vpxenc.c 65 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) || !CONFIG_OS_SUPPORT
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 <assert.h>
27
#include "vpx/vpx_encoder.h"
John Koleszar's avatar
John Koleszar committed
28
29
30
31
32
33
34
#if USE_POSIX_MMAP
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include <fcntl.h>
#include <unistd.h>
#endif
35
#include "vpx_config.h"
John Koleszar's avatar
John Koleszar committed
36
#include "vpx_version.h"
37
#include "vpx/vp8cx.h"
John Koleszar's avatar
John Koleszar committed
38
39
#include "vpx_ports/mem_ops.h"
#include "vpx_ports/vpx_timer.h"
John Koleszar's avatar
John Koleszar committed
40
#include "tools_common.h"
41
#include "y4minput.h"
John Koleszar's avatar
John Koleszar committed
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
#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
62

63
64
65
66
67
68
69
70
/* We should use 32-bit file operations in WebM file format
 * when building ARM executable file (.axf) with RVCT */
#if !CONFIG_OS_SUPPORT
typedef long off_t;
#define fseeko fseek
#define ftello ftell
#endif

John Koleszar's avatar
John Koleszar committed
71
72
73
74
75
76
77
78
79
static const char *exec_name;

static const struct codec_item
{
    char const              *name;
    const vpx_codec_iface_t *iface;
    unsigned int             fourcc;
} codecs[] =
{
80
81
82
#if CONFIG_EXPERIMENTAL && CONFIG_VP8_ENCODER
    {"vp8x",  &vpx_codec_vp8x_cx_algo, 0x78385056},
#endif
John Koleszar's avatar
John Koleszar committed
83
84
85
86
87
88
89
90
91
92
93
#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);
94
95
    vfprintf(stderr, fmt, ap);
    fprintf(stderr, "\n");
John Koleszar's avatar
John Koleszar committed
96
97
98
99
100
101
102
103
104
    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);

105
        fprintf(stderr, "%s: %s\n", s, vpx_codec_error(ctx));
John Koleszar's avatar
John Koleszar committed
106
107

        if (detail)
108
            fprintf(stderr, "    %s\n", detail);
John Koleszar's avatar
John Koleszar committed
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

        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)
        {
Ralph Giles's avatar
Ralph Giles committed
171
172
            fprintf(stderr, "Failed to allocate first-pass stats buffer (%lu bytes)\n",
                    (unsigned long)stats->buf_alloc_sz);
John Koleszar's avatar
John Koleszar committed
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
            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;
}


James Berry's avatar
James Berry committed
202
void stats_close(stats_io_t *stats, int last_pass)
John Koleszar's avatar
John Koleszar committed
203
204
205
{
    if (stats->file)
    {
James Berry's avatar
James Berry committed
206
        if (stats->pass == last_pass)
John Koleszar's avatar
John Koleszar committed
207
208
209
210
211
212
213
214
215
216
217
218
219
220
        {
#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
    {
James Berry's avatar
James Berry committed
221
        if (stats->pass == last_pass)
John Koleszar's avatar
John Koleszar committed
222
223
224
225
226
227
228
229
            free(stats->buf.buf);
    }
}

void stats_write(stats_io_t *stats, const void *pkt, size_t len)
{
    if (stats->file)
    {
230
        if(fwrite(pkt, 1, len, stats->file));
John Koleszar's avatar
John Koleszar committed
231
232
233
234
235
236
237
238
239
240
241
242
243
    }
    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;
244
245
246
247
248
249
250
            }
            else
            {
                fprintf(stderr,
                        "\nFailed to realloc firstpass stats buffer.\n");
                exit(EXIT_FAILURE);
            }
John Koleszar's avatar
John Koleszar committed
251
252
253
254
255
256
257
258
259
260
261
262
263
        }

        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;
}

264
265
266
267
268
269
270
271
272
273
/* Stereo 3D packed frame format */
typedef enum stereo_format
{
    STEREO_FORMAT_MONO       = 0,
    STEREO_FORMAT_LEFT_RIGHT = 1,
    STEREO_FORMAT_BOTTOM_TOP = 2,
    STEREO_FORMAT_TOP_BOTTOM = 3,
    STEREO_FORMAT_RIGHT_LEFT = 11
} stereo_format_t;

274
275
276
277
278
279
280
enum video_file_type
{
    FILE_TYPE_RAW,
    FILE_TYPE_IVF,
    FILE_TYPE_Y4M
};

