proresenc_anatoliy.c 20.9 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
/*
 * Apple ProRes encoder
 *
 * Copyright (c) 2011 Anatoliy Wasserman
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
24 25
 * @file
 * Apple ProRes encoder (Anatoliy Wasserman version)
26 27 28 29
 * Known FOURCCs: 'apch' (HQ), 'apcn' (SD), 'apcs' (LT), 'acpo' (Proxy)
 */

#include "avcodec.h"
30
#include "dct.h"
31
#include "internal.h"
32
#include "put_bits.h"
33
#include "bytestream.h"
34
#include "fdctdsp.h"
35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147

#define DEFAULT_SLICE_MB_WIDTH 8

#define FF_PROFILE_PRORES_PROXY     0
#define FF_PROFILE_PRORES_LT        1
#define FF_PROFILE_PRORES_STANDARD  2
#define FF_PROFILE_PRORES_HQ        3

static const AVProfile profiles[] = {
    { FF_PROFILE_PRORES_PROXY,    "apco"},
    { FF_PROFILE_PRORES_LT,       "apcs"},
    { FF_PROFILE_PRORES_STANDARD, "apcn"},
    { FF_PROFILE_PRORES_HQ,       "apch"},
    { FF_PROFILE_UNKNOWN }
};

static const int qp_start_table[4] = { 4, 1, 1, 1 };
static const int qp_end_table[4]   = { 8, 9, 6, 6 };
static const int bitrate_table[5]  = { 1000, 2100, 3500, 5400 };

static const uint8_t progressive_scan[64] = {
     0,  1,  8,  9,  2,  3, 10, 11,
    16, 17, 24, 25, 18, 19, 26, 27,
     4,  5, 12, 20, 13,  6,  7, 14,
    21, 28, 29, 22, 15, 23, 30, 31,
    32, 33, 40, 48, 41, 34, 35, 42,
    49, 56, 57, 50, 43, 36, 37, 44,
    51, 58, 59, 52, 45, 38, 39, 46,
    53, 60, 61, 54, 47, 55, 62, 63
};

static const uint8_t QMAT_LUMA[4][64] = {
    {
         4,  7,  9, 11, 13, 14, 15, 63,
         7,  7, 11, 12, 14, 15, 63, 63,
         9, 11, 13, 14, 15, 63, 63, 63,
        11, 11, 13, 14, 63, 63, 63, 63,
        11, 13, 14, 63, 63, 63, 63, 63,
        13, 14, 63, 63, 63, 63, 63, 63,
        13, 63, 63, 63, 63, 63, 63, 63,
        63, 63, 63, 63, 63, 63, 63, 63
    }, {
         4,  5,  6,  7,  9, 11, 13, 15,
         5,  5,  7,  8, 11, 13, 15, 17,
         6,  7,  9, 11, 13, 15, 15, 17,
         7,  7,  9, 11, 13, 15, 17, 19,
         7,  9, 11, 13, 14, 16, 19, 23,
         9, 11, 13, 14, 16, 19, 23, 29,
         9, 11, 13, 15, 17, 21, 28, 35,
        11, 13, 16, 17, 21, 28, 35, 41
    }, {
         4,  4,  5,  5,  6,  7,  7,  9,
         4,  4,  5,  6,  7,  7,  9,  9,
         5,  5,  6,  7,  7,  9,  9, 10,
         5,  5,  6,  7,  7,  9,  9, 10,
         5,  6,  7,  7,  8,  9, 10, 12,
         6,  7,  7,  8,  9, 10, 12, 15,
         6,  7,  7,  9, 10, 11, 14, 17,
         7,  7,  9, 10, 11, 14, 17, 21
    }, {
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  5,
         4,  4,  4,  4,  4,  4,  5,  5,
         4,  4,  4,  4,  4,  5,  5,  6,
         4,  4,  4,  4,  5,  5,  6,  7,
         4,  4,  4,  4,  5,  6,  7,  7
    }
};

