Merge "remove more vp8 and vp9 only code"
diff --git a/examples.mk b/examples.mk
index f10bec6..eb77e73 100644
--- a/examples.mk
+++ b/examples.mk
@@ -97,22 +97,6 @@
endif
vpxenc.GUID = 548DEC74-7A15-4B2B-AFC3-AA102E7C25C1
vpxenc.DESCRIPTION = Full featured encoder
-ifeq ($(CONFIG_SPATIAL_SVC),yes)
- EXAMPLES-$(CONFIG_VP9_ENCODER) += vp9_spatial_svc_encoder.c
- vp9_spatial_svc_encoder.SRCS += args.c args.h
- vp9_spatial_svc_encoder.SRCS += ivfenc.c ivfenc.h
- vp9_spatial_svc_encoder.SRCS += tools_common.c tools_common.h
- vp9_spatial_svc_encoder.SRCS += video_common.h
- vp9_spatial_svc_encoder.SRCS += video_writer.h video_writer.c
- vp9_spatial_svc_encoder.SRCS += vpx_ports/msvc.h
- vp9_spatial_svc_encoder.SRCS += vpxstats.c vpxstats.h
- vp9_spatial_svc_encoder.GUID = 4A38598D-627D-4505-9C7B-D4020C84100D
- vp9_spatial_svc_encoder.DESCRIPTION = VP9 Spatial SVC Encoder
-endif
-
-ifneq ($(CONFIG_SHARED),yes)
-EXAMPLES-$(CONFIG_VP9_ENCODER) += resize_util.c
-endif
EXAMPLES-$(CONFIG_ENCODERS) += vpx_temporal_svc_encoder.c
vpx_temporal_svc_encoder.SRCS += ivfenc.c ivfenc.h
@@ -161,14 +145,6 @@
simple_encoder.SRCS += vpx_ports/msvc.h
simple_encoder.GUID = 4607D299-8A71-4D2C-9B1D-071899B6FBFD
simple_encoder.DESCRIPTION = Simplified encoder loop
-EXAMPLES-$(CONFIG_VP9_ENCODER) += vp9_lossless_encoder.c
-vp9_lossless_encoder.SRCS += ivfenc.h ivfenc.c
-vp9_lossless_encoder.SRCS += tools_common.h tools_common.c
-vp9_lossless_encoder.SRCS += video_common.h
-vp9_lossless_encoder.SRCS += video_writer.h video_writer.c
-vp9_lossless_encoder.SRCS += vpx_ports/msvc.h
-vp9_lossless_encoder.GUID = B63C7C88-5348-46DC-A5A6-CC151EF93366
-vp9_lossless_encoder.DESCRIPTION = Simplified lossless VP9 encoder
EXAMPLES-$(CONFIG_ENCODERS) += twopass_encoder.c
twopass_encoder.SRCS += ivfenc.h ivfenc.c
twopass_encoder.SRCS += tools_common.h tools_common.c
@@ -195,40 +171,17 @@
set_maps.SRCS += vpx_ports/msvc.h
set_maps.GUID = ECB2D24D-98B8-4015-A465-A4AF3DCC145F
set_maps.DESCRIPTION = Set active and ROI maps
-EXAMPLES-$(CONFIG_VP8_ENCODER) += vp8cx_set_ref.c
-vp8cx_set_ref.SRCS += ivfenc.h ivfenc.c
-vp8cx_set_ref.SRCS += tools_common.h tools_common.c
-vp8cx_set_ref.SRCS += video_common.h
-vp8cx_set_ref.SRCS += video_writer.h video_writer.c
-vp8cx_set_ref.SRCS += vpx_ports/msvc.h
-vp8cx_set_ref.GUID = C5E31F7F-96F6-48BD-BD3E-10EBF6E8057A
-vp8cx_set_ref.DESCRIPTION = VP8 set encoder reference frame
-ifeq ($(CONFIG_MULTI_RES_ENCODING),yes)
-ifeq ($(CONFIG_LIBYUV),yes)
-EXAMPLES-$(CONFIG_VP8_ENCODER) += vp8_multi_resolution_encoder.c
-vp8_multi_resolution_encoder.SRCS += ivfenc.h ivfenc.c
-vp8_multi_resolution_encoder.SRCS += tools_common.h tools_common.c
-vp8_multi_resolution_encoder.SRCS += video_writer.h video_writer.c
-vp8_multi_resolution_encoder.SRCS += vpx_ports/msvc.h
-vp8_multi_resolution_encoder.SRCS += $(LIBYUV_SRCS)
-vp8_multi_resolution_encoder.GUID = 04f8738e-63c8-423b-90fa-7c2703a374de
-vp8_multi_resolution_encoder.DESCRIPTION = VP8 Multiple-resolution Encoding
-endif
-endif
-
# Handle extra library flags depending on codec configuration
# We should not link to math library (libm) on RVCT
# when building for bare-metal targets
ifeq ($(CONFIG_OS_SUPPORT), yes)
-CODEC_EXTRA_LIBS-$(CONFIG_VP8) += m
-CODEC_EXTRA_LIBS-$(CONFIG_VP9) += m
+CODEC_EXTRA_LIBS-$(CONFIG_VP10) += m
else
ifeq ($(CONFIG_GCC), yes)
- CODEC_EXTRA_LIBS-$(CONFIG_VP8) += m
- CODEC_EXTRA_LIBS-$(CONFIG_VP9) += m
+ CODEC_EXTRA_LIBS-$(CONFIG_VP10) += m
endif
endif
#
@@ -245,10 +198,8 @@
INC_PATH-yes := $(SRC_PATH_BARE)/../include
else
LIB_PATH-yes += $(if $(BUILD_PFX),$(BUILD_PFX),.)
- INC_PATH-$(CONFIG_VP8_DECODER) += $(SRC_PATH_BARE)/vp8
- INC_PATH-$(CONFIG_VP8_ENCODER) += $(SRC_PATH_BARE)/vp8
- INC_PATH-$(CONFIG_VP9_DECODER) += $(SRC_PATH_BARE)/vp9
- INC_PATH-$(CONFIG_VP9_ENCODER) += $(SRC_PATH_BARE)/vp9
+ INC_PATH-$(CONFIG_VP10_DECODER) += $(SRC_PATH_BARE)/vp10
+ INC_PATH-$(CONFIG_VP10_ENCODER) += $(SRC_PATH_BARE)/vp10
endif
INC_PATH-$(CONFIG_LIBYUV) += $(SRC_PATH_BARE)/third_party/libyuv/include
LIB_PATH := $(call enabled,LIB_PATH)
diff --git a/examples/resize_util.c b/examples/resize_util.c
deleted file mode 100644
index e6fdd5b..0000000
--- a/examples/resize_util.c
+++ /dev/null
@@ -1,130 +0,0 @@
-/*
- * Copyright (c) 2014 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-#include <assert.h>
-#include <limits.h>
-#include <math.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "../tools_common.h"
-#include "../vp9/encoder/vp9_resize.h"
-
-static const char *exec_name = NULL;
-
-static void usage() {
- printf("Usage:\n");
- printf("%s <input_yuv> <width>x<height> <target_width>x<target_height> ",
- exec_name);
- printf("<output_yuv> [<frames>]\n");
-}
-
-void usage_exit(void) {
- usage();
- exit(EXIT_FAILURE);
-}
-
-static int parse_dim(char *v, int *width, int *height) {
- char *x = strchr(v, 'x');
- if (x == NULL)
- x = strchr(v, 'X');
- if (x == NULL)
- return 0;
- *width = atoi(v);
- *height = atoi(&x[1]);
- if (*width <= 0 || *height <= 0)
- return 0;
- else
- return 1;
-}
-
-int main(int argc, char *argv[]) {
- char *fin, *fout;
- FILE *fpin, *fpout;
- uint8_t *inbuf, *outbuf;
- uint8_t *inbuf_u, *outbuf_u;
- uint8_t *inbuf_v, *outbuf_v;
- int f, frames;
- int width, height, target_width, target_height;
-
- exec_name = argv[0];
-
- if (argc < 5) {
- printf("Incorrect parameters:\n");
- usage();
- return 1;
- }
-
- fin = argv[1];
- fout = argv[4];
- if (!parse_dim(argv[2], &width, &height)) {
- printf("Incorrect parameters: %s\n", argv[2]);
- usage();
- return 1;
- }
- if (!parse_dim(argv[3], &target_width, &target_height)) {
- printf("Incorrect parameters: %s\n", argv[3]);
- usage();
- return 1;
- }
-
- fpin = fopen(fin, "rb");
- if (fpin == NULL) {
- printf("Can't open file %s to read\n", fin);
- usage();
- return 1;
- }
- fpout = fopen(fout, "wb");
- if (fpout == NULL) {
- printf("Can't open file %s to write\n", fout);
- usage();
- return 1;
- }
- if (argc >= 6)
- frames = atoi(argv[5]);
- else
- frames = INT_MAX;
-
- printf("Input size: %dx%d\n",
- width, height);
- printf("Target size: %dx%d, Frames: ",
- target_width, target_height);
- if (frames == INT_MAX)
- printf("All\n");
- else
- printf("%d\n", frames);
-
- inbuf = (uint8_t*)malloc(width * height * 3 / 2);
- outbuf = (uint8_t*)malloc(target_width * target_height * 3 / 2);
- inbuf_u = inbuf + width * height;
- inbuf_v = inbuf_u + width * height / 4;
- outbuf_u = outbuf + target_width * target_height;
- outbuf_v = outbuf_u + target_width * target_height / 4;
- f = 0;
- while (f < frames) {
- if (fread(inbuf, width * height * 3 / 2, 1, fpin) != 1)
- break;
- vp9_resize_frame420(inbuf, width, inbuf_u, inbuf_v, width / 2,
- height, width,
- outbuf, target_width, outbuf_u, outbuf_v,
- target_width / 2,
- target_height, target_width);
- fwrite(outbuf, target_width * target_height * 3 / 2, 1, fpout);
- f++;
- }
- printf("%d frames processed\n", f);
- fclose(fpin);
- fclose(fpout);
-
- free(inbuf);
- free(outbuf);
- return 0;
-}
diff --git a/examples/vp8_multi_resolution_encoder.c b/examples/vp8_multi_resolution_encoder.c
deleted file mode 100644
index 0248ede..0000000
--- a/examples/vp8_multi_resolution_encoder.c
+++ /dev/null
@@ -1,729 +0,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-/*
- * This is an example demonstrating multi-resolution encoding in VP8.
