1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/media/libvpx/vpx/vpx_encoder.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,936 @@
1.4 +/*
1.5 + * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
1.6 + *
1.7 + * Use of this source code is governed by a BSD-style license
1.8 + * that can be found in the LICENSE file in the root of the source
1.9 + * tree. An additional intellectual property rights grant can be found
1.10 + * in the file PATENTS. All contributing project authors may
1.11 + * be found in the AUTHORS file in the root of the source tree.
1.12 + */
1.13 +#ifndef VPX_ENCODER_H
1.14 +#define VPX_ENCODER_H
1.15 +
1.16 +/*!\defgroup encoder Encoder Algorithm Interface
1.17 + * \ingroup codec
1.18 + * This abstraction allows applications using this encoder to easily support
1.19 + * multiple video formats with minimal code duplication. This section describes
1.20 + * the interface common to all encoders.
1.21 + * @{
1.22 + */
1.23 +
1.24 +/*!\file
1.25 + * \brief Describes the encoder algorithm interface to applications.
1.26 + *
1.27 + * This file describes the interface between an application and a
1.28 + * video encoder algorithm.
1.29 + *
1.30 + */
1.31 +#ifdef __cplusplus
1.32 +extern "C" {
1.33 +#endif
1.34 +
1.35 +#include "vpx_codec.h"
1.36 +
1.37 + /*! Temporal Scalability: Maximum length of the sequence defining frame
1.38 + * layer membership
1.39 + */
1.40 +#define VPX_TS_MAX_PERIODICITY 16
1.41 +
1.42 + /*! Temporal Scalability: Maximum number of coding layers */
1.43 +#define VPX_TS_MAX_LAYERS 5
1.44 +
1.45 + /*!\deprecated Use #VPX_TS_MAX_PERIODICITY instead. */
1.46 +#define MAX_PERIODICITY VPX_TS_MAX_PERIODICITY
1.47 +
1.48 + /*!\deprecated Use #VPX_TS_MAX_LAYERS instead. */
1.49 +#define MAX_LAYERS VPX_TS_MAX_LAYERS
1.50 +
1.51 +/*! Spatial Scalability: Maximum number of coding layers */
1.52 +#define VPX_SS_MAX_LAYERS 5
1.53 +
1.54 +/*! Spatial Scalability: Default number of coding layers */
1.55 +#define VPX_SS_DEFAULT_LAYERS 3
1.56 +
1.57 + /*!\brief Current ABI version number
1.58 + *
1.59 + * \internal
1.60 + * If this file is altered in any way that changes the ABI, this value
1.61 + * must be bumped. Examples include, but are not limited to, changing
1.62 + * types, removing or reassigning enums, adding/removing/rearranging
1.63 + * fields to structures
1.64 + */
1.65 +#define VPX_ENCODER_ABI_VERSION (3 + VPX_CODEC_ABI_VERSION) /**<\hideinitializer*/
1.66 +
1.67 +
1.68 + /*! \brief Encoder capabilities bitfield
1.69 + *
1.70 + * Each encoder advertises the capabilities it supports as part of its
1.71 + * ::vpx_codec_iface_t interface structure. Capabilities are extra
1.72 + * interfaces or functionality, and are not required to be supported
1.73 + * by an encoder.
1.74 + *
1.75 + * The available flags are specified by VPX_CODEC_CAP_* defines.
1.76 + */
1.77 +#define VPX_CODEC_CAP_PSNR 0x10000 /**< Can issue PSNR packets */
1.78 +
1.79 + /*! Can output one partition at a time. Each partition is returned in its
1.80 + * own VPX_CODEC_CX_FRAME_PKT, with the FRAME_IS_FRAGMENT flag set for
1.81 + * every partition but the last. In this mode all frames are always
1.82 + * returned partition by partition.
1.83 + */
1.84 +#define VPX_CODEC_CAP_OUTPUT_PARTITION 0x20000
1.85 +
1.86 +
1.87 + /*! \brief Initialization-time Feature Enabling
1.88 + *
1.89 + * Certain codec features must be known at initialization time, to allow
1.90 + * for proper memory allocation.
1.91 + *
1.92 + * The available flags are specified by VPX_CODEC_USE_* defines.
1.93 + */
1.94 +#define VPX_CODEC_USE_PSNR 0x10000 /**< Calculate PSNR on each frame */
1.95 +#define VPX_CODEC_USE_OUTPUT_PARTITION 0x20000 /**< Make the encoder output one
1.96 + partition at a time. */
1.97 +
1.98 +
1.99 + /*!\brief Generic fixed size buffer structure
1.100 + *
1.101 + * This structure is able to hold a reference to any fixed size buffer.
1.102 + */
1.103 + typedef struct vpx_fixed_buf {
1.104 + void *buf; /**< Pointer to the data */
1.105 + size_t sz; /**< Length of the buffer, in chars */
1.106 + } vpx_fixed_buf_t; /**< alias for struct vpx_fixed_buf */
1.107 +
1.108 +
1.109 + /*!\brief Time Stamp Type
1.110 + *
1.111 + * An integer, which when multiplied by the stream's time base, provides
1.112 + * the absolute time of a sample.
