

#define TEMPLATE_REMATRIX_S16
#include "rematrix_template.c"
#define TEMPLATE_CLIP
#include "rematrix_template.c"
#undef TEMPLATE_CLIP
#undef TEMPLATE_REMATRIX_S16

#define TEMPLATE_REMATRIX_S32
#include "rematrix_template.c"
#undef TEMPLATE_REMATRIX_S32

#define FRONT_LEFT             0
#define FRONT_RIGHT            1
#define FRONT_CENTER           2
#define LOW_FREQUENCY          3
#define BACK_LEFT              4
#define BACK_RIGHT             5
#define FRONT_LEFT_OF_CENTER   6
#define FRONT_RIGHT_OF_CENTER  7
#define BACK_CENTER            8
#define SIDE_LEFT              9
#define SIDE_RIGHT             10
#define TOP_CENTER             11
#define TOP_FRONT_LEFT         12
#define TOP_FRONT_CENTER       13
#define TOP_FRONT_RIGHT        14
#define TOP_BACK_LEFT          15
#define TOP_BACK_CENTER        16
#define TOP_BACK_RIGHT         17
#define NUM_NAMED_CHANNELS     18

int swr_set_matrix(struct SwrContext *s, const double *matrix, int stride)
{
    int nb_in, nb_out, in, out;

    if (!s || s->in_convert) // s needs to be allocated but not initialized
        return AVERROR(EINVAL);
    memset(s->matrix, 0, sizeof(s->matrix));
    memset(s->matrix_flt, 0, sizeof(s->matrix_flt));

    nb_in = s->user_in_chlayout.nb_channels;
    nb_out = s->user_out_chlayout.nb_channels;
    for (out = 0; out < nb_out; out++) {
        for (in = 0; in < nb_in; in++)
            s->matrix_flt[out][in] = s->matrix[out][in] = matrix[in];
        matrix += stride;
    }
    s->rematrix_custom = 1;
    return 0;
}

static int even(int64_t layout){
    if(!layout) return 1;
    if(layout&(layout-1)) return 1;
    return 0;
}

static int clean_layout(AVChannelLayout *out, const AVChannelLayout *in, void *s)
{
    int ret = 0;

    if (av_channel_layout_index_from_channel(in, AV_CHAN_FRONT_CENTER) < 0 && in->nb_channels == 1) {
        char buf[128];
        av_channel_layout_describe(in, buf, sizeof(buf));
        av_log(s, AV_LOG_VERBOSE, "Treating %s as mono\n", buf);
        *out = (AVChannelLayout)AV_CHANNEL_LAYOUT_MONO;
    } else
        ret = av_channel_layout_copy(out, in);

    return ret;
}

static int sane_layout(AVChannelLayout *ch_layout) {
    if(ch_layout->nb_channels >= SWR_CH_MAX)
        return 0;
    if(ch_layout->order == AV_CHANNEL_ORDER_CUSTOM)
        for (int i = 0; i < ch_layout->nb_channels; i++) {
            if (ch_layout->u.map[i].id >= 64)
                return 0;
        }
    else if (ch_layout->order != AV_CHANNEL_ORDER_NATIVE)
        return 0;
    if(!av_channel_layout_subset(ch_layout, AV_CH_LAYOUT_SURROUND)) // at least 1 front speaker
        return 0;
    if(!even(av_channel_layout_subset(ch_layout, (AV_CH_FRONT_LEFT | AV_CH_FRONT_RIGHT)))) // no asymetric front
        return 0;
    if(!even(av_channel_layout_subset(ch_layout, (AV_CH_SIDE_LEFT | AV_CH_SIDE_RIGHT))))   // no asymetric side
        return 0;
    if(!even(av_channel_layout_subset(ch_layout, (AV_CH_BACK_LEFT | AV_CH_BACK_RIGHT))))
        return 0;
    if(!even(av_channel_layout_subset(ch_layout, (AV_CH_FRONT_LEFT_OF_CENTER | AV_CH_FRONT_RIGHT_OF_CENTER))))
        return 0;
    if(!even(av_channel_layout_subset(ch_layout, (AV_CH_TOP_FRONT_LEFT | AV_CH_TOP_FRONT_RIGHT))))
        return 0;