- * High-resolution input video is down-sampled to lower-resolutions. The
- * encoder then encodes the video and outputs multiple bitstreams with
- * different resolutions.
- *
- * This test also allows for settings temporal layers for each spatial layer.
- * Different number of temporal layers per spatial stream may be used.
- * Currently up to 3 temporal layers per spatial stream (encoder) are supported
- * in this test.
- */
-
-#include "./vpx_config.h"
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdarg.h>
-#include <string.h>
-#include <math.h>
-#include <assert.h>
-#include <sys/time.h>
-#include "vpx_ports/vpx_timer.h"
-#include "vpx/vpx_encoder.h"
-#include "vpx/vp8cx.h"
-#include "vpx_ports/mem_ops.h"
-#include "../tools_common.h"
-#define interface (vpx_codec_vp8_cx())
-#define fourcc 0x30385056
-
-void usage_exit(void) {
- exit(EXIT_FAILURE);
-}
-
-/*
- * The input video frame is downsampled several times to generate a multi-level
- * hierarchical structure. NUM_ENCODERS is defined as the number of encoding
- * levels required. For example, if the size of input video is 1280x720,
- * NUM_ENCODERS is 3, and down-sampling factor is 2, the encoder outputs 3
- * bitstreams with resolution of 1280x720(level 0), 640x360(level 1), and
- * 320x180(level 2) respectively.
- */
-
-/* Number of encoders (spatial resolutions) used in this test. */
-#define NUM_ENCODERS 3
-
-/* Maximum number of temporal layers allowed for this test. */
-#define MAX_NUM_TEMPORAL_LAYERS 3
-
-/* This example uses the scaler function in libyuv. */
-#include "third_party/libyuv/include/libyuv/basic_types.h"
-#include "third_party/libyuv/include/libyuv/scale.h"
-#include "third_party/libyuv/include/libyuv/cpu_id.h"
-
-int (*read_frame_p)(FILE *f, vpx_image_t *img);
-
-static int read_frame(FILE *f, vpx_image_t *img) {
- size_t nbytes, to_read;
- int res = 1;
-
- to_read = img->w*img->h*3/2;
- nbytes = fread(img->planes[0], 1, to_read, f);
- if(nbytes != to_read) {
- res = 0;
- if(nbytes > 0)
- printf("Warning: Read partial frame. Check your width & height!\n");
- }
- return res;
-}
-
-static int read_frame_by_row(FILE *f, vpx_image_t *img) {
- size_t nbytes, to_read;
- int res = 1;
- int plane;
-
- for (plane = 0; plane < 3; plane++)
- {
- 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++)
- {
- to_read = w;
-
- nbytes = fread(ptr, 1, to_read, f);
- if(nbytes != to_read) {
- res = 0;
- if(nbytes > 0)
- printf("Warning: Read partial frame. Check your width & height!\n");
- break;
- }
-
- ptr += img->stride[plane];
- }
- if (!res)
- break;
- }
-
- return res;
-}
-
-static void write_ivf_file_header(FILE *outfile,
- const vpx_codec_enc_cfg_t *cfg,
- 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 */
-
- (void) fwrite(header, 1, 32, outfile);
-}
-
-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);
-
- (void) fwrite(header, 1, 12, outfile);
-}
-
-/* Temporal scaling parameters */
-/* This sets all the temporal layer parameters given |num_temporal_layers|,
- * including the target bit allocation across temporal layers. Bit allocation
- * parameters will be passed in as user parameters in another version.
- */
-static void set_temporal_layer_pattern(int num_temporal_layers,
- vpx_codec_enc_cfg_t *cfg,
- int bitrate,
- int *layer_flags)
-{
- assert(num_temporal_layers <= MAX_NUM_TEMPORAL_LAYERS);
- switch (num_temporal_layers)
- {
- case 1:
- {
- /* 1-layer */
- cfg->ts_number_layers = 1;
- cfg->ts_periodicity = 1;
- cfg->ts_rate_decimator[0] = 1;
- cfg->ts_layer_id[0] = 0;
- cfg->ts_target_bitrate[0] = bitrate;
-
- // Update L only.
- layer_flags[0] = VP8_EFLAG_NO_UPD_GF | VP8_EFLAG_NO_UPD_ARF;
- break;
- }
-
- case 2:
- {
- /* 2-layers, with sync point at first frame of layer 1. */
- cfg->ts_number_layers = 2;
- cfg->ts_periodicity = 2;
- cfg->ts_rate_decimator[0] = 2;
- cfg->ts_rate_decimator[1] = 1;
- cfg->ts_layer_id[0] = 0;
- cfg->ts_layer_id[1] = 1;
- // Use 60/40 bit allocation as example.
- cfg->ts_target_bitrate[0] = 0.6f * bitrate;
- cfg->ts_target_bitrate[1] = bitrate;
-
- /* 0=L, 1=GF */
- // ARF is used as predictor for all frames, and is only updated on
- // key frame. Sync point every 8 frames.
-
- // Layer 0: predict from L and ARF, update L and G.
- layer_flags[0] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_UPD_ARF;
-
- // Layer 1: sync point: predict from L and ARF, and update G.
- layer_flags[1] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_ARF;
-
- // Layer 0, predict from L and ARF, update L.
- layer_flags[2] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF;
-
- // Layer 1: predict from L, G and ARF, and update G.
- layer_flags[3] = VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_ENTROPY;
-
- // Layer 0
- layer_flags[4] = layer_flags[2];
-
- // Layer 1
- layer_flags[5] = layer_flags[3];
-
- // Layer 0
- layer_flags[6] = layer_flags[4];
-
- // Layer 1
- layer_flags[7] = layer_flags[5];
- break;
- }
-
- case 3:
- default:
- {
- // 3-layers structure where ARF is used as predictor for all frames,
- // and is only updated on key frame.
- // Sync points for layer 1 and 2 every 8 frames.
- cfg->ts_number_layers = 3;
- cfg->ts_periodicity = 4;
- cfg->ts_rate_decimator[0] = 4;
- cfg->ts_rate_decimator[1] = 2;
- cfg->ts_rate_decimator[2] = 1;
- cfg->ts_layer_id[0] = 0;
- cfg->ts_layer_id[1] = 2;
- cfg->ts_layer_id[2] = 1;
- cfg->ts_layer_id[3] = 2;
- // Use 40/20/40 bit allocation as example.
- cfg->ts_target_bitrate[0] = 0.4f * bitrate;
- cfg->ts_target_bitrate[1] = 0.6f * bitrate;
- cfg->ts_target_bitrate[2] = bitrate;
-
- /* 0=L, 1=GF, 2=ARF */
-
- // Layer 0: predict from L and ARF; update L and G.
- layer_flags[0] = VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_REF_GF;
-
- // Layer 2: sync point: predict from L and ARF; update none.