1.113 + */
1.114 + typedef int64_t vpx_codec_pts_t;
1.115 +
1.116 +
1.117 + /*!\brief Compressed Frame Flags
1.118 + *
1.119 + * This type represents a bitfield containing information about a compressed
1.120 + * frame that may be useful to an application. The most significant 16 bits
1.121 + * can be used by an algorithm to provide additional detail, for example to
1.122 + * support frame types that are codec specific (MPEG-1 D-frames for example)
1.123 + */
1.124 + typedef uint32_t vpx_codec_frame_flags_t;
1.125 +#define VPX_FRAME_IS_KEY 0x1 /**< frame is the start of a GOP */
1.126 +#define VPX_FRAME_IS_DROPPABLE 0x2 /**< frame can be dropped without affecting
1.127 + the stream (no future frame depends on
1.128 + this one) */
1.129 +#define VPX_FRAME_IS_INVISIBLE 0x4 /**< frame should be decoded but will not
1.130 + be shown */
1.131 +#define VPX_FRAME_IS_FRAGMENT 0x8 /**< this is a fragment of the encoded
1.132 + frame */
1.133 +
1.134 + /*!\brief Error Resilient flags
1.135 + *
1.136 + * These flags define which error resilient features to enable in the
1.137 + * encoder. The flags are specified through the
1.138 + * vpx_codec_enc_cfg::g_error_resilient variable.
1.139 + */
1.140 + typedef uint32_t vpx_codec_er_flags_t;
1.141 +#define VPX_ERROR_RESILIENT_DEFAULT 0x1 /**< Improve resiliency against
1.142 + losses of whole frames */
1.143 +#define VPX_ERROR_RESILIENT_PARTITIONS 0x2 /**< The frame partitions are
1.144 + independently decodable by the
1.145 + bool decoder, meaning that
1.146 + partitions can be decoded even
1.147 + though earlier partitions have
1.148 + been lost. Note that intra
1.149 + predicition is still done over
1.150 + the partition boundary. */
1.151 +
1.152 + /*!\brief Encoder output packet variants
1.153 + *
1.154 + * This enumeration lists the different kinds of data packets that can be
1.155 + * returned by calls to vpx_codec_get_cx_data(). Algorithms \ref MAY
1.156 + * extend this list to provide additional functionality.
1.157 + */
1.158 + enum vpx_codec_cx_pkt_kind {
1.159 + VPX_CODEC_CX_FRAME_PKT, /**< Compressed video frame */
1.160 + VPX_CODEC_STATS_PKT, /**< Two-pass statistics for this frame */
1.161 + VPX_CODEC_PSNR_PKT, /**< PSNR statistics for this frame */
1.162 + VPX_CODEC_CUSTOM_PKT = 256 /**< Algorithm extensions */
1.163 + };
1.164 +
1.165 +
1.166 + /*!\brief Encoder output packet
1.167 + *
1.168 + * This structure contains the different kinds of output data the encoder
1.169 + * may produce while compressing a frame.
1.170 + */
1.171 + typedef struct vpx_codec_cx_pkt {
1.172 + enum vpx_codec_cx_pkt_kind kind; /**< packet variant */
1.173 + union {
1.174 + struct {
1.175 + void *buf; /**< compressed data buffer */
1.176 + size_t sz; /**< length of compressed data */
1.177 + vpx_codec_pts_t pts; /**< time stamp to show frame
1.178 + (in timebase units) */
1.179 + unsigned long duration; /**< duration to show frame
1.180 + (in timebase units) */
1.181 + vpx_codec_frame_flags_t flags; /**< flags for this frame */
1.182 + int partition_id; /**< the partition id
1.183 + defines the decoding order
1.184 + of the partitions. Only
1.185 + applicable when "output partition"
1.186 + mode is enabled. First partition
1.187 + has id 0.*/
1.188 +
1.189 + } frame; /**< data for compressed frame packet */
1.190 + struct vpx_fixed_buf twopass_stats; /**< data for two-pass packet */
1.191 + struct vpx_psnr_pkt {
1.192 + unsigned int samples[4]; /**< Number of samples, total/y/u/v */
1.193 + uint64_t sse[4]; /**< sum squared error, total/y/u/v */
1.194 + double psnr[4]; /**< PSNR, total/y/u/v */
1.195 + } psnr; /**< data for PSNR packet */
1.196 + struct vpx_fixed_buf raw; /**< data for arbitrary packets */
1.197 +
1.198 + /* This packet size is fixed to allow codecs to extend this
1.199 + * interface without having to manage storage for raw packets,
1.200 + * i.e., if it's smaller than 128 bytes, you can store in the
1.201 + * packet list directly.
1.202 + */
1.203 + char pad[128 - sizeof(enum vpx_codec_cx_pkt_kind)]; /**< fixed sz */
1.204 + } data; /**< packet data */
1.205 + } vpx_codec_cx_pkt_t; /**< alias for struct vpx_codec_cx_pkt */
1.206 +
1.207 +
1.208 + /*!\brief Rational Number
1.209 + *
1.210 + * This structure holds a fractional value.
1.211 + */
1.212 + typedef struct vpx_rational {
1.213 + int num; /**< fraction numerator */
1.214 + int den; /**< fraction denominator */
1.215 + } vpx_rational_t; /**< alias for struct vpx_rational */
1.216 +
1.217 +
1.218 + /*!\brief Multi-pass Encoding Pass */
1.219 + enum vpx_enc_pass {
1.220 + VPX_RC_ONE_PASS, /**< Single pass mode */
1.221 + VPX_RC_FIRST_PASS, /**< First pass of multi-pass mode */
1.222 + VPX_RC_LAST_PASS /**< Final pass of multi-pass mode */
1.223 + };
1.224 +
1.225 +
1.226 + /*!\brief Rate control mode */
1.227 + enum vpx_rc_mode {
1.228 + VPX_VBR, /**< Variable Bit Rate (VBR) mode */
1.229 + VPX_CBR, /**< Constant Bit Rate (CBR) mode */
1.230 + VPX_CQ, /**< Constrained Quality (CQ) mode */
1.231 + VPX_Q, /**< Constant Quality (Q) mode */
1.232 + };
1.233 +
1.234 +
1.235 + /*!\brief Keyframe placement mode.