    return 1;
}

static void build_matrix(const AVChannelLayout *in_ch_layout, const AVChannelLayout *out_ch_layout,
                         double center_mix_level, double surround_mix_level,
                         double lfe_mix_level, double maxval, double rematrix_volume, double *matrix_param,
                         ptrdiff_t stride, enum AVMatrixEncoding matrix_encoding)
{
    double matrix[NUM_NAMED_CHANNELS][NUM_NAMED_CHANNELS]={{0}};
    uint64_t unaccounted =  av_channel_layout_subset(in_ch_layout,  UINT64_MAX) &
                           ~av_channel_layout_subset(out_ch_layout, UINT64_MAX);
    double maxcoef=0;
    int i, j;

    for(i=0; i<FF_ARRAY_ELEMS(matrix); i++){
        if(   av_channel_layout_index_from_channel(in_ch_layout, i) >= 0
           && av_channel_layout_index_from_channel(out_ch_layout, i) >= 0)
            matrix[i][i]= 1.0;
    }

//FIXME implement dolby surround
//FIXME implement full ac3

    if(unaccounted & AV_CH_FRONT_CENTER){
        if (av_channel_layout_subset(out_ch_layout, AV_CH_LAYOUT_STEREO) == AV_CH_LAYOUT_STEREO) {
            if (av_channel_layout_subset(in_ch_layout, AV_CH_LAYOUT_STEREO)) {
                matrix[ FRONT_LEFT][FRONT_CENTER]+= center_mix_level;
                matrix[FRONT_RIGHT][FRONT_CENTER]+= center_mix_level;
            } else {
                matrix[ FRONT_LEFT][FRONT_CENTER]+= M_SQRT1_2;
                matrix[FRONT_RIGHT][FRONT_CENTER]+= M_SQRT1_2;
            }
        }else
            av_assert0(0);
    }
    if(unaccounted & AV_CH_LAYOUT_STEREO){
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[FRONT_CENTER][ FRONT_LEFT]+= M_SQRT1_2;
            matrix[FRONT_CENTER][FRONT_RIGHT]+= M_SQRT1_2;
            if (av_channel_layout_index_from_channel(in_ch_layout, AV_CHAN_FRONT_CENTER) >= 0)
                matrix[FRONT_CENTER][ FRONT_CENTER] = center_mix_level*sqrt(2);
        }else
            av_assert0(0);
    }

    if(unaccounted & AV_CH_BACK_CENTER){
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_BACK_LEFT) >= 0) {
            matrix[ BACK_LEFT][BACK_CENTER]+= M_SQRT1_2;
            matrix[BACK_RIGHT][BACK_CENTER]+= M_SQRT1_2;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_SIDE_LEFT) >= 0) {
            matrix[ SIDE_LEFT][BACK_CENTER]+= M_SQRT1_2;
            matrix[SIDE_RIGHT][BACK_CENTER]+= M_SQRT1_2;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY ||
                matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
                if (unaccounted & (AV_CH_BACK_LEFT | AV_CH_SIDE_LEFT)) {
                    matrix[FRONT_LEFT ][BACK_CENTER] -= surround_mix_level * M_SQRT1_2;
                    matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
                } else {
                    matrix[FRONT_LEFT ][BACK_CENTER] -= surround_mix_level;
                    matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level;
                }
            } else {
                matrix[ FRONT_LEFT][BACK_CENTER]+= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][BACK_CENTER]+= surround_mix_level * M_SQRT1_2;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[ FRONT_CENTER][BACK_CENTER]+= surround_mix_level * M_SQRT1_2;
        }else
            av_assert0(0);
    }
    if(unaccounted & AV_CH_BACK_LEFT){
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_BACK_CENTER) >= 0) {
            matrix[BACK_CENTER][ BACK_LEFT]+= M_SQRT1_2;
            matrix[BACK_CENTER][BACK_RIGHT]+= M_SQRT1_2;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_SIDE_LEFT) >= 0) {
            if (av_channel_layout_index_from_channel(in_ch_layout, AV_CHAN_SIDE_LEFT) >= 0) {
                matrix[ SIDE_LEFT][ BACK_LEFT]+= M_SQRT1_2;
                matrix[SIDE_RIGHT][BACK_RIGHT]+= M_SQRT1_2;
            }else{
            matrix[ SIDE_LEFT][ BACK_LEFT]+= 1.0;
            matrix[SIDE_RIGHT][BACK_RIGHT]+= 1.0;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
                matrix[FRONT_LEFT ][BACK_LEFT ] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_LEFT ][BACK_RIGHT] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][BACK_LEFT ] += surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level * M_SQRT1_2;
            } else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
                matrix[FRONT_LEFT ][BACK_LEFT ] -= surround_mix_level * SQRT3_2;
                matrix[FRONT_LEFT ][BACK_RIGHT] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][BACK_LEFT ] += surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level * SQRT3_2;
            } else {
                matrix[ FRONT_LEFT][ BACK_LEFT] += surround_mix_level;
                matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[ FRONT_CENTER][BACK_LEFT ]+= surround_mix_level*M_SQRT1_2;
            matrix[ FRONT_CENTER][BACK_RIGHT]+= surround_mix_level*M_SQRT1_2;
        }else
            av_assert0(0);
    }