281
282
struct detect_buffer {
    char buf[4];
283
284
    size_t buf_read;
    size_t position;
285
286
287
};


John Koleszar's avatar
John Koleszar committed
288
#define IVF_FRAME_HDR_SZ (4+8) /* 4 byte size + 8 byte timestamp */
289
static int read_frame(FILE *f, vpx_image_t *img, unsigned int file_type,
290
                      y4m_input *y4m, struct detect_buffer *detect)
John Koleszar's avatar
John Koleszar committed
291
292
{
    int plane = 0;
293
    int shortread = 0;
John Koleszar's avatar
John Koleszar committed
294

295
    if (file_type == FILE_TYPE_Y4M)
John Koleszar's avatar
John Koleszar committed
296
    {
297
        if (y4m_input_fetch_frame(y4m, f, img) < 1)
298
           return 0;
John Koleszar's avatar
John Koleszar committed
299
    }
300
    else
John Koleszar's avatar
John Koleszar committed
301
    {
302
        if (file_type == FILE_TYPE_IVF)
John Koleszar's avatar
John Koleszar committed
303
        {
304
305
306
307
308
309
            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.
             */
310
            if(fread(junk, 1, IVF_FRAME_HDR_SZ, f));
John Koleszar's avatar
John Koleszar committed
311
312
        }

313
        for (plane = 0; plane < 3; plane++)
John Koleszar's avatar
John Koleszar committed
314
        {
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
            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++)
            {
338
339
340
341
                size_t needed = w;
                size_t buf_position = 0;
                const size_t left = detect->buf_read - detect->position;
                if (left > 0)
342
                {
343
344
345
346
347
348
349
350
351
                    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);
352
353
                }

354
355
                ptr += img->stride[plane];
            }
John Koleszar's avatar
John Koleszar committed
356
357
358
        }
    }

359
    return !shortread;
John Koleszar's avatar
John Koleszar committed
360
361
362
}


363
unsigned int file_is_y4m(FILE      *infile,
364
365
                         y4m_input *y4m,
                         char       detect[4])
366
{
367
    if(memcmp(detect, "YUV4", 4) == 0)
368
369
370
371
372
373
    {
        return 1;
    }
    return 0;
}

John Koleszar's avatar
John Koleszar committed
374
375
376
377
#define IVF_FILE_HDR_SZ (32)
unsigned int file_is_ivf(FILE *infile,
                         unsigned int *fourcc,
                         unsigned int *width,
378
                         unsigned int *height,
379
                         struct detect_buffer *detect)
John Koleszar's avatar
John Koleszar committed
380
381
382
383
{
    char raw_hdr[IVF_FILE_HDR_SZ];
    int is_ivf = 0;

384
    if(memcmp(detect->buf, "DKIF", 4) != 0)
385
386
        return 0;

John Koleszar's avatar
John Koleszar committed
387
388
389
    /* See write_ivf_file_header() for more documentation on the file header
     * layout.
     */
390
391
    if (fread(raw_hdr + 4, 1, IVF_FILE_HDR_SZ - 4, infile)
        == IVF_FILE_HDR_SZ - 4)
John Koleszar's avatar
John Koleszar committed
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
    {
        {
            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);
408
        detect->position = 4;
John Koleszar's avatar
John Koleszar committed
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
    }

    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 */

439
    if(fwrite(header, 1, 32, outfile));
John Koleszar's avatar
John Koleszar committed
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
}


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);

457
    if(fwrite(header, 1, 12, outfile));
John Koleszar's avatar
John Koleszar committed
458
459
}

John Koleszar's avatar
John Koleszar committed
460
461
462
463
464
465
466
467
468
469
470
471
472

typedef off_t EbmlLoc;


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


struct EbmlGlobal
{
473
474
    int debug;

John Koleszar's avatar
John Koleszar committed
475
    FILE    *stream;
476
    int64_t last_pts_ms;
John Koleszar's avatar
John Koleszar committed
477
478
479
480
481
482
483
484
485
486
    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
487
488
489
    /* This pointer is to a specific element to be serialized */
    off_t    track_id_pos;

John Koleszar's avatar
John Koleszar committed
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
    /* 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)
{
505
    if(fwrite(buffer_in, 1, len, glob->stream));
John Koleszar's avatar
John Koleszar committed
506
507
}

