xref: /third_party/ffmpeg/libavcodec/libxavs.c (revision cabdff1a)
1/*
2 * AVS encoding using the xavs library
3 * Copyright (C) 2010 Amanda, Y.N. Wu <amanda11192003@gmail.com>
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 <stdio.h>
23#include <stdlib.h>
24#include <string.h>
25#include <math.h>
26#include <stdint.h>
27#include <float.h>
28#include <xavs.h>
29#include "avcodec.h"
30#include "codec_internal.h"
31#include "encode.h"
32#include "packet_internal.h"
33#include "libavutil/internal.h"
34#include "libavutil/mem.h"
35#include "libavutil/opt.h"
36
37#define END_OF_STREAM 0x001
38
39#define XAVS_PART_I8X8 0x002 /* Analyze i8x8 (requires 8x8 transform) */
40#define XAVS_PART_P8X8 0x010 /* Analyze p16x8, p8x16 and p8x8 */
41#define XAVS_PART_B8X8 0x100 /* Analyze b16x8, b*/
42
43typedef struct XavsContext {
44    AVClass        *class;
45    xavs_param_t    params;
46    xavs_t         *enc;
47    xavs_picture_t  pic;
48    uint8_t        *sei;
49    int             sei_size;
50    int             end_of_stream;
51    float crf;
52    int cqp;
53    int b_bias;
54    float cplxblur;
55    int direct_pred;
56    int aud;
57    int fast_pskip;
58    int motion_est;
59    int mbtree;
60    int mixed_refs;
61    int b_frame_strategy;
62    int chroma_offset;
63    int scenechange_threshold;
64    int noise_reduction;
65
66    int64_t *pts_buffer;
67    int out_frame_count;
68} XavsContext;
69
70static void XAVS_log(void *p, int level, const char *fmt, va_list args)
71{
72    static const int level_map[] = {
73        [XAVS_LOG_ERROR]   = AV_LOG_ERROR,
74        [XAVS_LOG_WARNING] = AV_LOG_WARNING,
75        [XAVS_LOG_INFO]    = AV_LOG_INFO,
76        [XAVS_LOG_DEBUG]   = AV_LOG_DEBUG
77    };
78
79    if (level < 0 || level > XAVS_LOG_DEBUG)
80        return;
81
82    av_vlog(p, level_map[level], fmt, args);
83}
84
85static int encode_nals(AVCodecContext *ctx, AVPacket *pkt,
86                       xavs_nal_t *nals, int nnal)
87{
88    XavsContext *x4 = ctx->priv_data;
89    int64_t size = x4->sei_size;
90    uint8_t *p, *p_end;
91    int i, s, ret;
92
93    if (!nnal)
94        return 0;
95
96    for (i = 0; i < nnal; i++)
97        size += 3U + nals[i].i_payload;
98
99    if ((ret = ff_get_encode_buffer(ctx, pkt, size, 0)) < 0)
100        return ret;
101    p = pkt->data;
102    p_end = pkt->data + size;
103
104    /* Write the SEI as part of the first frame. */
105    if (x4->sei_size > 0 && nnal > 0) {
106        memcpy(p, x4->sei, x4->sei_size);
107        p += x4->sei_size;
108        x4->sei_size = 0;
109    }
110
111    for (i = 0; i < nnal; i++) {
112        int size = p_end - p;
113        s = xavs_nal_encode(p, &size, 1, nals + i);
114        if (s < 0)
115            return AVERROR_EXTERNAL;
116        if (s != 3U + nals[i].i_payload)
117            return AVERROR_EXTERNAL;
118        p += s;
119    }
120
121    return 1;
122}
123
124static int XAVS_frame(AVCodecContext *avctx, AVPacket *pkt,
125                      const AVFrame *frame, int *got_packet)
126{
127    XavsContext *x4 = avctx->priv_data;
128    xavs_nal_t *nal;
129    int nnal, i, ret;
130    xavs_picture_t pic_out;
131    int pict_type;
132
133    x4->pic.img.i_csp   = XAVS_CSP_I420;
134    x4->pic.img.i_plane = 3;
135
136    if (frame) {
137       for (i = 0; i < 3; i++) {
138            x4->pic.img.plane[i] = frame->data[i];
139            x4->pic.img.i_stride[i] = frame->linesize[i];
140       }
141