    if(unaccounted & AV_CH_SIDE_LEFT){
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_BACK_LEFT) >= 0) {
            /* if back channels do not exist in the input, just copy side
               channels to back channels, otherwise mix side into back */
            if (av_channel_layout_index_from_channel(in_ch_layout, AV_CHAN_BACK_LEFT) >= 0) {
                matrix[BACK_LEFT ][SIDE_LEFT ] += M_SQRT1_2;
                matrix[BACK_RIGHT][SIDE_RIGHT] += M_SQRT1_2;
            } else {
                matrix[BACK_LEFT ][SIDE_LEFT ] += 1.0;
                matrix[BACK_RIGHT][SIDE_RIGHT] += 1.0;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_BACK_CENTER) >= 0) {
            matrix[BACK_CENTER][ SIDE_LEFT]+= M_SQRT1_2;
            matrix[BACK_CENTER][SIDE_RIGHT]+= M_SQRT1_2;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
                matrix[FRONT_LEFT ][SIDE_LEFT ] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_LEFT ][SIDE_RIGHT] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][SIDE_LEFT ] += surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level * M_SQRT1_2;
            } else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
                matrix[FRONT_LEFT ][SIDE_LEFT ] -= surround_mix_level * SQRT3_2;
                matrix[FRONT_LEFT ][SIDE_RIGHT] -= surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][SIDE_LEFT ] += surround_mix_level * M_SQRT1_2;
                matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level * SQRT3_2;
            } else {
                matrix[ FRONT_LEFT][ SIDE_LEFT] += surround_mix_level;
                matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[ FRONT_CENTER][SIDE_LEFT ]+= surround_mix_level * M_SQRT1_2;
            matrix[ FRONT_CENTER][SIDE_RIGHT]+= surround_mix_level * M_SQRT1_2;
        }else
            av_assert0(0);
    }

    if(unaccounted & AV_CH_FRONT_LEFT_OF_CENTER){
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            matrix[ FRONT_LEFT][ FRONT_LEFT_OF_CENTER]+= 1.0;
            matrix[FRONT_RIGHT][FRONT_RIGHT_OF_CENTER]+= 1.0;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[ FRONT_CENTER][ FRONT_LEFT_OF_CENTER]+= M_SQRT1_2;
            matrix[ FRONT_CENTER][FRONT_RIGHT_OF_CENTER]+= M_SQRT1_2;
        }else
            av_assert0(0);
    }

    if (unaccounted & AV_CH_TOP_FRONT_LEFT) {
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_TOP_FRONT_CENTER) >= 0) {
            matrix[TOP_FRONT_CENTER][TOP_FRONT_LEFT ] += M_SQRT1_2;
            matrix[TOP_FRONT_CENTER][TOP_FRONT_RIGHT] += M_SQRT1_2;
            if (av_channel_layout_index_from_channel(in_ch_layout, AV_CHAN_TOP_FRONT_CENTER) >= 0)
                matrix[TOP_FRONT_CENTER][TOP_FRONT_CENTER] = center_mix_level * sqrt(2);
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            if (av_channel_layout_index_from_channel(in_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
                matrix[FRONT_LEFT ][TOP_FRONT_LEFT ] += M_SQRT1_2;
                matrix[FRONT_RIGHT][TOP_FRONT_RIGHT] += M_SQRT1_2;
            } else {
                matrix[FRONT_LEFT ][TOP_FRONT_LEFT ] += 1.0;
                matrix[FRONT_RIGHT][TOP_FRONT_RIGHT] += 1.0;
            }
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[FRONT_CENTER][TOP_FRONT_LEFT ] += M_SQRT1_2;
            matrix[FRONT_CENTER][TOP_FRONT_RIGHT] += M_SQRT1_2;
        } else
            av_assert0(0);
    }

