vpxenc.c 52.7 KB
Newer Older
John Koleszar's avatar
John Koleszar committed
1
/*
2
 *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
John Koleszar's avatar
John Koleszar committed
3
 *
4
 *  Use of this source code is governed by a BSD-style license
5
6
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
7
 *  in the file PATENTS.  All contributing project authors may
8
 *  be found in the AUTHORS file in the root of the source tree.
John Koleszar's avatar
John Koleszar committed
9
10
11
12
13
14
 */


/* This is a simple program that encodes YV12 files and generates ivf
 * files using the new interface.
 */
15
#if defined(_WIN32)
16
17
18
19
#define USE_POSIX_MMAP 0
#else
#define USE_POSIX_MMAP 1
#endif
John Koleszar's avatar
John Koleszar committed
20
21
22
23
24

#include <stdio.h>
#include <stdlib.h>
#include <stdarg.h>
#include <string.h>
John Koleszar's avatar
John Koleszar committed
25
#include <limits.h>
26
#include "vpx/vpx_encoder.h"
John Koleszar's avatar
John Koleszar committed
27
28
29
30
31
32
33
#if USE_POSIX_MMAP
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <fcntl.h>
#include <unistd.h>
#endif
John Koleszar's avatar
John Koleszar committed
34
#include "vpx_version.h"
35
#include "vpx/vp8cx.h"
John Koleszar's avatar
John Koleszar committed
36
37
#include "vpx_ports/mem_ops.h"
#include "vpx_ports/vpx_timer.h"
John Koleszar's avatar
John Koleszar committed
38
#include "tools_common.h"
39
#include "y4minput.h"
John Koleszar's avatar
John Koleszar committed
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
#include "libmkv/EbmlWriter.h"
#include "libmkv/EbmlIDs.h"

/* Need special handling of these functions on Windows */
#if defined(_MSC_VER)
/* MSVS doesn't define off_t, and uses _f{seek,tell}i64 */
typedef __int64 off_t;
#define fseeko _fseeki64
#define ftello _ftelli64
#elif defined(_WIN32)
/* MinGW defines off_t, and uses f{seek,tell}o64 */
#define fseeko fseeko64
#define ftello ftello64
#endif

#if defined(_MSC_VER)
#define LITERALU64(n) n
#else
#define LITERALU64(n) n##LLU
#endif
John Koleszar's avatar
John Koleszar committed
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80

static const char *exec_name;

static const struct codec_item
{
    char const              *name;
    const vpx_codec_iface_t *iface;
    unsigned int             fourcc;
} codecs[] =
{
#if CONFIG_VP8_ENCODER
    {"vp8",  &vpx_codec_vp8_cx_algo, 0x30385056},
#endif
};

static void usage_exit();

void die(const char *fmt, ...)
{
    va_list ap;
    va_start(ap, fmt);
81
82
    vfprintf(stderr, fmt, ap);
    fprintf(stderr, "\n");
John Koleszar's avatar
John Koleszar committed
83
84
85
86
87
88
89
90
91
    usage_exit();
}

static void ctx_exit_on_error(vpx_codec_ctx_t *ctx, const char *s)
{
    if (ctx->err)
    {
        const char *detail = vpx_codec_error_detail(ctx);

92
        fprintf(stderr, "%s: %s\n", s, vpx_codec_error(ctx));
John Koleszar's avatar
John Koleszar committed
93
94

        if (detail)
95
            fprintf(stderr, "    %s\n", detail);
John Koleszar's avatar
John Koleszar committed
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
148
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
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216

        exit(EXIT_FAILURE);
    }
}

/* This structure is used to abstract the different ways of handling
 * first pass statistics.
 */
typedef struct
{
    vpx_fixed_buf_t buf;
    int             pass;
    FILE           *file;
    char           *buf_ptr;
    size_t          buf_alloc_sz;
} stats_io_t;

int stats_open_file(stats_io_t *stats, const char *fpf, int pass)
{
    int res;

    stats->pass = pass;

    if (pass == 0)
    {
        stats->file = fopen(fpf, "wb");
        stats->buf.sz = 0;
        stats->buf.buf = NULL,
                   res = (stats->file != NULL);
    }
    else
    {
#if 0
#elif USE_POSIX_MMAP
        struct stat stat_buf;
        int fd;

        fd = open(fpf, O_RDONLY);
        stats->file = fdopen(fd, "rb");
        fstat(fd, &stat_buf);
        stats->buf.sz = stat_buf.st_size;
        stats->buf.buf = mmap(NULL, stats->buf.sz, PROT_READ, MAP_PRIVATE,
                              fd, 0);
        res = (stats->buf.buf != NULL);
#else
        size_t nbytes;

        stats->file = fopen(fpf, "rb");

        if (fseek(stats->file, 0, SEEK_END))
        {
            fprintf(stderr, "First-pass stats file must be seekable!\n");
            exit(EXIT_FAILURE);
        }

        stats->buf.sz = stats->buf_alloc_sz = ftell(stats->file);
        rewind(stats->file);

        stats->buf.buf = malloc(stats->buf_alloc_sz);

        if (!stats->buf.buf)
        {
            fprintf(stderr, "Failed to allocate first-pass stats buffer (%d bytes)\n",
                    stats->buf_alloc_sz);
            exit(EXIT_FAILURE);
        }

        nbytes = fread(stats->buf.buf, 1, stats->buf.sz, stats->file);
        res = (nbytes == stats->buf.sz);
#endif
    }

    return res;
}

int stats_open_mem(stats_io_t *stats, int pass)
{
    int res;
    stats->pass = pass;

    if (!pass)
    {
        stats->buf.sz = 0;
        stats->buf_alloc_sz = 64 * 1024;
        stats->buf.buf = malloc(stats->buf_alloc_sz);
    }

    stats->buf_ptr = stats->buf.buf;
    res = (stats->buf.buf != NULL);
    return res;
}


void stats_close(stats_io_t *stats)
{
    if (stats->file)
    {
        if (stats->pass == 1)
        {
#if 0
#elif USE_POSIX_MMAP
            munmap(stats->buf.buf, stats->buf.sz);
#else
            free(stats->buf.buf);
#endif
        }

        fclose(stats->file);
        stats->file = NULL;
    }
    else
    {
        if (stats->pass == 1)
            free(stats->buf.buf);
    }
}

void stats_write(stats_io_t *stats, const void *pkt, size_t len)
{
    if (stats->file)
    {
217
        if(fwrite(pkt, 1, len, stats->file));
John Koleszar's avatar
John Koleszar committed
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
    }
    else
    {
        if (stats->buf.sz + len > stats->buf_alloc_sz)
        {
            size_t  new_sz = stats->buf_alloc_sz + 64 * 1024;
            char   *new_ptr = realloc(stats->buf.buf, new_sz);

            if (new_ptr)
            {
                stats->buf_ptr = new_ptr + (stats->buf_ptr - (char *)stats->buf.buf);
                stats->buf.buf = new_ptr;
                stats->buf_alloc_sz = new_sz;
            } /* else ... */
        }

        memcpy(stats->buf_ptr, pkt, len);
        stats->buf.sz += len;
        stats->buf_ptr += len;
    }
}

vpx_fixed_buf_t stats_get(stats_io_t *stats)
{
    return stats->buf;
}

245
246
247
248
249
250
251
enum video_file_type
{
    FILE_TYPE_RAW,
    FILE_TYPE_IVF,
    FILE_TYPE_Y4M
};

252
253
struct detect_buffer {
    char buf[4];
254
255
    size_t buf_read;
    size_t position;
256
257
258
};


John Koleszar's avatar
John Koleszar committed
259
#define IVF_FRAME_HDR_SZ (4+8) /* 4 byte size + 8 byte timestamp */
260
static int read_frame(FILE *f, vpx_image_t *img, unsigned int file_type,
261
                      y4m_input *y4m, struct detect_buffer *detect)
John Koleszar's avatar
John Koleszar committed
262
263
{
    int plane = 0;
264
    int shortread = 0;
John Koleszar's avatar
John Koleszar committed
265

266
    if (file_type == FILE_TYPE_Y4M)
John Koleszar's avatar
John Koleszar committed
267
    {
268
        if (y4m_input_fetch_frame(y4m, f, img) < 1)
269
           return 0;
John Koleszar's avatar
John Koleszar committed
270
    }
271
    else
John Koleszar's avatar
John Koleszar committed
272
    {
273
        if (file_type == FILE_TYPE_IVF)
John Koleszar's avatar
John Koleszar committed
274
        {
275
276
277
278
279
280
            char junk[IVF_FRAME_HDR_SZ];