142        x4->pic.i_pts  = frame->pts;
143        x4->pic.i_type = XAVS_TYPE_AUTO;
144        x4->pts_buffer[avctx->frame_number % (avctx->max_b_frames+1)] = frame->pts;
145    }
146
147    if (xavs_encoder_encode(x4->enc, &nal, &nnal,
148                            frame? &x4->pic: NULL, &pic_out) < 0)
149        return AVERROR_EXTERNAL;
150
151    ret = encode_nals(avctx, pkt, nal, nnal);
152
153    if (ret < 0)
154        return ret;
155
156    if (!ret) {
157        if (!frame && !(x4->end_of_stream)) {
158            if ((ret = ff_get_encode_buffer(avctx, pkt, 4, 0)) < 0)
159                return ret;
160
161            pkt->data[0] = 0x0;
162            pkt->data[1] = 0x0;
163            pkt->data[2] = 0x01;
164            pkt->data[3] = 0xb1;
165            pkt->dts = 2*x4->pts_buffer[(x4->out_frame_count-1)%(avctx->max_b_frames+1)] -
166                         x4->pts_buffer[(x4->out_frame_count-2)%(avctx->max_b_frames+1)];
167            x4->end_of_stream = END_OF_STREAM;
168            *got_packet = 1;
169        }
170        return 0;
171    }
172
173    pkt->pts = pic_out.i_pts;
174    if (avctx->has_b_frames) {
175        if (!x4->out_frame_count)
176            pkt->dts = pkt->pts - (x4->pts_buffer[1] - x4->pts_buffer[0]);
177        else
178            pkt->dts = x4->pts_buffer[(x4->out_frame_count-1)%(avctx->max_b_frames+1)];
179    } else
180        pkt->dts = pkt->pts;
181
182    switch (pic_out.i_type) {
183    case XAVS_TYPE_IDR:
184    case XAVS_TYPE_I:
185        pict_type = AV_PICTURE_TYPE_I;
186        break;
187    case XAVS_TYPE_P:
188        pict_type = AV_PICTURE_TYPE_P;
189        break;
190    case XAVS_TYPE_B:
191    case XAVS_TYPE_BREF:
192        pict_type = AV_PICTURE_TYPE_B;
193        break;
194    default:
195        pict_type = AV_PICTURE_TYPE_NONE;
196    }
197
198    /* There is no IDR frame in AVS JiZhun */
199    /* Sequence header is used as a flag */
200    if (pic_out.i_type == XAVS_TYPE_I) {
201        pkt->flags |= AV_PKT_FLAG_KEY;
202    }
203
204    ff_side_data_set_encoder_stats(pkt, (pic_out.i_qpplus1 - 1) * FF_QP2LAMBDA, NULL, 0, pict_type);
205
206    x4->out_frame_count++;
207    *got_packet = ret;
208    return 0;
209}
210
211static av_cold int XAVS_close(AVCodecContext *avctx)
212{
213    XavsContext *x4 = avctx->priv_data;
214
215    av_freep(&x4->sei);
216    av_freep(&x4->pts_buffer);
217
218    if (x4->enc)
219        xavs_encoder_close(x4->enc);
220
221    return 0;
222}
223
224static av_cold int XAVS_init(AVCodecContext *avctx)
225{
226    XavsContext *x4 = avctx->priv_data;
227
228    x4->sei_size = 0;
229    xavs_param_default(&x4->params);
230
231    x4->params.pf_log               = XAVS_log;
232    x4->params.p_log_private        = avctx;
233    x4->params.i_keyint_max         = avctx->gop_size;
234    if (avctx->bit_rate) {
235        x4->params.rc.i_bitrate   = avctx->bit_rate / 1000;
236        x4->params.rc.i_rc_method = XAVS_RC_ABR;
237    }
238    x4->params.rc.i_vbv_buffer_size = avctx->rc_buffer_size / 1000;
239    x4->params.rc.i_vbv_max_bitrate = avctx->rc_max_rate    / 1000;
240    x4->params.rc.b_stat_write      = avctx->flags & AV_CODEC_FLAG_PASS1;
241    if (avctx->flags & AV_CODEC_FLAG_PASS2) {
242        x4->params.rc.b_stat_read = 1;
243    } else {
244        if (x4->crf >= 0) {
245            x4->params.rc.i_rc_method   = XAVS_RC_CRF;
246            x4->params.rc.f_rf_constant = x4->crf;
247        } else if (x4->cqp >= 0) {