508
509
510
511
512
513
514
#define WRITE_BUFFER(s) \
for(i = len-1; i>=0; i--)\
{ \
    x = *(const s *)buffer_in >> (i * CHAR_BIT); \
    Ebml_Write(glob, &x, 1); \
}
void Ebml_Serialize(EbmlGlobal *glob, const void *buffer_in, int buffer_size, unsigned long len)
John Koleszar's avatar
John Koleszar committed
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
    char x;
    int i;

    /* buffer_size:
     * 1 - int8_t;
     * 2 - int16_t;
     * 3 - int32_t;
     * 4 - int64_t;
     */
    switch (buffer_size)
    {
        case 1:
            WRITE_BUFFER(int8_t)
            break;
        case 2:
            WRITE_BUFFER(int16_t)
            break;
        case 4:
            WRITE_BUFFER(int32_t)
            break;
        case 8:
            WRITE_BUFFER(int64_t)
            break;
        default:
            break;
    }
John Koleszar's avatar
John Koleszar committed
542
}
543
#undef WRITE_BUFFER
John Koleszar's avatar
John Koleszar committed
544

545
/* Need a fixed size serializer for the track ID. libmkv provides a 64 bit
John Koleszar's avatar
John Koleszar committed
546
547
548
549
550
551
 * 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);
552
553
    Ebml_Serialize(glob, &sizeSerialized, sizeof(sizeSerialized), 1);
    Ebml_Serialize(glob, &ui, sizeof(ui), 4);
John Koleszar's avatar
John Koleszar committed
554
555
556
}


John Koleszar's avatar
John Koleszar committed
557
558
559
560
561
static void
Ebml_StartSubElement(EbmlGlobal *glob, EbmlLoc *ebmlLoc,
                          unsigned long class_id)
{
    //todo this is always taking 8 bytes, this may need later optimization
562
    //this is a key that says length unknown
563
    uint64_t unknownLen =  LITERALU64(0x01FFFFFFFFFFFFFF);
John Koleszar's avatar
John Koleszar committed
564
565
566

    Ebml_WriteID(glob, class_id);
    *ebmlLoc = ftello(glob->stream);
567
    Ebml_Serialize(glob, &unknownLen, sizeof(unknownLen), 8);
John Koleszar's avatar
John Koleszar committed
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
}

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);
585
    Ebml_Serialize(glob, &size, sizeof(size), 8);
John Koleszar's avatar
John Koleszar committed
586
587
588

    /* Reset the stream pointer */
    fseeko(glob->stream, pos, SEEK_SET);
John Koleszar's avatar
John Koleszar committed
589
590
}

John Koleszar's avatar
John Koleszar committed
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
626
627
628
629
630
631
632
633
634
635
636
637
638

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);
639
640
641
642
        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
643
644
645
646
647
648
649
650
        Ebml_EndSubElement(ebml, &startInfo);
    }
}


static void
write_webm_file_header(EbmlGlobal                *glob,
                       const vpx_codec_enc_cfg_t *cfg,
651
652
                       const struct vpx_rational *fps,
                       stereo_format_t            stereo_fmt)
John Koleszar's avatar
John Koleszar committed
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
{
    {
        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
683
684
                glob->track_id_pos = ftello(glob->stream);
                Ebml_SerializeUnsigned32(glob, TrackUID, trackID);
John Koleszar's avatar
John Koleszar committed
685
686
687
688
689
690
691
692
693
694
695
                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);
696
                    Ebml_SerializeUnsigned(glob, StereoMode, stereo_fmt);
John Koleszar's avatar
John Koleszar committed
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
                    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;
718
    int64_t        pts_ms;
John Koleszar's avatar
John Koleszar committed
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
    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)
        {
751
752
753
754
755
756
757
758
759
760
761
            struct cue_entry *cue, *new_cue_list;

            new_cue_list = realloc(glob->cue_list,
                                   (glob->cues+1) * sizeof(struct cue_entry));
            if(new_cue_list)
                glob->cue_list = new_cue_list;
            else
            {
                fprintf(stderr, "\nFailed to realloc cue list.\n");
                exit(EXIT_FAILURE);
            }
John Koleszar's avatar
John Koleszar committed
762
763
764
765
766
767
768
769
770
771
772
773
774

            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;
775
    Ebml_Serialize(glob, &block_length, sizeof(block_length), 4);
John Koleszar's avatar
John Koleszar committed
776
777
778
779
780

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

781
    Ebml_Serialize(glob, &block_timecode, sizeof(block_timecode), 2);
John Koleszar's avatar
John Koleszar committed
782
783
784
785
786
787
788
789
790
791
792
793
794

