/src/ffmpeg/libavcodec/roqaudioenc.c
Line | Count | Source (jump to first uncovered line) |
1 | | /* |
2 | | * RoQ audio encoder |
3 | | * |
4 | | * Copyright (c) 2005 Eric Lasota |
5 | | * Based on RoQ specs (c)2001 Tim Ferguson |
6 | | * |
7 | | * This file is part of FFmpeg. |
8 | | * |
9 | | * FFmpeg is free software; you can redistribute it and/or |
10 | | * modify it under the terms of the GNU Lesser General Public |
11 | | * License as published by the Free Software Foundation; either |
12 | | * version 2.1 of the License, or (at your option) any later version. |
13 | | * |
14 | | * FFmpeg is distributed in the hope that it will be useful, |
15 | | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
16 | | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
17 | | * Lesser General Public License for more details. |
18 | | * |
19 | | * You should have received a copy of the GNU Lesser General Public |
20 | | * License along with FFmpeg; if not, write to the Free Software |
21 | | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
22 | | */ |
23 | | |
24 | | #include "libavutil/mem.h" |
25 | | #include "avcodec.h" |
26 | | #include "bytestream.h" |
27 | | #include "codec_internal.h" |
28 | | #include "encode.h" |
29 | | #include "mathops.h" |
30 | | |
31 | 0 | #define ROQ_FRAME_SIZE 735 |
32 | 0 | #define ROQ_HEADER_SIZE 8 |
33 | | |
34 | 0 | #define MAX_DPCM (127*127) |
35 | | |
36 | | |
37 | | typedef struct ROQDPCMContext { |
38 | | short lastSample[2]; |
39 | | int input_frames; |
40 | | int buffered_samples; |
41 | | int16_t *frame_buffer; |
42 | | int64_t first_pts; |
43 | | } ROQDPCMContext; |
44 | | |
45 | | |
46 | | static av_cold int roq_dpcm_encode_close(AVCodecContext *avctx) |
47 | 0 | { |
48 | 0 | ROQDPCMContext *context = avctx->priv_data; |
49 | |
|
50 | 0 | av_freep(&context->frame_buffer); |
51 | |
|
52 | 0 | return 0; |
53 | 0 | } |
54 | | |
55 | | static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx) |
56 | 0 | { |
57 | 0 | ROQDPCMContext *context = avctx->priv_data; |
58 | 0 | int channels = avctx->ch_layout.nb_channels; |
59 | |
|
60 | 0 | if (channels > 2) { |
61 | 0 | av_log(avctx, AV_LOG_ERROR, "Audio must be mono or stereo\n"); |
62 | 0 | return AVERROR(EINVAL); |
63 | 0 | } |
64 | 0 | if (avctx->sample_rate != 22050) { |
65 | 0 | av_log(avctx, AV_LOG_ERROR, "Audio must be 22050 Hz\n"); |
66 | 0 | return AVERROR(EINVAL); |
67 | 0 | } |
68 | | |
69 | 0 | avctx->frame_size = ROQ_FRAME_SIZE; |
70 | 0 | avctx->bit_rate = (ROQ_HEADER_SIZE + ROQ_FRAME_SIZE * channels) * |
71 | 0 | (22050 / ROQ_FRAME_SIZE) * 8; |
72 | |
|
73 | 0 | context->frame_buffer = av_malloc(8 * ROQ_FRAME_SIZE * channels * |
74 | 0 | sizeof(*context->frame_buffer)); |
75 | 0 | if (!context->frame_buffer) |
76 | 0 | return AVERROR(ENOMEM); |
77 | | |
78 | 0 | context->lastSample[0] = context->lastSample[1] = 0; |
79 | |
|
80 | 0 | return 0; |
81 | 0 | } |
82 | | |
83 | | static unsigned char dpcm_predict(short *previous, short current) |
84 | 0 | { |
85 | 0 | int diff; |
86 | 0 | int negative; |
87 | 0 | int result; |
88 | 0 | int predicted; |
89 | |
|
90 | 0 | diff = current - *previous; |
91 | |
|
92 | 0 | negative = diff<0; |
93 | 0 | diff = FFABS(diff); |
94 | |
|
95 | 0 | if (diff >= MAX_DPCM) |
96 | 0 | result = 127; |
97 | 0 | else { |
98 | 0 | result = ff_sqrt(diff); |
99 | 0 | result += diff > result*result+result; |
100 | 0 | } |
101 | | |
102 | | /* See if this overflows */ |
103 | 0 | retry: |
104 | 0 | diff = result*result; |
105 | 0 | if (negative) |
106 | 0 | diff = -diff; |
107 | 0 | predicted = *previous + diff; |