1.236 + *
1.237 + * This enumeration determines whether keyframes are placed automatically by
1.238 + * the encoder or whether this behavior is disabled. Older releases of this
1.239 + * SDK were implemented such that VPX_KF_FIXED meant keyframes were disabled.
1.240 + * This name is confusing for this behavior, so the new symbols to be used
1.241 + * are VPX_KF_AUTO and VPX_KF_DISABLED.
1.242 + */
1.243 + enum vpx_kf_mode {
1.244 + VPX_KF_FIXED, /**< deprecated, implies VPX_KF_DISABLED */
1.245 + VPX_KF_AUTO, /**< Encoder determines optimal placement automatically */
1.246 + VPX_KF_DISABLED = 0 /**< Encoder does not place keyframes. */
1.247 + };
1.248 +
1.249 +
1.250 + /*!\brief Encoded Frame Flags
1.251 + *
1.252 + * This type indicates a bitfield to be passed to vpx_codec_encode(), defining
1.253 + * per-frame boolean values. By convention, bits common to all codecs will be
1.254 + * named VPX_EFLAG_*, and bits specific to an algorithm will be named
1.255 + * /algo/_eflag_*. The lower order 16 bits are reserved for common use.
1.256 + */
1.257 + typedef long vpx_enc_frame_flags_t;
1.258 +#define VPX_EFLAG_FORCE_KF (1<<0) /**< Force this frame to be a keyframe */
1.259 +
1.260 +
1.261 + /*!\brief Encoder configuration structure
1.262 + *
1.263 + * This structure contains the encoder settings that have common representations
1.264 + * across all codecs. This doesn't imply that all codecs support all features,
1.265 + * however.
1.266 + */
1.267 + typedef struct vpx_codec_enc_cfg {
1.268 + /*
1.269 + * generic settings (g)
1.270 + */
1.271 +
1.272 + /*!\brief Algorithm specific "usage" value
1.273 + *
1.274 + * Algorithms may define multiple values for usage, which may convey the
1.275 + * intent of how the application intends to use the stream. If this value
1.276 + * is non-zero, consult the documentation for the codec to determine its
1.277 + * meaning.
1.278 + */
1.279 + unsigned int g_usage;
1.280 +
1.281 +
1.282 + /*!\brief Maximum number of threads to use
1.283 + *
1.284 + * For multi-threaded implementations, use no more than this number of
1.285 + * threads. The codec may use fewer threads than allowed. The value
1.286 + * 0 is equivalent to the value 1.
1.287 + */
1.288 + unsigned int g_threads;
1.289 +
1.290 +
1.291 + /*!\brief Bitstream profile to use
1.292 + *
1.293 + * Some codecs support a notion of multiple bitstream profiles. Typically
1.294 + * this maps to a set of features that are turned on or off. Often the
1.295 + * profile to use is determined by the features of the intended decoder.
1.296 + * Consult the documentation for the codec to determine the valid values
1.297 + * for this parameter, or set to zero for a sane default.
1.298 + */
1.299 + unsigned int g_profile; /**< profile of bitstream to use */
1.300 +
1.301 +
1.302 +
1.303 + /*!\brief Width of the frame
1.304 + *
1.305 + * This value identifies the presentation resolution of the frame,
1.306 + * in pixels. Note that the frames passed as input to the encoder must
1.307 + * have this resolution. Frames will be presented by the decoder in this
1.308 + * resolution, independent of any spatial resampling the encoder may do.
1.309 + */
1.310 + unsigned int g_w;
1.311 +
1.312 +
1.313 + /*!\brief Height of the frame
1.314 + *
1.315 + * This value identifies the presentation resolution of the frame,
1.316 + * in pixels. Note that the frames passed as input to the encoder must
1.317 + * have this resolution. Frames will be presented by the decoder in this
1.318 + * resolution, independent of any spatial resampling the encoder may do.
1.319 + */
1.320 + unsigned int g_h;
1.321 +
1.322 +
1.323 + /*!\brief Stream timebase units
1.324 + *
1.325 + * Indicates the smallest interval of time, in seconds, used by the stream.
1.326 + * For fixed frame rate material, or variable frame rate material where
1.327 + * frames are timed at a multiple of a given clock (ex: video capture),
1.328 + * the \ref RECOMMENDED method is to set the timebase to the reciprocal
1.329 + * of the frame rate (ex: 1001/30000 for 29.970 Hz NTSC). This allows the
1.330 + * pts to correspond to the frame number, which can be handy. For
1.331 + * re-encoding video from containers with absolute time timestamps, the
1.332 + * \ref RECOMMENDED method is to set the timebase to that of the parent
1.333 + * container or multimedia framework (ex: 1/1000 for ms, as in FLV).
1.334 + */
1.335 + struct vpx_rational g_timebase;
1.336 +
1.337 +
1.338 + /*!\brief Enable error resilient modes.
1.339 + *
1.340 + * The error resilient bitfield indicates to the encoder which features
1.341 + * it should enable to take measures for streaming over lossy or noisy
1.342 + * links.
1.343 + */
1.344 + vpx_codec_er_flags_t g_error_resilient;
1.345 +
1.346 +
1.347 + /*!\brief Multi-pass Encoding Mode
1.348 + *
1.349 + * This value should be set to the current phase for multi-pass encoding.
1.350 + * For single pass, set to #VPX_RC_ONE_PASS.