- layer_flags[1] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_ENTROPY;
-
- // Layer 1: sync point: predict from L and ARF; update G.
- layer_flags[2] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_UPD_LAST;
-
- // Layer 2: predict from L, G, ARF; update none.
- layer_flags[3] = VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_ENTROPY;
-
- // Layer 0: predict from L and ARF; update L.
- layer_flags[4] = VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_REF_GF;
-
- // Layer 2: predict from L, G, ARF; update none.
- layer_flags[5] = layer_flags[3];
-
- // Layer 1: predict from L, G, ARF; update G.
- layer_flags[6] = VP8_EFLAG_NO_UPD_ARF |
- VP8_EFLAG_NO_UPD_LAST;
-
- // Layer 2: predict from L, G, ARF; update none.
- layer_flags[7] = layer_flags[3];
- break;
- }
- }
-}
-
-/* The periodicity of the pattern given the number of temporal layers. */
-static int periodicity_to_num_layers[MAX_NUM_TEMPORAL_LAYERS] = {1, 8, 8};
-
-int main(int argc, char **argv)
-{
- FILE *infile, *outfile[NUM_ENCODERS];
- FILE *downsampled_input[NUM_ENCODERS - 1];
- char filename[50];
- vpx_codec_ctx_t codec[NUM_ENCODERS];
- vpx_codec_enc_cfg_t cfg[NUM_ENCODERS];
- int frame_cnt = 0;
- vpx_image_t raw[NUM_ENCODERS];
- vpx_codec_err_t res[NUM_ENCODERS];
-
- int i;
- long width;
- long height;
- int length_frame;
- int frame_avail;
- int got_data;
- int flags = 0;
- int layer_id = 0;
-
- int layer_flags[VPX_TS_MAX_PERIODICITY * NUM_ENCODERS]
- = {0};
- int flag_periodicity;
-
- /*Currently, only realtime mode is supported in multi-resolution encoding.*/
- int arg_deadline = VPX_DL_REALTIME;
-
- /* Set show_psnr to 1/0 to show/not show PSNR. Choose show_psnr=0 if you
- don't need to know PSNR, which will skip PSNR calculation and save
- encoding time. */
- int show_psnr = 0;
- int key_frame_insert = 0;
- uint64_t psnr_sse_total[NUM_ENCODERS] = {0};
- uint64_t psnr_samples_total[NUM_ENCODERS] = {0};
- double psnr_totals[NUM_ENCODERS][4] = {{0,0}};
- int psnr_count[NUM_ENCODERS] = {0};
-
- double cx_time = 0;
- struct timeval tv1, tv2, difftv;
-
- /* Set the required target bitrates for each resolution level.
- * If target bitrate for highest-resolution level is set to 0,
- * (i.e. target_bitrate[0]=0), we skip encoding at that level.
- */
- unsigned int target_bitrate[NUM_ENCODERS]={1000, 500, 100};
-
- /* Enter the frame rate of the input video */
- int framerate = 30;
-
- /* Set down-sampling factor for each resolution level.
- dsf[0] controls down sampling from level 0 to level 1;
- dsf[1] controls down sampling from level 1 to level 2;
- dsf[2] is not used. */
- vpx_rational_t dsf[NUM_ENCODERS] = {{2, 1}, {2, 1}, {1, 1}};
-
- /* Set the number of temporal layers for each encoder/resolution level,
- * starting from highest resoln down to lowest resoln. */
- unsigned int num_temporal_layers[NUM_ENCODERS] = {3, 3, 3};
-
- if(argc!= (7 + 3 * NUM_ENCODERS))
- die("Usage: %s <width> <height> <frame_rate> <infile> <outfile(s)> "
- "<rate_encoder(s)> <temporal_layer(s)> <key_frame_insert> <output psnr?> \n",
- argv[0]);
-
- printf("Using %s\n",vpx_codec_iface_name(interface));
-
- width = strtol(argv[1], NULL, 0);
- height = strtol(argv[2], NULL, 0);
- framerate = strtol(argv[3], NULL, 0);
-
- if(width < 16 || width%2 || height <16 || height%2)
- die("Invalid resolution: %ldx%ld", width, height);
-
- /* Open input video file for encoding */
- if(!(infile = fopen(argv[4], "rb")))
- die("Failed to open %s for reading", argv[4]);
-
- /* Open output file for each encoder to output bitstreams */
- for (i=0; i< NUM_ENCODERS; i++)
- {
- if(!target_bitrate[i])
- {
- outfile[i] = NULL;
- continue;
- }
-
- if(!(outfile[i] = fopen(argv[i+5], "wb")))
- die("Failed to open %s for writing", argv[i+4]);
- }
-
- // Bitrates per spatial layer: overwrite default rates above.
- for (i=0; i< NUM_ENCODERS; i++)
- {
- target_bitrate[i] = strtol(argv[NUM_ENCODERS + 5 + i], NULL, 0);
- }
-
- // Temporal layers per spatial layers: overwrite default settings above.
- for (i=0; i< NUM_ENCODERS; i++)
- {
- num_temporal_layers[i] = strtol(argv[2 * NUM_ENCODERS + 5 + i], NULL, 0);
- if (num_temporal_layers[i] < 1 || num_temporal_layers[i] > 3)
- die("Invalid temporal layers: %d, Must be 1, 2, or 3. \n",
- num_temporal_layers);
- }
-
- /* Open file to write out each spatially downsampled input stream. */
- for (i=0; i< NUM_ENCODERS - 1; i++)
- {
- // Highest resoln is encoder 0.
- if (sprintf(filename,"ds%d.yuv",NUM_ENCODERS - i) < 0)
- {
- return EXIT_FAILURE;
- }
- downsampled_input[i] = fopen(filename,"wb");
- }
-
- key_frame_insert = strtol(argv[3 * NUM_ENCODERS + 5], NULL, 0);
-
- show_psnr = strtol(argv[3 * NUM_ENCODERS + 6], NULL, 0);
-
-
- /* Populate default encoder configuration */
- for (i=0; i< NUM_ENCODERS; i++)
- {
- res[i] = vpx_codec_enc_config_default(interface, &cfg[i], 0);
- if(res[i]) {
- printf("Failed to get config: %s\n", vpx_codec_err_to_string(res[i]));
- return EXIT_FAILURE;
- }
- }
-
- /*
- * Update the default configuration according to needs of the application.
- */
- /* Highest-resolution encoder settings */
- cfg[0].g_w = width;
- cfg[0].g_h = height;
- cfg[0].rc_dropframe_thresh = 0;
- cfg[0].rc_end_usage = VPX_CBR;
- cfg[0].rc_resize_allowed = 0;
- cfg[0].rc_min_quantizer = 2;
- cfg[0].rc_max_quantizer = 56;
- cfg[0].rc_undershoot_pct = 100;
- cfg[0].rc_overshoot_pct = 15;
- cfg[0].rc_buf_initial_sz = 500;
- cfg[0].rc_buf_optimal_sz = 600;
- cfg[0].rc_buf_sz = 1000;
- cfg[0].g_error_resilient = 1; /* Enable error resilient mode */
- cfg[0].g_lag_in_frames = 0;
-
- /* Disable automatic keyframe placement */
- /* Note: These 3 settings are copied to all levels. But, except the lowest
- * resolution level, all other levels are set to VPX_KF_DISABLED internally.
- */
- cfg[0].kf_mode = VPX_KF_AUTO;
- cfg[0].kf_min_dist = 3000;
- cfg[0].kf_max_dist = 3000;
-
- cfg[0].rc_target_bitrate = target_bitrate[0]; /* Set target bitrate */
- cfg[0].g_timebase.num = 1; /* Set fps */
- cfg[0].g_timebase.den = framerate;
-
- /* Other-resolution encoder settings */
- for (i=1; i< NUM_ENCODERS; i++)
- {
- memcpy(&cfg[i], &cfg[0], sizeof(vpx_codec_enc_cfg_t));
-
- cfg[i].rc_target_bitrate = target_bitrate[i];
-
- /* Note: Width & height of other-resolution encoders are calculated
- * from the highest-resolution encoder's size and the corresponding
- * down_sampling_factor.
- */
- {
- unsigned int iw = cfg[i-1].g_w*dsf[i-1].den + dsf[i-1].num - 1;
- unsigned int ih = cfg[i-1].g_h*dsf[i-1].den + dsf[i-1].num - 1;
- cfg[i].g_w = iw/dsf[i-1].num;
- cfg[i].g_h = ih/dsf[i-1].num;
- }
-
- /* Make width & height to be multiplier of 2. */
- // Should support odd size ???