static const uint8_t QMAT_CHROMA[4][64] = {
    {
         4,  7,  9, 11, 13, 14, 63, 63,
         7,  7, 11, 12, 14, 63, 63, 63,
         9, 11, 13, 14, 63, 63, 63, 63,
        11, 11, 13, 14, 63, 63, 63, 63,
        11, 13, 14, 63, 63, 63, 63, 63,
        13, 14, 63, 63, 63, 63, 63, 63,
        13, 63, 63, 63, 63, 63, 63, 63,
        63, 63, 63, 63, 63, 63, 63, 63
    }, {
         4,  5,  6,  7,  9, 11, 13, 15,
         5,  5,  7,  8, 11, 13, 15, 17,
         6,  7,  9, 11, 13, 15, 15, 17,
         7,  7,  9, 11, 13, 15, 17, 19,
         7,  9, 11, 13, 14, 16, 19, 23,
         9, 11, 13, 14, 16, 19, 23, 29,
         9, 11, 13, 15, 17, 21, 28, 35,
        11, 13, 16, 17, 21, 28, 35, 41
    }, {
         4,  4,  5,  5,  6,  7,  7,  9,
         4,  4,  5,  6,  7,  7,  9,  9,
         5,  5,  6,  7,  7,  9,  9, 10,
         5,  5,  6,  7,  7,  9,  9, 10,
         5,  6,  7,  7,  8,  9, 10, 12,
         6,  7,  7,  8,  9, 10, 12, 15,
         6,  7,  7,  9, 10, 11, 14, 17,
         7,  7,  9, 10, 11, 14, 17, 21
    }, {
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  4,
         4,  4,  4,  4,  4,  4,  4,  5,
         4,  4,  4,  4,  4,  4,  5,  5,
         4,  4,  4,  4,  4,  5,  5,  6,
         4,  4,  4,  4,  5,  5,  6,  7,
         4,  4,  4,  4,  5,  6,  7,  7
    }
};


typedef struct {
148
    FDCTDSPContext fdsp;
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
    uint8_t* fill_y;
    uint8_t* fill_u;
    uint8_t* fill_v;

    int qmat_luma[16][64];
    int qmat_chroma[16][64];
} ProresContext;

static void encode_codeword(PutBitContext *pb, int val, int codebook)
{
    unsigned int rice_order, exp_order, switch_bits, first_exp, exp, zeros,
            mask;

    /* number of bits to switch between rice and exp golomb */
    switch_bits = codebook & 3;
    rice_order  = codebook >> 5;
    exp_order   = (codebook >> 2) & 7;

    first_exp = ((switch_bits + 1) << rice_order);

    if (val >= first_exp) { /* exp golomb */
        val -= first_exp;
        val += (1 << exp_order);
        exp = av_log2(val);
        zeros = exp - exp_order + switch_bits + 1;
        put_bits(pb, zeros, 0);
175
        put_bits(pb, exp + 1, val);
176 177 178 179 180 181 182 183 184 185 186
    } else if (rice_order) {
        mask = (1 << rice_order) - 1;
        put_bits(pb, (val >> rice_order), 0);
        put_bits(pb, 1, 1);
        put_bits(pb, rice_order, val & mask);
    } else {
        put_bits(pb, val, 0);
        put_bits(pb, 1, 1);
    }
}

187 188 189 190 191
#define QSCALE(qmat,ind,val) ((val) / ((qmat)[ind]))
#define TO_GOLOMB(val) (((val) << 1) ^ ((val) >> 31))
#define DIFF_SIGN(val, sign) (((val) >> 31) ^ (sign))
#define IS_NEGATIVE(val) ((((val) >> 31) ^ -1) + 1)
#define TO_GOLOMB2(val,sign) ((val)==0 ? 0 : ((val) << 1) + (sign))
192 193 194 195 196 197 198 199 200 201 202

static av_always_inline int get_level(int val)
{
    int sign = (val >> 31);
    return (val ^ sign) - sign;
}

#define FIRST_DC_CB 0xB8

static const uint8_t dc_codebook[7] = { 0x04, 0x28, 0x28, 0x4D, 0x4D, 0x70, 0x70};