1.351 + */
1.352 + enum vpx_enc_pass g_pass;
1.353 +
1.354 +
1.355 + /*!\brief Allow lagged encoding
1.356 + *
1.357 + * If set, this value allows the encoder to consume a number of input
1.358 + * frames before producing output frames. This allows the encoder to
1.359 + * base decisions for the current frame on future frames. This does
1.360 + * increase the latency of the encoding pipeline, so it is not appropriate
1.361 + * in all situations (ex: realtime encoding).
1.362 + *
1.363 + * Note that this is a maximum value -- the encoder may produce frames
1.364 + * sooner than the given limit. Set this value to 0 to disable this
1.365 + * feature.
1.366 + */
1.367 + unsigned int g_lag_in_frames;
1.368 +
1.369 +
1.370 + /*
1.371 + * rate control settings (rc)
1.372 + */
1.373 +
1.374 + /*!\brief Temporal resampling configuration, if supported by the codec.
1.375 + *
1.376 + * Temporal resampling allows the codec to "drop" frames as a strategy to
1.377 + * meet its target data rate. This can cause temporal discontinuities in
1.378 + * the encoded video, which may appear as stuttering during playback. This
1.379 + * trade-off is often acceptable, but for many applications is not. It can
1.380 + * be disabled in these cases.
1.381 + *
1.382 + * Note that not all codecs support this feature. All vpx VPx codecs do.
1.383 + * For other codecs, consult the documentation for that algorithm.
1.384 + *
1.385 + * This threshold is described as a percentage of the target data buffer.
1.386 + * When the data buffer falls below this percentage of fullness, a
1.387 + * dropped frame is indicated. Set the threshold to zero (0) to disable
1.388 + * this feature.
1.389 + */
1.390 + unsigned int rc_dropframe_thresh;
1.391 +
1.392 +
1.393 + /*!\brief Enable/disable spatial resampling, if supported by the codec.
1.394 + *
1.395 + * Spatial resampling allows the codec to compress a lower resolution
1.396 + * version of the frame, which is then upscaled by the encoder to the
1.397 + * correct presentation resolution. This increases visual quality at
1.398 + * low data rates, at the expense of CPU time on the encoder/decoder.
1.399 + */
1.400 + unsigned int rc_resize_allowed;
1.401 +
1.402 +
1.403 + /*!\brief Spatial resampling up watermark.
1.404 + *
1.405 + * This threshold is described as a percentage of the target data buffer.
1.406 + * When the data buffer rises above this percentage of fullness, the
1.407 + * encoder will step up to a higher resolution version of the frame.
1.408 + */
1.409 + unsigned int rc_resize_up_thresh;
1.410 +
1.411 +
1.412 + /*!\brief Spatial resampling down watermark.
1.413 + *
1.414 + * This threshold is described as a percentage of the target data buffer.
1.415 + * When the data buffer falls below this percentage of fullness, the
1.416 + * encoder will step down to a lower resolution version of the frame.
1.417 + */
1.418 + unsigned int rc_resize_down_thresh;
1.419 +
1.420 +
1.421 + /*!\brief Rate control algorithm to use.
1.422 + *
1.423 + * Indicates whether the end usage of this stream is to be streamed over
1.424 + * a bandwidth constrained link, indicating that Constant Bit Rate (CBR)
1.425 + * mode should be used, or whether it will be played back on a high
1.426 + * bandwidth link, as from a local disk, where higher variations in
1.427 + * bitrate are acceptable.
1.428 + */
1.429 + enum vpx_rc_mode rc_end_usage;
1.430 +
1.431 +
1.432 + /*!\brief Two-pass stats buffer.
1.433 + *
1.434 + * A buffer containing all of the stats packets produced in the first
1.435 + * pass, concatenated.
1.436 + */
1.437 + struct vpx_fixed_buf rc_twopass_stats_in;
1.438 +
1.439 +
1.440 + /*!\brief Target data rate
1.441 + *
1.442 + * Target bandwidth to use for this stream, in kilobits per second.
1.443 + */
1.444 + unsigned int rc_target_bitrate;
1.445 +
1.446 +
1.447 + /*
1.448 + * quantizer settings
1.449 + */
1.450 +
1.451 +
1.452 + /*!\brief Minimum (Best Quality) Quantizer
1.453 + *
1.454 + * The quantizer is the most direct control over the quality of the
1.455 + * encoded image. The range of valid values for the quantizer is codec
1.456 + * specific. Consult the documentation for the codec to determine the
1.457 + * values to use. To determine the range programmatically, call
1.458 + * vpx_codec_enc_config_default() with a usage value of 0.
1.459 + */
1.460 + unsigned int rc_min_quantizer;
1.461 +
1.462 +
1.463 + /*!\brief Maximum (Worst Quality) Quantizer
1.464 + *
1.465 + * The quantizer is the most direct control over the quality of the
1.466 + * encoded image. The range of valid values for the quantizer is codec
1.467 + * specific. Consult the documentation for the codec to determine the
1.468 + * values to use. To determine the range programmatically, call
1.469 + * vpx_codec_enc_config_default() with a usage value of 0.
1.470 + */
1.471 + unsigned int rc_max_quantizer;
1.472 +
1.473 +
1.474 + /*
1.475 + * bitrate tolerance
1.476 + */
1.477 +
1.478 +
1.479 + /*!\brief Rate control adaptation undershoot control
1.480 + *
1.481 + * This value, expressed as a percentage of the target bitrate,
1.482 + * controls the maximum allowed adaptation speed of the codec.
1.483 + * This factor controls the maximum amount of bits that can
1.484 + * be subtracted from the target bitrate in order to compensate
1.485 + * for prior overshoot.