            /* Skip the frame header. We know how big the frame should be. See
             * write_ivf_frame_header() for documentation on the frame header
             * layout.
             */
281
            if(fread(junk, 1, IVF_FRAME_HDR_SZ, f));
John Koleszar's avatar
John Koleszar committed
282
283
        }

284
        for (plane = 0; plane < 3; plane++)
John Koleszar's avatar
John Koleszar committed
285
        {
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
            unsigned char *ptr;
            int w = (plane ? (1 + img->d_w) / 2 : img->d_w);
            int h = (plane ? (1 + img->d_h) / 2 : img->d_h);
            int r;

            /* Determine the correct plane based on the image format. The for-loop
             * always counts in Y,U,V order, but this may not match the order of
             * the data on disk.
             */
            switch (plane)
            {
            case 1:
                ptr = img->planes[img->fmt==VPX_IMG_FMT_YV12? VPX_PLANE_V : VPX_PLANE_U];
                break;
            case 2:
                ptr = img->planes[img->fmt==VPX_IMG_FMT_YV12?VPX_PLANE_U : VPX_PLANE_V];
                break;
            default:
                ptr = img->planes[plane];
            }

            for (r = 0; r < h; r++)
            {
309
310
311
312
                size_t needed = w;
                size_t buf_position = 0;
                const size_t left = detect->buf_read - detect->position;
                if (left > 0)
313
                {
314
315
316
317
318
319
320
321
322
                    const size_t more = (left < needed) ? left : needed;
                    memcpy(ptr, detect->buf + detect->position, more);
                    buf_position = more;
                    needed -= more;
                    detect->position += more;
                }
                if (needed > 0)
                {
                    shortread |= (fread(ptr + buf_position, 1, needed, f) < needed);
323
324
                }

325
326
                ptr += img->stride[plane];
            }
John Koleszar's avatar
John Koleszar committed
327
328
329
        }
    }

330
    return !shortread;
John Koleszar's avatar
John Koleszar committed
331
332
333
}


334
unsigned int file_is_y4m(FILE      *infile,
335
336
                         y4m_input *y4m,
                         char       detect[4])
337
{
338
    if(memcmp(detect, "YUV4", 4) == 0)
339
340
341
342
343
344
    {
        return 1;
    }
    return 0;
}

John Koleszar's avatar
John Koleszar committed
345
346
347
348
#define IVF_FILE_HDR_SZ (32)
unsigned int file_is_ivf(FILE *infile,
                         unsigned int *fourcc,
                         unsigned int *width,
349
                         unsigned int *height,
350
                         struct detect_buffer *detect)
John Koleszar's avatar
John Koleszar committed
351
352
353
354
{
    char raw_hdr[IVF_FILE_HDR_SZ];
    int is_ivf = 0;

355
    if(memcmp(detect->buf, "DKIF", 4) != 0)
356
357
        return 0;

John Koleszar's avatar
John Koleszar committed
358
359
360
    /* See write_ivf_file_header() for more documentation on the file header
     * layout.
     */
361
362
    if (fread(raw_hdr + 4, 1, IVF_FILE_HDR_SZ - 4, infile)
        == IVF_FILE_HDR_SZ - 4)
John Koleszar's avatar
John Koleszar committed
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
    {
        {
            is_ivf = 1;

            if (mem_get_le16(raw_hdr + 4) != 0)
                fprintf(stderr, "Error: Unrecognized IVF version! This file may not"
                        " decode properly.");

            *fourcc = mem_get_le32(raw_hdr + 8);
        }
    }

    if (is_ivf)
    {
        *width = mem_get_le16(raw_hdr + 12);
        *height = mem_get_le16(raw_hdr + 14);
379
        detect->position = 4;
John Koleszar's avatar
John Koleszar committed
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
    }

    return is_ivf;
}


static void write_ivf_file_header(FILE *outfile,
                                  const vpx_codec_enc_cfg_t *cfg,
                                  unsigned int fourcc,
                                  int frame_cnt)
{
    char header[32];

    if (cfg->g_pass != VPX_RC_ONE_PASS && cfg->g_pass != VPX_RC_LAST_PASS)
        return;

    header[0] = 'D';
    header[1] = 'K';
    header[2] = 'I';
    header[3] = 'F';
    mem_put_le16(header + 4,  0);                 /* version */
    mem_put_le16(header + 6,  32);                /* headersize */
    mem_put_le32(header + 8,  fourcc);            /* headersize */
    mem_put_le16(header + 12, cfg->g_w);          /* width */
    mem_put_le16(header + 14, cfg->g_h);          /* height */
    mem_put_le32(header + 16, cfg->g_timebase.den); /* rate */
    mem_put_le32(header + 20, cfg->g_timebase.num); /* scale */
    mem_put_le32(header + 24, frame_cnt);         /* length */
    mem_put_le32(header + 28, 0);                 /* unused */

410
    if(fwrite(header, 1, 32, outfile));
John Koleszar's avatar
John Koleszar committed
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
}


static void write_ivf_frame_header(FILE *outfile,
                                   const vpx_codec_cx_pkt_t *pkt)
{
    char             header[12];
    vpx_codec_pts_t  pts;

    if (pkt->kind != VPX_CODEC_CX_FRAME_PKT)
        return;

    pts = pkt->data.frame.pts;
    mem_put_le32(header, pkt->data.frame.sz);
    mem_put_le32(header + 4, pts & 0xFFFFFFFF);
    mem_put_le32(header + 8, pts >> 32);

428
    if(fwrite(header, 1, 12, outfile));
John Koleszar's avatar
John Koleszar committed
429
430
}

John Koleszar's avatar
John Koleszar committed
431
432
433
434
435
436
437
438
439
440
441
442
443

typedef off_t EbmlLoc;


struct cue_entry
{
    unsigned int time;
    uint64_t     loc;
};


struct EbmlGlobal
{
444
445
    int debug;

John Koleszar's avatar
John Koleszar committed
446
    FILE    *stream;
447
    int64_t last_pts_ms;
John Koleszar's avatar
John Koleszar committed
448
449
450
451
452
453
454
455
456
457
    vpx_rational_t  framerate;

    /* These pointers are to the start of an element */
    off_t    position_reference;
    off_t    seek_info_pos;
    off_t    segment_info_pos;
    off_t    track_pos;
    off_t    cue_pos;
    off_t    cluster_pos;

John Koleszar's avatar
John Koleszar committed
458
459
460
    /* This pointer is to a specific element to be serialized */
    off_t    track_id_pos;

John Koleszar's avatar
John Koleszar committed
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
    /* These pointers are to the size field of the element */
    EbmlLoc  startSegment;
    EbmlLoc  startCluster;

    uint32_t cluster_timecode;
    int      cluster_open;

    struct cue_entry *cue_list;
    unsigned int      cues;

};


void Ebml_Write(EbmlGlobal *glob, const void *buffer_in, unsigned long len)
{
476
    if(fwrite(buffer_in, 1, len, glob->stream));
John Koleszar's avatar
John Koleszar committed
477
478
479
480
481
482
483
484
485
486
487
488
}


void Ebml_Serialize(EbmlGlobal *glob, const void *buffer_in, unsigned long len)
{
    const unsigned char *q = (const unsigned char *)buffer_in + len - 1;

    for(; len; len--)
        Ebml_Write(glob, q--, 1);
}


John Koleszar's avatar
John Koleszar committed
489
490
491
492
493
494
495
496
497
498
499
500
/* Need a fixed size serializer for the track ID. libmkv provdes a 64 bit
 * one, but not a 32 bit one.
 */
static void Ebml_SerializeUnsigned32(EbmlGlobal *glob, unsigned long class_id, uint64_t ui)
{
    unsigned char sizeSerialized = 4 | 0x80;
    Ebml_WriteID(glob, class_id);
    Ebml_Serialize(glob, &sizeSerialized, 1);
    Ebml_Serialize(glob, &ui, 4);
}


John Koleszar's avatar
John Koleszar committed
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
static void
Ebml_StartSubElement(EbmlGlobal *glob, EbmlLoc *ebmlLoc,
                          unsigned long class_id)
{
    //todo this is always taking 8 bytes, this may need later optimization
    //this is a key that says lenght unknown
    unsigned long long unknownLen =  LITERALU64(0x01FFFFFFFFFFFFFF);

    Ebml_WriteID(glob, class_id);
    *ebmlLoc = ftello(glob->stream);
    Ebml_Serialize(glob, &unknownLen, 8);
}

static void
Ebml_EndSubElement(EbmlGlobal *glob, EbmlLoc *ebmlLoc)
{
    off_t pos;
    uint64_t size;

