Hugging Face's logo

camenduru
/
ffmpeg-cuda

Model card Files Community
ffmpeg-cuda
File size: 5,072 Bytes 
8ead80b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
 
/*
 * Copyright (C) 2012-2013 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 "libavutil/avassert.h"
#include "swresample_internal.h"

#include "noise_shaping_data.c"

int swri_get_dither(SwrContext *s, void *dst, int len, unsigned seed, enum AVSampleFormat noise_fmt) {
    double scale = s->dither.noise_scale;
#define TMP_EXTRA 2
    double *tmp = av_malloc_array(len + TMP_EXTRA, sizeof(double));
    int i;

    if (!tmp)
        return AVERROR(ENOMEM);

    for(i=0; i<len + TMP_EXTRA; i++){
        double v;
        seed = seed* 1664525 + 1013904223;

        switch(s->dither.method){
            case SWR_DITHER_RECTANGULAR: v= ((double)seed) / UINT_MAX - 0.5; break;
            default:
                av_assert0(s->dither.method < SWR_DITHER_NB);
                v = ((double)seed) / UINT_MAX;
                seed = seed*1664525 + 1013904223;
                v-= ((double)seed) / UINT_MAX;
                break;
        }
        tmp[i] = v;
    }

    for(i=0; i<len; i++){
        double v;

        switch(s->dither.method){
            default:
                av_assert0(s->dither.method < SWR_DITHER_NB);
                v = tmp[i];
                break;
            case SWR_DITHER_TRIANGULAR_HIGHPASS :
                v = (- tmp[i] + 2*tmp[i+1] - tmp[i+2]) / sqrt(6);
                break;
        }

        v*= scale;

        switch(noise_fmt){
            case AV_SAMPLE_FMT_S16P: ((int16_t*)dst)[i] = v; break;
            case AV_SAMPLE_FMT_S32P: ((int32_t*)dst)[i] = v; break;
            case AV_SAMPLE_FMT_FLTP: ((float  *)dst)[i] = v; break;
            case AV_SAMPLE_FMT_DBLP: ((double *)dst)[i] = v; break;
            default: av_assert0(0);
        }
    }

    av_free(tmp);
    return 0;
}

av_cold int swri_dither_init(SwrContext *s, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt)
{
    int i;
    double scale = 0;

    if (s->dither.method > SWR_DITHER_TRIANGULAR_HIGHPASS && s->dither.method <= SWR_DITHER_NS)
        return AVERROR(EINVAL);

    out_fmt = av_get_packed_sample_fmt(out_fmt);
    in_fmt  = av_get_packed_sample_fmt( in_fmt);

    if(in_fmt == AV_SAMPLE_FMT_FLT || in_fmt == AV_SAMPLE_FMT_DBL){
        if(out_fmt == AV_SAMPLE_FMT_S32) scale = 1.0/(1LL<<31);
        if(out_fmt == AV_SAMPLE_FMT_S16) scale = 1.0/(1LL<<15);
        if(out_fmt == AV_SAMPLE_FMT_U8 ) scale = 1.0/(1LL<< 7);
    }
    if(in_fmt == AV_SAMPLE_FMT_S32 && out_fmt == AV_SAMPLE_FMT_S32 && (s->dither.output_sample_bits&31)) scale = 1;
    if(in_fmt == AV_SAMPLE_FMT_S32 && out_fmt == AV_SAMPLE_FMT_S16) scale = 1<<16;
    if(in_fmt == AV_SAMPLE_FMT_S32 && out_fmt == AV_SAMPLE_FMT_U8 ) scale = 1<<24;
    if(in_fmt == AV_SAMPLE_FMT_S16 && out_fmt == AV_SAMPLE_FMT_U8 ) scale = 1<<8;

    scale *= s->dither.scale;

    if (out_fmt == AV_SAMPLE_FMT_S32 && s->dither.output_sample_bits)
        scale *= 1<<(32-s->dither.output_sample_bits);

    if (scale == 0) {
        s->dither.method = 0;
        return 0;
    }

    s->dither.ns_pos = 0;
    s->dither.noise_scale=   scale;
    s->dither.ns_scale   =   scale;
    s->dither.ns_scale_1 = scale ? 1/scale : 0;
    memset(s->dither.ns_errors, 0, sizeof(s->dither.ns_errors));
    for (i=0; filters[i].coefs; i++) {
        const filter_t *f = &filters[i];
        if (llabs(s->out_sample_rate - f->rate)*20 <= f->rate && f->name == s->dither.method) {
            int j;
            s->dither.ns_taps = f->len;
            for (j=0; j<f->len; j++)
                s->dither.ns_coeffs[j] = f->coefs[j];
            s->dither.ns_scale_1 *= 1 - exp(f->gain_cB * M_LN10 * 0.005) * 2 / (1<<(8*av_get_bytes_per_sample(out_fmt)));
            break;
        }
    }
    if (!filters[i].coefs && s->dither.method > SWR_DITHER_NS) {
        av_log(s, AV_LOG_WARNING, "Requested noise shaping dither not available at this sampling rate, using triangular hp dither\n");
        s->dither.method = SWR_DITHER_TRIANGULAR_HIGHPASS;
    }

    return 0;
}

#define TEMPLATE_DITHER_S16
#include "dither_template.c"
#undef TEMPLATE_DITHER_S16

#define TEMPLATE_DITHER_S32
#include "dither_template.c"
#undef TEMPLATE_DITHER_S32

#define TEMPLATE_DITHER_FLT
#include "dither_template.c"
#undef TEMPLATE_DITHER_FLT

#define TEMPLATE_DITHER_DBL
#include "dither_template.c"
#undef TEMPLATE_DITHER_DBL