203
static void encode_dc_coeffs(PutBitContext *pb, int16_t *in,
204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234
        int blocks_per_slice, int *qmat)
{
    int prev_dc, code;
    int i, sign, idx;
    int new_dc, delta, diff_sign, new_code;

    prev_dc = QSCALE(qmat, 0, in[0] - 16384);
    code = TO_GOLOMB(prev_dc);
    encode_codeword(pb, code, FIRST_DC_CB);

    code = 5; sign = 0; idx = 64;
    for (i = 1; i < blocks_per_slice; i++, idx += 64) {
        new_dc    = QSCALE(qmat, 0, in[idx] - 16384);
        delta     = new_dc - prev_dc;
        diff_sign = DIFF_SIGN(delta, sign);
        new_code  = TO_GOLOMB2(get_level(delta), diff_sign);

        encode_codeword(pb, new_code, dc_codebook[FFMIN(code, 6)]);

        code      = new_code;
        sign      = delta >> 31;
        prev_dc   = new_dc;
    }
}

static const uint8_t run_to_cb[16] = { 0x06, 0x06, 0x05, 0x05, 0x04, 0x29,
        0x29, 0x29, 0x29, 0x28, 0x28, 0x28, 0x28, 0x28, 0x28, 0x4C };
static const uint8_t lev_to_cb[10] = { 0x04, 0x0A, 0x05, 0x06, 0x04, 0x28,
        0x28, 0x28, 0x28, 0x4C };

static void encode_ac_coeffs(AVCodecContext *avctx, PutBitContext *pb,
235
        int16_t *in, int blocks_per_slice, int *qmat)
236 237 238 239
{
    int prev_run = 4;
    int prev_level = 2;

Michael Niedermayer's avatar
Michael Niedermayer committed
240 241
    int run = 0, level, code, i, j;
    for (i = 1; i < 64; i++) {
242
        int indp = progressive_scan[i];
Michael Niedermayer's avatar
Michael Niedermayer committed
243
        for (j = 0; j < blocks_per_slice; j++) {
244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
            int val = QSCALE(qmat, indp, in[(j << 6) + indp]);
            if (val) {
                encode_codeword(pb, run, run_to_cb[FFMIN(prev_run, 15)]);

                prev_run   = run;
                run        = 0;
                level      = get_level(val);
                code       = level - 1;

                encode_codeword(pb, code, lev_to_cb[FFMIN(prev_level, 9)]);

                prev_level = level;

                put_bits(pb, 1, IS_NEGATIVE(val));
            } else {
                ++run;
            }
        }
    }
}

265
static void get(uint8_t *pixels, int stride, int16_t* block)
266
{
267
    int i;
268 269

    for (i = 0; i < 8; i++) {
270 271 272
        AV_WN64(block, AV_RN64(pixels));
        AV_WN64(block+4, AV_RN64(pixels+8));
        pixels += stride;
273 274 275 276
        block += 8;
    }
}

277
static void fdct_get(FDCTDSPContext *fdsp, uint8_t *pixels, int stride, int16_t* block)
278 279
{
    get(pixels, stride, block);
280
    fdsp->fdct(block);
281 282 283 284 285 286
}

static int encode_slice_plane(AVCodecContext *avctx, int mb_count,
        uint8_t *src, int src_stride, uint8_t *buf, unsigned buf_size,
        int *qmat, int chroma)
{
287
    ProresContext* ctx = avctx->priv_data;
288
    FDCTDSPContext *fdsp = &ctx->fdsp;
289
    DECLARE_ALIGNED(16, int16_t, blocks)[DEFAULT_SLICE_MB_WIDTH << 8], *block;
290 291 292 293 294
    int i, blocks_per_slice;
    PutBitContext pb;