    /* Save the current stream pointer */
    pos = ftello(glob->stream);

    /* Calculate the size of this element */
    size = pos - *ebmlLoc - 8;
    size |=  LITERALU64(0x0100000000000000);

    /* Seek back to the beginning of the element and write the new size */
    fseeko(glob->stream, *ebmlLoc, SEEK_SET);
    Ebml_Serialize(glob, &size, 8);

    /* Reset the stream pointer */
    fseeko(glob->stream, pos, SEEK_SET);
}


static void
write_webm_seek_element(EbmlGlobal *ebml, unsigned long id, off_t pos)
{
    uint64_t offset = pos - ebml->position_reference;
    EbmlLoc start;
    Ebml_StartSubElement(ebml, &start, Seek);
    Ebml_SerializeBinary(ebml, SeekID, id);
    Ebml_SerializeUnsigned64(ebml, SeekPosition, offset);
    Ebml_EndSubElement(ebml, &start);
}


static void
write_webm_seek_info(EbmlGlobal *ebml)
{

    off_t pos;

    /* Save the current stream pointer */
    pos = ftello(ebml->stream);

    if(ebml->seek_info_pos)
        fseeko(ebml->stream, ebml->seek_info_pos, SEEK_SET);
    else
        ebml->seek_info_pos = pos;

    {
        EbmlLoc start;

        Ebml_StartSubElement(ebml, &start, SeekHead);
        write_webm_seek_element(ebml, Tracks, ebml->track_pos);
        write_webm_seek_element(ebml, Cues,   ebml->cue_pos);
        write_webm_seek_element(ebml, Info,   ebml->segment_info_pos);
        Ebml_EndSubElement(ebml, &start);
    }
    {
        //segment info
        EbmlLoc startInfo;
        uint64_t frame_time;

        frame_time = (uint64_t)1000 * ebml->framerate.den
                     / ebml->framerate.num;
        ebml->segment_info_pos = ftello(ebml->stream);
        Ebml_StartSubElement(ebml, &startInfo, Info);
        Ebml_SerializeUnsigned(ebml, TimecodeScale, 1000000);
        Ebml_SerializeFloat(ebml, Segment_Duration,
                            ebml->last_pts_ms + frame_time);
583
584
585
586
        Ebml_SerializeString(ebml, 0x4D80,
            ebml->debug ? "vpxenc" : "vpxenc" VERSION_STRING);
        Ebml_SerializeString(ebml, 0x5741,
            ebml->debug ? "vpxenc" : "vpxenc" VERSION_STRING);
John Koleszar's avatar
John Koleszar committed
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
        Ebml_EndSubElement(ebml, &startInfo);
    }
}


static void
write_webm_file_header(EbmlGlobal                *glob,
                       const vpx_codec_enc_cfg_t *cfg,
                       const struct vpx_rational *fps)
{
    {
        EbmlLoc start;
        Ebml_StartSubElement(glob, &start, EBML);
        Ebml_SerializeUnsigned(glob, EBMLVersion, 1);
        Ebml_SerializeUnsigned(glob, EBMLReadVersion, 1); //EBML Read Version
        Ebml_SerializeUnsigned(glob, EBMLMaxIDLength, 4); //EBML Max ID Length
        Ebml_SerializeUnsigned(glob, EBMLMaxSizeLength, 8); //EBML Max Size Length
        Ebml_SerializeString(glob, DocType, "webm"); //Doc Type
        Ebml_SerializeUnsigned(glob, DocTypeVersion, 2); //Doc Type Version
        Ebml_SerializeUnsigned(glob, DocTypeReadVersion, 2); //Doc Type Read Version
        Ebml_EndSubElement(glob, &start);
    }
    {
        Ebml_StartSubElement(glob, &glob->startSegment, Segment); //segment
        glob->position_reference = ftello(glob->stream);
        glob->framerate = *fps;
        write_webm_seek_info(glob);

        {
            EbmlLoc trackStart;
            glob->track_pos = ftello(glob->stream);
            Ebml_StartSubElement(glob, &trackStart, Tracks);
            {
                unsigned int trackNumber = 1;
                uint64_t     trackID = 0;

                EbmlLoc start;
                Ebml_StartSubElement(glob, &start, TrackEntry);
                Ebml_SerializeUnsigned(glob, TrackNumber, trackNumber);
John Koleszar's avatar
John Koleszar committed
626
627
                glob->track_id_pos = ftello(glob->stream);
                Ebml_SerializeUnsigned32(glob, TrackUID, trackID);
John Koleszar's avatar
John Koleszar committed
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
                Ebml_SerializeUnsigned(glob, TrackType, 1); //video is always 1
                Ebml_SerializeString(glob, CodecID, "V_VP8");
                {
                    unsigned int pixelWidth = cfg->g_w;
                    unsigned int pixelHeight = cfg->g_h;
                    float        frameRate   = (float)fps->num/(float)fps->den;

                    EbmlLoc videoStart;
                    Ebml_StartSubElement(glob, &videoStart, Video);
                    Ebml_SerializeUnsigned(glob, PixelWidth, pixelWidth);
                    Ebml_SerializeUnsigned(glob, PixelHeight, pixelHeight);
                    Ebml_SerializeFloat(glob, FrameRate, frameRate);
                    Ebml_EndSubElement(glob, &videoStart); //Video
                }
                Ebml_EndSubElement(glob, &start); //Track Entry
            }
            Ebml_EndSubElement(glob, &trackStart);
        }
        // segment element is open
    }
}


static void
write_webm_block(EbmlGlobal                *glob,
                 const vpx_codec_enc_cfg_t *cfg,
                 const vpx_codec_cx_pkt_t  *pkt)
{
    unsigned long  block_length;
    unsigned char  track_number;
    unsigned short block_timecode = 0;
    unsigned char  flags;
660
    int64_t        pts_ms;
John Koleszar's avatar
John Koleszar committed
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
    int            start_cluster = 0, is_keyframe;

    /* Calculate the PTS of this frame in milliseconds */
    pts_ms = pkt->data.frame.pts * 1000
             * (uint64_t)cfg->g_timebase.num / (uint64_t)cfg->g_timebase.den;
    if(pts_ms <= glob->last_pts_ms)
        pts_ms = glob->last_pts_ms + 1;
    glob->last_pts_ms = pts_ms;

    /* Calculate the relative time of this block */
    if(pts_ms - glob->cluster_timecode > SHRT_MAX)
        start_cluster = 1;
    else
        block_timecode = pts_ms - glob->cluster_timecode;

    is_keyframe = (pkt->data.frame.flags & VPX_FRAME_IS_KEY);
    if(start_cluster || is_keyframe)
    {
        if(glob->cluster_open)
            Ebml_EndSubElement(glob, &glob->startCluster);

        /* Open the new cluster */
        block_timecode = 0;
        glob->cluster_open = 1;
        glob->cluster_timecode = pts_ms;
        glob->cluster_pos = ftello(glob->stream);
        Ebml_StartSubElement(glob, &glob->startCluster, Cluster); //cluster
        Ebml_SerializeUnsigned(glob, Timecode, glob->cluster_timecode);

        /* Save a cue point if this is a keyframe. */
        if(is_keyframe)
        {
            struct cue_entry *cue;

            glob->cue_list = realloc(glob->cue_list,
                                     (glob->cues+1) * sizeof(struct cue_entry));
            cue = &glob->cue_list[glob->cues];
            cue->time = glob->cluster_timecode;
            cue->loc = glob->cluster_pos;
            glob->cues++;
        }
    }

    /* Write the Simple Block */
    Ebml_WriteID(glob, SimpleBlock);

    block_length = pkt->data.frame.sz + 4;
    block_length |= 0x10000000;
    Ebml_Serialize(glob, &block_length, 4);

    track_number = 1;
    track_number |= 0x80;
    Ebml_Write(glob, &track_number, 1);

    Ebml_Serialize(glob, &block_timecode, 2);

    flags = 0;
    if(is_keyframe)
        flags |= 0x80;
    if(pkt->data.frame.flags & VPX_FRAME_IS_INVISIBLE)
        flags |= 0x08;
    Ebml_Write(glob, &flags, 1);

    Ebml_Write(glob, pkt->data.frame.buf, pkt->data.frame.sz);
}


static void
John Koleszar's avatar
John Koleszar committed
729
write_webm_file_footer(EbmlGlobal *glob, long hash)
John Koleszar's avatar
John Koleszar committed
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
{

    if(glob->cluster_open)
        Ebml_EndSubElement(glob, &glob->startCluster);