- if((cfg[i].g_w)%2)cfg[i].g_w++;
- if((cfg[i].g_h)%2)cfg[i].g_h++;
- }
-
-
- // Set the number of threads per encode/spatial layer.
- // (1, 1, 1) means no encoder threading.
- cfg[0].g_threads = 2;
- cfg[1].g_threads = 1;
- cfg[2].g_threads = 1;
-
- /* Allocate image for each encoder */
- for (i=0; i< NUM_ENCODERS; i++)
- if(!vpx_img_alloc(&raw[i], VPX_IMG_FMT_I420, cfg[i].g_w, cfg[i].g_h, 32))
- die("Failed to allocate image", cfg[i].g_w, cfg[i].g_h);
-
- if (raw[0].stride[VPX_PLANE_Y] == raw[0].d_w)
- read_frame_p = read_frame;
- else
- read_frame_p = read_frame_by_row;
-
- for (i=0; i< NUM_ENCODERS; i++)
- if(outfile[i])
- write_ivf_file_header(outfile[i], &cfg[i], 0);
-
- /* Temporal layers settings */
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- set_temporal_layer_pattern(num_temporal_layers[i],
- &cfg[i],
- cfg[i].rc_target_bitrate,
- &layer_flags[i * VPX_TS_MAX_PERIODICITY]);
- }
-
- /* Initialize multi-encoder */
- if(vpx_codec_enc_init_multi(&codec[0], interface, &cfg[0], NUM_ENCODERS,
- (show_psnr ? VPX_CODEC_USE_PSNR : 0), &dsf[0]))
- die_codec(&codec[0], "Failed to initialize encoder");
-
- /* The extra encoding configuration parameters can be set as follows. */
- /* Set encoding speed */
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- int speed = -6;
- /* Lower speed for the lowest resolution. */
- if (i == NUM_ENCODERS - 1) speed = -4;
- if(vpx_codec_control(&codec[i], VP8E_SET_CPUUSED, speed))
- die_codec(&codec[i], "Failed to set cpu_used");
- }
-
- /* Set static threshold = 1 for all encoders */
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- if(vpx_codec_control(&codec[i], VP8E_SET_STATIC_THRESHOLD, 1))
- die_codec(&codec[i], "Failed to set static threshold");
- }
-
- /* Set NOISE_SENSITIVITY to do TEMPORAL_DENOISING */
- /* Enable denoising for the highest-resolution encoder. */
- if(vpx_codec_control(&codec[0], VP8E_SET_NOISE_SENSITIVITY, 1))
- die_codec(&codec[0], "Failed to set noise_sensitivity");
- for ( i=1; i< NUM_ENCODERS; i++)
- {
- if(vpx_codec_control(&codec[i], VP8E_SET_NOISE_SENSITIVITY, 0))
- die_codec(&codec[i], "Failed to set noise_sensitivity");
- }
-
- /* Set the number of token partitions */
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- if(vpx_codec_control(&codec[i], VP8E_SET_TOKEN_PARTITIONS, 1))
- die_codec(&codec[i], "Failed to set static threshold");
- }
-
- /* Set the max intra target bitrate */
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- unsigned int max_intra_size_pct =
- (int)(((double)cfg[0].rc_buf_optimal_sz * 0.5) * framerate / 10);
- if(vpx_codec_control(&codec[i], VP8E_SET_MAX_INTRA_BITRATE_PCT,
- max_intra_size_pct))
- die_codec(&codec[i], "Failed to set static threshold");
- //printf("%d %d \n",i,max_intra_size_pct);
- }
-
- frame_avail = 1;
- got_data = 0;
-
- while(frame_avail || got_data)
- {
- vpx_codec_iter_t iter[NUM_ENCODERS]={NULL};
- const vpx_codec_cx_pkt_t *pkt[NUM_ENCODERS];
-
- flags = 0;
- frame_avail = read_frame_p(infile, &raw[0]);
-
- if(frame_avail)
- {
- for ( i=1; i<NUM_ENCODERS; i++)
- {
- /*Scale the image down a number of times by downsampling factor*/
- /* FilterMode 1 or 2 give better psnr than FilterMode 0. */
- I420Scale(raw[i-1].planes[VPX_PLANE_Y], raw[i-1].stride[VPX_PLANE_Y],
- raw[i-1].planes[VPX_PLANE_U], raw[i-1].stride[VPX_PLANE_U],
- raw[i-1].planes[VPX_PLANE_V], raw[i-1].stride[VPX_PLANE_V],
- raw[i-1].d_w, raw[i-1].d_h,
- raw[i].planes[VPX_PLANE_Y], raw[i].stride[VPX_PLANE_Y],
- raw[i].planes[VPX_PLANE_U], raw[i].stride[VPX_PLANE_U],
- raw[i].planes[VPX_PLANE_V], raw[i].stride[VPX_PLANE_V],
- raw[i].d_w, raw[i].d_h, 1);
- /* Write out down-sampled input. */
- length_frame = cfg[i].g_w * cfg[i].g_h *3/2;
- if (fwrite(raw[i].planes[0], 1, length_frame,
- downsampled_input[NUM_ENCODERS - i - 1]) !=
- length_frame)
- {
- return EXIT_FAILURE;
- }
- }
- }
-
- /* Set the flags (reference and update) for all the encoders.*/
- for ( i=0; i<NUM_ENCODERS; i++)
- {
- layer_id = cfg[i].ts_layer_id[frame_cnt % cfg[i].ts_periodicity];
- flags = 0;
- flag_periodicity = periodicity_to_num_layers
- [num_temporal_layers[i] - 1];
- flags = layer_flags[i * VPX_TS_MAX_PERIODICITY +
- frame_cnt % flag_periodicity];
- // Key frame flag for first frame.
- if (frame_cnt == 0)
- {
- flags |= VPX_EFLAG_FORCE_KF;
- }
- if (frame_cnt > 0 && frame_cnt == key_frame_insert)
- {
- flags = VPX_EFLAG_FORCE_KF;
- }
-
- vpx_codec_control(&codec[i], VP8E_SET_FRAME_FLAGS, flags);
- vpx_codec_control(&codec[i], VP8E_SET_TEMPORAL_LAYER_ID, layer_id);
- }
-
- gettimeofday(&tv1, NULL);
- /* Encode each frame at multi-levels */
- /* Note the flags must be set to 0 in the encode call if they are set
- for each frame with the vpx_codec_control(), as done above. */
- if(vpx_codec_encode(&codec[0], frame_avail? &raw[0] : NULL,
- frame_cnt, 1, 0, arg_deadline))
- {
- die_codec(&codec[0], "Failed to encode frame");
- }
- gettimeofday(&tv2, NULL);
- timersub(&tv2, &tv1, &difftv);
- cx_time += (double)(difftv.tv_sec * 1000000 + difftv.tv_usec);
- for (i=NUM_ENCODERS-1; i>=0 ; i--)
- {
- got_data = 0;
- while( (pkt[i] = vpx_codec_get_cx_data(&codec[i], &iter[i])) )
- {
- got_data = 1;
- switch(pkt[i]->kind) {
- case VPX_CODEC_CX_FRAME_PKT:
- write_ivf_frame_header(outfile[i], pkt[i]);
- (void) fwrite(pkt[i]->data.frame.buf, 1,
- pkt[i]->data.frame.sz, outfile[i]);
- break;
- case VPX_CODEC_PSNR_PKT:
- if (show_psnr)
- {
- int j;
-
- psnr_sse_total[i] += pkt[i]->data.psnr.sse[0];
- psnr_samples_total[i] += pkt[i]->data.psnr.samples[0];
- for (j = 0; j < 4; j++)
- {
- psnr_totals[i][j] += pkt[i]->data.psnr.psnr[j];
- }
- psnr_count[i]++;
- }
-
- break;
- default:
- break;
- }
- printf(pkt[i]->kind == VPX_CODEC_CX_FRAME_PKT
- && (pkt[i]->data.frame.flags & VPX_FRAME_IS_KEY)? "K":"");
- fflush(stdout);
- }
- }
- frame_cnt++;
- }
- printf("\n");
- printf("FPS for encoding %d %f %f \n", frame_cnt, (float)cx_time / 1000000,
- 1000000 * (double)frame_cnt / (double)cx_time);
-
- fclose(infile);
-
- printf("Processed %ld frames.\n",(long int)frame_cnt-1);
- for (i=0; i< NUM_ENCODERS; i++)
- {
- /* Calculate PSNR and print it out */
- if ( (show_psnr) && (psnr_count[i]>0) )
- {
- int j;
- double ovpsnr = sse_to_psnr(psnr_samples_total[i], 255.0,
- psnr_sse_total[i]);
-
- fprintf(stderr, "\n ENC%d PSNR (Overall/Avg/Y/U/V)", i);
-
- fprintf(stderr, " %.3lf", ovpsnr);
- for (j = 0; j < 4; j++)
- {
- fprintf(stderr, " %.3lf", psnr_totals[i][j]/psnr_count[i]);
- }
- }
-
- if(vpx_codec_destroy(&codec[i]))
- die_codec(&codec[i], "Failed to destroy codec");
-
- vpx_img_free(&raw[i]);
-
- if(!outfile[i])
- continue;
-
- /* Try to rewrite the file header with the actual frame count */
- if(!fseek(outfile[i], 0, SEEK_SET))
- write_ivf_file_header(outfile[i], &cfg[i], frame_cnt-1);
- fclose(outfile[i]);
- }
- printf("\n");
-
- return EXIT_SUCCESS;
-}
diff --git a/examples/vp8cx_set_ref.c b/examples/vp8cx_set_ref.c
deleted file mode 100644
index 8b4cc30..0000000
--- a/examples/vp8cx_set_ref.c
+++ /dev/null
@@ -1,194 +0,0 @@
-/*
- * Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-
-// VP8 Set Reference Frame
-// =======================
-//
-// This is an example demonstrating how to overwrite the VP8 encoder's
-// internal reference frame. In the sample we set the last frame to the
-// current frame. If this is done at a cut scene it will avoid a keyframe.
