FFmpeg
dcaadpcm.c
Go to the documentation of this file.
1 /*
2  * DCA ADPCM engine
3  * Copyright (C) 2017 Daniil Cherednik
4  *
5  * This file is part of FFmpeg.
6  *
7  * FFmpeg is free software; you can redistribute it and/or
8  * modify it under the terms of the GNU Lesser General Public
9  * License as published by the Free Software Foundation; either
10  * version 2.1 of the License, or (at your option) any later version.
11  *
12  * FFmpeg is distributed in the hope that it will be useful,
13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15  * Lesser General Public License for more details.
16  *
17  * You should have received a copy of the GNU Lesser General Public
18  * License along with FFmpeg; if not, write to the Free Software
19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20  */
21 
22 
23 #include "dcaadpcm.h"
24 #include "dcaenc.h"
25 #include "dca_core.h"
26 #include "mathops.h"
27 
29 
30 //assume we have DCA_ADPCM_COEFFS values before x
31 static inline int64_t calc_corr(const int32_t *x, int len, int j, int k)
32 {
33  int n;
34  int64_t s = 0;
35  for (n = 0; n < len; n++)
36  s += MUL64(x[n-j], x[n-k]);
37  return s;
38 }
39 
40 static inline int64_t apply_filter(const int16_t a[DCA_ADPCM_COEFFS], const int64_t corr[15], const int32_t aa[10])
41 {
42  int64_t err = 0;
43  int64_t tmp = 0;
44 
45  err = corr[0];
46 
47  tmp += MUL64(a[0], corr[1]);
48  tmp += MUL64(a[1], corr[2]);
49  tmp += MUL64(a[2], corr[3]);
50  tmp += MUL64(a[3], corr[4]);
51 
52  tmp = norm__(tmp, 13);
53  tmp += tmp;
54 
55  err -= tmp;
56  tmp = 0;
57 
58  tmp += MUL64(corr[5], aa[0]);
59  tmp += MUL64(corr[6], aa[1]);
60  tmp += MUL64(corr[7], aa[2]);
61  tmp += MUL64(corr[8], aa[3]);
62 
63  tmp += MUL64(corr[9], aa[4]);
64  tmp += MUL64(corr[10], aa[5]);
65  tmp += MUL64(corr[11], aa[6]);
66 
67  tmp += MUL64(corr[12], aa[7]);
68  tmp += MUL64(corr[13], aa[8]);
69 
70  tmp += MUL64(corr[14], aa[9]);
71 
72  tmp = norm__(tmp, 26);
73 
74  err += tmp;
75 
76  return llabs(err);
77 }
78 
79 static int64_t find_best_filter(const DCAADPCMEncContext *s, const int32_t *in, int len)
80 {
81  const premultiplied_coeffs *precalc_data = s->private_data;
82  int i, j, k = 0;
83  int vq = -1;
84  int64_t err;
85  int64_t min_err = 1ll << 62;
86  int64_t corr[15];
87 
88  for (i = 0; i <= DCA_ADPCM_COEFFS; i++)
89  for (j = i; j <= DCA_ADPCM_COEFFS; j++)
90  corr[k++] = calc_corr(in+4, len, i, j);
91 
92  for (i = 0; i < DCA_ADPCM_VQCODEBOOK_SZ; i++) {
93  err = apply_filter(ff_dca_adpcm_vb[i], corr, *precalc_data);
94  if (err < min_err) {
95  min_err = err;
96  vq = i;
97  }
98  precalc_data++;
99  }
100 
101  return vq;
102 }
103 
104 static inline int64_t calc_prediction_gain(int pred_vq, const int32_t *in, int32_t *out, int len)
105 {
106  int i;
107  int32_t error;
108 
109  int64_t signal_energy = 0;
110  int64_t error_energy = 0;
111 
112  for (i = 0; i < len; i++) {
113  error = in[DCA_ADPCM_COEFFS + i] - ff_dcaadpcm_predict(pred_vq, in + i);
114  out[i] = error;
115  signal_energy += MUL64(in[DCA_ADPCM_COEFFS + i], in[DCA_ADPCM_COEFFS + i]);
116  error_energy += MUL64(error, error);
117  }
118 
119  if (!error_energy)
120  return -1;
121 
122  return signal_energy / error_energy;
123 }
124 
126 {
127  int pred_vq, i;
128  int32_t input_buffer[16 + DCA_ADPCM_COEFFS];
129  int32_t input_buffer2[16 + DCA_ADPCM_COEFFS];
130 
131  int32_t max = 0;
132  int shift_bits;
133  uint64_t pg = 0;
134 
135  for (i = 0; i < len + DCA_ADPCM_COEFFS; i++)
136  max |= FFABS(in[i]);
137 
138  // normalize input to simplify apply_filter
139  shift_bits = av_log2(max) - 11;
140 
141  for (i = 0; i < len + DCA_ADPCM_COEFFS; i++) {
142  input_buffer[i] = norm__(in[i], 7);
143  input_buffer2[i] = norm__(in[i], shift_bits);
144  }
145 
146  pred_vq = find_best_filter(s, input_buffer2, len);
147 
148  if (pred_vq < 0)
149  return -1;
150 
151  pg = calc_prediction_gain(pred_vq, input_buffer, diff, len);
152 
153  // Greater than 10db (10*log(10)) prediction gain to use ADPCM.