    {
        EbmlLoc start;
        int i;

        glob->cue_pos = ftello(glob->stream);
        Ebml_StartSubElement(glob, &start, Cues);
        for(i=0; i<glob->cues; i++)
        {
            struct cue_entry *cue = &glob->cue_list[i];
            EbmlLoc start;

            Ebml_StartSubElement(glob, &start, CuePoint);
            {
                EbmlLoc start;

                Ebml_SerializeUnsigned(glob, CueTime, cue->time);

                Ebml_StartSubElement(glob, &start, CueTrackPositions);
                Ebml_SerializeUnsigned(glob, CueTrack, 1);
                Ebml_SerializeUnsigned64(glob, CueClusterPosition,
                                         cue->loc - glob->position_reference);
                //Ebml_SerializeUnsigned(glob, CueBlockNumber, cue->blockNumber);
                Ebml_EndSubElement(glob, &start);
            }
            Ebml_EndSubElement(glob, &start);
        }
        Ebml_EndSubElement(glob, &start);
    }

    Ebml_EndSubElement(glob, &glob->startSegment);

    /* Patch up the seek info block */
    write_webm_seek_info(glob);
John Koleszar's avatar
John Koleszar committed
768
769
770
771
772

    /* Patch up the track id */
    fseeko(glob->stream, glob->track_id_pos, SEEK_SET);
    Ebml_SerializeUnsigned32(glob, TrackUID, glob->debug ? 0xDEADBEEF : hash);

John Koleszar's avatar
John Koleszar committed
773
774
775
776
    fseeko(glob->stream, 0, SEEK_END);
}


John Koleszar's avatar
John Koleszar committed
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
/* Murmur hash derived from public domain reference implementation at
 *   http://sites.google.com/site/murmurhash/
 */
static unsigned int murmur ( const void * key, int len, unsigned int seed )
{
    const unsigned int m = 0x5bd1e995;
    const int r = 24;

    unsigned int h = seed ^ len;

    const unsigned char * data = (const unsigned char *)key;

    while(len >= 4)
    {
        unsigned int k;

        k  = data[0];
        k |= data[1] << 8;
        k |= data[2] << 16;
        k |= data[3] << 24;

        k *= m;
        k ^= k >> r;
        k *= m;

        h *= m;
        h ^= k;

        data += 4;
        len -= 4;
    }

    switch(len)
    {
    case 3: h ^= data[2] << 16;
    case 2: h ^= data[1] << 8;
    case 1: h ^= data[0];
            h *= m;
    };

    h ^= h >> 13;
    h *= m;
    h ^= h >> 15;

    return h;
}

824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
#include "math.h"

static double vp8_mse2psnr(double Samples, double Peak, double Mse)
{
    double psnr;

    if ((double)Mse > 0.0)
        psnr = 10.0 * log10(Peak * Peak * Samples / Mse);
    else
        psnr = 60;      // Limit to prevent / 0

    if (psnr > 60)
        psnr = 60;

    return psnr;
}

John Koleszar's avatar
John Koleszar committed
841

John Koleszar's avatar
John Koleszar committed
842
843
#include "args.h"

844
845
static const arg_def_t debugmode = ARG_DEF("D", "debug", 0,
        "Debug mode (makes output deterministic)");
846
847
static const arg_def_t outputfile = ARG_DEF("o", "output", 1,
        "Output filename");
John Koleszar's avatar
John Koleszar committed
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
static const arg_def_t use_yv12 = ARG_DEF(NULL, "yv12", 0,
                                  "Input file is YV12 ");
static const arg_def_t use_i420 = ARG_DEF(NULL, "i420", 0,
                                  "Input file is I420 (default)");
static const arg_def_t codecarg = ARG_DEF(NULL, "codec", 1,
                                  "Codec to use");
static const arg_def_t passes           = ARG_DEF("p", "passes", 1,
        "Number of passes (1/2)");
static const arg_def_t pass_arg         = ARG_DEF(NULL, "pass", 1,
        "Pass to execute (1/2)");
static const arg_def_t fpf_name         = ARG_DEF(NULL, "fpf", 1,
        "First pass statistics file name");
static const arg_def_t limit = ARG_DEF(NULL, "limit", 1,
                                       "Stop encoding after n input frames");
static const arg_def_t deadline         = ARG_DEF("d", "deadline", 1,
        "Deadline per frame (usec)");
static const arg_def_t best_dl          = ARG_DEF(NULL, "best", 0,
        "Use Best Quality Deadline");
static const arg_def_t good_dl          = ARG_DEF(NULL, "good", 0,
        "Use Good Quality Deadline");
static const arg_def_t rt_dl            = ARG_DEF(NULL, "rt", 0,
        "Use Realtime Quality Deadline");
static const arg_def_t verbosearg       = ARG_DEF("v", "verbose", 0,
        "Show encoder parameters");
static const arg_def_t psnrarg          = ARG_DEF(NULL, "psnr", 0,
        "Show PSNR in status line");
874
static const arg_def_t framerate        = ARG_DEF(NULL, "fps", 1,
875
        "Stream frame rate (rate/scale)");
John Koleszar's avatar
John Koleszar committed
876
877
static const arg_def_t use_ivf          = ARG_DEF(NULL, "ivf", 0,
        "Output IVF (default is WebM)");
John Koleszar's avatar
John Koleszar committed
878
879
static const arg_def_t *main_args[] =
{
880
    &debugmode,
881
882
    &outputfile, &codecarg, &passes, &pass_arg, &fpf_name, &limit, &deadline,
    &best_dl, &good_dl, &rt_dl,
John Koleszar's avatar
John Koleszar committed
883
    &verbosearg, &psnrarg, &use_ivf, &framerate,
John Koleszar's avatar
John Koleszar committed
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
    NULL
};

static const arg_def_t usage            = ARG_DEF("u", "usage", 1,
        "Usage profile number to use");
static const arg_def_t threads          = ARG_DEF("t", "threads", 1,
        "Max number of threads to use");
static const arg_def_t profile          = ARG_DEF(NULL, "profile", 1,
        "Bitstream profile number to use");
static const arg_def_t width            = ARG_DEF("w", "width", 1,
        "Frame width");
static const arg_def_t height           = ARG_DEF("h", "height", 1,
        "Frame height");
static const arg_def_t timebase         = ARG_DEF(NULL, "timebase", 1,
        "Stream timebase (frame duration)");
static const arg_def_t error_resilient  = ARG_DEF(NULL, "error-resilient", 1,
        "Enable error resiliency features");
static const arg_def_t lag_in_frames    = ARG_DEF(NULL, "lag-in-frames", 1,
        "Max number of frames to lag");

static const arg_def_t *global_args[] =
{
    &use_yv12, &use_i420, &usage, &threads, &profile,
907
    &width, &height, &timebase, &framerate, &error_resilient,
John Koleszar's avatar
John Koleszar committed
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
    &lag_in_frames, NULL
};

static const arg_def_t dropframe_thresh   = ARG_DEF(NULL, "drop-frame", 1,
        "Temporal resampling threshold (buf %)");
static const arg_def_t resize_allowed     = ARG_DEF(NULL, "resize-allowed", 1,
        "Spatial resampling enabled (bool)");
static const arg_def_t resize_up_thresh   = ARG_DEF(NULL, "resize-up", 1,
        "Upscale threshold (buf %)");
static const arg_def_t resize_down_thresh = ARG_DEF(NULL, "resize-down", 1,
        "Downscale threshold (buf %)");
static const arg_def_t end_usage          = ARG_DEF(NULL, "end-usage", 1,
        "VBR=0 | CBR=1");
static const arg_def_t target_bitrate     = ARG_DEF(NULL, "target-bitrate", 1,
        "Bitrate (kbps)");
static const arg_def_t min_quantizer      = ARG_DEF(NULL, "min-q", 1,
        "Minimum (best) quantizer");
static const arg_def_t max_quantizer      = ARG_DEF(NULL, "max-q", 1,
        "Maximum (worst) quantizer");
static const arg_def_t undershoot_pct     = ARG_DEF(NULL, "undershoot-pct", 1,
        "Datarate undershoot (min) target (%)");
static const arg_def_t overshoot_pct      = ARG_DEF(NULL, "overshoot-pct", 1,
        "Datarate overshoot (max) target (%)");
static const arg_def_t buf_sz             = ARG_DEF(NULL, "buf-sz", 1,
        "Client buffer size (ms)");
static const arg_def_t buf_initial_sz     = ARG_DEF(NULL, "buf-initial-sz", 1,
        "Client initial buffer size (ms)");
static const arg_def_t buf_optimal_sz     = ARG_DEF(NULL, "buf-optimal-sz", 1,
        "Client optimal buffer size (ms)");
static const arg_def_t *rc_args[] =
{
    &dropframe_thresh, &resize_allowed, &resize_up_thresh, &resize_down_thresh,
    &end_usage, &target_bitrate, &min_quantizer, &max_quantizer,
    &undershoot_pct, &overshoot_pct, &buf_sz, &buf_initial_sz, &buf_optimal_sz,
    NULL
};