    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
795
write_webm_file_footer(EbmlGlobal *glob, long hash)
John Koleszar's avatar
John Koleszar committed
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
824
825
826
827
828
829
830
831
832
833
{

    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
834
835
836
837
838

    /* 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
839
840
841
842
    fseeko(glob->stream, 0, SEEK_END);
}


John Koleszar's avatar
John Koleszar committed
843
844
845
846
847
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
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
/* 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;
}

890
#include "math.h"
891
#define MAX_PSNR 100
892
893
894
895
896
897
898
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
899
        psnr = MAX_PSNR;      // Limit to prevent / 0
900

901
902
    if (psnr > MAX_PSNR)
        psnr = MAX_PSNR;
903
904
905
906

    return psnr;
}

John Koleszar's avatar
John Koleszar committed
907

John Koleszar's avatar
John Koleszar committed
908
909
#include "args.h"

910
911
static const arg_def_t debugmode = ARG_DEF("D", "debug", 0,
        "Debug mode (makes output deterministic)");
912
913
static const arg_def_t outputfile = ARG_DEF("o", "output", 1,
        "Output filename");
John Koleszar's avatar
John Koleszar committed
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
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");
940
static const arg_def_t framerate        = ARG_DEF(NULL, "fps", 1,
941
        "Stream frame rate (rate/scale)");
John Koleszar's avatar
John Koleszar committed
942
943
static const arg_def_t use_ivf          = ARG_DEF(NULL, "ivf", 0,
        "Output IVF (default is WebM)");
944
945
static const arg_def_t q_hist_n         = ARG_DEF(NULL, "q-hist", 1,
        "Show quantizer histogram (n-buckets)");
946
947
static const arg_def_t rate_hist_n         = ARG_DEF(NULL, "rate-hist", 1,
        "Show rate histogram (n-buckets)");
John Koleszar's avatar
John Koleszar committed
948
949
static const arg_def_t *main_args[] =
{
950
    &debugmode,
951
952
    &outputfile, &codecarg, &passes, &pass_arg, &fpf_name, &limit, &deadline,
    &best_dl, &good_dl, &rt_dl,
953
    &verbosearg, &psnrarg, &use_ivf, &q_hist_n, &rate_hist_n,
John Koleszar's avatar
John Koleszar committed
954
955
956
957
958
959
960
961
962
963
964
965
966
    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");
967
968
969
970
971
972
973
974
975
976
static const struct arg_enum_list stereo_mode_enum[] = {
    {"mono"      , STEREO_FORMAT_MONO},
    {"left-right", STEREO_FORMAT_LEFT_RIGHT},
    {"bottom-top", STEREO_FORMAT_BOTTOM_TOP},
    {"top-bottom", STEREO_FORMAT_TOP_BOTTOM},
    {"right-left", STEREO_FORMAT_RIGHT_LEFT},
    {NULL, 0}
};
static const arg_def_t stereo_mode      = ARG_DEF_ENUM(NULL, "stereo-mode", 1,
        "Stereo 3D video format", stereo_mode_enum);
John Koleszar's avatar
John Koleszar committed
977
static const arg_def_t timebase         = ARG_DEF(NULL, "timebase", 1,
978
        "Output timestamp precision (fractional seconds)");
John Koleszar's avatar
John Koleszar committed
979
980
981
982
983
984
985
986
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,
987
    &width, &height, &stereo_mode, &timebase, &framerate, &error_resilient,
John Koleszar's avatar
John Koleszar committed
988
989
990
991
992
993
994
995
996
997
998
    &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 %)");
999
1000
1001
1002
1003
1004
1005
1006
static const struct arg_enum_list end_usage_enum[] = {
    {"vbr", VPX_VBR},
    {"cbr", VPX_CBR},
    {"cq",  VPX_CQ},
    {NULL, 0}
};
static const arg_def_t end_usage          = ARG_DEF_ENUM(NULL, "end-usage", 1,
        "Rate control mode", end_usage_enum);
John Koleszar's avatar
John Koleszar committed
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
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)");
1048
1049
static const arg_def_t kf_disabled = ARG_DEF(NULL, "disable-kf", 0,
                                     "Disable keyframe placement");
John Koleszar's avatar
John Koleszar committed
1050
1051
static const arg_def_t *kf_args[] =
{
1052
    &kf_min_dist, &kf_max_dist, &kf_disabled, NULL
John Koleszar's avatar
John Koleszar committed
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
};


