| 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 "internal.h" |
| 27 | #include "mathops.h" |
| 28 | |
| 29 | #define ROQ_FRAME_SIZE 735 |
| 30 | #define ROQ_HEADER_SIZE 8 |
| 31 | |
| 32 | #define MAX_DPCM (127*127) |
| 33 | |
| 34 | |
| 35 | typedef struct |
| 36 | { |
| 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 | |
| 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 | |
| 54 | static av_cold int roq_dpcm_encode_init(AVCodecContext *avctx) |
| 55 | { |
| 56 | ROQDPCMContext *context = avctx->priv_data; |
| 57 | int ret; |
| 58 | |
| 59 | if (avctx->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 * avctx->channels) * |
| 70 | (22050 / ROQ_FRAME_SIZE) * 8; |
| 71 | |
| 72 | context->frame_buffer = av_malloc(8 * ROQ_FRAME_SIZE * avctx->channels * |
| 73 | sizeof(*context->frame_buffer)); |
| 74 | if (!context->frame_buffer) { |
| 75 | ret = AVERROR(ENOMEM); |
| 76 | goto error; |
| 77 | } |
| 78 | |
| 79 | context->lastSample[0] = context->lastSample[1] = 0; |
| 80 | |
| 81 | return 0; |
| 82 | error: |
| 83 | roq_dpcm_encode_close(avctx); |
| 84 | return ret; |
| 85 | } |
| 86 | |
| 87 | static unsigned char dpcm_predict(short *previous, short current) |
| 88 | { |
| 89 | int diff; |
| 90 | int negative; |
| 91 | int result; |
| 92 | int predicted; |
| 93 | |
| 94 | diff = current - *previous; |
| 95 | |
| 96 | negative = diff<0; |
| 97 | diff = FFABS(diff); |
| 98 | |
| 99 | if (diff >= MAX_DPCM) |
| 100 | result = 127; |
| 101 | else { |
| 102 | result = ff_sqrt(diff); |
| 103 | result += diff > result*result+result; |
| 104 | } |
| 105 | |
| 106 | /* See if this overflows */ |
| 107 | retry: |
| 108 | diff = result*result; |
| 109 | if (negative) |
| 110 | diff = -diff; |
| 111 | predicted = *previous + diff; |
| 112 | |
| 113 | /* If it overflows, back off a step */ |
| 114 | if (predicted > 32767 || predicted < -32768) { |
| 115 | result--; |
| 116 | goto retry; |
| 117 | } |
| 118 | |
| 119 | /* Add the sign bit */ |
| 120 | result |= negative << 7; //if (negative) result |= 128; |
| 121 | |
| 122 | *previous = predicted; |
| 123 | |
| 124 | return result; |
| 125 | } |
| 126 | |
| 127 | static int roq_dpcm_encode_frame(AVCodecContext *avctx, AVPacket *avpkt, |
| 128 | const AVFrame *frame, int *got_packet_ptr) |
| 129 | { |
| 130 | int i, stereo, data_size, ret; |
| 131 | const int16_t *in = frame ? (const int16_t *)frame->data[0] : NULL; |
| 132 | uint8_t *out; |
| 133 | ROQDPCMContext *context = avctx->priv_data; |
| 134 | |
| 135 | stereo = (avctx->channels == 2); |
| 136 | |
| 137 | if (!in && context->input_frames >= 8) |
| 138 | return 0; |
| 139 | |
| 140 | if (in && context->input_frames < 8) { |
| 141 | memcpy(&context->frame_buffer[context->buffered_samples * avctx->channels], |
| 142 | in, avctx->frame_size * avctx->channels * sizeof(*in)); |
| 143 | context->buffered_samples += avctx->frame_size; |
| 144 | if (context->input_frames == 0) |
| 145 | context->first_pts = frame->pts; |
| 146 | if (context->input_frames < 7) { |
| 147 | context->input_frames++; |
| 148 | return 0; |
| 149 | } |
| 150 | } |
| 151 | if (context->input_frames < 8) { |
| 152 | in = context->frame_buffer; |
| 153 | } |
| 154 | |
| 155 | if (stereo) { |
| 156 | context->lastSample[0] &= 0xFF00; |
| 157 | context->lastSample[1] &= 0xFF00; |
| 158 | } |
| 159 | |
| 160 | if (context->input_frames == 7) |
| 161 | data_size = avctx->channels * context->buffered_samples; |
| 162 | else |
| 163 | data_size = avctx->channels * avctx->frame_size; |
| 164 | |
| 165 | if ((ret = ff_alloc_packet2(avctx, avpkt, ROQ_HEADER_SIZE + data_size)) < 0) |
| 166 | return ret; |
| 167 | out = avpkt->data; |
| 168 | |
| 169 | bytestream_put_byte(&out, stereo ? 0x21 : 0x20); |
| 170 | bytestream_put_byte(&out, 0x10); |
| 171 | bytestream_put_le32(&out, data_size); |
| 172 | |
| 173 | if (stereo) { |
| 174 | bytestream_put_byte(&out, (context->lastSample[1])>>8); |
| 175 | bytestream_put_byte(&out, (context->lastSample[0])>>8); |
| 176 | } else |
| 177 | bytestream_put_le16(&out, context->lastSample[0]); |
| 178 | |
| 179 | /* Write the actual samples */ |
| 180 | for (i = 0; i < data_size; i++) |
| 181 | *out++ = dpcm_predict(&context->lastSample[i & 1], *in++); |
| 182 | |
| 183 | avpkt->pts = context->input_frames <= 7 ? context->first_pts : frame->pts; |
| 184 | avpkt->duration = data_size / avctx->channels; |
| 185 | |
| 186 | context->input_frames++; |
| 187 | if (!in) |
| 188 | context->input_frames = FFMAX(context->input_frames, 8); |
| 189 | |
| 190 | *got_packet_ptr = 1; |
| 191 | return 0; |
| 192 | } |
| 193 | |
| 194 | AVCodec ff_roq_dpcm_encoder = { |
| 195 | .name = "roq_dpcm", |
| 196 | .long_name = NULL_IF_CONFIG_SMALL("id RoQ DPCM"), |
| 197 | .type = AVMEDIA_TYPE_AUDIO, |
| 198 | .id = AV_CODEC_ID_ROQ_DPCM, |
| 199 | .priv_data_size = sizeof(ROQDPCMContext), |
| 200 | .init = roq_dpcm_encode_init, |
| 201 | .encode2 = roq_dpcm_encode_frame, |
| 202 | .close = roq_dpcm_encode_close, |
| 203 | .capabilities = CODEC_CAP_DELAY, |
| 204 | .sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16, |
| 205 | AV_SAMPLE_FMT_NONE }, |
| 206 | }; |