1 /* 2 * Copyright (c) 2010 The WebM project authors. All Rights Reserved. 3 * 4 * Use of this source code is governed by a BSD-style license 5 * that can be found in the LICENSE file in the root of the source 6 * 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. 9 */ 10 11 12 /*!\defgroup encoder Encoder Algorithm Interface 13 * \ingroup codec 14 * This abstraction allows applications using this encoder to easily support 15 * multiple video formats with minimal code duplication. This section describes 16 * the interface common to all encoders. 17 * @{ 18 */ 19 20 /*!\file 21 * \brief Describes the encoder algorithm interface to applications. 22 * 23 * This file describes the interface between an application and a 24 * video encoder algorithm. 25 * 26 */ 27 #ifdef __cplusplus 28 extern "C" { 29 #endif 30 31 #ifndef VPX_ENCODER_H 32 #define VPX_ENCODER_H 33 #include "vpx_codec.h" 34 35 /*! Temporal Scalability: Maximum length of the sequence defining frame 36 * layer membership 37 */ 38 #define VPX_TS_MAX_PERIODICITY 16 39 40 /*! Temporal Scalability: Maximum number of coding layers */ 41 #define VPX_TS_MAX_LAYERS 5 42 43 /*!\deprecated Use #VPX_TS_MAX_PERIODICITY instead. */ 44 #define MAX_PERIODICITY VPX_TS_MAX_PERIODICITY 45 46 /*!\deprecated Use #VPX_TS_MAX_LAYERS instead. */ 47 #define MAX_LAYERS VPX_TS_MAX_LAYERS 48 49 /*!\brief Current ABI version number 50 * 51 * \internal 52 * If this file is altered in any way that changes the ABI, this value 53 * must be bumped. Examples include, but are not limited to, changing 54 * types, removing or reassigning enums, adding/removing/rearranging 55 * fields to structures 56 */ 57 #define VPX_ENCODER_ABI_VERSION (3 + VPX_CODEC_ABI_VERSION) /**<\hideinitializer*/ 58 59 60 /*! \brief Encoder capabilities bitfield 61 * 62 * Each encoder advertises the capabilities it supports as part of its 63 * ::vpx_codec_iface_t interface structure. Capabilities are extra 64 * interfaces or functionality, and are not required to be supported 65 * by an encoder. 66 * 67 * The available flags are specified by VPX_CODEC_CAP_* defines. 68 */ 69 #define VPX_CODEC_CAP_PSNR 0x10000 /**< Can issue PSNR packets */ 70 71 /*! Can output one partition at a time. Each partition is returned in its 72 * own VPX_CODEC_CX_FRAME_PKT, with the FRAME_IS_FRAGMENT flag set for 73 * every partition but the last. In this mode all frames are always 74 * returned partition by partition. 75 */ 76 #define VPX_CODEC_CAP_OUTPUT_PARTITION 0x20000 77 78 79 /*! \brief Initialization-time Feature Enabling 80 * 81 * Certain codec features must be known at initialization time, to allow 82 * for proper memory allocation. 83 * 84 * The available flags are specified by VPX_CODEC_USE_* defines. 85 */ 86 #define VPX_CODEC_USE_PSNR 0x10000 /**< Calculate PSNR on each frame */ 87 #define VPX_CODEC_USE_OUTPUT_PARTITION 0x20000 /**< Make the encoder output one 88 partition at a time. */ 89 90 91 /*!\brief Generic fixed size buffer structure 92 * 93 * This structure is able to hold a reference to any fixed size buffer. 94 */ 95 typedef struct vpx_fixed_buf 96 { 97 void *buf; /**< Pointer to the data */ 98 size_t sz; /**< Length of the buffer, in chars */ 99 } vpx_fixed_buf_t; /**< alias for struct vpx_fixed_buf */ 100 101 102 /*!\brief Time Stamp Type 103 * 104 * An integer, which when multiplied by the stream's time base, provides 105 * the absolute time of a sample. 106 */ 107 typedef int64_t vpx_codec_pts_t; 108 109 110 /*!\brief Compressed Frame Flags 111 * 112 * This type represents a bitfield containing information about a compressed 113 * frame that may be useful to an application. The most significant 16 bits 114 * can be used by an algorithm to provide additional detail, for example to 115 * support frame types that are codec specific (MPEG-1 D-frames for example) 116 */ 117 typedef uint32_t vpx_codec_frame_flags_t; 118 #define VPX_FRAME_IS_KEY 0x1 /**< frame is the start of a GOP */ 119 #define VPX_FRAME_IS_DROPPABLE 0x2 /**< frame can be dropped without affecting 120 the stream (no future frame depends on 121 this one) */ 122 #define VPX_FRAME_IS_INVISIBLE 0x4 /**< frame should be decoded but will not 123 be shown */ 124 #define VPX_FRAME_IS_FRAGMENT 0x8 /**< this is a fragment of the encoded 125 frame */ 126 127 /*!