1.486 + *
1.487 + * Valid values in the range 0-1000.
1.488 + */
1.489 + unsigned int rc_undershoot_pct;
1.490 +
1.491 +
1.492 + /*!\brief Rate control adaptation overshoot control
1.493 + *
1.494 + * This value, expressed as a percentage of the target bitrate,
1.495 + * controls the maximum allowed adaptation speed of the codec.
1.496 + * This factor controls the maximum amount of bits that can
1.497 + * be added to the target bitrate in order to compensate for
1.498 + * prior undershoot.
1.499 + *
1.500 + * Valid values in the range 0-1000.
1.501 + */
1.502 + unsigned int rc_overshoot_pct;
1.503 +
1.504 +
1.505 + /*
1.506 + * decoder buffer model parameters
1.507 + */
1.508 +
1.509 +
1.510 + /*!\brief Decoder Buffer Size
1.511 + *
1.512 + * This value indicates the amount of data that may be buffered by the
1.513 + * decoding application. Note that this value is expressed in units of
1.514 + * time (milliseconds). For example, a value of 5000 indicates that the
1.515 + * client will buffer (at least) 5000ms worth of encoded data. Use the
1.516 + * target bitrate (#rc_target_bitrate) to convert to bits/bytes, if
1.517 + * necessary.
1.518 + */
1.519 + unsigned int rc_buf_sz;
1.520 +
1.521 +
1.522 + /*!\brief Decoder Buffer Initial Size
1.523 + *
1.524 + * This value indicates the amount of data that will be buffered by the
1.525 + * decoding application prior to beginning playback. This value is
1.526 + * expressed in units of time (milliseconds). Use the target bitrate
1.527 + * (#rc_target_bitrate) to convert to bits/bytes, if necessary.
1.528 + */
1.529 + unsigned int rc_buf_initial_sz;
1.530 +
1.531 +
1.532 + /*!\brief Decoder Buffer Optimal Size
1.533 + *
1.534 + * This value indicates the amount of data that the encoder should try
1.535 + * to maintain in the decoder's buffer. This value is expressed in units
1.536 + * of time (milliseconds). Use the target bitrate (#rc_target_bitrate)
1.537 + * to convert to bits/bytes, if necessary.
1.538 + */
1.539 + unsigned int rc_buf_optimal_sz;
1.540 +
1.541 +
1.542 + /*
1.543 + * 2 pass rate control parameters
1.544 + */
1.545 +
1.546 +
1.547 + /*!\brief Two-pass mode CBR/VBR bias
1.548 + *
1.549 + * Bias, expressed on a scale of 0 to 100, for determining target size
1.550 + * for the current frame. The value 0 indicates the optimal CBR mode
1.551 + * value should be used. The value 100 indicates the optimal VBR mode
1.552 + * value should be used. Values in between indicate which way the
1.553 + * encoder should "lean."
1.554 + */
1.555 + unsigned int rc_2pass_vbr_bias_pct; /**< RC mode bias between CBR and VBR(0-100: 0->CBR, 100->VBR) */
1.556 +
1.557 +
1.558 + /*!\brief Two-pass mode per-GOP minimum bitrate
1.559 + *
1.560 + * This value, expressed as a percentage of the target bitrate, indicates
1.561 + * the minimum bitrate to be used for a single GOP (aka "section")
1.562 + */
1.563 + unsigned int rc_2pass_vbr_minsection_pct;
1.564 +
1.565 +
1.566 + /*!\brief Two-pass mode per-GOP maximum bitrate
1.567 + *
1.568 + * This value, expressed as a percentage of the target bitrate, indicates
1.569 + * the maximum bitrate to be used for a single GOP (aka "section")
1.570 + */
1.571 + unsigned int rc_2pass_vbr_maxsection_pct;
1.572 +
1.573 +
1.574 + /*
1.575 + * keyframing settings (kf)
1.576 + */
1.577 +
1.578 + /*!\brief Keyframe placement mode
1.579 + *
1.580 + * This value indicates whether the encoder should place keyframes at a
1.581 + * fixed interval, or determine the optimal placement automatically
1.582 + * (as governed by the #kf_min_dist and #kf_max_dist parameters)
1.583 + */
1.584 + enum vpx_kf_mode kf_mode;
1.585 +
1.586 +
1.587 + /*!\brief Keyframe minimum interval
1.588 + *
1.589 + * This value, expressed as a number of frames, prevents the encoder from
1.590 + * placing a keyframe nearer than kf_min_dist to the previous keyframe. At
1.591 + * least kf_min_dist frames non-keyframes will be coded before the next
1.592 + * keyframe. Set kf_min_dist equal to kf_max_dist for a fixed interval.
1.593 + */
1.594 + unsigned int kf_min_dist;
1.595 +
1.596 +
1.597 + /*!\brief Keyframe maximum interval
1.598 + *
1.599 + * This value, expressed as a number of frames, forces the encoder to code
1.600 + * a keyframe if one has not been coded in the last kf_max_dist frames.
1.601 + * A value of 0 implies all frames will be keyframes. Set kf_min_dist
1.602 + * equal to kf_max_dist for a fixed interval.
1.603 + */
1.604 + unsigned int kf_max_dist;
1.605 +
1.606 + /*
1.607 + * Spatial scalability settings (ss)
1.608 + */
1.609 +
1.610 + /*!\brief Number of coding layers (spatial)
1.611 + *
1.612 + * This value specifies the number of coding layers to be used.
1.613 + */
1.614 + unsigned int ss_number_layers;
1.615 +
1.616 + /*!\brief Number of coding layers
1.617 + *
1.618 + * This value specifies the number of coding layers to be used.