    /* mix LFE into front left/right or center */
    if (unaccounted & AV_CH_LOW_FREQUENCY) {
        if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_CENTER) >= 0) {
            matrix[FRONT_CENTER][LOW_FREQUENCY] += lfe_mix_level;
        } else if (av_channel_layout_index_from_channel(out_ch_layout, AV_CHAN_FRONT_LEFT) >= 0) {
            matrix[FRONT_LEFT ][LOW_FREQUENCY] += lfe_mix_level * M_SQRT1_2;
            matrix[FRONT_RIGHT][LOW_FREQUENCY] += lfe_mix_level * M_SQRT1_2;
        } else
            av_assert0(0);
    }


    for (i = 0; i < 64; i++) {
        double sum=0;
        int out_i = av_channel_layout_index_from_channel(out_ch_layout, i);
        if (out_i < 0)
            continue;
        for(j=0; j<64; j++){
            int in_i = av_channel_layout_index_from_channel(in_ch_layout, j);
            if (in_i < 0)
                continue;
            if (i < FF_ARRAY_ELEMS(matrix) && j < FF_ARRAY_ELEMS(matrix[0]))
                matrix_param[stride*out_i + in_i] = matrix[i][j];
            else
                matrix_param[stride*out_i + in_i] = i == j &&
                (   av_channel_layout_index_from_channel(in_ch_layout, i) >= 0
                 && av_channel_layout_index_from_channel(out_ch_layout, i) >= 0);
            sum += fabs(matrix_param[stride*out_i + in_i]);
        }
        maxcoef= FFMAX(maxcoef, sum);
    }
    if(rematrix_volume  < 0)
        maxcoef = -rematrix_volume;

    if(maxcoef > maxval || rematrix_volume  < 0){
        maxcoef /= maxval;
        for(i=0; i<SWR_CH_MAX; i++)
            for(j=0; j<SWR_CH_MAX; j++){
                matrix_param[stride*i + j] /= maxcoef;
            }
    }
}

av_cold int swr_build_matrix2(const AVChannelLayout *in_layout, const AVChannelLayout *out_layout,
                              double center_mix_level, double surround_mix_level,
                              double lfe_mix_level, double maxval,
                              double rematrix_volume, double *matrix_param,
                              ptrdiff_t stride, enum AVMatrixEncoding matrix_encoding, void *log_context)
{
    int i, j, ret;
    AVChannelLayout in_ch_layout = { 0 }, out_ch_layout = { 0 };
    char buf[128];

    ret  = clean_layout(&in_ch_layout, in_layout, log_context);
    ret |= clean_layout(&out_ch_layout, out_layout, log_context);
    if (ret < 0)
        goto fail;

    if(   !av_channel_layout_compare(&out_ch_layout, &(AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO_DOWNMIX)
       && !av_channel_layout_subset(&in_ch_layout, AV_CH_LAYOUT_STEREO_DOWNMIX)
    ) {
        av_channel_layout_uninit(&out_ch_layout);
        out_ch_layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO;
    }
    if(   !av_channel_layout_compare(&in_ch_layout, &(AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO_DOWNMIX)
       && !av_channel_layout_subset(&out_ch_layout, AV_CH_LAYOUT_STEREO_DOWNMIX)
    ) {
        av_channel_layout_uninit(&in_ch_layout);
        in_ch_layout = (AVChannelLayout)AV_CHANNEL_LAYOUT_STEREO;
    }
    if (!av_channel_layout_compare(&in_ch_layout, &(AVChannelLayout)AV_CHANNEL_LAYOUT_22POINT2) &&
        av_channel_layout_compare(&out_ch_layout, &(AVChannelLayout)AV_CHANNEL_LAYOUT_22POINT2)) {
        av_channel_layout_from_mask(&in_ch_layout, (AV_CH_LAYOUT_7POINT1_WIDE_BACK|AV_CH_BACK_CENTER));
        av_channel_layout_describe(&in_ch_layout, buf, sizeof(buf));
        av_log(log_context, AV_LOG_WARNING,
               "Full-on remixing from 22.2 has not yet been implemented! "
               "Processing the input as '%s'\n",
               buf);
    }