\brief Error Resilient flags 128 * 129 * These flags define which error resilient features to enable in the 130 * encoder. The flags are specified through the 131 * vpx_codec_enc_cfg::g_error_resilient variable. 132 */ 133 typedef uint32_t vpx_codec_er_flags_t; 134 #define VPX_ERROR_RESILIENT_DEFAULT 0x1 /**< Improve resiliency against 135 losses of whole frames */ 136 #define VPX_ERROR_RESILIENT_PARTITIONS 0x2 /**< The frame partitions are 137 independently decodable by the 138 bool decoder, meaning that 139 partitions can be decoded even 140 though earlier partitions have 141 been lost. Note that intra 142 predicition is still done over 143 the partition boundary. */ 144 145 /*!\brief Encoder output packet variants 146 * 147 * This enumeration lists the different kinds of data packets that can be 148 * returned by calls to vpx_codec_get_cx_data(). Algorithms \ref MAY 149 * extend this list to provide additional functionality. 150 */ 151 enum vpx_codec_cx_pkt_kind 152 { 153 VPX_CODEC_CX_FRAME_PKT, /**< Compressed video frame */ 154 VPX_CODEC_STATS_PKT, /**< Two-pass statistics for this frame */ 155 VPX_CODEC_PSNR_PKT, /**< PSNR statistics for this frame */ 156 VPX_CODEC_CUSTOM_PKT = 256 /**< Algorithm extensions */ 157 }; 158 159 160 /*!\brief Encoder output packet 161 * 162 * This structure contains the different kinds of output data the encoder 163 * may produce while compressing a frame. 164 */ 165 typedef struct vpx_codec_cx_pkt 166 { 167 enum vpx_codec_cx_pkt_kind kind; /**< packet variant */ 168 union 169 { 170 struct 171 { 172 void *buf; /**< compressed data buffer */ 173 size_t sz; /**< length of compressed data */ 174 vpx_codec_pts_t pts; /**< time stamp to show frame 175 (in timebase units) */ 176 unsigned long duration; /**< duration to show frame 177 (in timebase units) */ 178 vpx_codec_frame_flags_t flags; /**< flags for this frame */ 179 int partition_id; /**< the partition id 180 defines the decoding order 181 of the partitions. Only 182 applicable when "output partition" 183 mode is enabled. First partition 184 has id 0.*/ 185 186 } frame; /**< data for compressed frame packet */ 187 struct vpx_fixed_buf twopass_stats; /**< data for two-pass packet */ 188 struct vpx_psnr_pkt 189 { 190 unsigned int samples[4]; /**< Number of samples, total/y/u/v */ 191 uint64_t sse[4]; /**< sum squared error, total/y/u/v */ 192 double psnr[4]; /**< PSNR, total/y/u/v */ 193 } psnr; /**< data for PSNR packet */ 194 struct vpx_fixed_buf raw; /**< data for arbitrary packets */ 195 196 /* This packet size is fixed to allow codecs to extend this 197 * interface without having to manage storage for raw packets, 198 * i.e., if it's smaller than 128 bytes, you can store in the 199 * packet list directly. 200 */ 201 char pad[128 - sizeof(enum vpx_codec_cx_pkt_kind)]; /**< fixed sz */ 202 } data; /**< packet data */ 203 } vpx_codec_cx_pkt_t; /**< alias for struct vpx_codec_cx_pkt */ 204 205 206 /*!\brief Rational Number 207 * 208 * This structure holds a fractional value. 209 */ 210 typedef struct vpx_rational 211 { 212 int num; /**< fraction numerator */ 213 int den; /**< fraction denominator */ 214 } vpx_rational_t; /**< alias for struct vpx_rational */ 215 216 217 /*!\brief Multi-pass Encoding Pass */ 218 enum vpx_enc_pass 219 { 220 VPX_RC_ONE_PASS, /**< Single pass mode */ 221 VPX_RC_FIRST_PASS, /**< First pass of multi-pass mode */ 222 VPX_RC_LAST_PASS /**< Final pass of multi-pass mode */ 223 }; 224 225 226 /*!\brief Rate control mode */ 227 enum vpx_rc_mode 228 { 229 VPX_VBR, /**< Variable Bit Rate (VBR) mode */ 230 VPX_CBR, /**< Constant Bit Rate (CBR) mode */ 231 VPX_CQ /**< Constant Quality (CQ) mode */ 232 }; 233 234 235 /*!\brief Keyframe placement mode. 236 * 237 * This enumeration determines whether keyframes are placed automatically by 238 * the encoder or whether this behavior is disabled. Older releases of this 239 * SDK were implemented such that VPX_KF_FIXED meant keyframes were disabled. 240 * This name is confusing for this behavior, so the new symbols to be used 241 * are VPX_KF_AUTO and VPX_KF_DISABLED. 242 */ 243 enum vpx_kf_mode 244 { 245 VPX_KF_FIXED, /**< deprecated, implies VPX_KF_DISABLED */ 246 VPX_KF_AUTO, /**< Encoder determines optimal placement automatically */ 247 VPX_KF_DISABLED = 0 /**< Encoder does not place keyframes. */ 248 }; 249 250 251 /*!