108 | | |
109 | | /* If it overflows, back off a step */ |
110 | 0 | if (predicted > 32767 || predicted < -32768) { |
111 | 0 | result--; |
112 | 0 | goto retry; |
113 | 0 | } |
114 | | |
115 | | /* Add the sign bit */ |
116 | 0 | result |= negative << 7; //if (negative) result |= 128; |
117 | |
|
118 | 0 | *previous = predicted; |
119 | |
|
120 | 0 | return result; |
121 | 0 | } |
122 | | |
123 | | static int roq_dpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt, |
124 | | const AVFrame *frame, int *got_packet_ptr) |
125 | 0 | { |
126 | 0 | int i, stereo, data_size, ret; |
127 | 0 | const int16_t *in = frame ? (const int16_t *)frame->data[0] : NULL; |
128 | 0 | int channels = avctx->ch_layout.nb_channels; |
129 | 0 | uint8_t *out; |
130 | 0 | ROQDPCMContext *context = avctx->priv_data; |
131 | |
|
132 | 0 | stereo = (channels == 2); |
133 | |
|
134 | 0 | if (!in && context->input_frames >= 8) |
135 | 0 | return 0; |
136 | | |
137 | 0 | if (in && context->input_frames < 8) { |
138 | 0 | memcpy(&context->frame_buffer[context->buffered_samples * channels], |
139 | 0 | in, avctx->frame_size * channels * sizeof(*in)); |
140 | 0 | context->buffered_samples += avctx->frame_size; |
141 | 0 | if (context->input_frames == 0) |
142 | 0 | context->first_pts = frame->pts; |
143 | 0 | if (context->input_frames < 7) { |
144 | 0 | context->input_frames++; |
145 | 0 | return 0; |
146 | 0 | } |
147 | 0 | } |
148 | 0 | if (context->input_frames < 8) |
149 | 0 | in = context->frame_buffer; |
150 | |
|
151 | 0 | if (stereo) { |
152 | 0 | context->lastSample[0] &= 0xFF00; |
153 | 0 | context->lastSample[1] &= 0xFF00; |
154 | 0 | } |
155 | |
|
156 | 0 | if (context->input_frames == 7) |
157 | 0 | data_size = channels * context->buffered_samples; |
158 | 0 | else |
159 | 0 | data_size = channels * avctx->frame_size; |
160 | |
|
161 | 0 | ret = ff_get_encode_buffer(avctx, avpkt, ROQ_HEADER_SIZE + data_size, 0); |
162 | 0 | if (ret < 0) |
163 | 0 | return ret; |
164 | 0 | out = avpkt->data; |
165 | |
|
166 | 0 | bytestream_put_byte(&out, stereo ? 0x21 : 0x20); |
167 | 0 | bytestream_put_byte(&out, 0x10); |
168 | 0 | bytestream_put_le32(&out, data_size); |
169 | |
|
170 | 0 | if (stereo) { |
171 | 0 | bytestream_put_byte(&out, (context->lastSample[1])>>8); |
172 | 0 | bytestream_put_byte(&out, (context->lastSample[0])>>8); |
173 | 0 | } else |
174 | 0 | bytestream_put_le16(&out, context->lastSample[0]); |
175 | | |
176 | | /* Write the actual samples */ |
177 | 0 | for (i = 0; i < data_size; i++) |
178 | 0 | *out++ = dpcm_predict(&context->lastSample[(i & 1) & stereo], *in++); |
179 | |
|
180 | 0 | avpkt->pts = context->input_frames <= 7 ? context->first_pts : frame->pts; |
181 | 0 | avpkt->duration = data_size / channels; |
182 | |
|
183 | 0 | context->input_frames++; |
184 | 0 | if (!in) |
185 | 0 | context->input_frames = FFMAX(context->input_frames, 8); |
186 | |
|
187 | 0 | *got_packet_ptr = 1; |
188 | 0 | return 0; |
189 | 0 | } |
190 | | |
191 | | const FFCodec ff_roq_dpcm_encoder = { |
192 | | .p.name = "roq_dpcm", |
193 | | CODEC_LONG_NAME("id RoQ DPCM"), |
194 | | .p.type = AVMEDIA_TYPE_AUDIO, |
195 | | .p.id = AV_CODEC_ID_ROQ_DPCM, |
196 | | .p.capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_DELAY, |
197 | | .priv_data_size = sizeof(ROQDPCMContext), |
198 | | .init = roq_dpcm_encode_init, |
199 | | FF_CODEC_ENCODE_CB(roq_dpcm_encode_frame), |
200 | | .close = roq_dpcm_encode_close, |
201 | | .p.sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16, |
202 | | AV_SAMPLE_FMT_NONE }, |
203 | | }; |