    block = blocks;
    for (i = 0; i < mb_count; i++) {
295 296
        fdct_get(fdsp, src,                  src_stride, block + (0 << 6));
        fdct_get(fdsp, src + 8 * src_stride, src_stride, block + ((2 - chroma) << 6));
297
        if (!chroma) {
298 299
            fdct_get(fdsp, src + 16,                  src_stride, block + (1 << 6));
            fdct_get(fdsp, src + 16 + 8 * src_stride, src_stride, block + (3 << 6));
300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
        }

        block += (256 >> chroma);
        src   += (32  >> chroma);
    }

    blocks_per_slice = mb_count << (2 - chroma);
    init_put_bits(&pb, buf, buf_size << 3);

    encode_dc_coeffs(&pb, blocks, blocks_per_slice, qmat);
    encode_ac_coeffs(avctx, &pb, blocks, blocks_per_slice, qmat);

    flush_put_bits(&pb);
    return put_bits_ptr(&pb) - pb.buf;
}

static av_always_inline unsigned encode_slice_data(AVCodecContext *avctx,
        uint8_t *dest_y, uint8_t *dest_u, uint8_t *dest_v, int luma_stride,
        int chroma_stride, unsigned mb_count, uint8_t *buf, unsigned data_size,
        unsigned* y_data_size, unsigned* u_data_size, unsigned* v_data_size,
        int qp)
{
322
    ProresContext* ctx = avctx->priv_data;
323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370

    *y_data_size = encode_slice_plane(avctx, mb_count, dest_y, luma_stride,
            buf, data_size, ctx->qmat_luma[qp - 1], 0);

    if (!(avctx->flags & CODEC_FLAG_GRAY)) {
        *u_data_size = encode_slice_plane(avctx, mb_count, dest_u,
                chroma_stride, buf + *y_data_size, data_size - *y_data_size,
                ctx->qmat_chroma[qp - 1], 1);

        *v_data_size = encode_slice_plane(avctx, mb_count, dest_v,
                chroma_stride, buf + *y_data_size + *u_data_size,
                data_size - *y_data_size - *u_data_size,
                ctx->qmat_chroma[qp - 1], 1);
    }

    return *y_data_size + *u_data_size + *v_data_size;
}

static void subimage_with_fill(uint16_t *src, unsigned x, unsigned y,
        unsigned stride, unsigned width, unsigned height, uint16_t *dst,
        unsigned dst_width, unsigned dst_height)
{

    int box_width = FFMIN(width - x, dst_width);
    int box_height = FFMIN(height - y, dst_height);
    int i, j, src_stride = stride >> 1;
    uint16_t last_pix, *last_line;

    src += y * src_stride + x;
    for (i = 0; i < box_height; ++i) {
        for (j = 0; j < box_width; ++j) {
            dst[j] = src[j];
        }
        last_pix = dst[j - 1];
        for (; j < dst_width; j++)
            dst[j] = last_pix;
        src += src_stride;
        dst += dst_width;
    }
    last_line = dst - dst_width;
    for (; i < dst_height; i++) {
        for (j = 0; j < dst_width; ++j) {
            dst[j] = last_line[j];
        }
        dst += dst_width;
    }
}

371
static int encode_slice(AVCodecContext *avctx, const AVFrame *pic, int mb_x,
372 373 374 375 376 377 378
        int mb_y, unsigned mb_count, uint8_t *buf, unsigned data_size,
        int unsafe, int *qp)
{
    int luma_stride, chroma_stride;
    int hdr_size = 6, slice_size;
    uint8_t *dest_y, *dest_u, *dest_v;
    unsigned y_data_size = 0, u_data_size = 0, v_data_size = 0;
379
    ProresContext* ctx = avctx->priv_data;
380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
    int tgt_bits   = (mb_count * bitrate_table[avctx->profile]) >> 2;
    int low_bytes  = (tgt_bits - (tgt_bits >> 3)) >> 3; // 12% bitrate fluctuation
    int high_bytes = (tgt_bits + (tgt_bits >> 3)) >> 3;

    luma_stride   = pic->linesize[0];
    chroma_stride = pic->linesize[1];