\brief Encoded Frame Flags 252 * 253 * This type indicates a bitfield to be passed to vpx_codec_encode(), defining 254 * per-frame boolean values. By convention, bits common to all codecs will be 255 * named VPX_EFLAG_*, and bits specific to an algorithm will be named 256 * /algo/_eflag_*. The lower order 16 bits are reserved for common use. 257 */ 258 typedef long vpx_enc_frame_flags_t; 259 #define VPX_EFLAG_FORCE_KF (1<<0) /**< Force this frame to be a keyframe */ 260 261 262 /*!\brief Encoder configuration structure 263 * 264 * This structure contains the encoder settings that have common representations 265 * across all codecs. This doesn't imply that all codecs support all features, 266 * however. 267 */ 268 typedef struct vpx_codec_enc_cfg 269 { 270 /* 271 * generic settings (g) 272 */ 273 274 /*!\brief Algorithm specific "usage" value 275 * 276 * Algorithms may define multiple values for usage, which may convey the 277 * intent of how the application intends to use the stream. If this value 278 * is non-zero, consult the documentation for the codec to determine its 279 * meaning. 280 */ 281 unsigned int g_usage; 282 283 284 /*!\brief Maximum number of threads to use 285 * 286 * For multi-threaded implementations, use no more than this number of 287 * threads. The codec may use fewer threads than allowed. The value 288 * 0 is equivalent to the value 1. 289 */ 290 unsigned int g_threads; 291 292 293 /*!\brief Bitstream profile to use 294 * 295 * Some codecs support a notion of multiple bitstream profiles. Typically 296 * this maps to a set of features that are turned on or off. Often the 297 * profile to use is determined by the features of the intended decoder. 298 * Consult the documentation for the codec to determine the valid values 299 * for this parameter, or set to zero for a sane default. 300 */ 301 unsigned int g_profile; /**< profile of bitstream to use */ 302 303 304 305 /*!\brief Width of the frame 306 * 307 * This value identifies the presentation resolution of the frame, 308 * in pixels. Note that the frames passed as input to the encoder must 309 * have this resolution. Frames will be presented by the decoder in this 310 * resolution, independent of any spatial resampling the encoder may do. 311 */ 312 unsigned int g_w; 313 314 315 /*!\brief Height of the frame 316 * 317 * This value identifies the presentation resolution of the frame, 318 * in pixels. Note that the frames passed as input to the encoder must 319 * have this resolution. Frames will be presented by the decoder in this 320 * resolution, independent of any spatial resampling the encoder may do. 321 */ 322 unsigned int g_h; 323 324 325 /*!\brief Stream timebase units 326 * 327 * Indicates the smallest interval of time, in seconds, used by the stream. 328 * For fixed frame rate material, or variable frame rate material where 329 * frames are timed at a multiple of a given clock (ex: video capture), 330 * the \ref RECOMMENDED method is to set the timebase to the reciprocal 331 * of the frame rate (ex: 1001/30000 for 29.970 Hz NTSC). This allows the 332 * pts to correspond to the frame number, which can be handy. For 333 * re-encoding video from containers with absolute time timestamps, the 334 * \ref RECOMMENDED method is to set the timebase to that of the parent 335 * container or multimedia framework (ex: 1/1000 for ms, as in FLV). 336 */ 337 struct vpx_rational g_timebase; 338 339 340 /*!\brief Enable error resilient modes. 341 * 342 * The error resilient bitfield indicates to the encoder which features 343 * it should enable to take measures for streaming over lossy or noisy 344 * links. 345 */ 346 vpx_codec_er_flags_t g_error_resilient; 347 348 349 /*!\brief Multi-pass Encoding Mode 350 * 351 * This value should be set to the current phase for multi-pass encoding. 352 * For single pass, set to #VPX_RC_ONE_PASS. 353 */ 354 enum vpx_enc_pass g_pass; 355 356 357 /*!\brief Allow lagged encoding 358 * 359 * If set, this value allows the encoder to consume a number of input 360 * frames before producing output frames. This allows the encoder to 361 * base decisions for the current frame on future frames. This does 362 * increase the latency of the encoding pipeline, so it is not appropriate 363 * in all situations (ex: realtime encoding). 364 * 365 * Note that this is a maximum value -- the encoder may produce frames 366 * sooner than the given limit. Set this value to 0 to disable this 367 * feature. 