#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,
1077
                                        "AltRef Max Frames");
John Koleszar's avatar
John Koleszar committed
1078
static const arg_def_t arnr_strength = ARG_DEF(NULL, "arnr-strength", 1,
1079
                                       "AltRef Strength");
John Koleszar's avatar
John Koleszar committed
1080
static const arg_def_t arnr_type = ARG_DEF(NULL, "arnr-type", 1,
1081
                                   "AltRef Type");
John Koleszar's avatar
John Koleszar committed
1082
1083
1084
1085
1086
1087
1088
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);
Paul Wilkins's avatar
CQ Mode    
Paul Wilkins committed
1089
1090
static const arg_def_t cq_level = ARG_DEF(NULL, "cq-level", 1,
                                   "Constrained Quality Level");
1091
1092
static const arg_def_t max_intra_rate_pct = ARG_DEF(NULL, "max-intra-rate", 1,
        "Max I-frame bitrate (pct)");
John Koleszar's avatar
John Koleszar committed
1093
1094
1095
1096

static const arg_def_t *vp8_args[] =
{
    &cpu_used, &auto_altref, &noise_sens, &sharpness, &static_thresh,
John Koleszar's avatar
John Koleszar committed
1097
    &token_parts, &arnr_maxframes, &arnr_strength, &arnr_type,
1098
    &tune_ssim, &cq_level, &max_intra_rate_pct, NULL
John Koleszar's avatar
John Koleszar committed
1099
1100
1101
1102
1103
1104
};
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
1105
    VP8E_SET_ARNR_MAXFRAMES, VP8E_SET_ARNR_STRENGTH , VP8E_SET_ARNR_TYPE,
1106
    VP8E_SET_TUNING, VP8E_SET_CQ_LEVEL, VP8E_SET_MAX_INTRA_BITRATE_PCT, 0
John Koleszar's avatar
John Koleszar committed
1107
1108
1109
1110
1111
1112
1113
1114
1115
};
#endif

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

static void usage_exit()
{
    int i;

1116
1117
    fprintf(stderr, "Usage: %s <options> -o dst_filename src_filename \n",
            exec_name);
John Koleszar's avatar
John Koleszar committed
1118

1119
    fprintf(stderr, "\nOptions:\n");
John Koleszar's avatar
John Koleszar committed
1120
    arg_show_usage(stdout, main_args);
1121
    fprintf(stderr, "\nEncoder Global Options:\n");
John Koleszar's avatar
John Koleszar committed
1122
    arg_show_usage(stdout, global_args);
1123
    fprintf(stderr, "\nRate Control Options:\n");
John Koleszar's avatar
John Koleszar committed
1124
    arg_show_usage(stdout, rc_args);
1125
    fprintf(stderr, "\nTwopass Rate Control Options:\n");
John Koleszar's avatar
John Koleszar committed
1126
    arg_show_usage(stdout, rc_twopass_args);
1127
    fprintf(stderr, "\nKeyframe Placement Options:\n");
John Koleszar's avatar
John Koleszar committed
1128
1129
    arg_show_usage(stdout, kf_args);
#if CONFIG_VP8_ENCODER
1130
    fprintf(stderr, "\nVP8 Specific Options:\n");
John Koleszar's avatar
John Koleszar committed
1131
1132
    arg_show_usage(stdout, vp8_args);
#endif
1133
    fprintf(stderr, "\nStream timebase (--timebase):\n"
1134
            "  The desired precision of timestamps in the output, expressed\n"
1135
            "  in fractional seconds. Default is 1/1000.\n");
1136
    fprintf(stderr, "\n"
John Koleszar's avatar
John Koleszar committed
1137
1138
1139
1140
           "Included encoders:\n"
           "\n");

    for (i = 0; i < sizeof(codecs) / sizeof(codecs[0]); i++)
1141
        fprintf(stderr, "    %-6s - %s\n",
John Koleszar's avatar
John Koleszar committed
1142
1143
1144
1145
1146
1147
               codecs[i].name,
               vpx_codec_iface_name(codecs[i].iface));

    exit(EXIT_FAILURE);
}

1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173

#define HIST_BAR_MAX 40
struct hist_bucket
{
    int low, high, count;
};


static int merge_hist_buckets(struct hist_bucket *bucket,
                              int *buckets_,
                              int max_buckets)
{
    int small_bucket = 0, merge_bucket = INT_MAX, big_bucket=0;
    int buckets = *buckets_;
    int i;