    dest_y = pic->data[0] + (mb_y << 4) * luma_stride   + (mb_x << 5);
    dest_u = pic->data[1] + (mb_y << 4) * chroma_stride + (mb_x << 4);
    dest_v = pic->data[2] + (mb_y << 4) * chroma_stride + (mb_x << 4);

    if (unsafe) {

        subimage_with_fill((uint16_t *) pic->data[0], mb_x << 4, mb_y << 4,
                luma_stride, avctx->width, avctx->height,
                (uint16_t *) ctx->fill_y, mb_count << 4, 16);
        subimage_with_fill((uint16_t *) pic->data[1], mb_x << 3, mb_y << 4,
                chroma_stride, avctx->width >> 1, avctx->height,
                (uint16_t *) ctx->fill_u, mb_count << 3, 16);
        subimage_with_fill((uint16_t *) pic->data[2], mb_x << 3, mb_y << 4,
                chroma_stride, avctx->width >> 1, avctx->height,
                (uint16_t *) ctx->fill_v, mb_count << 3, 16);

        encode_slice_data(avctx, ctx->fill_y, ctx->fill_u, ctx->fill_v,
                mb_count << 5, mb_count << 4, mb_count, buf + hdr_size,
                data_size - hdr_size, &y_data_size, &u_data_size, &v_data_size,
                *qp);
    } else {
        slice_size = encode_slice_data(avctx, dest_y, dest_u, dest_v,
                luma_stride, chroma_stride, mb_count, buf + hdr_size,
                data_size - hdr_size, &y_data_size, &u_data_size, &v_data_size,
                *qp);

        if (slice_size > high_bytes && *qp < qp_end_table[avctx->profile]) {
            do {
                *qp += 1;
                slice_size = encode_slice_data(avctx, dest_y, dest_u, dest_v,
                        luma_stride, chroma_stride, mb_count, buf + hdr_size,
                        data_size - hdr_size, &y_data_size, &u_data_size,
                        &v_data_size, *qp);
            } while (slice_size > high_bytes && *qp < qp_end_table[avctx->profile]);
        } else if (slice_size < low_bytes && *qp
                > qp_start_table[avctx->profile]) {
            do {
                *qp -= 1;
                slice_size = encode_slice_data(avctx, dest_y, dest_u, dest_v,
                        luma_stride, chroma_stride, mb_count, buf + hdr_size,
                        data_size - hdr_size, &y_data_size, &u_data_size,
                        &v_data_size, *qp);
            } while (slice_size < low_bytes && *qp > qp_start_table[avctx->profile]);
        }
    }

    buf[0] = hdr_size << 3;
    buf[1] = *qp;
    AV_WB16(buf + 2, y_data_size);
    AV_WB16(buf + 4, u_data_size);

    return hdr_size + y_data_size + u_data_size + v_data_size;
}

441
static int prores_encode_picture(AVCodecContext *avctx, const AVFrame *pic,
442 443 444 445
        uint8_t *buf, const int buf_size)
{
    int mb_width = (avctx->width + 15) >> 4;
    int mb_height = (avctx->height + 15) >> 4;
Michael Niedermayer's avatar
Michael Niedermayer committed
446
    int hdr_size, sl_size, i;
447
    int mb_y, sl_data_size, qp;
448
    int unsafe_bot, unsafe_right;
449
    uint8_t *sl_data, *sl_data_sizes;
450 451
    int slice_per_line = 0, rem = mb_width;

Michael Niedermayer's avatar
Michael Niedermayer committed
452
    for (i = av_log2(DEFAULT_SLICE_MB_WIDTH); i >= 0; --i) {
453 454 455 456 457
        slice_per_line += rem >> i;
        rem &= (1 << i) - 1;
    }

    qp = qp_start_table[avctx->profile];
458 459 460
    hdr_size = 8; sl_data_size = buf_size - hdr_size;
    sl_data_sizes = buf + hdr_size;
    sl_data = sl_data_sizes + (slice_per_line * mb_height * 2);
461 462 463 464 465 466 467 468 469 470 471 472 473
    for (mb_y = 0; mb_y < mb_height; mb_y++) {
        int mb_x = 0;
        int slice_mb_count = DEFAULT_SLICE_MB_WIDTH;
        while (mb_x < mb_width) {
            while (mb_width - mb_x < slice_mb_count)
                slice_mb_count >>= 1;