368 */ 369 unsigned int g_lag_in_frames; 370 371 372 /* 373 * rate control settings (rc) 374 */ 375 376 /*!\brief Temporal resampling configuration, if supported by the codec. 377 * 378 * Temporal resampling allows the codec to "drop" frames as a strategy to 379 * meet its target data rate. This can cause temporal discontinuities in 380 * the encoded video, which may appear as stuttering during playback. This 381 * trade-off is often acceptable, but for many applications is not. It can 382 * be disabled in these cases. 383 * 384 * Note that not all codecs support this feature. All vpx VPx codecs do. 385 * For other codecs, consult the documentation for that algorithm. 386 * 387 * This threshold is described as a percentage of the target data buffer. 388 * When the data buffer falls below this percentage of fullness, a 389 * dropped frame is indicated. Set the threshold to zero (0) to disable 390 * this feature. 391 */ 392 unsigned int rc_dropframe_thresh; 393 394 395 /*!\brief Enable/disable spatial resampling, if supported by the codec. 396 * 397 * Spatial resampling allows the codec to compress a lower resolution 398 * version of the frame, which is then upscaled by the encoder to the 399 * correct presentation resolution. This increases visual quality at 400 * low data rates, at the expense of CPU time on the encoder/decoder. 401 */ 402 unsigned int rc_resize_allowed; 403 404 405 /*!\brief Spatial resampling up watermark. 406 * 407 * This threshold is described as a percentage of the target data buffer. 408 * When the data buffer rises above this percentage of fullness, the 409 * encoder will step up to a higher resolution version of the frame. 410 */ 411 unsigned int rc_resize_up_thresh; 412 413 414 /*!\brief Spatial resampling down watermark. 415 * 416 * This threshold is described as a percentage of the target data buffer. 417 * When the data buffer falls below this percentage of fullness, the 418 * encoder will step down to a lower resolution version of the frame. 419 */ 420 unsigned int rc_resize_down_thresh; 421 422 423 /*!\brief Rate control algorithm to use. 424 * 425 * Indicates whether the end usage of this stream is to be streamed over 426 * a bandwidth constrained link, indicating that Constant Bit Rate (CBR) 427 * mode should be used, or whether it will be played back on a high 428 * bandwidth link, as from a local disk, where higher variations in 429 * bitrate are acceptable. 430 */ 431 enum vpx_rc_mode rc_end_usage; 432 433 434 /*!\brief Two-pass stats buffer. 435 * 436 * A buffer containing all of the stats packets produced in the first 437 * pass, concatenated. 438 */ 439 struct vpx_fixed_buf rc_twopass_stats_in; 440 441 442 /*!\brief Target data rate 443 * 444 * Target bandwidth to use for this stream, in kilobits per second. 445 */ 446 unsigned int rc_target_bitrate; 447 448 449 /* 450 * quantizer settings 451 */ 452 453 454 /*!\brief Minimum (Best Quality) Quantizer 455 * 456 * The quantizer is the most direct control over the quality of the 457 * encoded image. The range of valid values for the quantizer is codec 458 * specific. Consult the documentation for the codec to determine the 459 * values to use. To determine the range programmatically, call 460 * vpx_codec_enc_config_default() with a usage value of 0. 461 */ 462 unsigned int rc_min_quantizer; 463 464 465 /*!\brief Maximum (Worst Quality) Quantizer 466 * 467 * The quantizer is the most direct control over the quality of the 468 * encoded image. The range of valid values for the quantizer is codec 469 * specific. Consult the documentation for the codec to determine the 470 * values to use. To determine the range programmatically, call 471 * vpx_codec_enc_config_default() with a usage value of 0. 472 */ 473 unsigned int rc_max_quantizer; 474 475 476 /* 477 * bitrate tolerance 478 */ 479 480 481 /*!\brief Rate control adaptation undershoot control 482 * 483 * This value, expressed as a percentage of the target bitrate, 484 * controls the maximum allowed adaptation speed of the codec. 485 * This factor controls the maximum amount of bits that can 486 * be subtracted from the target bitrate in order to compensate 487 * for prior overshoot. 488 * 489 * Valid values in the range 0-1000. 