1.619 + */
1.620 + unsigned int ts_number_layers;
1.621 +
1.622 + /*!\brief Target bitrate for each layer
1.623 + *
1.624 + * These values specify the target coding bitrate for each coding layer.
1.625 + */
1.626 + unsigned int ts_target_bitrate[VPX_TS_MAX_LAYERS];
1.627 +
1.628 + /*!\brief Frame rate decimation factor for each layer
1.629 + *
1.630 + * These values specify the frame rate decimation factors to apply
1.631 + * to each layer.
1.632 + */
1.633 + unsigned int ts_rate_decimator[VPX_TS_MAX_LAYERS];
1.634 +
1.635 + /*!\brief Length of the sequence defining frame layer membership
1.636 + *
1.637 + * This value specifies the length of the sequence that defines the
1.638 + * membership of frames to layers. For example, if ts_periodicity=8 then
1.639 + * frames are assigned to coding layers with a repeated sequence of
1.640 + * length 8.
1.641 + */
1.642 + unsigned int ts_periodicity;
1.643 +
1.644 + /*!\brief Template defining the membership of frames to coding layers
1.645 + *
1.646 + * This array defines the membership of frames to coding layers. For a
1.647 + * 2-layer encoding that assigns even numbered frames to one layer (0)
1.648 + * and odd numbered frames to a second layer (1) with ts_periodicity=8,
1.649 + * then ts_layer_id = (0,1,0,1,0,1,0,1).
1.650 + */
1.651 + unsigned int ts_layer_id[VPX_TS_MAX_PERIODICITY];
1.652 + } vpx_codec_enc_cfg_t; /**< alias for struct vpx_codec_enc_cfg */
1.653 +
1.654 +
1.655 + /*!\brief Initialize an encoder instance
1.656 + *
1.657 + * Initializes a encoder context using the given interface. Applications
1.658 + * should call the vpx_codec_enc_init convenience macro instead of this
1.659 + * function directly, to ensure that the ABI version number parameter
1.660 + * is properly initialized.
1.661 + *
1.662 + * If the library was configured with --disable-multithread, this call
1.663 + * is not thread safe and should be guarded with a lock if being used
1.664 + * in a multithreaded context.
1.665 + *
1.666 + * In XMA mode (activated by setting VPX_CODEC_USE_XMA in the flags
1.667 + * parameter), the storage pointed to by the cfg parameter must be
1.668 + * kept readable and stable until all memory maps have been set.
1.669 + *
1.670 + * \param[in] ctx Pointer to this instance's context.
1.671 + * \param[in] iface Pointer to the algorithm interface to use.
1.672 + * \param[in] cfg Configuration to use, if known. May be NULL.
1.673 + * \param[in] flags Bitfield of VPX_CODEC_USE_* flags
1.674 + * \param[in] ver ABI version number. Must be set to
1.675 + * VPX_ENCODER_ABI_VERSION
1.676 + * \retval #VPX_CODEC_OK
1.677 + * The decoder algorithm initialized.
1.678 + * \retval #VPX_CODEC_MEM_ERROR
1.679 + * Memory allocation failed.
1.680 + */
1.681 + vpx_codec_err_t vpx_codec_enc_init_ver(vpx_codec_ctx_t *ctx,
1.682 + vpx_codec_iface_t *iface,
1.683 + vpx_codec_enc_cfg_t *cfg,
1.684 + vpx_codec_flags_t flags,
1.685 + int ver);
1.686 +
1.687 +
1.688 + /*!\brief Convenience macro for vpx_codec_enc_init_ver()
1.689 + *
1.690 + * Ensures the ABI version parameter is properly set.
1.691 + */
1.692 +#define vpx_codec_enc_init(ctx, iface, cfg, flags) \
1.693 + vpx_codec_enc_init_ver(ctx, iface, cfg, flags, VPX_ENCODER_ABI_VERSION)
1.694 +
1.695 +
1.696 + /*!\brief Initialize multi-encoder instance
1.697 + *
1.698 + * Initializes multi-encoder context using the given interface.
1.699 + * Applications should call the vpx_codec_enc_init_multi convenience macro
1.700 + * instead of this function directly, to ensure that the ABI version number
1.701 + * parameter is properly initialized.
1.702 + *
1.703 + * In XMA mode (activated by setting VPX_CODEC_USE_XMA in the flags
1.704 + * parameter), the storage pointed to by the cfg parameter must be
1.705 + * kept readable and stable until all memory maps have been set.
1.706 + *
1.707 + * \param[in] ctx Pointer to this instance's context.
1.708 + * \param[in] iface Pointer to the algorithm interface to use.
1.709 + * \param[in] cfg Configuration to use, if known. May be NULL.
1.710 + * \param[in] num_enc Total number of encoders.
1.711 + * \param[in] flags Bitfield of VPX_CODEC_USE_* flags
1.712 + * \param[in] dsf Pointer to down-sampling factors.
1.713 + * \param[in] ver ABI version number. Must be set to
1.714 + * VPX_ENCODER_ABI_VERSION
1.715 + * \retval #VPX_CODEC_OK
1.716 + * The decoder algorithm initialized.
1.717 + * \retval #VPX_CODEC_MEM_ERROR
1.718 + * Memory allocation failed.