248            x4->params.rc.i_rc_method   = XAVS_RC_CQP;
249            x4->params.rc.i_qp_constant = x4->cqp;
250        }
251    }
252
253    if (x4->aud >= 0)
254        x4->params.b_aud                      = x4->aud;
255    if (x4->mbtree >= 0)
256        x4->params.rc.b_mb_tree               = x4->mbtree;
257    if (x4->direct_pred >= 0)
258        x4->params.analyse.i_direct_mv_pred   = x4->direct_pred;
259    if (x4->fast_pskip >= 0)
260        x4->params.analyse.b_fast_pskip       = x4->fast_pskip;
261    if (x4->motion_est >= 0)
262        x4->params.analyse.i_me_method        = x4->motion_est;
263    if (x4->mixed_refs >= 0)
264        x4->params.analyse.b_mixed_references = x4->mixed_refs;
265    if (x4->b_bias != INT_MIN)
266        x4->params.i_bframe_bias              = x4->b_bias;
267    if (x4->cplxblur >= 0)
268        x4->params.rc.f_complexity_blur = x4->cplxblur;
269
270    x4->params.i_bframe          = avctx->max_b_frames;
271    /* cabac is not included in AVS JiZhun Profile */
272    x4->params.b_cabac           = 0;
273
274    x4->params.i_bframe_adaptive = x4->b_frame_strategy;
275
276    avctx->has_b_frames          = !!avctx->max_b_frames;
277
278    /* AVS doesn't allow B picture as reference */
279    /* The max allowed reference frame number of B is 2 */
280    x4->params.i_keyint_min      = avctx->keyint_min;
281    if (x4->params.i_keyint_min > x4->params.i_keyint_max)
282        x4->params.i_keyint_min = x4->params.i_keyint_max;
283
284    x4->params.i_scenecut_threshold = x4->scenechange_threshold;
285
286   // x4->params.b_deblocking_filter       = avctx->flags & AV_CODEC_FLAG_LOOP_FILTER;
287
288    x4->params.rc.i_qp_min                 = avctx->qmin;
289    x4->params.rc.i_qp_max                 = avctx->qmax;
290    x4->params.rc.i_qp_step                = avctx->max_qdiff;
291
292    x4->params.rc.f_qcompress       = avctx->qcompress; /* 0.0 => cbr, 1.0 => constant qp */
293    x4->params.rc.f_qblur           = avctx->qblur;     /* temporally blur quants */
294
295    x4->params.i_frame_reference    = avctx->refs;
296
297    x4->params.i_width              = avctx->width;
298    x4->params.i_height             = avctx->height;
299    x4->params.vui.i_sar_width      = avctx->sample_aspect_ratio.num;
300    x4->params.vui.i_sar_height     = avctx->sample_aspect_ratio.den;
301    /* This is only used for counting the fps */
302    x4->params.i_fps_num            = avctx->time_base.den;
303    x4->params.i_fps_den            = avctx->time_base.num;
304    x4->params.analyse.inter        = XAVS_ANALYSE_I8x8 |XAVS_ANALYSE_PSUB16x16| XAVS_ANALYSE_BSUB16x16;
305
306    x4->params.analyse.i_me_range = avctx->me_range;
307    x4->params.analyse.i_subpel_refine    = avctx->me_subpel_quality;
308
309    x4->params.analyse.b_chroma_me        = avctx->me_cmp & FF_CMP_CHROMA;
310    /* AVS P2 only enables 8x8 transform */
311    x4->params.analyse.b_transform_8x8    = 1; //avctx->flags2 & AV_CODEC_FLAG2_8X8DCT;
312
313    x4->params.analyse.i_trellis          = avctx->trellis;
314
315    x4->params.analyse.i_noise_reduction  = x4->noise_reduction;
316
317    if (avctx->level > 0)
318        x4->params.i_level_idc = avctx->level;
319
320    if (avctx->bit_rate > 0)
321        x4->params.rc.f_rate_tolerance =
322            (float)avctx->bit_rate_tolerance / avctx->bit_rate;
323
324    if ((avctx->rc_buffer_size) &&
325        (avctx->rc_initial_buffer_occupancy <= avctx->rc_buffer_size)) {