    if(!av_channel_layout_check(&in_ch_layout)) {
        av_log(log_context, AV_LOG_ERROR, "Input channel layout is invalid\n");
        ret = AVERROR(EINVAL);
        goto fail;
    }
    if(!sane_layout(&in_ch_layout)) {
        av_channel_layout_describe(&in_ch_layout, buf, sizeof(buf));
        av_log(log_context, AV_LOG_ERROR, "Input channel layout '%s' is not supported\n", buf);
        ret = AVERROR(EINVAL);
        goto fail;
    }

    if(!av_channel_layout_check(&out_ch_layout)) {
        av_log(log_context, AV_LOG_ERROR, "Output channel layout is invalid\n");
        ret = AVERROR(EINVAL);
        goto fail;
    }
    if(!sane_layout(&out_ch_layout)) {
        av_channel_layout_describe(&out_ch_layout, buf, sizeof(buf));
        av_log(log_context, AV_LOG_ERROR, "Output channel layout '%s' is not supported\n", buf);
        ret = AVERROR(EINVAL);
        goto fail;
    }

    build_matrix(&in_ch_layout, &out_ch_layout, center_mix_level,
                 surround_mix_level, lfe_mix_level, maxval, rematrix_volume,
                 matrix_param, stride, matrix_encoding);

    if(rematrix_volume > 0){
        for(i=0; i<SWR_CH_MAX; i++)
            for(j=0; j<SWR_CH_MAX; j++){
                matrix_param[stride*i + j] *= rematrix_volume;
            }
    }

    av_log(log_context, AV_LOG_DEBUG, "Matrix coefficients:\n");
    for (i = 0; i < out_ch_layout.nb_channels; i++){
        av_channel_name(buf, sizeof(buf), av_channel_layout_channel_from_index(&out_ch_layout, i));
        av_log(log_context, AV_LOG_DEBUG, "%s: ", buf);
        for (j = 0; j < in_ch_layout.nb_channels; j++){
            av_channel_name(buf, sizeof(buf), av_channel_layout_channel_from_index(&in_ch_layout, j));
            av_log(log_context, AV_LOG_DEBUG, "%s:%f ", buf, matrix_param[stride*i + j]);
        }
        av_log(log_context, AV_LOG_DEBUG, "\n");
    }

    ret = 0;
fail:
    av_channel_layout_uninit(&in_ch_layout);
    av_channel_layout_uninit(&out_ch_layout);

    return ret;
}

av_cold static int auto_matrix(SwrContext *s)
{
    double maxval;
    int ret;

    if (s->rematrix_maxval > 0) {
        maxval = s->rematrix_maxval;
    } else if (   av_get_packed_sample_fmt(s->out_sample_fmt) < AV_SAMPLE_FMT_FLT
               || av_get_packed_sample_fmt(s->int_sample_fmt) < AV_SAMPLE_FMT_FLT) {
        maxval = 1.0;
    } else
        maxval = INT_MAX;

    memset(s->matrix, 0, sizeof(s->matrix));
    ret = swr_build_matrix2(&s->in_ch_layout, &s->out_ch_layout,
                           s->clev, s->slev, s->lfe_mix_level,
                           maxval, s->rematrix_volume, (double*)s->matrix,
                           s->matrix[1] - s->matrix[0], s->matrix_encoding, s);

    if (ret >= 0 && s->int_sample_fmt == AV_SAMPLE_FMT_FLTP) {
        int i, j;
        for (i = 0; i < FF_ARRAY_ELEMS(s->matrix[0]); i++)
            for (j = 0; j < FF_ARRAY_ELEMS(s->matrix[0]); j++)
                s->matrix_flt[i][j] = s->matrix[i][j];
    }

    return ret;
}

av_cold int swri_rematrix_init(SwrContext *s){
    int i, j;
    int nb_in  = s->used_ch_layout.nb_channels;
    int nb_out = s->out.ch_count;

    s->mix_any_f = NULL;

    if (!s->rematrix_custom) {
        int r = auto_matrix(s);
        if (r)
            return r;
    }
    if (s->midbuf.fmt == AV_SAMPLE_FMT_S16P){
        int maxsum = 0;
        s->native_matrix = av_calloc(nb_in * nb_out, sizeof(int));
        s->native_one    = av_mallocz(sizeof(int));
        if (!s->native_matrix || !s->native_one)
            return AVERROR(ENOMEM);
        for (i = 0; i < nb_out; i++) {
            double rem = 0;
            int sum = 0;