            unsafe_bot = (avctx->height & 0xf) && (mb_y == mb_height - 1);
            unsafe_right = (avctx->width & 0xf) && (mb_x + slice_mb_count == mb_width);

            sl_size = encode_slice(avctx, pic, mb_x, mb_y, slice_mb_count,
                    sl_data, sl_data_size, unsafe_bot || unsafe_right, &qp);

474
            bytestream_put_be16(&sl_data_sizes, sl_size);
475 476 477 478 479 480 481 482 483 484 485 486 487 488
            sl_data           += sl_size;
            sl_data_size      -= sl_size;
            mb_x              += slice_mb_count;
        }
    }

    buf[0] = hdr_size << 3;
    AV_WB32(buf + 1, sl_data - buf);
    AV_WB16(buf + 5, slice_per_line * mb_height);
    buf[7] = av_log2(DEFAULT_SLICE_MB_WIDTH) << 4;

    return sl_data - buf;
}

489 490
static int prores_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
                               const AVFrame *pict, int *got_packet)
491 492
{
    int header_size = 148;
493 494 495 496 497
    uint8_t *buf;
    int pic_size, ret;
    int frame_size = FFALIGN(avctx->width, 16) * FFALIGN(avctx->height, 16)*16 + 500 + FF_MIN_BUFFER_SIZE; //FIXME choose tighter limit


498
    if ((ret = ff_alloc_packet2(avctx, pkt, frame_size + FF_MIN_BUFFER_SIZE)) < 0)
499 500 501 502 503
        return ret;

    buf = pkt->data;
    pic_size = prores_encode_picture(avctx, pict, buf + header_size + 8,
            pkt->size - header_size - 8);
504

505 506 507 508 509 510
    bytestream_put_be32(&buf, pic_size + 8 + header_size);
    bytestream_put_buffer(&buf, "icpf", 4);

    bytestream_put_be16(&buf, header_size);
    bytestream_put_be16(&buf, 0);
    bytestream_put_buffer(&buf, "fmpg", 4);
511 512
    bytestream_put_be16(&buf, avctx->width);
    bytestream_put_be16(&buf, avctx->height);
513 514 515 516 517 518 519 520 521 522 523
    *buf++ = 0x83; // {10}(422){00}{00}(frame){11}
    *buf++ = 0;
    *buf++ = 2;
    *buf++ = 2;
    *buf++ = 6;
    *buf++ = 32;
    *buf++ = 0;
    *buf++ = 3;

    bytestream_put_buffer(&buf, QMAT_LUMA[avctx->profile],   64);
    bytestream_put_buffer(&buf, QMAT_CHROMA[avctx->profile], 64);
524

525 526 527 528 529
    pkt->flags |= AV_PKT_FLAG_KEY;
    pkt->size = pic_size + 8 + header_size;
    *got_packet = 1;

    return 0;
530 531 532 533 534 535 536 537 538 539 540 541
}

static void scale_mat(const uint8_t* src, int* dst, int scale)
{
    int i;
    for (i = 0; i < 64; i++)
        dst[i] = src[i] * scale;
}

static av_cold int prores_encode_init(AVCodecContext *avctx)
{
    int i;
542
    ProresContext* ctx = avctx->priv_data;
543

544
    if (avctx->pix_fmt != AV_PIX_FMT_YUV422P10) {
545 546 547
        av_log(avctx, AV_LOG_ERROR, "need YUV422P10\n");
        return -1;
    }
548 549
    avctx->bits_per_raw_sample = 10;

550 551 552 553 554 555
    if (avctx->width & 0x1) {
        av_log(avctx, AV_LOG_ERROR,
                "frame width needs to be multiple of 2\n");
        return -1;
    }