326        x4->params.rc.f_vbv_buffer_init =
327            (float)avctx->rc_initial_buffer_occupancy / avctx->rc_buffer_size;
328    } else
329        x4->params.rc.f_vbv_buffer_init = 0.9;
330
331    /* TAG:do we have MB tree RC method */
332    /* what is the RC method we are now using? Default NO */
333    x4->params.rc.f_ip_factor             = 1 / fabs(avctx->i_quant_factor);
334    x4->params.rc.f_pb_factor             = avctx->b_quant_factor;
335
336    x4->params.analyse.i_chroma_qp_offset = x4->chroma_offset;
337
338    x4->params.analyse.b_psnr = avctx->flags & AV_CODEC_FLAG_PSNR;
339    x4->params.i_log_level    = XAVS_LOG_DEBUG;
340    x4->params.i_threads      = avctx->thread_count;
341    x4->params.b_interlaced   = avctx->flags & AV_CODEC_FLAG_INTERLACED_DCT;
342
343    if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER)
344        x4->params.b_repeat_headers = 0;
345
346    x4->enc = xavs_encoder_open(&x4->params);
347    if (!x4->enc)
348        return AVERROR_EXTERNAL;
349
350    if (!FF_ALLOCZ_TYPED_ARRAY(x4->pts_buffer, avctx->max_b_frames + 1))
351        return AVERROR(ENOMEM);
352
353    /* TAG: Do we have GLOBAL HEADER in AVS */
354    /* We Have PPS and SPS in AVS */
355    if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER && 0) {
356        xavs_nal_t *nal;
357        int nnal, s, i, size;
358        uint8_t *p;
359
360        s = xavs_encoder_headers(x4->enc, &nal, &nnal);
361
362        avctx->extradata = p = av_malloc(s);
363        for (i = 0; i < nnal; i++) {
364            /* Don't put the SEI in extradata. */
365            if (nal[i].i_type == NAL_SEI) {
366                x4->sei = av_malloc( 5 + nal[i].i_payload * 4 / 3 );
367                if (xavs_nal_encode(x4->sei, &x4->sei_size, 1, nal + i) < 0)
368                    return -1;
369
370                continue;
371            }
372            size = xavs_nal_encode(p, &s, 1, nal + i);
373            if (size < 0)
374                return -1;
375            p += size;
376        }
377        avctx->extradata_size = p - avctx->extradata;
378    }
379    return 0;
380}
381
382#define OFFSET(x) offsetof(XavsContext, x)
383#define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
384static const AVOption options[] = {
385    { "crf",           "Select the quality for constant quality mode",    OFFSET(crf),           AV_OPT_TYPE_FLOAT,  {.dbl = -1 }, -1, FLT_MAX, VE },
386    { "qp",            "Constant quantization parameter rate control method",OFFSET(cqp),        AV_OPT_TYPE_INT,    {.i64 = -1 }, -1, INT_MAX, VE },
387    { "b-bias",        "Influences how often B-frames are used",          OFFSET(b_bias),        AV_OPT_TYPE_INT,    {.i64 = INT_MIN}, INT_MIN, INT_MAX, VE },
388    { "cplxblur",      "Reduce fluctuations in QP (before curve compression)", OFFSET(cplxblur), AV_OPT_TYPE_FLOAT,  {.dbl = -1 }, -1, FLT_MAX, VE},
389    { "direct-pred",   "Direct MV prediction mode",                       OFFSET(direct_pred),   AV_OPT_TYPE_INT,    {.i64 = -1 }, -1, INT_MAX, VE, "direct-pred" },
390    { "none",          NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_DIRECT_PRED_NONE },     0, 0, VE, "direct-pred" },
391    { "spatial",       NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_DIRECT_PRED_SPATIAL },  0, 0, VE, "direct-pred" },
392    { "temporal",      NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_DIRECT_PRED_TEMPORAL }, 0, 0, VE, "direct-pred" },