            for (j = 0; j < nb_in; j++) {
                double target = s->matrix[i][j] * 32768 + rem;
                ((int*)s->native_matrix)[i * nb_in + j] = lrintf(target);
                rem += target - ((int*)s->native_matrix)[i * nb_in + j];
                sum += FFABS(((int*)s->native_matrix)[i * nb_in + j]);
            }
            maxsum = FFMAX(maxsum, sum);
        }
        *((int*)s->native_one) = 32768;
        if (maxsum <= 32768) {
            s->mix_1_1_f = (mix_1_1_func_type*)copy_s16;
            s->mix_2_1_f = (mix_2_1_func_type*)sum2_s16;
            s->mix_any_f = (mix_any_func_type*)get_mix_any_func_s16(s);
        } else {
            s->mix_1_1_f = (mix_1_1_func_type*)copy_clip_s16;
            s->mix_2_1_f = (mix_2_1_func_type*)sum2_clip_s16;
            s->mix_any_f = (mix_any_func_type*)get_mix_any_func_clip_s16(s);
        }
    }else if(s->midbuf.fmt == AV_SAMPLE_FMT_FLTP){
        s->native_matrix = av_calloc(nb_in * nb_out, sizeof(float));
        s->native_one    = av_mallocz(sizeof(float));
        if (!s->native_matrix || !s->native_one)
            return AVERROR(ENOMEM);
        for (i = 0; i < nb_out; i++)
            for (j = 0; j < nb_in; j++)
                ((float*)s->native_matrix)[i * nb_in + j] = s->matrix[i][j];
        *((float*)s->native_one) = 1.0;
        s->mix_1_1_f = (mix_1_1_func_type*)copy_float;
        s->mix_2_1_f = (mix_2_1_func_type*)sum2_float;
        s->mix_any_f = (mix_any_func_type*)get_mix_any_func_float(s);
    }else if(s->midbuf.fmt == AV_SAMPLE_FMT_DBLP){
        s->native_matrix = av_calloc(nb_in * nb_out, sizeof(double));
        s->native_one    = av_mallocz(sizeof(double));
        if (!s->native_matrix || !s->native_one)
            return AVERROR(ENOMEM);
        for (i = 0; i < nb_out; i++)
            for (j = 0; j < nb_in; j++)
                ((double*)s->native_matrix)[i * nb_in + j] = s->matrix[i][j];
        *((double*)s->native_one) = 1.0;
        s->mix_1_1_f = (mix_1_1_func_type*)copy_double;
        s->mix_2_1_f = (mix_2_1_func_type*)sum2_double;
        s->mix_any_f = (mix_any_func_type*)get_mix_any_func_double(s);
    }else if(s->midbuf.fmt == AV_SAMPLE_FMT_S32P){
        s->native_one    = av_mallocz(sizeof(int));
        if (!s->native_one)
            return AVERROR(ENOMEM);
        s->native_matrix = av_calloc(nb_in * nb_out, sizeof(int));
        if (!s->native_matrix) {
            av_freep(&s->native_one);
            return AVERROR(ENOMEM);
        }
        for (i = 0; i < nb_out; i++) {
            double rem = 0;

            for (j = 0; j < nb_in; j++) {
                double target = s->matrix[i][j] * 32768 + rem;
                ((int*)s->native_matrix)[i * nb_in + j] = lrintf(target);
                rem += target - ((int*)s->native_matrix)[i * nb_in + j];
            }
        }
        *((int*)s->native_one) = 32768;
        s->mix_1_1_f = (mix_1_1_func_type*)copy_s32;
        s->mix_2_1_f = (mix_2_1_func_type*)sum2_s32;
        s->mix_any_f = (mix_any_func_type*)get_mix_any_func_s32(s);
    }else
        av_assert0(0);
    //FIXME quantize for integeres
    for (i = 0; i < SWR_CH_MAX; i++) {
        int ch_in=0;
        for (j = 0; j < SWR_CH_MAX; j++) {
            s->matrix32[i][j]= lrintf(s->matrix[i][j] * 32768);
            if(s->matrix[i][j])
                s->matrix_ch[i][++ch_in]= j;
        }
        s->matrix_ch[i][0]= ch_in;
    }

