• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * Chromaprint fingerprinting muxer
3  * Copyright (c) 2015 rcombs
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 #include "avformat.h"
23 #include "internal.h"
24 #include "libavutil/opt.h"
25 #include <chromaprint.h>
26 
27 #define CPR_VERSION_INT AV_VERSION_INT(CHROMAPRINT_VERSION_MAJOR, \
28                                        CHROMAPRINT_VERSION_MINOR, \
29                                        CHROMAPRINT_VERSION_PATCH)
30 
31 typedef enum FingerprintFormat {
32     FINGERPRINT_RAW,
33     FINGERPRINT_COMPRESSED,
34     FINGERPRINT_BASE64,
35 } FingerprintFormat;
36 
37 typedef struct ChromaprintMuxContext {
38     const AVClass *class;
39     int silence_threshold;
40     int algorithm;
41     FingerprintFormat fp_format;
42 #if CPR_VERSION_INT >= AV_VERSION_INT(1, 4, 0)
43     ChromaprintContext *ctx;
44 #else
45     ChromaprintContext ctx;
46 #endif
47 } ChromaprintMuxContext;
48 
deinit(AVFormatContext * s)49 static void deinit(AVFormatContext *s)
50 {
51     ChromaprintMuxContext *const cpr = s->priv_data;
52 
53     if (cpr->ctx) {
54         ff_lock_avformat();
55         chromaprint_free(cpr->ctx);
56         ff_unlock_avformat();
57     }
58 }
59 
write_header(AVFormatContext * s)60 static int write_header(AVFormatContext *s)
61 {
62     ChromaprintMuxContext *cpr = s->priv_data;
63     AVStream *st;
64 
65     ff_lock_avformat();
66     cpr->ctx = chromaprint_new(cpr->algorithm);
67     ff_unlock_avformat();
68 
69     if (!cpr->ctx) {
70         av_log(s, AV_LOG_ERROR, "Failed to create chromaprint context.\n");
71         return AVERROR_EXTERNAL;
72     }
73 
74     if (cpr->silence_threshold != -1) {
75 #if CPR_VERSION_INT >= AV_VERSION_INT(0, 7, 0)
76         if (!chromaprint_set_option(cpr->ctx, "silence_threshold", cpr->silence_threshold)) {
77             av_log(s, AV_LOG_ERROR, "Failed to set silence threshold. Setting silence_threshold requires -algorithm 3 option.\n");
78             return AVERROR_EXTERNAL;
79         }
80 #else
81         av_log(s, AV_LOG_ERROR, "Setting the silence threshold requires Chromaprint "
82                                 "version 0.7.0 or later.\n");
83         return AVERROR(ENOSYS);
84 #endif
85     }
86 
87     if (s->nb_streams != 1) {
88         av_log(s, AV_LOG_ERROR, "Only one stream is supported\n");
89         return AVERROR(EINVAL);
90     }
91 
92     st = s->streams[0];
93 
94     if (st->codecpar->ch_layout.nb_channels > 2) {
95         av_log(s, AV_LOG_ERROR, "Only up to 2 channels are supported\n");
96         return AVERROR(EINVAL);
97     }
98 
99     if (st->codecpar->sample_rate < 1000) {
100         av_log(s, AV_LOG_ERROR, "Sampling rate must be at least 1000\n");
101         return AVERROR(EINVAL);
102     }
103 
104     if (!chromaprint_start(cpr->ctx, st->codecpar->sample_rate, st->codecpar->ch_layout.nb_channels)) {
105         av_log(s, AV_LOG_ERROR, "Failed to start chromaprint\n");
106         return AVERROR_EXTERNAL;
107     }
108 
109     return 0;
110 }
111 
write_packet(AVFormatContext * s,AVPacket * pkt)112 static int write_packet(AVFormatContext *s, AVPacket *pkt)
113 {
114     ChromaprintMuxContext *cpr = s->priv_data;
115     return chromaprint_feed(cpr->ctx, (const int16_t *)pkt->data, pkt->size / 2) ? 0 : AVERROR(EINVAL);
116 }