-// This technique could be used to bounce between two cameras.
-//
-// Note that the decoder would also have to set the reference frame to the
-// same value on the same frame, or the video will become corrupt.
-//
-// Usage
-// -----
-// This example adds a single argument to the `simple_encoder` example,
-// which specifies the frame number to update the reference frame on.
-// The parameter is parsed as follows:
-//
-//
-// Extra Variables
-// ---------------
-// This example maintains the frame number passed on the command line
-// in the `update_frame_num` variable.
-//
-//
-// Configuration
-// -------------
-//
-// The reference frame is updated on the frame specified on the command
-// line.
-//
-// Observing The Effects
-// ---------------------
-// Use the `simple_encoder` example to encode a sample with a cut scene.
-// Determine the frame number of the cut scene by looking for a generated
-// key-frame (indicated by a 'K'). Supply that frame number as an argument
-// to this example, and observe that no key-frame is generated.
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "vpx/vp8cx.h"
-#include "vpx/vpx_encoder.h"
-
-#include "../tools_common.h"
-#include "../video_writer.h"
-
-static const char *exec_name;
-
-void usage_exit(void) {
- fprintf(stderr, "Usage: %s <width> <height> <infile> <outfile> <frame>\n",
- exec_name);
- exit(EXIT_FAILURE);
-}
-
-static int encode_frame(vpx_codec_ctx_t *codec,
- vpx_image_t *img,
- int frame_index,
- VpxVideoWriter *writer) {
- int got_pkts = 0;
- vpx_codec_iter_t iter = NULL;
- const vpx_codec_cx_pkt_t *pkt = NULL;
- const vpx_codec_err_t res = vpx_codec_encode(codec, img, frame_index, 1, 0,
- VPX_DL_GOOD_QUALITY);
- if (res != VPX_CODEC_OK)
- die_codec(codec, "Failed to encode frame");
-
- while ((pkt = vpx_codec_get_cx_data(codec, &iter)) != NULL) {
- got_pkts = 1;
-
- if (pkt->kind == VPX_CODEC_CX_FRAME_PKT) {
- const int keyframe = (pkt->data.frame.flags & VPX_FRAME_IS_KEY) != 0;
- if (!vpx_video_writer_write_frame(writer,
- pkt->data.frame.buf,
- pkt->data.frame.sz,
- pkt->data.frame.pts)) {
- die_codec(codec, "Failed to write compressed frame");
- }
-
- printf(keyframe ? "K" : ".");
- fflush(stdout);
- }
- }
-
- return got_pkts;
-}
-
-int main(int argc, char **argv) {
- FILE *infile = NULL;
- vpx_codec_ctx_t codec = {0};
- vpx_codec_enc_cfg_t cfg = {0};
- int frame_count = 0;
- vpx_image_t raw;
- vpx_codec_err_t res;
- VpxVideoInfo info = {0};
- VpxVideoWriter *writer = NULL;
- const VpxInterface *encoder = NULL;
- int update_frame_num = 0;
- const int fps = 30; // TODO(dkovalev) add command line argument
- const int bitrate = 200; // kbit/s TODO(dkovalev) add command line argument
-
- exec_name = argv[0];
-
- if (argc != 6)
- die("Invalid number of arguments");
-
- // TODO(dkovalev): add vp9 support and rename the file accordingly
- encoder = get_vpx_encoder_by_name("vp8");
- if (!encoder)
- die("Unsupported codec.");
-
- update_frame_num = atoi(argv[5]);
- if (!update_frame_num)
- die("Couldn't parse frame number '%s'\n", argv[5]);
-
- info.codec_fourcc = encoder->fourcc;
- info.frame_width = strtol(argv[1], NULL, 0);
- info.frame_height = strtol(argv[2], NULL, 0);
- info.time_base.numerator = 1;
- info.time_base.denominator = fps;
-
- if (info.frame_width <= 0 ||
- info.frame_height <= 0 ||
- (info.frame_width % 2) != 0 ||
- (info.frame_height % 2) != 0) {
- die("Invalid frame size: %dx%d", info.frame_width, info.frame_height);
- }
-
- if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, info.frame_width,
- info.frame_height, 1)) {
- die("Failed to allocate image.");
- }
-
- printf("Using %s\n", vpx_codec_iface_name(encoder->codec_interface()));
-
- res = vpx_codec_enc_config_default(encoder->codec_interface(), &cfg, 0);
- if (res)
- die_codec(&codec, "Failed to get default codec config.");
-
- cfg.g_w = info.frame_width;
- cfg.g_h = info.frame_height;
- cfg.g_timebase.num = info.time_base.numerator;
- cfg.g_timebase.den = info.time_base.denominator;
- cfg.rc_target_bitrate = bitrate;
-
- writer = vpx_video_writer_open(argv[4], kContainerIVF, &info);
- if (!writer)
- die("Failed to open %s for writing.", argv[4]);
-
- if (!(infile = fopen(argv[3], "rb")))
- die("Failed to open %s for reading.", argv[3]);
-
- if (vpx_codec_enc_init(&codec, encoder->codec_interface(), &cfg, 0))
- die_codec(&codec, "Failed to initialize encoder");
-
- // Encode frames.
- while (vpx_img_read(&raw, infile)) {
- if (frame_count + 1 == update_frame_num) {
- vpx_ref_frame_t ref;
- ref.frame_type = VP8_LAST_FRAME;
- ref.img = raw;
- if (vpx_codec_control(&codec, VP8_SET_REFERENCE, &ref))
- die_codec(&codec, "Failed to set reference frame");
- }
-
- encode_frame(&codec, &raw, frame_count++, writer);
- }
-
- // Flush encoder.