static const arg_def_t bias_pct = ARG_DEF(NULL, "bias-pct", 1,
                                  "CBR/VBR bias (0=CBR, 100=VBR)");
static const arg_def_t minsection_pct = ARG_DEF(NULL, "minsection-pct", 1,
                                        "GOP min bitrate (% of target)");
static const arg_def_t maxsection_pct = ARG_DEF(NULL, "maxsection-pct", 1,
                                        "GOP max bitrate (% of target)");
static const arg_def_t *rc_twopass_args[] =
{
    &bias_pct, &minsection_pct, &maxsection_pct, NULL
};


static const arg_def_t kf_min_dist = ARG_DEF(NULL, "kf-min-dist", 1,
                                     "Minimum keyframe interval (frames)");
static const arg_def_t kf_max_dist = ARG_DEF(NULL, "kf-max-dist", 1,
                                     "Maximum keyframe interval (frames)");
962
963
static const arg_def_t kf_disabled = ARG_DEF(NULL, "disable-kf", 0,
                                     "Disable keyframe placement");
John Koleszar's avatar
John Koleszar committed
964
965
static const arg_def_t *kf_args[] =
{
966
    &kf_min_dist, &kf_max_dist, &kf_disabled, NULL
John Koleszar's avatar
John Koleszar committed
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
};


#if CONFIG_VP8_ENCODER
static const arg_def_t noise_sens = ARG_DEF(NULL, "noise-sensitivity", 1,
                                    "Noise sensitivity (frames to blur)");
static const arg_def_t sharpness = ARG_DEF(NULL, "sharpness", 1,
                                   "Filter sharpness (0-7)");
static const arg_def_t static_thresh = ARG_DEF(NULL, "static-thresh", 1,
                                       "Motion detection threshold");
#endif

#if CONFIG_VP8_ENCODER
static const arg_def_t cpu_used = ARG_DEF(NULL, "cpu-used", 1,
                                  "CPU Used (-16..16)");
#endif


#if CONFIG_VP8_ENCODER
static const arg_def_t token_parts = ARG_DEF(NULL, "token-parts", 1,
                                     "Number of token partitions to use, log2");
static const arg_def_t auto_altref = ARG_DEF(NULL, "auto-alt-ref", 1,
                                     "Enable automatic alt reference frames");
static const arg_def_t arnr_maxframes = ARG_DEF(NULL, "arnr-maxframes", 1,
991
                                        "AltRef Max Frames");
John Koleszar's avatar
John Koleszar committed
992
static const arg_def_t arnr_strength = ARG_DEF(NULL, "arnr-strength", 1,
993
                                       "AltRef Strength");
John Koleszar's avatar
John Koleszar committed
994
static const arg_def_t arnr_type = ARG_DEF(NULL, "arnr-type", 1,
995
                                   "AltRef Type");
John Koleszar's avatar
John Koleszar committed
996
997
998
999
1000
1001
1002
static const struct arg_enum_list tuning_enum[] = {
    {"psnr", VP8_TUNE_PSNR},
    {"ssim", VP8_TUNE_SSIM},
    {NULL, 0}
};
static const arg_def_t tune_ssim = ARG_DEF_ENUM(NULL, "tune", 1,
                                   "Material to favor", tuning_enum);
John Koleszar's avatar
John Koleszar committed
1003
1004
1005
1006

static const arg_def_t *vp8_args[] =
{
    &cpu_used, &auto_altref, &noise_sens, &sharpness, &static_thresh,
John Koleszar's avatar
John Koleszar committed
1007
1008
    &token_parts, &arnr_maxframes, &arnr_strength, &arnr_type,
    &tune_ssim, NULL
John Koleszar's avatar
John Koleszar committed
1009
1010
1011
1012
1013
1014
};
static const int vp8_arg_ctrl_map[] =
{
    VP8E_SET_CPUUSED, VP8E_SET_ENABLEAUTOALTREF,
    VP8E_SET_NOISE_SENSITIVITY, VP8E_SET_SHARPNESS, VP8E_SET_STATIC_THRESHOLD,
    VP8E_SET_TOKEN_PARTITIONS,
John Koleszar's avatar
John Koleszar committed
1015
1016
    VP8E_SET_ARNR_MAXFRAMES, VP8E_SET_ARNR_STRENGTH , VP8E_SET_ARNR_TYPE,
    VP8E_SET_TUNING, 0
John Koleszar's avatar
John Koleszar committed
1017
1018
1019
1020
1021
1022
1023
1024
1025
};
#endif

static const arg_def_t *no_args[] = { NULL };

static void usage_exit()
{
    int i;

1026
1027
    fprintf(stderr, "Usage: %s <options> -o dst_filename src_filename \n",
            exec_name);
John Koleszar's avatar
John Koleszar committed
1028

1029
    fprintf(stderr, "\nOptions:\n");
John Koleszar's avatar
John Koleszar committed
1030
    arg_show_usage(stdout, main_args);
1031
    fprintf(stderr, "\nEncoder Global Options:\n");
John Koleszar's avatar
John Koleszar committed
1032
    arg_show_usage(stdout, global_args);
1033
    fprintf(stderr, "\nRate Control Options:\n");
John Koleszar's avatar
John Koleszar committed
1034
    arg_show_usage(stdout, rc_args);
1035
    fprintf(stderr, "\nTwopass Rate Control Options:\n");
John Koleszar's avatar
John Koleszar committed
1036
    arg_show_usage(stdout, rc_twopass_args);
1037
    fprintf(stderr, "\nKeyframe Placement Options:\n");
John Koleszar's avatar
John Koleszar committed
1038
1039
    arg_show_usage(stdout, kf_args);
#if CONFIG_VP8_ENCODER
1040
    fprintf(stderr, "\nVP8 Specific Options:\n");
John Koleszar's avatar
John Koleszar committed
1041
1042
    arg_show_usage(stdout, vp8_args);
#endif
1043
    fprintf(stderr, "\n"
John Koleszar's avatar
John Koleszar committed
1044
1045
1046
1047
           "Included encoders:\n"
           "\n");

    for (i = 0; i < sizeof(codecs) / sizeof(codecs[0]); i++)
1048
        fprintf(stderr, "    %-6s - %s\n",
John Koleszar's avatar
John Koleszar committed
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
               codecs[i].name,
               vpx_codec_iface_name(codecs[i].iface));

    exit(EXIT_FAILURE);
}

#define ARG_CTRL_CNT_MAX 10


int main(int argc, const char **argv_)
{
    vpx_codec_ctx_t        encoder;
    const char                  *in_fn = NULL, *out_fn = NULL, *stats_fn = NULL;
    int                    i;
    FILE                  *infile, *outfile;
    vpx_codec_enc_cfg_t    cfg;
    vpx_codec_err_t        res;
    int                    pass, one_pass_only = 0;
    stats_io_t             stats;
    vpx_image_t            raw;
    const struct codec_item  *codec = codecs;
    int                    frame_avail, got_data;

    struct arg               arg;
    char                   **argv, **argi, **argj;
    int                      arg_usage = 0, arg_passes = 1, arg_deadline = 0;
    int                      arg_ctrls[ARG_CTRL_CNT_MAX][2], arg_ctrl_cnt = 0;
    int                      arg_limit = 0;
    static const arg_def_t **ctrl_args = no_args;
    static const int        *ctrl_args_map = NULL;
    int                      verbose = 0, show_psnr = 0;
    int                      arg_use_i420 = 1;
    unsigned long            cx_time = 0;
1082
1083
    unsigned int             file_type, fourcc;
    y4m_input                y4m;
1084
1085
    struct vpx_rational      arg_framerate = {30, 1};
    int                      arg_have_framerate = 0;
John Koleszar's avatar
John Koleszar committed
1086
    int                      write_webm = 1;
John Koleszar's avatar
John Koleszar committed
1087
    EbmlGlobal               ebml = {0};
John Koleszar's avatar
John Koleszar committed
1088
    uint32_t                 hash = 0;
1089
1090
1091
1092
    uint64_t                 psnr_sse_total = 0;
    uint64_t                 psnr_samples_total = 0;
    double                   psnr_totals[4] = {0, 0, 0, 0};
    int                      psnr_count = 0;
John Koleszar's avatar
John Koleszar committed
1093
1094

    exec_name = argv_[0];
1095
    ebml.last_pts_ms = -1;
John Koleszar's avatar
John Koleszar committed
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164

    if (argc < 3)
        usage_exit();