154  // TODO: Tune it.
155  if (pg < 10)
156  return -1;
157 
158  for (i = 0; i < len; i++)
159  diff[i] <<= 7;
160 
161  return pred_vq;
162 }
163 
165 {
166  int i, j, k;
167 
168  for (i = 0; i < DCA_ADPCM_VQCODEBOOK_SZ; i++) {
169  int id = 0;
170  int32_t t = 0;
171  for (j = 0; j < DCA_ADPCM_COEFFS; j++) {
172  for (k = j; k < DCA_ADPCM_COEFFS; k++) {
174  if (j != k)
175  t *= 2;
176  (*data)[id++] = t;
177  }
178  }
179  data++;
180  }
181 }
182 
183 int ff_dcaadpcm_do_real(int pred_vq_index,
184  softfloat quant, int32_t scale_factor, int32_t step_size,
185  const int32_t *prev_hist, const int32_t *in, int32_t *next_hist, int32_t *out,
186  int len, int32_t peak)
187 {
188  int i;
189  int64_t delta;
190  int32_t dequant_delta;
191  int32_t work_bufer[16 + DCA_ADPCM_COEFFS];
192 
193  memcpy(work_bufer, prev_hist, sizeof(int32_t) * DCA_ADPCM_COEFFS);
194 
195  for (i = 0; i < len; i++) {
196  work_bufer[DCA_ADPCM_COEFFS + i] = ff_dcaadpcm_predict(pred_vq_index, &work_bufer[i]);
197 
198  delta = (int64_t)in[i] - ((int64_t)work_bufer[DCA_ADPCM_COEFFS + i] << 7);
199 
200  out[i] = quantize_value(av_clip64(delta, -peak, peak), quant);
201 
202  ff_dca_core_dequantize(&dequant_delta, &out[i], step_size, scale_factor, 0, 1);
203 
204  work_bufer[DCA_ADPCM_COEFFS+i] += dequant_delta;
205  }
206 
207  memcpy(next_hist, &work_bufer[len], sizeof(int32_t) * DCA_ADPCM_COEFFS);
208 
209  return 0;
210 }
211 
213 {
214  if (!s)
215  return -1;
216 
217  s->private_data = av_malloc(sizeof(premultiplied_coeffs) * DCA_ADPCM_VQCODEBOOK_SZ);
218  if (!s->private_data)
219  return AVERROR(ENOMEM);
220 
221  precalc(s->private_data);
222  return 0;
223 }
224 
226 {
227  if (!s)
228  return;
229 
230  av_freep(&s->private_data);
231 }
error
static void error(const char *err)
Definition: target_bsf_fuzzer.c:30
find_best_filter
static int64_t find_best_filter(const DCAADPCMEncContext *s, const int32_t *in, int len)
Definition: dcaadpcm.c:79
AVERROR
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later. That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another. Frame references ownership and permissions
ff_dcaadpcm_do_real
int ff_dcaadpcm_do_real(int pred_vq_index, softfloat quant, int32_t scale_factor, int32_t step_size, const int32_t *prev_hist, const int32_t *in, int32_t *next_hist, int32_t *out, int len, int32_t peak)
Definition: dcaadpcm.c:183
DCAADPCMEncContext
Definition: dcaadpcm.h:29
out
FILE * out
Definition: movenc.c:54
ff_dcaadpcm_free
av_cold void ff_dcaadpcm_free(DCAADPCMEncContext *s)
Definition: dcaadpcm.c:225
tmp
static uint8_t tmp[11]
Definition: aes_ctr.c:27
data
const char data[16]
Definition: mxf.c:142
max
#define max(a, b)
Definition: cuda_runtime.h:33
quantize_value
static int32_t quantize_value(int32_t value, softfloat quant)
Definition: dcaenc.h:149
av_malloc
#define av_malloc(s)
Definition: tableprint_vlc.h:31
dca_core.h
av_clip64
#define av_clip64
Definition: common.h:125
av_cold
#define av_cold
Definition: attributes.h:90
ff_dcaadpcm_predict
static int64_t ff_dcaadpcm_predict(int pred_vq_index, const int32_t *input)
Definition: dcaadpcm.h:33
s
#define s(width, name)
Definition: cbs_vp9.c:257
DCA_ADPCM_COEFFS
#define DCA_ADPCM_COEFFS
Definition: dcadata.h:28
apply_filter
static int64_t apply_filter(const int16_t a[DCA_ADPCM_COEFFS], const int64_t corr[15], const int32_t aa[10])
Definition: dcaadpcm.c:40
int32_t
int32_t
Definition: audio_convert.c:194
FFABS
#define FFABS(a)
Absolute value, Note, INT_MIN / INT64_MIN result in undefined behavior as they are not representable ...