490 */ 491 unsigned int rc_undershoot_pct; 492 493 494 /*!\brief Rate control adaptation overshoot control 495 * 496 * This value, expressed as a percentage of the target bitrate, 497 * controls the maximum allowed adaptation speed of the codec. 498 * This factor controls the maximum amount of bits that can 499 * be added to the target bitrate in order to compensate for 500 * prior undershoot. 501 * 502 * Valid values in the range 0-1000. 503 */ 504 unsigned int rc_overshoot_pct; 505 506 507 /* 508 * decoder buffer model parameters 509 */ 510 511 512 /*!\brief Decoder Buffer Size 513 * 514 * This value indicates the amount of data that may be buffered by the 515 * decoding application. Note that this value is expressed in units of 516 * time (milliseconds). For example, a value of 5000 indicates that the 517 * client will buffer (at least) 5000ms worth of encoded data. Use the 518 * target bitrate (#rc_target_bitrate) to convert to bits/bytes, if 519 * necessary. 520 */ 521 unsigned int rc_buf_sz; 522 523 524 /*!\brief Decoder Buffer Initial Size 525 * 526 * This value indicates the amount of data that will be buffered by the 527 * decoding application prior to beginning playback. This value is 528 * expressed in units of time (milliseconds). Use the target bitrate 529 * (#rc_target_bitrate) to convert to bits/bytes, if necessary. 530 */ 531 unsigned int rc_buf_initial_sz; 532 533 534 /*!\brief Decoder Buffer Optimal Size 535 * 536 * This value indicates the amount of data that the encoder should try 537 * to maintain in the decoder's buffer. This value is expressed in units 538 * of time (milliseconds). Use the target bitrate (#rc_target_bitrate) 539 * to convert to bits/bytes, if necessary. 540 */ 541 unsigned int rc_buf_optimal_sz; 542 543 544 /* 545 * 2 pass rate control parameters 546 */ 547 548 549 /*!\brief Two-pass mode CBR/VBR bias 550 * 551 * Bias, expressed on a scale of 0 to 100, for determining target size 552 * for the current frame. The value 0 indicates the optimal CBR mode 553 * value should be used. The value 100 indicates the optimal VBR mode 554 * value should be used. Values in between indicate which way the 555 * encoder should "lean." 556 */ 557 unsigned int rc_2pass_vbr_bias_pct; /**< RC mode bias between CBR and VBR(0-100: 0->CBR, 100->VBR) */ 558 559 560 /*!\brief Two-pass mode per-GOP minimum bitrate 561 * 562 * This value, expressed as a percentage of the target bitrate, indicates 563 * the minimum bitrate to be used for a single GOP (aka "section") 564 */ 565 unsigned int rc_2pass_vbr_minsection_pct; 566 567 568 /*!\brief Two-pass mode per-GOP maximum bitrate 569 * 570 * This value, expressed as a percentage of the target bitrate, indicates 571 * the maximum bitrate to be used for a single GOP (aka "section") 572 */ 573 unsigned int rc_2pass_vbr_maxsection_pct; 574 575 576 /* 577 * keyframing settings (kf) 578 */ 579 580 /*!\brief Keyframe placement mode 581 * 582 * This value indicates whether the encoder should place keyframes at a 583 * fixed interval, or determine the optimal placement automatically 584 * (as governed by the #kf_min_dist and #kf_max_dist parameters) 585 */ 586 enum vpx_kf_mode kf_mode; 587 588 589 /*!\brief Keyframe minimum interval 590 * 591 * This value, expressed as a number of frames, prevents the encoder from 592 * placing a keyframe nearer than kf_min_dist to the previous keyframe. At 593 * least kf_min_dist frames non-keyframes will be coded before the next 594 * keyframe. Set kf_min_dist equal to kf_max_dist for a fixed interval. 595 */ 596 unsigned int kf_min_dist; 597 598 599 /*!\brief Keyframe maximum interval 600 * 601 * This value, expressed as a number of frames, forces the encoder to code 602 * a keyframe if one has not been coded in the last kf_max_dist frames. 603 * A value of 0 implies all frames will be keyframes. Set kf_min_dist 604 * equal to kf_max_dist for a fixed interval. 605 */ 606 unsigned int kf_max_dist; 607 608 /* 609 * Temporal scalability settings (ts) 610 */ 611 612 /*!\brief Number of coding layers 613 * 614 * This value specifies the number of coding layers to be used. 615 */ 616 unsigned int ts_number_layers; 617 618 /*!\brief Target bitrate for each layer 619 * 620 * These values specify the target coding bitrate for each coding layer. 621 */ 622 unsigned int ts_target_bitrate[VPX_TS_MAX_LAYERS]; 623 624 /*!