    /* First parse the codec and usage values, because we want to apply other
     * parameters on top of the default configuration provided by the codec.
     */
    argv = argv_dup(argc - 1, argv_ + 1);

    for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
    {
        arg.argv_step = 1;

        if (arg_match(&arg, &codecarg, argi))
        {
            int j, k = -1;

            for (j = 0; j < sizeof(codecs) / sizeof(codecs[0]); j++)
                if (!strcmp(codecs[j].name, arg.val))
                    k = j;

            if (k >= 0)
                codec = codecs + k;
            else
                die("Error: Unrecognized argument (%s) to --codec\n",
                    arg.val);

        }
        else if (arg_match(&arg, &passes, argi))
        {
            arg_passes = arg_parse_uint(&arg);

            if (arg_passes < 1 || arg_passes > 2)
                die("Error: Invalid number of passes (%d)\n", arg_passes);
        }
        else if (arg_match(&arg, &pass_arg, argi))
        {
            one_pass_only = arg_parse_uint(&arg);

            if (one_pass_only < 1 || one_pass_only > 2)
                die("Error: Invalid pass selected (%d)\n", one_pass_only);
        }
        else if (arg_match(&arg, &fpf_name, argi))
            stats_fn = arg.val;
        else if (arg_match(&arg, &usage, argi))
            arg_usage = arg_parse_uint(&arg);
        else if (arg_match(&arg, &deadline, argi))
            arg_deadline = arg_parse_uint(&arg);
        else if (arg_match(&arg, &best_dl, argi))
            arg_deadline = VPX_DL_BEST_QUALITY;
        else if (arg_match(&arg, &good_dl, argi))
            arg_deadline = VPX_DL_GOOD_QUALITY;
        else if (arg_match(&arg, &rt_dl, argi))
            arg_deadline = VPX_DL_REALTIME;
        else if (arg_match(&arg, &use_yv12, argi))
        {
            arg_use_i420 = 0;
        }
        else if (arg_match(&arg, &use_i420, argi))
        {
            arg_use_i420 = 1;
        }
        else if (arg_match(&arg, &verbosearg, argi))
            verbose = 1;
        else if (arg_match(&arg, &limit, argi))
            arg_limit = arg_parse_uint(&arg);
        else if (arg_match(&arg, &psnrarg, argi))
            show_psnr = 1;
1165
1166
1167
1168
1169
        else if (arg_match(&arg, &framerate, argi))
        {
            arg_framerate = arg_parse_rational(&arg);
            arg_have_framerate = 1;
        }
John Koleszar's avatar
John Koleszar committed
1170
1171
        else if (arg_match(&arg, &use_ivf, argi))
            write_webm = 0;
1172
1173
        else if (arg_match(&arg, &outputfile, argi))
            out_fn = arg.val;
1174
1175
        else if (arg_match(&arg, &debugmode, argi))
            ebml.debug = 1;
John Koleszar's avatar
John Koleszar committed
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
        else
            argj++;
    }

    /* Ensure that --passes and --pass are consistent. If --pass is set and --passes=2,
     * ensure --fpf was set.
     */
    if (one_pass_only)
    {
        /* DWIM: Assume the user meant passes=2 if pass=2 is specified */
        if (one_pass_only > arg_passes)
        {
1188
            fprintf(stderr, "Warning: Assuming --pass=%d implies --passes=%d\n",
John Koleszar's avatar
John Koleszar committed
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
                   one_pass_only, one_pass_only);
            arg_passes = one_pass_only;
        }

        if (arg_passes == 2 && !stats_fn)
            die("Must specify --fpf when --pass=%d and --passes=2\n", one_pass_only);
    }

    /* Populate encoder configuration */
    res = vpx_codec_enc_config_default(codec->iface, &cfg, arg_usage);

    if (res)
    {
1202
1203
        fprintf(stderr, "Failed to get config: %s\n",
                vpx_codec_err_to_string(res));
John Koleszar's avatar
John Koleszar committed
1204
1205
1206
        return EXIT_FAILURE;
    }

1207
1208
1209
    /* Change the default timebase to a high enough value so that the encoder
     * will always create strictly increasing timestamps.
     */
1210
    cfg.g_timebase.den = 1000;
1211

1212
1213
1214
1215
1216
1217
    /* Never use the library's default resolution, require it be parsed
     * from the file or set on the command line.
     */
    cfg.g_w = 0;
    cfg.g_h = 0;

John Koleszar's avatar
John Koleszar committed
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
    /* Now parse the remainder of the parameters. */
    for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
    {
        arg.argv_step = 1;

        if (0);
        else if (arg_match(&arg, &threads, argi))
            cfg.g_threads = arg_parse_uint(&arg);
        else if (arg_match(&arg, &profile, argi))
            cfg.g_profile = arg_parse_uint(&arg);
        else if (arg_match(&arg, &width, argi))
            cfg.g_w = arg_parse_uint(&arg);
        else if (arg_match(&arg, &height, argi))
            cfg.g_h = arg_parse_uint(&arg);
        else if (arg_match(&arg, &timebase, argi))
            cfg.g_timebase = arg_parse_rational(&arg);
        else if (arg_match(&arg, &error_resilient, argi))
            cfg.g_error_resilient = arg_parse_uint(&arg);
        else if (arg_match(&arg, &lag_in_frames, argi))
            cfg.g_lag_in_frames = arg_parse_uint(&arg);
        else if (arg_match(&arg, &dropframe_thresh, argi))
            cfg.rc_dropframe_thresh = arg_parse_uint(&arg);
        else if (arg_match(&arg, &resize_allowed, argi))
            cfg.rc_resize_allowed = arg_parse_uint(&arg);
        else if (arg_match(&arg, &resize_up_thresh, argi))
            cfg.rc_resize_up_thresh = arg_parse_uint(&arg);
        else if (arg_match(&arg, &resize_down_thresh, argi))
            cfg.rc_resize_down_thresh = arg_parse_uint(&arg);
        else if (arg_match(&arg, &resize_down_thresh, argi))
            cfg.rc_resize_down_thresh = arg_parse_uint(&arg);
        else if (arg_match(&arg, &end_usage, argi))
            cfg.rc_end_usage = arg_parse_uint(&arg);
        else if (arg_match(&arg, &target_bitrate, argi))
            cfg.rc_target_bitrate = arg_parse_uint(&arg);
        else if (arg_match(&arg, &min_quantizer, argi))
            cfg.rc_min_quantizer = arg_parse_uint(&arg);
        else if (arg_match(&arg, &max_quantizer, argi))
            cfg.rc_max_quantizer = arg_parse_uint(&arg);
        else if (arg_match(&arg, &undershoot_pct, argi))
            cfg.rc_undershoot_pct = arg_parse_uint(&arg);
        else if (arg_match(&arg, &overshoot_pct, argi))
            cfg.rc_overshoot_pct = arg_parse_uint(&arg);
        else if (arg_match(&arg, &buf_sz, argi))
            cfg.rc_buf_sz = arg_parse_uint(&arg);
        else if (arg_match(&arg, &buf_initial_sz, argi))
            cfg.rc_buf_initial_sz = arg_parse_uint(&arg);
        else if (arg_match(&arg, &buf_optimal_sz, argi))
            cfg.rc_buf_optimal_sz = arg_parse_uint(&arg);
        else if (arg_match(&arg, &bias_pct, argi))
        {
            cfg.rc_2pass_vbr_bias_pct = arg_parse_uint(&arg);

            if (arg_passes < 2)
1271
1272
1273
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1274
1275
1276
1277
1278
1279
        }
        else if (arg_match(&arg, &minsection_pct, argi))
        {
            cfg.rc_2pass_vbr_minsection_pct = arg_parse_uint(&arg);

            if (arg_passes < 2)
1280
1281
1282
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1283
1284
1285
1286
1287
1288
        }
        else if (arg_match(&arg, &maxsection_pct, argi))
        {
            cfg.rc_2pass_vbr_maxsection_pct = arg_parse_uint(&arg);

            if (arg_passes < 2)
1289
1290
1291
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1292
1293
1294
1295
1296
        }
        else if (arg_match(&arg, &kf_min_dist, argi))
            cfg.kf_min_dist = arg_parse_uint(&arg);
        else if (arg_match(&arg, &kf_max_dist, argi))
            cfg.kf_max_dist = arg_parse_uint(&arg);
1297
1298
        else if (arg_match(&arg, &kf_disabled, argi))
            cfg.kf_mode = VPX_KF_DISABLED;
John Koleszar's avatar
John Koleszar committed
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
        else
            argj++;
    }