Definition: common.h:72
softfloat
Definition: dcaenc.h:29
mathops.h
DCA_ADPCM_VQCODEBOOK_SZ
#define DCA_ADPCM_VQCODEBOOK_SZ
Definition: dcadata.h:29
ff_dcaadpcm_subband_analysis
int ff_dcaadpcm_subband_analysis(const DCAADPCMEncContext *s, const int32_t *in, int len, int *diff)
Definition: dcaadpcm.c:125
dcaadpcm.h
dcaenc.h
a
The reader does not expect b to be semantically here and if the code is changed by maybe adding a a division or other the signedness will almost certainly be mistaken To avoid this confusion a new type was SUINT is the C unsigned type but it holds a signed int to use the same example SUINT a
Definition: undefined.txt:41
in
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi - 0x80) *(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(const uint8_t *) pi - 0x80) *(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(const int16_t *) pi >> 8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t, *(const int16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t, *(const int16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(const int32_t *) pi >> 24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t, *(const int32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t, *(const int32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(const float *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(const float *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(const float *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(const double *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(const double *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(const double *) pi *(1U<< 31)))) #define SET_CONV_FUNC_GROUP(ofmt, ifmt) static void set_generic_function(AudioConvert *ac) { } void ff_audio_convert_free(AudioConvert **ac) { if(! *ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);} AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enum AVSampleFormat out_fmt, enum AVSampleFormat in_fmt, int channels, int sample_rate, int apply_map) { AudioConvert *ac;int in_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) return NULL;ac->avr=avr;ac->out_fmt=out_fmt;ac->in_fmt=in_fmt;ac->channels=channels;ac->apply_map=apply_map;if(avr->dither_method !=AV_RESAMPLE_DITHER_NONE &&av_get_packed_sample_fmt(out_fmt)==AV_SAMPLE_FMT_S16 &&av_get_bytes_per_sample(in_fmt) > 2) { ac->dc=ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate, apply_map);if(!ac->dc) { av_free(ac);return NULL;} return ac;} in_planar=ff_sample_fmt_is_planar(in_fmt, channels);out_planar=ff_sample_fmt_is_planar(out_fmt, channels);if(in_planar==out_planar) { ac->func_type=CONV_FUNC_TYPE_FLAT;ac->planes=in_planar ? ac->channels :1;} else if(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;else ac->func_type=CONV_FUNC_TYPE_DEINTERLEAVE;set_generic_function(ac);if(ARCH_AARCH64) ff_audio_convert_init_aarch64(ac);if(ARCH_ARM) ff_audio_convert_init_arm(ac);if(ARCH_X86) ff_audio_convert_init_x86(ac);return ac;} int ff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in) { int use_generic=1;int len=in->nb_samples;int p;if(ac->dc) { av_log(ac->avr, AV_LOG_TRACE, "%d samples - audio_convert: %s to %s (dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));return ff_convert_dither(ac-> in
Definition: audio_convert.c:326
i
int i
Definition: input.c:407
delta
float delta
Definition: vorbis_enc_data.h:457
len
int len
Definition: vorbis_enc_data.h:452
norm__
static int32_t norm__(int64_t a, int bits)
Definition: dcamath.h:27
ff_dca_adpcm_vb
const int16_t ff_dca_adpcm_vb[DCA_ADPCM_VQCODEBOOK_SZ][DCA_ADPCM_COEFFS]
Definition: dcadata.c:60
ff_dca_core_dequantize
static void ff_dca_core_dequantize(int32_t *output, const int32_t *input, int32_t step_size, int32_t scale, int residual, int len)
Definition: dca_core.h:227
ff_dcaadpcm_init
av_cold int ff_dcaadpcm_init(DCAADPCMEncContext *s)
Definition: dcaadpcm.c:212
MUL64
#define MUL64(a, b)
Definition: mathops.h:54
precalc
static void precalc(premultiplied_coeffs *data)
Definition: dcaadpcm.c:164
quant
const uint8_t * quant
Definition: vorbis_enc_data.h:458
diff
static av_always_inline int diff(const uint32_t a, const uint32_t b)
Definition: vf_palettegen.c:136
premultiplied_coeffs
int32_t premultiplied_coeffs[10]
Definition: dcaadpcm.c:28
av_freep
#define av_freep(p)
Definition: tableprint_vlc.h:35
av_log2
int av_log2(unsigned v)
Definition: intmath.c:26
calc_prediction_gain
static int64_t calc_prediction_gain(int pred_vq, const int32_t *in, int32_t *out, int len)
Definition: dcaadpcm.c:104
calc_corr
static int64_t calc_corr(const int32_t *x, int len, int j, int k)
Definition: dcaadpcm.c:31