\brief Frame rate decimation factor for each layer 625 * 626 * These values specify the frame rate decimation factors to apply 627 * to each layer. 628 */ 629 unsigned int ts_rate_decimator[VPX_TS_MAX_LAYERS]; 630 631 /*!\brief Length of the sequence defining frame layer membership 632 * 633 * This value specifies the length of the sequence that defines the 634 * membership of frames to layers. For example, if ts_periodicity=8 then 635 * frames are assigned to coding layers with a repeated sequence of 636 * length 8. 637 */ 638 unsigned int ts_periodicity; 639 640 /*!\brief Template defining the membership of frames to coding layers 641 * 642 * This array defines the membership of frames to coding layers. For a 643 * 2-layer encoding that assigns even numbered frames to one layer (0) 644 * and odd numbered frames to a second layer (1) with ts_periodicity=8, 645 * then ts_layer_id = (0,1,0,1,0,1,0,1). 646 */ 647 unsigned int ts_layer_id[VPX_TS_MAX_PERIODICITY]; 648 } vpx_codec_enc_cfg_t; /**< alias for struct vpx_codec_enc_cfg */ 649 650 651 /*!\brief Initialize an encoder instance 652 * 653 * Initializes a encoder context using the given interface. Applications 654 * should call the vpx_codec_enc_init convenience macro instead of this 655 * function directly, to ensure that the ABI version number parameter 656 * is properly initialized. 657 * 658 * If the library was configured with --disable-multithread, this call 659 * is not thread safe and should be guarded with a lock if being used 660 * in a multithreaded context. 661 * 662 * In XMA mode (activated by setting VPX_CODEC_USE_XMA in the flags 663 * parameter), the storage pointed to by the cfg parameter must be 664 * kept readable and stable until all memory maps have been set. 665 * 666 * \param[in] ctx Pointer to this instance's context. 667 * \param[in] iface Pointer to the algorithm interface to use. 668 * \param[in] cfg Configuration to use, if known. May be NULL. 669 * \param[in] flags Bitfield of VPX_CODEC_USE_* flags 670 * \param[in] ver ABI version number. Must be set to 671 * VPX_ENCODER_ABI_VERSION 672 * \retval #VPX_CODEC_OK 673 * The decoder algorithm initialized. 674 * \retval #VPX_CODEC_MEM_ERROR 675 * Memory allocation failed. 676 */ 677 vpx_codec_err_t vpx_codec_enc_init_ver(vpx_codec_ctx_t *ctx, 678 vpx_codec_iface_t *iface, 679 vpx_codec_enc_cfg_t *cfg, 680 vpx_codec_flags_t flags, 681 int ver); 682 683 684 /*!\brief Convenience macro for vpx_codec_enc_init_ver() 685 * 686 * Ensures the ABI version parameter is properly set. 687 */ 688 #define vpx_codec_enc_init(ctx, iface, cfg, flags) \ 689 vpx_codec_enc_init_ver(ctx, iface, cfg, flags, VPX_ENCODER_ABI_VERSION) 690 691 692 /*!\brief Initialize multi-encoder instance 693 * 694 * Initializes multi-encoder context using the given interface. 695 * Applications should call the vpx_codec_enc_init_multi convenience macro 696 * instead of this function directly, to ensure that the ABI version number 697 * parameter is properly initialized. 698 * 699 * In XMA mode (activated by setting VPX_CODEC_USE_XMA in the flags 700 * parameter), the storage pointed to by the cfg parameter must be 701 * kept readable and stable until all memory maps have been set. 702 * 703 * \param[in] ctx Pointer to this instance's context. 704 * \param[in] iface Pointer to the algorithm interface to use. 705 * \param[in] cfg Configuration to use, if known. May be NULL. 706 * \param[in] num_enc Total number of encoders. 707 * \param[in] flags Bitfield of VPX_CODEC_USE_* flags 708 * \param[in] dsf Pointer to down-sampling factors. 709 * \param[in] ver ABI version number. Must be set to 710 * VPX_ENCODER_ABI_VERSION 711 * \retval #VPX_CODEC_OK 712 * The decoder algorithm initialized. 713 * \retval #VPX_CODEC_MEM_ERROR 714 * Memory allocation failed. 715 */ 716 vpx_codec_err_t vpx_codec_enc_init_multi_ver(vpx_codec_ctx_t *ctx, 717 vpx_codec_iface_t *iface, 718 vpx_codec_enc_cfg_t *cfg, 719 int num_enc, 720 vpx_codec_flags_t flags, 721 vpx_rational_t *dsf, 722 int ver); 723 724 725 /*!\brief Convenience macro for vpx_codec_enc_init_multi_ver() 726 * 727 * Ensures the ABI version parameter is properly set. 728 */ 729 #define vpx_codec_enc_init_multi(ctx, iface, cfg, num_enc, flags, dsf) \ 730 vpx_codec_enc_init_multi_ver(ctx, iface, cfg, num_enc, flags, dsf, \ 731 VPX_ENCODER_ABI_VERSION) 732 733 734 /*!