    /* Handle codec specific options */
#if CONFIG_VP8_ENCODER

    if (codec->iface == &vpx_codec_vp8_cx_algo)
    {
        ctrl_args = vp8_args;
        ctrl_args_map = vp8_arg_ctrl_map;
    }

#endif

    for (argi = argj = argv; (*argj = *argi); argi += arg.argv_step)
    {
        int match = 0;

        arg.argv_step = 1;

        for (i = 0; ctrl_args[i]; i++)
        {
            if (arg_match(&arg, ctrl_args[i], argi))
            {
                match = 1;

                if (arg_ctrl_cnt < ARG_CTRL_CNT_MAX)
                {
                    arg_ctrls[arg_ctrl_cnt][0] = ctrl_args_map[i];
John Koleszar's avatar
John Koleszar committed
1329
                    arg_ctrls[arg_ctrl_cnt][1] = arg_parse_enum_or_int(&arg);
John Koleszar's avatar
John Koleszar committed
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
                    arg_ctrl_cnt++;
                }
            }
        }

        if (!match)
            argj++;
    }

    /* Check for unrecognized options */
    for (argi = argv; *argi; argi++)
1341
        if (argi[0][0] == '-' && argi[0][1])
John Koleszar's avatar
John Koleszar committed
1342
1343
1344
1345
1346
            die("Error: Unrecognized option %s\n", *argi);

    /* Handle non-option arguments */
    in_fn = argv[0];

1347
    if (!in_fn)
John Koleszar's avatar
John Koleszar committed
1348
1349
        usage_exit();

1350
1351
1352
    if(!out_fn)
        die("Error: Output file is required (specify with -o)\n");

John Koleszar's avatar
John Koleszar committed
1353
1354
1355
1356
1357
1358
    memset(&stats, 0, sizeof(stats));

    for (pass = one_pass_only ? one_pass_only - 1 : 0; pass < arg_passes; pass++)
    {
        int frames_in = 0, frames_out = 0;
        unsigned long nbytes = 0;
1359
        struct detect_buffer detect;
John Koleszar's avatar
John Koleszar committed
1360

1361
        /* Parse certain options from the input file, if possible */
John Koleszar's avatar
John Koleszar committed
1362
1363
        infile = strcmp(in_fn, "-") ? fopen(in_fn, "rb")
                                    : set_binary_mode(stdin);
John Koleszar's avatar
John Koleszar committed
1364
1365
1366

        if (!infile)
        {
1367
            fprintf(stderr, "Failed to open input file\n");
John Koleszar's avatar
John Koleszar committed
1368
1369
1370
            return EXIT_FAILURE;
        }

1371
1372
1373
        /* For RAW input sources, these bytes will applied on the first frame
         *  in read_frame().
         */
1374
1375
        detect.buf_read = fread(detect.buf, 1, 4, infile);
        detect.position = 0;
1376

1377
        if (detect.buf_read == 4 && file_is_y4m(infile, &y4m, detect.buf))
1378
        {
1379
1380
1381
1382
1383
            if (y4m_input_open(&y4m, infile, detect.buf, 4) >= 0)
            {
                file_type = FILE_TYPE_Y4M;
                cfg.g_w = y4m.pic_w;
                cfg.g_h = y4m.pic_h;
1384

1385
1386
1387
                /* Use the frame rate from the file only if none was specified
                 * on the command-line.
                 */
1388
                if (!arg_have_framerate)
1389
                {
1390
1391
                    arg_framerate.num = y4m.fps_n;
                    arg_framerate.den = y4m.fps_d;
1392
                }
1393

1394
1395
1396
                arg_use_i420 = 0;
            }
            else
1397
            {
1398
1399
                fprintf(stderr, "Unsupported Y4M stream.\n");
                return EXIT_FAILURE;
1400
1401
            }
        }
1402
        else if (detect.buf_read == 4 &&
1403
                 file_is_ivf(infile, &fourcc, &cfg.g_w, &cfg.g_h, &detect))
1404
        {
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
            file_type = FILE_TYPE_IVF;
            switch (fourcc)
            {
            case 0x32315659:
                arg_use_i420 = 0;
                break;
            case 0x30323449:
                arg_use_i420 = 1;
                break;
            default:
                fprintf(stderr, "Unsupported fourcc (%08x) in IVF\n", fourcc);
                return EXIT_FAILURE;
            }
1418
        }
1419
1420
1421
1422
        else
        {
            file_type = FILE_TYPE_RAW;
        }
1423
1424
1425
1426
1427
1428
1429
1430

        if(!cfg.g_w || !cfg.g_h)
        {
            fprintf(stderr, "Specify stream dimensions with --width (-w) "
                            " and --height (-h).\n");
            return EXIT_FAILURE;
        }

1431
1432
1433
#define SHOW(field) fprintf(stderr, "    %-28s = %d\n", #field, cfg.field)

        if (verbose && pass == 0)
1434
        {
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
            fprintf(stderr, "Codec: %s\n", vpx_codec_iface_name(codec->iface));
            fprintf(stderr, "Source file: %s Format: %s\n", in_fn,
                    arg_use_i420 ? "I420" : "YV12");
            fprintf(stderr, "Destination file: %s\n", out_fn);
            fprintf(stderr, "Encoder parameters:\n");

            SHOW(g_usage);
            SHOW(g_threads);
            SHOW(g_profile);
            SHOW(g_w);
            SHOW(g_h);
            SHOW(g_timebase.num);
            SHOW(g_timebase.den);
            SHOW(g_error_resilient);
            SHOW(g_pass);
            SHOW(g_lag_in_frames);
            SHOW(rc_dropframe_thresh);
            SHOW(rc_resize_allowed);
            SHOW(rc_resize_up_thresh);
            SHOW(rc_resize_down_thresh);
            SHOW(rc_end_usage);
            SHOW(rc_target_bitrate);
            SHOW(rc_min_quantizer);
            SHOW(rc_max_quantizer);
            SHOW(rc_undershoot_pct);
            SHOW(rc_overshoot_pct);
            SHOW(rc_buf_sz);
            SHOW(rc_buf_initial_sz);
            SHOW(rc_buf_optimal_sz);
            SHOW(rc_2pass_vbr_bias_pct);
            SHOW(rc_2pass_vbr_minsection_pct);
            SHOW(rc_2pass_vbr_maxsection_pct);
            SHOW(kf_mode);
            SHOW(kf_min_dist);
            SHOW(kf_max_dist);
        }

1472
        if(pass == (one_pass_only ? one_pass_only - 1 : 0)) {
1473
1474
1475
1476
1477
1478
1479
1480
            if (file_type == FILE_TYPE_Y4M)
                /*The Y4M reader does its own allocation.
                  Just initialize this here to avoid problems if we never read any
                   frames.*/
                memset(&raw, 0, sizeof(raw));
            else
                vpx_img_alloc(&raw, arg_use_i420 ? VPX_IMG_FMT_I420 : VPX_IMG_FMT_YV12,
                              cfg.g_w, cfg.g_h, 1);
1481
1482
        }

John Koleszar's avatar
John Koleszar committed
1483
1484
        outfile = strcmp(out_fn, "-") ? fopen(out_fn, "wb")
                                      : set_binary_mode(stdout);
John Koleszar's avatar
John Koleszar committed
1485
1486
1487

        if (!outfile)
        {
1488
            fprintf(stderr, "Failed to open output file\n");
John Koleszar's avatar
John Koleszar committed
1489
1490
1491
            return EXIT_FAILURE;
        }

1492
1493
1494
1495
1496
1497
        if(write_webm && fseek(outfile, 0, SEEK_CUR))
        {
            fprintf(stderr, "WebM output to pipes not supported.\n");
            return EXIT_FAILURE;
        }

John Koleszar's avatar
John Koleszar committed
1498
1499
1500
1501
        if (stats_fn)
        {
            if (!stats_open_file(&stats, stats_fn, pass))
            {
1502
                fprintf(stderr, "Failed to open statistics store\n");
John Koleszar's avatar
John Koleszar committed
1503
1504
1505
1506
1507
1508
1509
                return EXIT_FAILURE;
            }
        }
        else
        {
            if (!stats_open_mem(&stats, pass))
            {
1510
                fprintf(stderr, "Failed to open statistics store\n");
John Koleszar's avatar
John Koleszar committed
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
                return EXIT_FAILURE;
            }
        }

        cfg.g_pass = arg_passes == 2
                     ? pass ? VPX_RC_LAST_PASS : VPX_RC_FIRST_PASS
                 : VPX_RC_ONE_PASS;
#if VPX_ENCODER_ABI_VERSION > (1 + VPX_CODEC_ABI_VERSION)

        if (pass)
        {
            cfg.rc_twopass_stats_in = stats_get(&stats);
        }