- while (encode_frame(&codec, NULL, -1, writer)) {}
-
- printf("\n");
- fclose(infile);
- printf("Processed %d frames.\n", frame_count);
-
- vpx_img_free(&raw);
- if (vpx_codec_destroy(&codec))
- die_codec(&codec, "Failed to destroy codec.");
-
- vpx_video_writer_close(writer);
-
- return EXIT_SUCCESS;
-}
diff --git a/examples/vp9_lossless_encoder.c b/examples/vp9_lossless_encoder.c
deleted file mode 100644
index 8272516..0000000
--- a/examples/vp9_lossless_encoder.c
+++ /dev/null
@@ -1,144 +0,0 @@
-/*
- * Copyright (c) 2014 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-#include "vpx/vpx_encoder.h"
-#include "vpx/vp8cx.h"
-
-#include "../tools_common.h"
-#include "../video_writer.h"
-
-static const char *exec_name;
-
-void usage_exit(void) {
- fprintf(stderr, "vp9_lossless_encoder: Example demonstrating VP9 lossless "
- "encoding feature. Supports raw input only.\n");
- fprintf(stderr, "Usage: %s <width> <height> <infile> <outfile>\n", exec_name);
- exit(EXIT_FAILURE);
-}
-
-static int encode_frame(vpx_codec_ctx_t *codec,
- vpx_image_t *img,
- int frame_index,
- int flags,
- VpxVideoWriter *writer) {
- int got_pkts = 0;
- vpx_codec_iter_t iter = NULL;
- const vpx_codec_cx_pkt_t *pkt = NULL;
- const vpx_codec_err_t res = vpx_codec_encode(codec, img, frame_index, 1,
- flags, VPX_DL_GOOD_QUALITY);
- if (res != VPX_CODEC_OK)
- die_codec(codec, "Failed to encode frame");
-
- while ((pkt = vpx_codec_get_cx_data(codec, &iter)) != NULL) {
- got_pkts = 1;
-
- if (pkt->kind == VPX_CODEC_CX_FRAME_PKT) {
- const int keyframe = (pkt->data.frame.flags & VPX_FRAME_IS_KEY) != 0;
- if (!vpx_video_writer_write_frame(writer,
- pkt->data.frame.buf,
- pkt->data.frame.sz,
- pkt->data.frame.pts)) {
- die_codec(codec, "Failed to write compressed frame");
- }
- printf(keyframe ? "K" : ".");
- fflush(stdout);
- }
- }
-
- return got_pkts;
-}
-
-int main(int argc, char **argv) {
- FILE *infile = NULL;
- vpx_codec_ctx_t codec;
- vpx_codec_enc_cfg_t cfg;
- int frame_count = 0;
- vpx_image_t raw;
- vpx_codec_err_t res;
- VpxVideoInfo info = {0};
- VpxVideoWriter *writer = NULL;
- const VpxInterface *encoder = NULL;
- const int fps = 30;
-
- exec_name = argv[0];
-
- if (argc < 5)
- die("Invalid number of arguments");
-
- encoder = get_vpx_encoder_by_name("vp9");
- if (!encoder)
- die("Unsupported codec.");
-
- info.codec_fourcc = encoder->fourcc;
- info.frame_width = strtol(argv[1], NULL, 0);
- info.frame_height = strtol(argv[2], NULL, 0);
- info.time_base.numerator = 1;
- info.time_base.denominator = fps;
-
- if (info.frame_width <= 0 ||
- info.frame_height <= 0 ||
- (info.frame_width % 2) != 0 ||
- (info.frame_height % 2) != 0) {
- die("Invalid frame size: %dx%d", info.frame_width, info.frame_height);
- }
-
- if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, info.frame_width,
- info.frame_height, 1)) {
- die("Failed to allocate image.");
- }
-
- printf("Using %s\n", vpx_codec_iface_name(encoder->codec_interface()));
-
- res = vpx_codec_enc_config_default(encoder->codec_interface(), &cfg, 0);
- if (res)
- die_codec(&codec, "Failed to get default codec config.");
-
- cfg.g_w = info.frame_width;
- cfg.g_h = info.frame_height;
- cfg.g_timebase.num = info.time_base.numerator;
- cfg.g_timebase.den = info.time_base.denominator;
-
- writer = vpx_video_writer_open(argv[4], kContainerIVF, &info);
- if (!writer)
- die("Failed to open %s for writing.", argv[4]);
-
- if (!(infile = fopen(argv[3], "rb")))
- die("Failed to open %s for reading.", argv[3]);
-
- if (vpx_codec_enc_init(&codec, encoder->codec_interface(), &cfg, 0))
- die_codec(&codec, "Failed to initialize encoder");
-
- if (vpx_codec_control_(&codec, VP9E_SET_LOSSLESS, 1))
- die_codec(&codec, "Failed to use lossless mode");
-
- // Encode frames.
- while (vpx_img_read(&raw, infile)) {
- encode_frame(&codec, &raw, frame_count++, 0, writer);
- }
-
- // Flush encoder.
- while (encode_frame(&codec, NULL, -1, 0, writer)) {}
-
- printf("\n");
- fclose(infile);
- printf("Processed %d frames.\n", frame_count);
-
- vpx_img_free(&raw);
- if (vpx_codec_destroy(&codec))
- die_codec(&codec, "Failed to destroy codec.");
-
- vpx_video_writer_close(writer);
-
- return EXIT_SUCCESS;
-}
diff --git a/examples/vp9_spatial_svc_encoder.c b/examples/vp9_spatial_svc_encoder.c
deleted file mode 100644
index 5fac82b..0000000
--- a/examples/vp9_spatial_svc_encoder.c
+++ /dev/null
@@ -1,914 +0,0 @@
-/*
- * Copyright (c) 2012 The WebM project authors. All Rights Reserved.
- *
- * Use of this source code is governed by a BSD-style license
- * that can be found in the LICENSE file in the root of the source
- * tree. An additional intellectual property rights grant can be found
- * in the file PATENTS. All contributing project authors may
- * be found in the AUTHORS file in the root of the source tree.
- */
-
-/*
- * This is an example demonstrating how to implement a multi-layer
- * VP9 encoding scheme based on spatial scalability for video applications
- * that benefit from a scalable bitstream.
- */
-
-#include <math.h>
-#include <stdarg.h>
-#include <stdlib.h>
-#include <string.h>
-#include <time.h>
-
-
-#include "../args.h"
-#include "../tools_common.h"
-#include "../video_writer.h"
-
-#include "../vpx_ports/vpx_timer.h"
-#include "vpx/svc_context.h"
-#include "vpx/vp8cx.h"
-#include "vpx/vpx_encoder.h"
-#include "../vpxstats.h"
-#define OUTPUT_RC_STATS 1
-
-static const arg_def_t skip_frames_arg =
- ARG_DEF("s", "skip-frames", 1, "input frames to skip");
-static const arg_def_t frames_arg =
- ARG_DEF("f", "frames", 1, "number of frames to encode");
-static const arg_def_t threads_arg =
- ARG_DEF("th", "threads", 1, "number of threads to use");
-#if OUTPUT_RC_STATS
-static const arg_def_t output_rc_stats_arg =
- ARG_DEF("rcstat", "output_rc_stats", 1, "output rc stats");
-#endif
-static const arg_def_t width_arg = ARG_DEF("w", "width", 1, "source width");
-static const arg_def_t height_arg = ARG_DEF("h", "height", 1, "source height");
-static const arg_def_t timebase_arg =
- ARG_DEF("t", "timebase", 1, "timebase (num/den)");
-static const arg_def_t bitrate_arg = ARG_DEF(
- "b", "target-bitrate", 1, "encoding bitrate, in kilobits per second");
-static const arg_def_t spatial_layers_arg =
- ARG_DEF("sl", "spatial-layers", 1, "number of spatial SVC layers");
-static const arg_def_t temporal_layers_arg =
- ARG_DEF("tl", "temporal-layers", 1, "number of temporal SVC layers");
-static const arg_def_t temporal_layering_mode_arg =
- ARG_DEF("tlm", "temporal-layering-mode", 1, "temporal layering scheme."
- "VP9E_TEMPORAL_LAYERING_MODE");
-static const arg_def_t kf_dist_arg =
- ARG_DEF("k", "kf-dist", 1, "number of frames between keyframes");
-static const arg_def_t scale_factors_arg =
- ARG_DEF("r", "scale-factors", 1, "scale factors (lowest to highest layer)");
-static const arg_def_t passes_arg =
- 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 =
- ARG_DEF(NULL, "fpf", 1, "First pass statistics file name");
-static const arg_def_t min_q_arg =
- ARG_DEF(NULL, "min-q", 1, "Minimum quantizer");
-static const arg_def_t max_q_arg =
- ARG_DEF(NULL, "max-q", 1, "Maximum quantizer");
-static const arg_def_t min_bitrate_arg =
- ARG_DEF(NULL, "min-bitrate", 1, "Minimum bitrate");
-static const arg_def_t max_bitrate_arg =
- ARG_DEF(NULL, "max-bitrate", 1, "Maximum bitrate");
-static const arg_def_t lag_in_frame_arg =
- ARG_DEF(NULL, "lag-in-frames", 1, "Number of frame to input before "
- "generating any outputs");
-static const arg_def_t rc_end_usage_arg =
- ARG_DEF(NULL, "rc-end-usage", 1, "0 - 3: VBR, CBR, CQ, Q");
-static const arg_def_t speed_arg =
- ARG_DEF("sp", "speed", 1, "speed configuration");
-static const arg_def_t aqmode_arg =
- ARG_DEF("aq", "aqmode", 1, "aq-mode off/on");
-
-#if CONFIG_VPX_HIGHBITDEPTH
-static const struct arg_enum_list bitdepth_enum[] = {
- {"8", VPX_BITS_8},
- {"10", VPX_BITS_10},
- {"12", VPX_BITS_12},
- {NULL, 0}
-};
-
-static const arg_def_t bitdepth_arg =
- ARG_DEF_ENUM("d", "bit-depth", 1, "Bit depth for codec 8, 10 or 12. ",
- bitdepth_enum);
-#endif // CONFIG_VPX_HIGHBITDEPTH
-
-
-static const arg_def_t *svc_args[] = {
- &frames_arg, &width_arg, &height_arg,
- &timebase_arg, &bitrate_arg, &skip_frames_arg, &spatial_layers_arg,
- &kf_dist_arg, &scale_factors_arg, &passes_arg, &pass_arg,
- &fpf_name_arg, &min_q_arg, &max_q_arg, &min_bitrate_arg,
- &max_bitrate_arg, &temporal_layers_arg, &temporal_layering_mode_arg,
- &lag_in_frame_arg, &threads_arg, &aqmode_arg,
-#if OUTPUT_RC_STATS
- &output_rc_stats_arg,
-#endif
-
-#if CONFIG_VPX_HIGHBITDEPTH
- &bitdepth_arg,
-#endif
- &speed_arg,
- &rc_end_usage_arg, NULL
-};
-
-static const uint32_t default_frames_to_skip = 0;
-static const uint32_t default_frames_to_code = 60 * 60;
-static const uint32_t default_width = 1920;
-static const uint32_t default_height = 1080;
-static const uint32_t default_timebase_num = 1;
-static const uint32_t default_timebase_den = 60;
-static const uint32_t default_bitrate = 1000;
-static const uint32_t default_spatial_layers = 5;
-static const uint32_t default_temporal_layers = 1;
-static const uint32_t default_kf_dist = 100;
-static const uint32_t default_temporal_layering_mode = 0;
-static const uint32_t default_output_rc_stats = 0;
-static const int32_t default_speed = -1; // -1 means use library default.
