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 "avcodec.h"
25 #include "bytestream.h"
26 #include "codec_internal.h"
27 #include "encode.h"
28 #include "mathops.h"
29
30 #define ROQ_FRAME_SIZE 735
31 #define ROQ_HEADER_SIZE 8
32
33 #define MAX_DPCM (127*127)
34
35
36 typedef struct ROQDPCMContext {
37 short lastSample[2];
38 int input_frames;
39 int buffered_samples;
40 int16_t *frame_buffer;
41 int64_t first_pts;
42 } ROQDPCMContext;
43
44
roq_dpcm_encode_close(AVCodecContext * avctx)45 static av_cold int roq_dpcm_encode_close(AVCodecContext *avctx)
46 {
47 ROQDPCMContext *context = avctx->priv_data;
48
49 av_freep(&context->frame_buffer);
50
51 return 0;
52 }
53
roq_dpcm_encode_init(AVCodecContext * avctx)54 static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx)
55 {
56 ROQDPCMContext *context = avctx->priv_data;
57 int channels = avctx->ch_layout.nb_channels;
58
59 if (channels > 2) {
60 av_log(avctx, AV_LOG_ERROR, "Audio must be mono or stereo\n");
61 return AVERROR(EINVAL);
62 }
63 if (avctx->sample_rate != 22050) {
64 av_log(avctx, AV_LOG_ERROR, "Audio must be 22050 Hz\n");
65 return AVERROR(EINVAL);
66 }
67
68 avctx->frame_size = ROQ_FRAME_SIZE;
69 avctx->bit_rate = (ROQ_HEADER_SIZE + ROQ_FRAME_SIZE * channels) *
70 (22050 / ROQ_FRAME_SIZE) * 8;
71
72 context->frame_buffer = av_malloc(8 * ROQ_FRAME_SIZE * channels *
73 sizeof(*context->frame_buffer));
74 if (!context->frame_buffer)
75 return AVERROR(ENOMEM);
76
77 context->lastSample[0] = context->lastSample[1] = 0;
78
79 return 0;
80 }
81
dpcm_predict(short * previous,short current)82 static unsigned char dpcm_predict(short *previous, short current)
83 {
84 int diff;
85 int negative;
86 int result;
87 int predicted;
88
89 diff = current - *previous;
90
91 negative = diff<0;
92 diff = FFABS(diff);
93
94 if (diff >= MAX_DPCM)
95 result = 127;
96 else {
97 result = ff_sqrt(diff);
98 result += diff > result*result+result;
99 }
100
101 /* See if this overflows */
102 retry:
103 diff = result*result;
104 if (negative)
105 diff = -diff;
106 predicted = *previous + diff;
107
108 /* If it overflows, back off a step */
109 if (predicted > 32767 || predicted < -32768) {
110 result--;
111 goto retry;
112 }
113
114 /* Add the sign bit */
115 result |= negative << 7; //if (negative) result |= 128;
116
117 *previous = predicted;
118
119 return result;
120 }
121
roq_dpcm_encode_frame(AVCodecContext * avctx,AVPacket * avpkt,const AVFrame * frame,int * got_packet_ptr)122 static int roq_dpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
123 const AVFrame *frame, int *got_packet_ptr)
124 {
125 int i, stereo, data_size, ret;
126 const int16_t *in = frame ? (const int16_t *)frame->data[0] : NULL;
127 int channels = avctx->ch_layout.nb_channels;
128 uint8_t *out;
129 ROQDPCMContext *context = avctx->priv_data;
130
131 stereo = (channels == 2);
132
133 if (!in && context->input_frames >= 8)
134 return 0;
135
136 if (in && context->input_frames < 8) {
137 memcpy(&context->frame_buffer[context->buffered_samples * channels],
138 in, avctx->frame_size * channels * sizeof(*in));
139 context->buffered_samples += avctx->frame_size;
140 if (context->input_frames == 0)
141 context->first_pts = frame->pts;
142 if (context->input_frames < 7) {
143 context->input_frames++;
144 return 0;
145 }
146 }
147 if (context->input_frames < 8)
148 in = context->frame_buffer;
149
150 if (stereo) {
151 context->lastSample[0] &= 0xFF00;
152 context->lastSample[1] &= 0xFF00;
153 }
154
155 if (context->input_frames == 7)
156 data_size = channels * context->buffered_samples;
157 else
158 data_size = channels * avctx->frame_size;
159
160 ret = ff_get_encode_buffer(avctx, avpkt, ROQ_HEADER_SIZE + data_size, 0);
161 if (ret < 0)
162 return ret;
163 out = avpkt->data;
164
165 bytestream_put_byte(&out, stereo ? 0x21 : 0x20);
166 bytestream_put_byte(&out, 0x10);
167 bytestream_put_le32(&out, data_size);
168
169 if (stereo) {
170 bytestream_put_byte(&out, (context->lastSample[1])>>8);
171 bytestream_put_byte(&out, (context->lastSample[0])>>8);
172 } else
173 bytestream_put_le16(&out, context->lastSample[0]);
174
175 /* Write the actual samples */
176 for (i = 0; i < data_size; i++)
177 *out++ = dpcm_predict(&context->lastSample[i & 1], *in++);
178
179 avpkt->pts = context->input_frames <= 7 ? context->first_pts : frame->pts;
180 avpkt->duration = data_size / channels;
181
182 context->input_frames++;
183 if (!in)
184 context->input_frames = FFMAX(context->input_frames, 8);
185
186 *got_packet_ptr = 1;
187 return 0;
188 }
189
190 const FFCodec ff_roq_dpcm_encoder = {
191 .p.name = "roq_dpcm",
192 .p.long_name = NULL_IF_CONFIG_SMALL("id RoQ DPCM"),
193 .p.type = AVMEDIA_TYPE_AUDIO,
194 .p.id = AV_CODEC_ID_ROQ_DPCM,
195 .p.capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_DELAY,
196 .priv_data_size = sizeof(ROQDPCMContext),
197 .init = roq_dpcm_encode_init,
198 FF_CODEC_ENCODE_CB(roq_dpcm_encode_frame),
199 .close = roq_dpcm_encode_close,
200 .p.sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16,
201 AV_SAMPLE_FMT_NONE },
202 .caps_internal = FF_CODEC_CAP_INIT_THREADSAFE,
203 };
204