#if ARCH_X86 && HAVE_X86ASM && HAVE_MMX
    return swri_rematrix_init_x86(s);
#endif

    return 0;
}

av_cold void swri_rematrix_free(SwrContext *s){
    av_freep(&s->native_matrix);
    av_freep(&s->native_one);
    av_freep(&s->native_simd_matrix);
    av_freep(&s->native_simd_one);
}

int swri_rematrix(SwrContext *s, AudioData *out, AudioData *in, int len, int mustcopy){
    int out_i, in_i, i, j;
    int len1 = 0;
    int off = 0;

    if(s->mix_any_f) {
        s->mix_any_f(out->ch, (const uint8_t **)in->ch, s->native_matrix, len);
        return 0;
    }

    if(s->mix_2_1_simd || s->mix_1_1_simd){
        len1= len&~15;
        off = len1 * out->bps;
    }

    av_assert0(s->out_ch_layout.order == AV_CHANNEL_ORDER_UNSPEC || out->ch_count == s->out_ch_layout.nb_channels);
    av_assert0(s-> in_ch_layout.order == AV_CHANNEL_ORDER_UNSPEC || in ->ch_count == s->in_ch_layout.nb_channels);

    for(out_i=0; out_i<out->ch_count; out_i++){
        switch(s->matrix_ch[out_i][0]){
        case 0:
            if(mustcopy)
                memset(out->ch[out_i], 0, len * av_get_bytes_per_sample(s->int_sample_fmt));
            break;
        case 1:
            in_i= s->matrix_ch[out_i][1];
            if(s->matrix[out_i][in_i]!=1.0){
                if(s->mix_1_1_simd && len1)
                    s->mix_1_1_simd(out->ch[out_i]    , in->ch[in_i]    , s->native_simd_matrix, in->ch_count*out_i + in_i, len1);
                if(len != len1)
                    s->mix_1_1_f   (out->ch[out_i]+off, in->ch[in_i]+off, s->native_matrix, in->ch_count*out_i + in_i, len-len1);
            }else if(mustcopy){
                memcpy(out->ch[out_i], in->ch[in_i], len*out->bps);
            }else{
                out->ch[out_i]= in->ch[in_i];
            }
            break;
        case 2: {
            int in_i1 = s->matrix_ch[out_i][1];
            int in_i2 = s->matrix_ch[out_i][2];
            if(s->mix_2_1_simd && len1)
                s->mix_2_1_simd(out->ch[out_i]    , in->ch[in_i1]    , in->ch[in_i2]    , s->native_simd_matrix, in->ch_count*out_i + in_i1, in->ch_count*out_i + in_i2, len1);
            else
                s->mix_2_1_f   (out->ch[out_i]    , in->ch[in_i1]    , in->ch[in_i2]    , s->native_matrix, in->ch_count*out_i + in_i1, in->ch_count*out_i + in_i2, len1);
            if(len != len1)
                s->mix_2_1_f   (out->ch[out_i]+off, in->ch[in_i1]+off, in->ch[in_i2]+off, s->native_matrix, in->ch_count*out_i + in_i1, in->ch_count*out_i + in_i2, len-len1);
            break;}
        default:
            if(s->int_sample_fmt == AV_SAMPLE_FMT_FLTP){
                for(i=0; i<len; i++){
                    float v=0;
                    for(j=0; j<s->matrix_ch[out_i][0]; j++){
                        in_i= s->matrix_ch[out_i][1+j];
                        v+= ((float*)in->ch[in_i])[i] * s->matrix_flt[out_i][in_i];
                    }
                    ((float*)out->ch[out_i])[i]= v;
                }
            }else if(s->int_sample_fmt == AV_SAMPLE_FMT_DBLP){
                for(i=0; i<len; i++){
                    double v=0;
                    for(j=0; j<s->matrix_ch[out_i][0]; j++){
                        in_i= s->matrix_ch[out_i][1+j];
                        v+= ((double*)in->ch[in_i])[i] * s->matrix[out_i][in_i];
                    }
                    ((double*)out->ch[out_i])[i]= v;
                }
            }else{
                for(i=0; i<len; i++){
                    int v=0;
                    for(j=0; j<s->matrix_ch[out_i][0]; j++){
                        in_i= s->matrix_ch[out_i][1+j];
                        v+= ((int16_t*)in->ch[in_i])[i] * s->matrix32[out_i][in_i];
                    }
                    ((int16_t*)out->ch[out_i])[i]= (v + 16384)>>15;
                }
            }
        }
    }
    return 0;
}