117 
write_trailer(AVFormatContext * s)118 static int write_trailer(AVFormatContext *s)
119 {
120     ChromaprintMuxContext *cpr = s->priv_data;
121     AVIOContext *pb = s->pb;
122     void *fp = NULL;
123     char *enc_fp = NULL;
124     int size, enc_size, ret = AVERROR_EXTERNAL;
125 
126     if (!chromaprint_finish(cpr->ctx)) {
127         av_log(s, AV_LOG_ERROR, "Failed to generate fingerprint\n");
128         goto fail;
129     }
130 
131     if (!chromaprint_get_raw_fingerprint(cpr->ctx, (uint32_t **)&fp, &size)) {
132         av_log(s, AV_LOG_ERROR, "Failed to retrieve fingerprint\n");
133         goto fail;
134     }
135 
136     switch (cpr->fp_format) {
137     case FINGERPRINT_RAW:
138         avio_write(pb, fp, size * 4); //fp points to array of uint32_t
139         break;
140     case FINGERPRINT_COMPRESSED:
141     case FINGERPRINT_BASE64:
142         if (!chromaprint_encode_fingerprint(fp, size, cpr->algorithm, &enc_fp, &enc_size,
143                                             cpr->fp_format == FINGERPRINT_BASE64)) {
144             av_log(s, AV_LOG_ERROR, "Failed to encode fingerprint\n");
145             goto fail;
146         }
147         avio_write(pb, enc_fp, enc_size);
148         break;
149     }
150 
151     ret = 0;
152 fail:
153     if (fp)
154         chromaprint_dealloc(fp);
155     if (enc_fp)
156         chromaprint_dealloc(enc_fp);
157     return ret;
158 }
159 
160 #define OFFSET(x) offsetof(ChromaprintMuxContext, x)
161 #define FLAGS AV_OPT_FLAG_ENCODING_PARAM
162 static const AVOption options[] = {
163     { "silence_threshold", "threshold for detecting silence", OFFSET(silence_threshold), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 32767, FLAGS },
164     { "algorithm", "version of the fingerprint algorithm", OFFSET(algorithm), AV_OPT_TYPE_INT, { .i64 = CHROMAPRINT_ALGORITHM_DEFAULT }, CHROMAPRINT_ALGORITHM_TEST1, INT_MAX, FLAGS },
165     { "fp_format", "fingerprint format to write", OFFSET(fp_format), AV_OPT_TYPE_INT, { .i64 = FINGERPRINT_BASE64 }, FINGERPRINT_RAW, FINGERPRINT_BASE64, FLAGS, "fp_format" },
166     { "raw", "binary raw fingerprint", 0, AV_OPT_TYPE_CONST, {.i64 = FINGERPRINT_RAW }, INT_MIN, INT_MAX, FLAGS, "fp_format"},
167     { "compressed", "binary compressed fingerprint", 0, AV_OPT_TYPE_CONST, {.i64 = FINGERPRINT_COMPRESSED }, INT_MIN, INT_MAX, FLAGS, "fp_format"},
168     { "base64", "Base64 compressed fingerprint", 0, AV_OPT_TYPE_CONST, {.i64 = FINGERPRINT_BASE64 }, INT_MIN, INT_MAX, FLAGS, "fp_format"},
169     { NULL },
170 };
171 
172 static const AVClass chromaprint_class = {
173     .class_name = "chromaprint muxer",
174     .item_name  = av_default_item_name,
175     .option     = options,
176     .version    = LIBAVUTIL_VERSION_INT,
177 };
178 
179 const AVOutputFormat ff_chromaprint_muxer = {
180     .name              = "chromaprint",
181     .long_name         = NULL_IF_CONFIG_SMALL("Chromaprint"),
182     .priv_data_size    = sizeof(ChromaprintMuxContext),
183     .audio_codec       = AV_NE(AV_CODEC_ID_PCM_S16BE, AV_CODEC_ID_PCM_S16LE),
184     .write_header      = write_header,
185     .write_packet      = write_packet,
186     .write_trailer     = write_trailer,
187     .deinit            = deinit,
188     .flags             = AVFMT_NOTIMESTAMPS,
189     .priv_class        = &chromaprint_class,
190 };
191