1.719 + */
1.720 + vpx_codec_err_t vpx_codec_enc_init_multi_ver(vpx_codec_ctx_t *ctx,
1.721 + vpx_codec_iface_t *iface,
1.722 + vpx_codec_enc_cfg_t *cfg,
1.723 + int num_enc,
1.724 + vpx_codec_flags_t flags,
1.725 + vpx_rational_t *dsf,
1.726 + int ver);
1.727 +
1.728 +
1.729 + /*!\brief Convenience macro for vpx_codec_enc_init_multi_ver()
1.730 + *
1.731 + * Ensures the ABI version parameter is properly set.
1.732 + */
1.733 +#define vpx_codec_enc_init_multi(ctx, iface, cfg, num_enc, flags, dsf) \
1.734 + vpx_codec_enc_init_multi_ver(ctx, iface, cfg, num_enc, flags, dsf, \
1.735 + VPX_ENCODER_ABI_VERSION)
1.736 +
1.737 +
1.738 + /*!\brief Get a default configuration
1.739 + *
1.740 + * Initializes a encoder configuration structure with default values. Supports
1.741 + * the notion of "usages" so that an algorithm may offer different default
1.742 + * settings depending on the user's intended goal. This function \ref SHOULD
1.743 + * be called by all applications to initialize the configuration structure
1.744 + * before specializing the configuration with application specific values.
1.745 + *
1.746 + * \param[in] iface Pointer to the algorithm interface to use.
1.747 + * \param[out] cfg Configuration buffer to populate
1.748 + * \param[in] usage End usage. Set to 0 or use codec specific values.
1.749 + *
1.750 + * \retval #VPX_CODEC_OK
1.751 + * The configuration was populated.
1.752 + * \retval #VPX_CODEC_INCAPABLE
1.753 + * Interface is not an encoder interface.
1.754 + * \retval #VPX_CODEC_INVALID_PARAM
1.755 + * A parameter was NULL, or the usage value was not recognized.
1.756 + */
1.757 + vpx_codec_err_t vpx_codec_enc_config_default(vpx_codec_iface_t *iface,
1.758 + vpx_codec_enc_cfg_t *cfg,
1.759 + unsigned int usage);
1.760 +
1.761 +
1.762 + /*!\brief Set or change configuration
1.763 + *
1.764 + * Reconfigures an encoder instance according to the given configuration.
1.765 + *
1.766 + * \param[in] ctx Pointer to this instance's context
1.767 + * \param[in] cfg Configuration buffer to use
1.768 + *
1.769 + * \retval #VPX_CODEC_OK
1.770 + * The configuration was populated.
1.771 + * \retval #VPX_CODEC_INCAPABLE
1.772 + * Interface is not an encoder interface.
1.773 + * \retval #VPX_CODEC_INVALID_PARAM
1.774 + * A parameter was NULL, or the usage value was not recognized.
1.775 + */
1.776 + vpx_codec_err_t vpx_codec_enc_config_set(vpx_codec_ctx_t *ctx,
1.777 + const vpx_codec_enc_cfg_t *cfg);
1.778 +
1.779 +
1.780 + /*!\brief Get global stream headers
1.781 + *
1.782 + * Retrieves a stream level global header packet, if supported by the codec.
1.783 + *
1.784 + * \param[in] ctx Pointer to this instance's context
1.785 + *
1.786 + * \retval NULL
1.787 + * Encoder does not support global header
1.788 + * \retval Non-NULL
1.789 + * Pointer to buffer containing global header packet
1.790 + */
1.791 + vpx_fixed_buf_t *vpx_codec_get_global_headers(vpx_codec_ctx_t *ctx);
1.792 +
1.793 +
1.794 +#define VPX_DL_REALTIME (1) /**< deadline parameter analogous to
1.795 + * VPx REALTIME mode. */
1.796 +#define VPX_DL_GOOD_QUALITY (1000000) /**< deadline parameter analogous to
1.797 + * VPx GOOD QUALITY mode. */
1.798 +#define VPX_DL_BEST_QUALITY (0) /**< deadline parameter analogous to
1.799 + * VPx BEST QUALITY mode. */
1.800 + /*!\brief Encode a frame
1.801 + *
1.802 + * Encodes a video frame at the given "presentation time." The presentation
1.803 + * time stamp (PTS) \ref MUST be strictly increasing.
1.804 + *
1.805 + * The encoder supports the notion of a soft real-time deadline. Given a
1.806 + * non-zero value to the deadline parameter, the encoder will make a "best
1.807 + * effort" guarantee to return before the given time slice expires. It is
1.808 + * implicit that limiting the available time to encode will degrade the
1.809 + * output quality. The encoder can be given an unlimited time to produce the
1.810 + * best possible frame by specifying a deadline of '0'. This deadline
1.811 + * supercedes the VPx notion of "best quality, good quality, realtime".
1.812 + * Applications that wish to map these former settings to the new deadline
1.813 + * based system can use the symbols #VPX_DL_REALTIME, #VPX_DL_GOOD_QUALITY,
1.814 + * and #VPX_DL_BEST_QUALITY.
1.815 + *
1.816 + * When the last frame has been passed to the encoder, this function should
1.817 + * continue to be called, with the img parameter set to NULL. This will
1.818 + * signal the end-of-stream condition to the encoder and allow it to encode
1.819 + * any held buffers. Encoding is complete when vpx_codec_encode() is called
1.820 + * and vpx_codec_get_cx_data() returns no data.
1.821 + *
1.822 + * \param[in] ctx Pointer to this instance's context
1.823 + * \param[in] img Image data to encode, NULL to flush.
1.824 + * \param[in] pts Presentation time stamp, in timebase units.
1.825 + * \param[in] duration Duration to show frame, in timebase units.
1.826 + * \param[in] flags Flags to use for encoding this frame.