\brief Get a default configuration 735 * 736 * Initializes a encoder configuration structure with default values. Supports 737 * the notion of "usages" so that an algorithm may offer different default 738 * settings depending on the user's intended goal. This function \ref SHOULD 739 * be called by all applications to initialize the configuration structure 740 * before specializing the configuration with application specific values. 741 * 742 * \param[in] iface Pointer to the algorithm interface to use. 743 * \param[out] cfg Configuration buffer to populate 744 * \param[in] usage End usage. Set to 0 or use codec specific values. 745 * 746 * \retval #VPX_CODEC_OK 747 * The configuration was populated. 748 * \retval #VPX_CODEC_INCAPABLE 749 * Interface is not an encoder interface. 750 * \retval #VPX_CODEC_INVALID_PARAM 751 * A parameter was NULL, or the usage value was not recognized. 752 */ 753 vpx_codec_err_t vpx_codec_enc_config_default(vpx_codec_iface_t *iface, 754 vpx_codec_enc_cfg_t *cfg, 755 unsigned int usage); 756 757 758 /*!\brief Set or change configuration 759 * 760 * Reconfigures an encoder instance according to the given configuration. 761 * 762 * \param[in] ctx Pointer to this instance's context 763 * \param[in] cfg Configuration buffer to use 764 * 765 * \retval #VPX_CODEC_OK 766 * The configuration was populated. 767 * \retval #VPX_CODEC_INCAPABLE 768 * Interface is not an encoder interface. 769 * \retval #VPX_CODEC_INVALID_PARAM 770 * A parameter was NULL, or the usage value was not recognized. 771 */ 772 vpx_codec_err_t vpx_codec_enc_config_set(vpx_codec_ctx_t *ctx, 773 const vpx_codec_enc_cfg_t *cfg); 774 775 776 /*!\brief Get global stream headers 777 * 778 * Retrieves a stream level global header packet, if supported by the codec. 779 * 780 * \param[in] ctx Pointer to this instance's context 781 * 782 * \retval NULL 783 * Encoder does not support global header 784 * \retval Non-NULL 785 * Pointer to buffer containing global header packet 786 */ 787 vpx_fixed_buf_t *vpx_codec_get_global_headers(vpx_codec_ctx_t *ctx); 788 789 790 #define VPX_DL_REALTIME (1) /**< deadline parameter analogous to 791 * VPx REALTIME mode. */ 792 #define VPX_DL_GOOD_QUALITY (1000000) /**< deadline parameter analogous to 793 * VPx GOOD QUALITY mode. */ 794 #define VPX_DL_BEST_QUALITY (0) /**< deadline parameter analogous to 795 * VPx BEST QUALITY mode. */ 796 /*!\brief Encode a frame 797 * 798 * Encodes a video frame at the given "presentation time." The presentation 799 * time stamp (PTS) \ref MUST be strictly increasing. 800 * 801 * The encoder supports the notion of a soft real-time deadline. Given a 802 * non-zero value to the deadline parameter, the encoder will make a "best 803 * effort" guarantee to return before the given time slice expires. It is 804 * implicit that limiting the available time to encode will degrade the 805 * output quality. The encoder can be given an unlimited time to produce the 806 * best possible frame by specifying a deadline of '0'. This deadline 807 * supercedes the VPx notion of "best quality, good quality, realtime". 808 * Applications that wish to map these former settings to the new deadline 809 * based system can use the symbols #VPX_DL_REALTIME, #VPX_DL_GOOD_QUALITY, 810 * and #VPX_DL_BEST_QUALITY. 811 * 812 * When the last frame has been passed to the encoder, this function should 813 * continue to be called, with the img parameter set to NULL. This will 814 * signal the end-of-stream condition to the encoder and allow it to encode 815 * any held buffers. Encoding is complete when vpx_codec_encode() is called 816 * and vpx_codec_get_cx_data() returns no data. 817 * 818 * \param[in] ctx Pointer to this instance's context 819 * \param[in] img Image data to encode, NULL to flush. 820 * \param[in] pts Presentation time stamp, in timebase units. 821 * \param[in] duration Duration to show frame, in timebase units. 822 * \param[in] flags Flags to use for encoding this frame. 823 * \param[in] deadline Time to spend encoding, in microseconds. (0=infinite) 824 * 825 * \retval #VPX_CODEC_OK 826 * The configuration was populated. 827 * \retval #VPX_CODEC_INCAPABLE 828 * Interface is not an encoder interface. 