    /* Find the extrema for this list of buckets */
    big_bucket = small_bucket = 0;
    for(i=0; i < buckets; i++)
    {
        if(bucket[i].count < bucket[small_bucket].count)
            small_bucket = i;
        if(bucket[i].count > bucket[big_bucket].count)
            big_bucket = i;
    }

1174
    /* If we have too many buckets, merge the smallest with an adjacent
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
     * bucket.
     */
    while(buckets > max_buckets)
    {
        int last_bucket = buckets - 1;

        // merge the small bucket with an adjacent one.
        if(small_bucket == 0)
            merge_bucket = 1;
        else if(small_bucket == last_bucket)
            merge_bucket = last_bucket - 1;
        else if(bucket[small_bucket - 1].count < bucket[small_bucket + 1].count)
            merge_bucket = small_bucket - 1;
        else
            merge_bucket = small_bucket + 1;

        assert(abs(merge_bucket - small_bucket) <= 1);
        assert(small_bucket < buckets);
        assert(big_bucket < buckets);
        assert(merge_bucket < buckets);

        if(merge_bucket < small_bucket)
        {
            bucket[merge_bucket].high = bucket[small_bucket].high;
            bucket[merge_bucket].count += bucket[small_bucket].count;
        }
        else
        {
            bucket[small_bucket].high = bucket[merge_bucket].high;
            bucket[small_bucket].count += bucket[merge_bucket].count;
            merge_bucket = small_bucket;
        }

        assert(bucket[merge_bucket].low != bucket[merge_bucket].high);

        buckets--;

        /* Remove the merge_bucket from the list, and find the new small
         * and big buckets while we're at it
         */
        big_bucket = small_bucket = 0;
        for(i=0; i < buckets; i++)
        {
            if(i > merge_bucket)
                bucket[i] = bucket[i+1];

            if(bucket[i].count < bucket[small_bucket].count)
                small_bucket = i;
            if(bucket[i].count > bucket[big_bucket].count)
                big_bucket = i;
        }

    }

    *buckets_ = buckets;
    return bucket[big_bucket].count;
}


static void show_histogram(const struct hist_bucket *bucket,
                           int                       buckets,
                           int                       total,
                           int                       scale)
{
1239
    const char *pat1, *pat2;
1240
1241
    int i;

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
1271
1272
1273
1274
    switch((int)(log(bucket[buckets-1].high)/log(10))+1)
    {
        case 1:
        case 2:
            pat1 = "%4d %2s: ";
            pat2 = "%4d-%2d: ";
            break;
        case 3:
            pat1 = "%5d %3s: ";
            pat2 = "%5d-%3d: ";
            break;
        case 4:
            pat1 = "%6d %4s: ";
            pat2 = "%6d-%4d: ";
            break;
        case 5:
            pat1 = "%7d %5s: ";
            pat2 = "%7d-%5d: ";
            break;
        case 6:
            pat1 = "%8d %6s: ";
            pat2 = "%8d-%6d: ";
            break;
        case 7:
            pat1 = "%9d %7s: ";
            pat2 = "%9d-%7d: ";
            break;
        default:
            pat1 = "%12d %10s: ";
            pat2 = "%12d-%10d: ";
            break;
    }

1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
    for(i=0; i<buckets; i++)
    {
        int len;
        int j;
        float pct;

        pct = 100.0 * (float)bucket[i].count / (float)total;
        len = HIST_BAR_MAX * bucket[i].count / scale;
        if(len < 1)
            len = 1;
        assert(len <= HIST_BAR_MAX);

        if(bucket[i].low == bucket[i].high)
1288
            fprintf(stderr, pat1, bucket[i].low, "");
1289
        else
1290
            fprintf(stderr, pat2, bucket[i].low, bucket[i].high);
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318

        for(j=0; j<HIST_BAR_MAX; j++)
            fprintf(stderr, j<len?"=":" ");
        fprintf(stderr, "\t%5d (%6.2f%%)\n",bucket[i].count,pct);
    }
}


static void show_q_histogram(const int counts[64], int max_buckets)
{
    struct hist_bucket bucket[64];
    int buckets = 0;
    int total = 0;
    int scale;
    int i;


    for(i=0; i<64; i++)
    {
        if(counts[i])
        {
            bucket[buckets].low = bucket[buckets].high = i;
            bucket[buckets].count = counts[i];
            buckets++;
            total += counts[i];
        }
    }