393    { "auto",          NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_DIRECT_PRED_AUTO },     0, 0, VE, "direct-pred" },
394    { "aud",           "Use access unit delimiters.",                     OFFSET(aud),           AV_OPT_TYPE_BOOL,    {.i64 = -1 }, -1, 1, VE},
395    { "mbtree",        "Use macroblock tree ratecontrol.",                OFFSET(mbtree),        AV_OPT_TYPE_BOOL,    {.i64 = -1 }, -1, 1, VE},
396    { "mixed-refs",    "One reference per partition, as opposed to one reference per macroblock", OFFSET(mixed_refs), AV_OPT_TYPE_BOOL, {.i64 = -1}, -1, 1, VE },
397    { "fast-pskip",    NULL,                                              OFFSET(fast_pskip),    AV_OPT_TYPE_BOOL,    {.i64 = -1 }, -1, 1, VE},
398    { "motion-est",   "Set motion estimation method",                     OFFSET(motion_est),    AV_OPT_TYPE_INT,    { .i64 = XAVS_ME_DIA }, -1, XAVS_ME_TESA, VE, "motion-est"},
399    { "dia",           NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_ME_DIA },               INT_MIN, INT_MAX, VE, "motion-est" },
400    { "hex",           NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_ME_HEX },               INT_MIN, INT_MAX, VE, "motion-est" },
401    { "umh",           NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_ME_UMH },               INT_MIN, INT_MAX, VE, "motion-est" },
402    { "esa",           NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_ME_ESA },               INT_MIN, INT_MAX, VE, "motion-est" },
403    { "tesa",          NULL,      0,    AV_OPT_TYPE_CONST, { .i64 = XAVS_ME_TESA },              INT_MIN, INT_MAX, VE, "motion-est" },
404    { "b_strategy",    "Strategy to choose between I/P/B-frames",         OFFSET(b_frame_strategy), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, 2, VE},
405    { "chromaoffset", "QP difference between chroma and luma",           OFFSET(chroma_offset), AV_OPT_TYPE_INT, {.i64 = 0 }, INT_MIN, INT_MAX, VE},
406    { "sc_threshold", "Scene change threshold",                           OFFSET(scenechange_threshold), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, INT_MAX, VE},
407    { "noise_reduction", "Noise reduction",                               OFFSET(noise_reduction), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, INT_MAX, VE},
408
409    { NULL },
410};
411
412static const AVClass xavs_class = {
413    .class_name = "libxavs",
414    .item_name  = av_default_item_name,
415    .option     = options,
416    .version    = LIBAVUTIL_VERSION_INT,
417};
418
419static const FFCodecDefault xavs_defaults[] = {
420    { "b",                "0" },
421    { NULL },
422};
423
424const FFCodec ff_libxavs_encoder = {
425    .p.name         = "libxavs",
426    .p.long_name    = NULL_IF_CONFIG_SMALL("libxavs Chinese AVS (Audio Video Standard)"),
427    .p.type         = AVMEDIA_TYPE_VIDEO,
428    .p.id           = AV_CODEC_ID_CAVS,
429    .p.capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_DELAY |
430                      AV_CODEC_CAP_OTHER_THREADS,
431    .priv_data_size = sizeof(XavsContext),
432    .init           = XAVS_init,
433    FF_CODEC_ENCODE_CB(XAVS_frame),
434    .close          = XAVS_close,
435    .caps_internal  = FF_CODEC_CAP_AUTO_THREADS,
436    .p.pix_fmts     = (const enum AVPixelFormat[]) { AV_PIX_FMT_YUV420P, AV_PIX_FMT_NONE },
437    .p.priv_class   = &xavs_class,
438    .defaults       = xavs_defaults,
439    .p.wrapper_name = "libxavs",
440};
441