-static const uint32_t default_threads = 0; // zero means use library default.
-
-typedef struct {
- const char *input_filename;
- const char *output_filename;
- uint32_t frames_to_code;
- uint32_t frames_to_skip;
- struct VpxInputContext input_ctx;
- stats_io_t rc_stats;
- int passes;
- int pass;
-} AppInput;
-
-static const char *exec_name;
-
-void usage_exit(void) {
- fprintf(stderr, "Usage: %s <options> input_filename output_filename\n",
- exec_name);
- fprintf(stderr, "Options:\n");
- arg_show_usage(stderr, svc_args);
- exit(EXIT_FAILURE);
-}
-
-static void parse_command_line(int argc, const char **argv_,
- AppInput *app_input, SvcContext *svc_ctx,
- vpx_codec_enc_cfg_t *enc_cfg) {
- struct arg arg = {0};
- char **argv = NULL;
- char **argi = NULL;
- char **argj = NULL;
- vpx_codec_err_t res;
- int passes = 0;
- int pass = 0;
- const char *fpf_file_name = NULL;
- unsigned int min_bitrate = 0;
- unsigned int max_bitrate = 0;
- char string_options[1024] = {0};
-
- // initialize SvcContext with parameters that will be passed to vpx_svc_init
- svc_ctx->log_level = SVC_LOG_DEBUG;
- svc_ctx->spatial_layers = default_spatial_layers;
- svc_ctx->temporal_layers = default_temporal_layers;
- svc_ctx->temporal_layering_mode = default_temporal_layering_mode;
-#if OUTPUT_RC_STATS
- svc_ctx->output_rc_stat = default_output_rc_stats;
-#endif
- svc_ctx->speed = default_speed;
- svc_ctx->threads = default_threads;
-
- // start with default encoder configuration
- res = vpx_codec_enc_config_default(vpx_codec_vp9_cx(), enc_cfg, 0);
- if (res) {
- die("Failed to get config: %s\n", vpx_codec_err_to_string(res));
- }
- // update enc_cfg with app default values
- enc_cfg->g_w = default_width;
- enc_cfg->g_h = default_height;
- enc_cfg->g_timebase.num = default_timebase_num;
- enc_cfg->g_timebase.den = default_timebase_den;
- enc_cfg->rc_target_bitrate = default_bitrate;
- enc_cfg->kf_min_dist = default_kf_dist;
- enc_cfg->kf_max_dist = default_kf_dist;
- enc_cfg->rc_end_usage = VPX_CQ;
-
- // initialize AppInput with default values
- app_input->frames_to_code = default_frames_to_code;
- app_input->frames_to_skip = default_frames_to_skip;
-
- // process command line options
- 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, &frames_arg, argi)) {
- app_input->frames_to_code = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &width_arg, argi)) {
- enc_cfg->g_w = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &height_arg, argi)) {
- enc_cfg->g_h = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &timebase_arg, argi)) {
- enc_cfg->g_timebase = arg_parse_rational(&arg);
- } else if (arg_match(&arg, &bitrate_arg, argi)) {
- enc_cfg->rc_target_bitrate = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &skip_frames_arg, argi)) {
- app_input->frames_to_skip = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &spatial_layers_arg, argi)) {
- svc_ctx->spatial_layers = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &temporal_layers_arg, argi)) {
- svc_ctx->temporal_layers = arg_parse_uint(&arg);
-#if OUTPUT_RC_STATS
- } else if (arg_match(&arg, &output_rc_stats_arg, argi)) {
- svc_ctx->output_rc_stat = arg_parse_uint(&arg);
-#endif
- } else if (arg_match(&arg, &speed_arg, argi)) {
- svc_ctx->speed = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &aqmode_arg, argi)) {
- svc_ctx->aqmode = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &threads_arg, argi)) {
- svc_ctx->threads = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &temporal_layering_mode_arg, argi)) {
- svc_ctx->temporal_layering_mode =
- enc_cfg->temporal_layering_mode = arg_parse_int(&arg);
- if (svc_ctx->temporal_layering_mode) {
- enc_cfg->g_error_resilient = 1;
- }
- } else if (arg_match(&arg, &kf_dist_arg, argi)) {
- enc_cfg->kf_min_dist = arg_parse_uint(&arg);
- enc_cfg->kf_max_dist = enc_cfg->kf_min_dist;
- } else if (arg_match(&arg, &scale_factors_arg, argi)) {
- snprintf(string_options, sizeof(string_options), "%s scale-factors=%s",
- string_options, arg.val);
- } else if (arg_match(&arg, &passes_arg, argi)) {
- passes = arg_parse_uint(&arg);
- if (passes < 1 || passes > 2) {
- die("Error: Invalid number of passes (%d)\n", passes);
- }
- } else if (arg_match(&arg, &pass_arg, argi)) {
- pass = arg_parse_uint(&arg);
- if (pass < 1 || pass > 2) {
- die("Error: Invalid pass selected (%d)\n", pass);
- }
- } else if (arg_match(&arg, &fpf_name_arg, argi)) {
- fpf_file_name = arg.val;
- } else if (arg_match(&arg, &min_q_arg, argi)) {
- snprintf(string_options, sizeof(string_options), "%s min-quantizers=%s",
- string_options, arg.val);
- } else if (arg_match(&arg, &max_q_arg, argi)) {
- snprintf(string_options, sizeof(string_options), "%s max-quantizers=%s",
- string_options, arg.val);
- } else if (arg_match(&arg, &min_bitrate_arg, argi)) {
- min_bitrate = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &max_bitrate_arg, argi)) {
- max_bitrate = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &lag_in_frame_arg, argi)) {
- enc_cfg->g_lag_in_frames = arg_parse_uint(&arg);
- } else if (arg_match(&arg, &rc_end_usage_arg, argi)) {
- enc_cfg->rc_end_usage = arg_parse_uint(&arg);
-#if CONFIG_VPX_HIGHBITDEPTH
- } else if (arg_match(&arg, &bitdepth_arg, argi)) {
- enc_cfg->g_bit_depth = arg_parse_enum_or_int(&arg);
- switch (enc_cfg->g_bit_depth) {
- case VPX_BITS_8:
- enc_cfg->g_input_bit_depth = 8;
- enc_cfg->g_profile = 0;
- break;
- case VPX_BITS_10:
- enc_cfg->g_input_bit_depth = 10;
- enc_cfg->g_profile = 2;
- break;
- case VPX_BITS_12:
- enc_cfg->g_input_bit_depth = 12;
- enc_cfg->g_profile = 2;
- break;
- default:
- die("Error: Invalid bit depth selected (%d)\n", enc_cfg->g_bit_depth);
- break;
- }
-#endif // CONFIG_VPX_HIGHBITDEPTH
- } else {
- ++argj;
- }
- }
-
- // There will be a space in front of the string options
- if (strlen(string_options) > 0)
- vpx_svc_set_options(svc_ctx, string_options + 1);
-
- if (passes == 0 || passes == 1) {
- if (pass) {
- fprintf(stderr, "pass is ignored since there's only one pass\n");
- }
- enc_cfg->g_pass = VPX_RC_ONE_PASS;
- } else {
- if (pass == 0) {
- die("pass must be specified when passes is 2\n");
- }
-
- if (fpf_file_name == NULL) {
- die("fpf must be specified when passes is 2\n");
- }
-
- if (pass == 1) {
- enc_cfg->g_pass = VPX_RC_FIRST_PASS;
- if (!stats_open_file(&app_input->rc_stats, fpf_file_name, 0)) {
- fatal("Failed to open statistics store");
- }
- } else {
- enc_cfg->g_pass = VPX_RC_LAST_PASS;
- if (!stats_open_file(&app_input->rc_stats, fpf_file_name, 1)) {
- fatal("Failed to open statistics store");
- }
- enc_cfg->rc_twopass_stats_in = stats_get(&app_input->rc_stats);
- }
- app_input->passes = passes;
- app_input->pass = pass;
- }
-
- if (enc_cfg->rc_target_bitrate > 0) {
- if (min_bitrate > 0) {
- enc_cfg->rc_2pass_vbr_minsection_pct =
- min_bitrate * 100 / enc_cfg->rc_target_bitrate;
- }
- if (max_bitrate > 0) {
- enc_cfg->rc_2pass_vbr_maxsection_pct =
- max_bitrate * 100 / enc_cfg->rc_target_bitrate;
- }
- }
-
- // Check for unrecognized options
- for (argi = argv; *argi; ++argi)
- if (argi[0][0] == '-' && strlen(argi[0]) > 1)
- die("Error: Unrecognized option %s\n", *argi);
-
- if (argv[0] == NULL || argv[1] == 0) {
- usage_exit();
- }
- app_input->input_filename = argv[0];
- app_input->output_filename = argv[1];
- free(argv);
-
- if (enc_cfg->g_w < 16 || enc_cfg->g_w % 2 || enc_cfg->g_h < 16 ||
- enc_cfg->g_h % 2)
- die("Invalid resolution: %d x %d\n", enc_cfg->g_w, enc_cfg->g_h);
-
- printf(
- "Codec %s\nframes: %d, skip: %d\n"
- "layers: %d\n"
- "width %d, height: %d,\n"
- "num: %d, den: %d, bitrate: %d,\n"
- "gop size: %d\n",
- vpx_codec_iface_name(vpx_codec_vp9_cx()), app_input->frames_to_code,
- app_input->frames_to_skip,
- svc_ctx->spatial_layers, enc_cfg->g_w, enc_cfg->g_h,
- enc_cfg->g_timebase.num, enc_cfg->g_timebase.den,
- enc_cfg->rc_target_bitrate, enc_cfg->kf_max_dist);
-}
-
-#if OUTPUT_RC_STATS
-// For rate control encoding stats.