1.827 + * \param[in] deadline Time to spend encoding, in microseconds. (0=infinite)
1.828 + *
1.829 + * \retval #VPX_CODEC_OK
1.830 + * The configuration was populated.
1.831 + * \retval #VPX_CODEC_INCAPABLE
1.832 + * Interface is not an encoder interface.
1.833 + * \retval #VPX_CODEC_INVALID_PARAM
1.834 + * A parameter was NULL, the image format is unsupported, etc.
1.835 + */
1.836 + vpx_codec_err_t vpx_codec_encode(vpx_codec_ctx_t *ctx,
1.837 + const vpx_image_t *img,
1.838 + vpx_codec_pts_t pts,
1.839 + unsigned long duration,
1.840 + vpx_enc_frame_flags_t flags,
1.841 + unsigned long deadline);
1.842 +
1.843 + /*!\brief Set compressed data output buffer
1.844 + *
1.845 + * Sets the buffer that the codec should output the compressed data
1.846 + * into. This call effectively sets the buffer pointer returned in the
1.847 + * next VPX_CODEC_CX_FRAME_PKT packet. Subsequent packets will be
1.848 + * appended into this buffer. The buffer is preserved across frames,
1.849 + * so applications must periodically call this function after flushing
1.850 + * the accumulated compressed data to disk or to the network to reset
1.851 + * the pointer to the buffer's head.
1.852 + *
1.853 + * `pad_before` bytes will be skipped before writing the compressed
1.854 + * data, and `pad_after` bytes will be appended to the packet. The size
1.855 + * of the packet will be the sum of the size of the actual compressed
1.856 + * data, pad_before, and pad_after. The padding bytes will be preserved
1.857 + * (not overwritten).
1.858 + *
1.859 + * Note that calling this function does not guarantee that the returned
1.860 + * compressed data will be placed into the specified buffer. In the
1.861 + * event that the encoded data will not fit into the buffer provided,
1.862 + * the returned packet \ref MAY point to an internal buffer, as it would
1.863 + * if this call were never used. In this event, the output packet will
1.864 + * NOT have any padding, and the application must free space and copy it
1.865 + * to the proper place. This is of particular note in configurations
1.866 + * that may output multiple packets for a single encoded frame (e.g., lagged
1.867 + * encoding) or if the application does not reset the buffer periodically.
1.868 + *
1.869 + * Applications may restore the default behavior of the codec providing
1.870 + * the compressed data buffer by calling this function with a NULL
1.871 + * buffer.
1.872 + *
1.873 + * Applications \ref MUSTNOT call this function during iteration of
1.874 + * vpx_codec_get_cx_data().
1.875 + *
1.876 + * \param[in] ctx Pointer to this instance's context
1.877 + * \param[in] buf Buffer to store compressed data into
1.878 + * \param[in] pad_before Bytes to skip before writing compressed data
1.879 + * \param[in] pad_after Bytes to skip after writing compressed data
1.880 + *
1.881 + * \retval #VPX_CODEC_OK
1.882 + * The buffer was set successfully.
1.883 + * \retval #VPX_CODEC_INVALID_PARAM
1.884 + * A parameter was NULL, the image format is unsupported, etc.
1.885 + */
1.886 + vpx_codec_err_t vpx_codec_set_cx_data_buf(vpx_codec_ctx_t *ctx,
1.887 + const vpx_fixed_buf_t *buf,
1.888 + unsigned int pad_before,
1.889 + unsigned int pad_after);
1.890 +
1.891 +
1.892 + /*!\brief Encoded data iterator
1.893 + *
1.894 + * Iterates over a list of data packets to be passed from the encoder to the
1.895 + * application. The different kinds of packets available are enumerated in
1.896 + * #vpx_codec_cx_pkt_kind.
1.897 + *
1.898 + * #VPX_CODEC_CX_FRAME_PKT packets should be passed to the application's
1.899 + * muxer. Multiple compressed frames may be in the list.
1.900 + * #VPX_CODEC_STATS_PKT packets should be appended to a global buffer.
1.901 + *
1.902 + * The application \ref MUST silently ignore any packet kinds that it does
1.903 + * not recognize or support.
1.904 + *
1.905 + * The data buffers returned from this function are only guaranteed to be
1.906 + * valid until the application makes another call to any vpx_codec_* function.
1.907 + *
1.908 + * \param[in] ctx Pointer to this instance's context
1.909 + * \param[in,out] iter Iterator storage, initialized to NULL
1.910 + *
1.911 + * \return Returns a pointer to an output data packet (compressed frame data,
1.912 + * two-pass statistics, etc.) or NULL to signal end-of-list.
1.913 + *
1.914 + */
1.915 + const vpx_codec_cx_pkt_t *vpx_codec_get_cx_data(vpx_codec_ctx_t *ctx,
1.916 + vpx_codec_iter_t *iter);
1.917 +
1.918 +
1.919 + /*!\brief Get Preview Frame
1.920 + *
1.921 + * Returns an image that can be used as a preview. Shows the image as it would
1.922 + * exist at the decompressor. The application \ref MUST NOT write into this
1.923 + * image buffer.
1.924 + *
1.925 + * \param[in] ctx Pointer to this instance's context
1.926 + *
1.927 + * \return Returns a pointer to a preview image, or NULL if no image is
1.928 + * available.
1.929 + *
1.930 + */
1.931 + const vpx_image_t *vpx_codec_get_preview_frame(vpx_codec_ctx_t *ctx);
1.932 +
1.933 +
1.934 + /*!@} - end defgroup encoder*/
1.935 +#ifdef __cplusplus
1.936 +}
1.937 +#endif
1.938 +#endif
1.939 +