556 557 558 559 560 561
    if (avctx->width > 65534 || avctx->height > 65535) {
        av_log(avctx, AV_LOG_ERROR,
                "The maximum dimensions are 65534x65535\n");
        return AVERROR(EINVAL);
    }

562
    if ((avctx->height & 0xf) || (avctx->width & 0xf)) {
563 564 565 566 567
        ctx->fill_y = av_malloc(4 * (DEFAULT_SLICE_MB_WIDTH << 8));
        if (!ctx->fill_y)
            return AVERROR(ENOMEM);
        ctx->fill_u = ctx->fill_y + (DEFAULT_SLICE_MB_WIDTH << 9);
        ctx->fill_v = ctx->fill_u + (DEFAULT_SLICE_MB_WIDTH << 8);
568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584
    }

    if (avctx->profile == FF_PROFILE_UNKNOWN) {
        avctx->profile = FF_PROFILE_PRORES_STANDARD;
        av_log(avctx, AV_LOG_INFO,
                "encoding with ProRes standard (apcn) profile\n");

    } else if (avctx->profile < FF_PROFILE_PRORES_PROXY
            || avctx->profile > FF_PROFILE_PRORES_HQ) {
        av_log(
                avctx,
                AV_LOG_ERROR,
                "unknown profile %d, use [0 - apco, 1 - apcs, 2 - apcn (default), 3 - apch]\n",
                avctx->profile);
        return -1;
    }

585
    ff_fdctdsp_init(&ctx->fdsp, avctx);
586

587 588 589 590 591 592 593
    avctx->codec_tag = AV_RL32((const uint8_t*)profiles[avctx->profile].name);

    for (i = 1; i <= 16; i++) {
        scale_mat(QMAT_LUMA[avctx->profile]  , ctx->qmat_luma[i - 1]  , i);
        scale_mat(QMAT_CHROMA[avctx->profile], ctx->qmat_chroma[i - 1], i);
    }

594
    avctx->coded_frame = av_frame_alloc();
595 596 597 598 599 600 601 602
    avctx->coded_frame->key_frame = 1;
    avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;

    return 0;
}

static av_cold int prores_encode_close(AVCodecContext *avctx)
{
603
    ProresContext* ctx = avctx->priv_data;
604
    av_freep(&avctx->coded_frame);
605
    av_freep(&ctx->fill_y);
606 607 608 609

    return 0;
}

610 611
AVCodec ff_prores_aw_encoder = {
    .name           = "prores_aw",
612
    .long_name      = NULL_IF_CONFIG_SMALL("Apple ProRes"),
613
    .type           = AVMEDIA_TYPE_VIDEO,
614
    .id             = AV_CODEC_ID_PRORES,
615 616 617
    .priv_data_size = sizeof(ProresContext),
    .init           = prores_encode_init,
    .close          = prores_encode_close,
618
    .encode2        = prores_encode_frame,
619
    .pix_fmts       = (const enum AVPixelFormat[]){AV_PIX_FMT_YUV422P10, AV_PIX_FMT_NONE},
620
    .capabilities   = CODEC_CAP_FRAME_THREADS | CODEC_CAP_INTRA_ONLY,
621 622 623
    .profiles       = profiles
};

624 625
AVCodec ff_prores_encoder = {
    .name           = "prores",
626
    .long_name      = NULL_IF_CONFIG_SMALL("Apple ProRes"),
627
    .type           = AVMEDIA_TYPE_VIDEO,
628
    .id             = AV_CODEC_ID_PRORES,
629 630 631
    .priv_data_size = sizeof(ProresContext),
    .init           = prores_encode_init,
    .close          = prores_encode_close,
632
    .encode2        = prores_encode_frame,
633
    .pix_fmts       = (const enum AVPixelFormat[]){AV_PIX_FMT_YUV422P10, AV_PIX_FMT_NONE},
634
    .capabilities   = CODEC_CAP_FRAME_THREADS | CODEC_CAP_INTRA_ONLY,
635 636
    .profiles       = profiles
};