1319
    fprintf(stderr, "\nQuantizer Selection:\n");
1320
1321
1322
1323
1324
    scale = merge_hist_buckets(bucket, &buckets, max_buckets);
    show_histogram(bucket, buckets, total, scale);
}


1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
#define RATE_BINS (100)
struct rate_hist
{
    int64_t            *pts;
    int                *sz;
    int                 samples;
    int                 frames;
    struct hist_bucket  bucket[RATE_BINS];
    int                 total;
};


static void init_rate_histogram(struct rate_hist          *hist,
1338
1339
                                const vpx_codec_enc_cfg_t *cfg,
                                const vpx_rational_t      *fps)
1340
1341
1342
{
    int i;

1343
1344
1345
1346
1347
    /* Determine the number of samples in the buffer. Use the file's framerate
     * to determine the number of frames in rc_buf_sz milliseconds, with an
     * adjustment (5/4) to account for alt-refs
     */
    hist->samples = cfg->rc_buf_sz * 5 / 4 * fps->num / fps->den / 1000;
1348
1349
1350
1351
1352

    // prevent division by zero
    if (hist->samples == 0)
      hist->samples=1;

1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
    hist->pts = calloc(hist->samples, sizeof(*hist->pts));
    hist->sz = calloc(hist->samples, sizeof(*hist->sz));
    for(i=0; i<RATE_BINS; i++)
    {
        hist->bucket[i].low = INT_MAX;
        hist->bucket[i].high = 0;
        hist->bucket[i].count = 0;
    }
}


static void destroy_rate_histogram(struct rate_hist *hist)
{
    free(hist->pts);
    free(hist->sz);
}


static void update_rate_histogram(struct rate_hist          *hist,
                                  const vpx_codec_enc_cfg_t *cfg,
                                  const vpx_codec_cx_pkt_t  *pkt)
{
    int i, idx;
    int64_t now, then, sum_sz = 0, avg_bitrate;

    now = pkt->data.frame.pts * 1000
          * (uint64_t)cfg->g_timebase.num / (uint64_t)cfg->g_timebase.den;

    idx = hist->frames++ % hist->samples;
    hist->pts[idx] = now;
    hist->sz[idx] = pkt->data.frame.sz;

    if(now < cfg->rc_buf_initial_sz)
        return;

Johann's avatar
Johann committed
1388
1389
    then = now;

1390
    /* Sum the size over the past rc_buf_sz ms */
1391
    for(i = hist->frames; i > 0 && hist->frames - i < hist->samples; i--)
1392
1393
1394
1395
1396
1397
1398
1399
1400
    {
        int i_idx = (i-1) % hist->samples;

        then = hist->pts[i_idx];
        if(now - then > cfg->rc_buf_sz)
            break;
        sum_sz += hist->sz[i_idx];
    }

Johann's avatar
Johann committed
1401
1402
1403
    if (now == then)
        return;

1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
    avg_bitrate = sum_sz * 8 * 1000 / (now - then);
    idx = avg_bitrate * (RATE_BINS/2) / (cfg->rc_target_bitrate * 1000);
    if(idx < 0)
        idx = 0;
    if(idx > RATE_BINS-1)
        idx = RATE_BINS-1;
    if(hist->bucket[idx].low > avg_bitrate)
        hist->bucket[idx].low = avg_bitrate;
    if(hist->bucket[idx].high < avg_bitrate)
        hist->bucket[idx].high = avg_bitrate;
    hist->bucket[idx].count++;
    hist->total++;
}


static void show_rate_histogram(struct rate_hist          *hist,
                                const vpx_codec_enc_cfg_t *cfg,
                                int                        max_buckets)
{
    int i, scale;
    int buckets = 0;

    for(i = 0; i < RATE_BINS; i++)
    {
        if(hist->bucket[i].low == INT_MAX)
            continue;
        hist->bucket[buckets++] = hist->bucket[i];
    }

    fprintf(stderr, "\nRate (over %dms window):\n", cfg->rc_buf_sz);
    scale = merge_hist_buckets(hist->bucket, &buckets, max_buckets);
    show_histogram(hist->bucket, buckets, hist->total, scale);
}