829 * \retval #VPX_CODEC_INVALID_PARAM 830 * A parameter was NULL, the image format is unsupported, etc. 831 */ 832 vpx_codec_err_t vpx_codec_encode(vpx_codec_ctx_t *ctx, 833 const vpx_image_t *img, 834 vpx_codec_pts_t pts, 835 unsigned long duration, 836 vpx_enc_frame_flags_t flags, 837 unsigned long deadline); 838 839 /*!\brief Set compressed data output buffer 840 * 841 * Sets the buffer that the codec should output the compressed data 842 * into. This call effectively sets the buffer pointer returned in the 843 * next VPX_CODEC_CX_FRAME_PKT packet. Subsequent packets will be 844 * appended into this buffer. The buffer is preserved across frames, 845 * so applications must periodically call this function after flushing 846 * the accumulated compressed data to disk or to the network to reset 847 * the pointer to the buffer's head. 848 * 849 * `pad_before` bytes will be skipped before writing the compressed 850 * data, and `pad_after` bytes will be appended to the packet. The size 851 * of the packet will be the sum of the size of the actual compressed 852 * data, pad_before, and pad_after. The padding bytes will be preserved 853 * (not overwritten). 854 * 855 * Note that calling this function does not guarantee that the returned 856 * compressed data will be placed into the specified buffer. In the 857 * event that the encoded data will not fit into the buffer provided, 858 * the returned packet \ref MAY point to an internal buffer, as it would 859 * if this call were never used. In this event, the output packet will 860 * NOT have any padding, and the application must free space and copy it 861 * to the proper place. This is of particular note in configurations 862 * that may output multiple packets for a single encoded frame (e.g., lagged 863 * encoding) or if the application does not reset the buffer periodically. 864 * 865 * Applications may restore the default behavior of the codec providing 866 * the compressed data buffer by calling this function with a NULL 867 * buffer. 868 * 869 * Applications \ref MUSTNOT call this function during iteration of 870 * vpx_codec_get_cx_data(). 871 * 872 * \param[in] ctx Pointer to this instance's context 873 * \param[in] buf Buffer to store compressed data into 874 * \param[in] pad_before Bytes to skip before writing compressed data 875 * \param[in] pad_after Bytes to skip after writing compressed data 876 * 877 * \retval #VPX_CODEC_OK 878 * The buffer was set successfully. 879 * \retval #VPX_CODEC_INVALID_PARAM 880 * A parameter was NULL, the image format is unsupported, etc. 881 */ 882 vpx_codec_err_t vpx_codec_set_cx_data_buf(vpx_codec_ctx_t *ctx, 883 const vpx_fixed_buf_t *buf, 884 unsigned int pad_before, 885 unsigned int pad_after); 886 887 888 /*!\brief Encoded data iterator 889 * 890 * Iterates over a list of data packets to be passed from the encoder to the 891 * application. The different kinds of packets available are enumerated in 892 * #vpx_codec_cx_pkt_kind. 893 * 894 * #VPX_CODEC_CX_FRAME_PKT packets should be passed to the application's 895 * muxer. Multiple compressed frames may be in the list. 896 * #VPX_CODEC_STATS_PKT packets should be appended to a global buffer. 897 * 898 * The application \ref MUST silently ignore any packet kinds that it does 899 * not recognize or support. 900 * 901 * The data buffers returned from this function are only guaranteed to be 902 * valid until the application makes another call to any vpx_codec_* function. 903 * 904 * \param[in] ctx Pointer to this instance's context 905 * \param[in,out] iter Iterator storage, initialized to NULL 906 * 907 * \return Returns a pointer to an output data packet (compressed frame data, 908 * two-pass statistics, etc.) or NULL to signal end-of-list. 909 * 910 */ 911 const vpx_codec_cx_pkt_t *vpx_codec_get_cx_data(vpx_codec_ctx_t *ctx, 912 vpx_codec_iter_t *iter); 913 914 915 /*!\brief Get Preview Frame 916 * 917 * Returns an image that can be used as a preview. Shows the image as it would 918 * exist at the decompressor. The application \ref MUST NOT write into this 919 * image buffer. 920 * 921 * \param[in] ctx Pointer to this instance's context 922 * 923 * \return Returns a pointer to a preview image, or NULL if no image is 924 * available. 925 * 926 */ 927 const vpx_image_t *vpx_codec_get_preview_frame(vpx_codec_ctx_t *ctx); 928 929 930 /*!@} - end defgroup encoder*/ 931 932 #endif 933 #ifdef __cplusplus 934 } 935 #endif 936