#endif

John Koleszar's avatar
John Koleszar committed
1527
1528
1529
1530
1531
1532
1533
        if(write_webm)
        {
            ebml.stream = outfile;
            write_webm_file_header(&ebml, &cfg, &arg_framerate);
        }
        else
            write_ivf_file_header(outfile, &cfg, codec->fourcc, 0);
John Koleszar's avatar
John Koleszar committed
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547


        /* Construct Encoder Context */
        vpx_codec_enc_init(&encoder, codec->iface, &cfg,
                           show_psnr ? VPX_CODEC_USE_PSNR : 0);
        ctx_exit_on_error(&encoder, "Failed to initialize encoder");

        /* Note that we bypass the vpx_codec_control wrapper macro because
         * we're being clever to store the control IDs in an array. Real
         * applications will want to make use of the enumerations directly
         */
        for (i = 0; i < arg_ctrl_cnt; i++)
        {
            if (vpx_codec_control_(&encoder, arg_ctrls[i][0], arg_ctrls[i][1]))
1548
1549
                fprintf(stderr, "Error: Tried to set control %d = %d\n",
                        arg_ctrls[i][0], arg_ctrls[i][1]);
John Koleszar's avatar
John Koleszar committed
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561

            ctx_exit_on_error(&encoder, "Failed to control codec");
        }

        frame_avail = 1;
        got_data = 0;

        while (frame_avail || got_data)
        {
            vpx_codec_iter_t iter = NULL;
            const vpx_codec_cx_pkt_t *pkt;
            struct vpx_usec_timer timer;
1562
            int64_t frame_start, next_frame_start;
John Koleszar's avatar
John Koleszar committed
1563
1564
1565

            if (!arg_limit || frames_in < arg_limit)
            {
1566
1567
                frame_avail = read_frame(infile, &raw, file_type, &y4m,
                                         &detect);
John Koleszar's avatar
John Koleszar committed
1568
1569
1570
1571

                if (frame_avail)
                    frames_in++;

1572
1573
1574
                fprintf(stderr,
                        "\rPass %d/%d frame %4d/%-4d %7ldB \033[K", pass + 1,
                        arg_passes, frames_in, frames_out, nbytes);
John Koleszar's avatar
John Koleszar committed
1575
1576
1577
1578
1579
1580
            }
            else
                frame_avail = 0;

            vpx_usec_timer_start(&timer);

1581
1582
            frame_start = (cfg.g_timebase.den * (int64_t)(frames_in - 1)
                          * arg_framerate.den) / cfg.g_timebase.num / arg_framerate.num;
1583
1584
1585
            next_frame_start = (cfg.g_timebase.den * (int64_t)(frames_in)
                                * arg_framerate.den)
                                / cfg.g_timebase.num / arg_framerate.num;
1586
            vpx_codec_encode(&encoder, frame_avail ? &raw : NULL, frame_start,
1587
                             next_frame_start - frame_start,
1588
                             0, arg_deadline);
John Koleszar's avatar
John Koleszar committed
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
            vpx_usec_timer_mark(&timer);
            cx_time += vpx_usec_timer_elapsed(&timer);
            ctx_exit_on_error(&encoder, "Failed to encode frame");
            got_data = 0;

            while ((pkt = vpx_codec_get_cx_data(&encoder, &iter)))
            {
                got_data = 1;

                switch (pkt->kind)
                {
                case VPX_CODEC_CX_FRAME_PKT:
                    frames_out++;
1602
1603
                    fprintf(stderr, " %6luF",
                            (unsigned long)pkt->data.frame.sz);
John Koleszar's avatar
John Koleszar committed
1604

John Koleszar's avatar
John Koleszar committed
1605
1606
                    if(write_webm)
                    {
John Koleszar's avatar
John Koleszar committed
1607
1608
1609
1610
1611
                        /* Update the hash */
                        if(!ebml.debug)
                            hash = murmur(pkt->data.frame.buf,
                                          pkt->data.frame.sz, hash);

John Koleszar's avatar
John Koleszar committed
1612
1613
1614
1615
1616
                        write_webm_block(&ebml, &cfg, pkt);
                    }
                    else
                    {
                        write_ivf_frame_header(outfile, pkt);
1617
1618
                        if(fwrite(pkt->data.frame.buf, 1,
                                  pkt->data.frame.sz, outfile));
John Koleszar's avatar
John Koleszar committed
1619
                    }
1620
                    nbytes += pkt->data.raw.sz;
John Koleszar's avatar
John Koleszar committed
1621
1622
1623
                    break;
                case VPX_CODEC_STATS_PKT:
                    frames_out++;
1624
                    fprintf(stderr, " %6luS",
John Koleszar's avatar
John Koleszar committed
1625
1626
1627
1628
                           (unsigned long)pkt->data.twopass_stats.sz);
                    stats_write(&stats,
                                pkt->data.twopass_stats.buf,
                                pkt->data.twopass_stats.sz);
1629
                    nbytes += pkt->data.raw.sz;
John Koleszar's avatar
John Koleszar committed
1630
1631
1632
1633
1634
1635
1636
                    break;
                case VPX_CODEC_PSNR_PKT:

                    if (show_psnr)
                    {
                        int i;

1637
1638
                        psnr_sse_total += pkt->data.psnr.sse[0];
                        psnr_samples_total += pkt->data.psnr.samples[0];
John Koleszar's avatar
John Koleszar committed
1639
                        for (i = 0; i < 4; i++)
1640
                        {
1641
                            fprintf(stderr, "%.3lf ", pkt->data.psnr.psnr[i]);
1642
1643
1644
                            psnr_totals[i] += pkt->data.psnr.psnr[i];
                        }
                        psnr_count++;
John Koleszar's avatar
John Koleszar committed
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
                    }

                    break;
                default:
                    break;
                }
            }

            fflush(stdout);
        }

1656
1657
        fprintf(stderr,
               "\rPass %d/%d frame %4d/%-4d %7ldB %7ldb/f %7"PRId64"b/s"
John Koleszar's avatar
John Koleszar committed
1658
1659
               " %7lu %s (%.2f fps)\033[K", pass + 1,
               arg_passes, frames_in, frames_out, nbytes, nbytes * 8 / frames_in,
1660
               nbytes * 8 *(int64_t)arg_framerate.num / arg_framerate.den / frames_in,
John Koleszar's avatar
John Koleszar committed
1661
1662
1663
1664
               cx_time > 9999999 ? cx_time / 1000 : cx_time,
               cx_time > 9999999 ? "ms" : "us",
               (float)frames_in * 1000000.0 / (float)cx_time);

1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
        if ( (show_psnr) && (psnr_count>0) )
        {
            int i;
            double ovpsnr = vp8_mse2psnr(psnr_samples_total, 255.0,
                                         psnr_sse_total);

            fprintf(stderr, "\nPSNR (Overall/Avg/Y/U/V)");

            fprintf(stderr, " %.3lf", ovpsnr);
            for (i = 0; i < 4; i++)
            {
                fprintf(stderr, " %.3lf", psnr_totals[i]/psnr_count);
            }
        }

John Koleszar's avatar
John Koleszar committed
1680
1681
1682
1683
        vpx_codec_destroy(&encoder);

        fclose(infile);

John Koleszar's avatar
John Koleszar committed
1684
1685
        if(write_webm)
        {
John Koleszar's avatar
John Koleszar committed
1686
            write_webm_file_footer(&ebml, hash);
John Koleszar's avatar
John Koleszar committed
1687
1688
1689
1690
1691
1692
        }
        else
        {
            if (!fseek(outfile, 0, SEEK_SET))
                write_ivf_file_header(outfile, &cfg, codec->fourcc, frames_out);
        }
John Koleszar's avatar
John Koleszar committed
1693
1694
1695

        fclose(outfile);
        stats_close(&stats);
1696
        fprintf(stderr, "\n");
John Koleszar's avatar
John Koleszar committed
1697
1698
1699
1700
1701
1702
1703
1704
1705

        if (one_pass_only)
            break;
    }

    vpx_img_free(&raw);
    free(argv);
    return EXIT_SUCCESS;
}