John Koleszar's avatar
John Koleszar committed
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
#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;
1464
1465
    unsigned int             file_type, fourcc;
    y4m_input                y4m;
1466
1467
    struct vpx_rational      arg_framerate = {30, 1};
    int                      arg_have_framerate = 0;
John Koleszar's avatar
John Koleszar committed
1468
    int                      write_webm = 1;
John Koleszar's avatar
John Koleszar committed
1469
    EbmlGlobal               ebml = {0};
John Koleszar's avatar
John Koleszar committed
1470
    uint32_t                 hash = 0;
1471
1472
1473
1474
    uint64_t                 psnr_sse_total = 0;
    uint64_t                 psnr_samples_total = 0;
    double                   psnr_totals[4] = {0, 0, 0, 0};
    int                      psnr_count = 0;
1475
    stereo_format_t          stereo_fmt = STEREO_FORMAT_MONO;
1476
1477
    int                      counts[64]={0};
    int                      show_q_hist_buckets=0;
1478
1479
    int                      show_rate_hist_buckets=0;
    struct rate_hist         rate_hist={0};
John Koleszar's avatar
John Koleszar committed
1480
1481

    exec_name = argv_[0];
1482
    ebml.last_pts_ms = -1;
John Koleszar's avatar
John Koleszar committed
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551

    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;
1552
1553
1554
1555
1556
        else if (arg_match(&arg, &framerate, argi))
        {
            arg_framerate = arg_parse_rational(&arg);
            arg_have_framerate = 1;
        }
John Koleszar's avatar
John Koleszar committed
1557
1558
        else if (arg_match(&arg, &use_ivf, argi))
            write_webm = 0;
1559
1560
        else if (arg_match(&arg, &outputfile, argi))
            out_fn = arg.val;
1561
1562
        else if (arg_match(&arg, &debugmode, argi))
            ebml.debug = 1;
1563
1564
        else if (arg_match(&arg, &q_hist_n, argi))
            show_q_hist_buckets = arg_parse_uint(&arg);
1565
1566
        else if (arg_match(&arg, &rate_hist_n, argi))
            show_rate_hist_buckets = arg_parse_uint(&arg);
John Koleszar's avatar
John Koleszar committed
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
        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)
        {
1579
            fprintf(stderr, "Warning: Assuming --pass=%d implies --passes=%d\n",
John Koleszar's avatar
John Koleszar committed
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
                   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)
    {
1593
1594
        fprintf(stderr, "Failed to get config: %s\n",
                vpx_codec_err_to_string(res));
John Koleszar's avatar
John Koleszar committed
1595
1596
1597
        return EXIT_FAILURE;
    }

1598
1599
1600
    /* Change the default timebase to a high enough value so that the encoder
     * will always create strictly increasing timestamps.
     */
1601
    cfg.g_timebase.den = 1000;
1602

1603
1604
1605
1606
1607
1608
    /* 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
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
    /* 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);
1623
1624
        else if (arg_match(&arg, &stereo_mode, argi))
            stereo_fmt = arg_parse_enum_or_int(&arg);
John Koleszar's avatar
John Koleszar committed
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
        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))
1642
            cfg.rc_end_usage = arg_parse_enum_or_int(&arg);
John Koleszar's avatar
John Koleszar committed
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
        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)
1664
1665
1666
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1667
1668
1669
1670
1671
1672
        }
        else if (arg_match(&arg, &minsection_pct, argi))
        {
            cfg.rc_2pass_vbr_minsection_pct = arg_parse_uint(&arg);

            if (arg_passes < 2)
1673
1674
1675
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1676
1677
1678
1679
1680
1681
        }
        else if (arg_match(&arg, &maxsection_pct, argi))
        {
            cfg.rc_2pass_vbr_maxsection_pct = arg_parse_uint(&arg);

            if (arg_passes < 2)
1682
1683
1684
                fprintf(stderr,
                        "Warning: option %s ignored in one-pass mode.\n",
                        arg.name);
John Koleszar's avatar
John Koleszar committed
1685
1686
1687
1688
1689
        }
        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);
1690
1691
        else if (arg_match(&arg, &kf_disabled, argi))
            cfg.kf_mode = VPX_KF_DISABLED;
John Koleszar's avatar
John Koleszar committed
1692
1693
1694
1695
1696
1697
1698
        else
            argj++;
    }

    /* Handle codec specific options */
#if CONFIG_VP8_ENCODER

1699
1700
1701
1702
1703
    if (codec->iface == &vpx_codec_vp8_cx_algo
#if CONFIG_EXPERIMENTAL
        || codec->iface == &vpx_codec_vp8x_cx_algo
#endif
        )
John Koleszar's avatar
John Koleszar committed
1704
1705
1706
1707
1708
1709
1710
1711