-struct RateControlStats {
- // Number of input frames per layer.
- int layer_input_frames[VPX_MAX_LAYERS];
- // Total (cumulative) number of encoded frames per layer.
- int layer_tot_enc_frames[VPX_MAX_LAYERS];
- // Number of encoded non-key frames per layer.
- int layer_enc_frames[VPX_MAX_LAYERS];
- // Framerate per layer (cumulative).
- double layer_framerate[VPX_MAX_LAYERS];
- // Target average frame size per layer (per-frame-bandwidth per layer).
- double layer_pfb[VPX_MAX_LAYERS];
- // Actual average frame size per layer.
- double layer_avg_frame_size[VPX_MAX_LAYERS];
- // Average rate mismatch per layer (|target - actual| / target).
- double layer_avg_rate_mismatch[VPX_MAX_LAYERS];
- // Actual encoding bitrate per layer (cumulative).
- double layer_encoding_bitrate[VPX_MAX_LAYERS];
- // Average of the short-time encoder actual bitrate.
- // TODO(marpan): Should we add these short-time stats for each layer?
- double avg_st_encoding_bitrate;
- // Variance of the short-time encoder actual bitrate.
- double variance_st_encoding_bitrate;
- // Window (number of frames) for computing short-time encoding bitrate.
- int window_size;
- // Number of window measurements.
- int window_count;
-};
-
-// Note: these rate control stats assume only 1 key frame in the
-// sequence (i.e., first frame only).
-static void set_rate_control_stats(struct RateControlStats *rc,
- vpx_codec_enc_cfg_t *cfg) {
- unsigned int sl, tl;
- // Set the layer (cumulative) framerate and the target layer (non-cumulative)
- // per-frame-bandwidth, for the rate control encoding stats below.
- const double framerate = cfg->g_timebase.den / cfg->g_timebase.num;
-
- for (sl = 0; sl < cfg->ss_number_layers; ++sl) {
- for (tl = 0; tl < cfg->ts_number_layers; ++tl) {
- const int layer = sl * cfg->ts_number_layers + tl;
- const int tlayer0 = sl * cfg->ts_number_layers;
- if (cfg->ts_number_layers == 1)
- rc->layer_framerate[layer] = framerate;
- else
- rc->layer_framerate[layer] =
- framerate / cfg->ts_rate_decimator[tl];
- if (tl > 0) {
- rc->layer_pfb[layer] = 1000.0 *
- (cfg->layer_target_bitrate[layer] -
- cfg->layer_target_bitrate[layer - 1]) /
- (rc->layer_framerate[layer] -
- rc->layer_framerate[layer - 1]);
- } else {
- rc->layer_pfb[tlayer0] = 1000.0 *
- cfg->layer_target_bitrate[tlayer0] /
- rc->layer_framerate[tlayer0];
- }
- rc->layer_input_frames[layer] = 0;
- rc->layer_enc_frames[layer] = 0;
- rc->layer_tot_enc_frames[layer] = 0;
- rc->layer_encoding_bitrate[layer] = 0.0;
- rc->layer_avg_frame_size[layer] = 0.0;
- rc->layer_avg_rate_mismatch[layer] = 0.0;
- }
- }
- rc->window_count = 0;
- rc->window_size = 15;
- rc->avg_st_encoding_bitrate = 0.0;
- rc->variance_st_encoding_bitrate = 0.0;
-}
-
-static void printout_rate_control_summary(struct RateControlStats *rc,
- vpx_codec_enc_cfg_t *cfg,
- int frame_cnt) {
- unsigned int sl, tl;
- int tot_num_frames = 0;
- double perc_fluctuation = 0.0;
- printf("Total number of processed frames: %d\n\n", frame_cnt - 1);
- printf("Rate control layer stats for sl%d tl%d layer(s):\n\n",
- cfg->ss_number_layers, cfg->ts_number_layers);
- for (sl = 0; sl < cfg->ss_number_layers; ++sl) {
- for (tl = 0; tl < cfg->ts_number_layers; ++tl) {
- const int layer = sl * cfg->ts_number_layers + tl;
- const int num_dropped = (tl > 0) ?
- (rc->layer_input_frames[layer] - rc->layer_enc_frames[layer]) :
- (rc->layer_input_frames[layer] - rc->layer_enc_frames[layer] - 1);
- if (!sl)
- tot_num_frames += rc->layer_input_frames[layer];
- rc->layer_encoding_bitrate[layer] = 0.001 * rc->layer_framerate[layer] *
- rc->layer_encoding_bitrate[layer] / tot_num_frames;
- rc->layer_avg_frame_size[layer] = rc->layer_avg_frame_size[layer] /
- rc->layer_enc_frames[layer];
- rc->layer_avg_rate_mismatch[layer] =
- 100.0 * rc->layer_avg_rate_mismatch[layer] /
- rc->layer_enc_frames[layer];
- printf("For layer#: sl%d tl%d \n", sl, tl);
- printf("Bitrate (target vs actual): %d %f.0 kbps\n",
- cfg->layer_target_bitrate[layer],
- rc->layer_encoding_bitrate[layer]);
- printf("Average frame size (target vs actual): %f %f bits\n",
- rc->layer_pfb[layer], rc->layer_avg_frame_size[layer]);
- printf("Average rate_mismatch: %f\n",
- rc->layer_avg_rate_mismatch[layer]);
- printf("Number of input frames, encoded (non-key) frames, "
- "and percent dropped frames: %d %d %f.0 \n",
- rc->layer_input_frames[layer], rc->layer_enc_frames[layer],
- 100.0 * num_dropped / rc->layer_input_frames[layer]);
- printf("\n");
- }
- }
- rc->avg_st_encoding_bitrate = rc->avg_st_encoding_bitrate / rc->window_count;
- rc->variance_st_encoding_bitrate =
- rc->variance_st_encoding_bitrate / rc->window_count -
- (rc->avg_st_encoding_bitrate * rc->avg_st_encoding_bitrate);
- perc_fluctuation = 100.0 * sqrt(rc->variance_st_encoding_bitrate) /
- rc->avg_st_encoding_bitrate;
- printf("Short-time stats, for window of %d frames: \n", rc->window_size);
- printf("Average, rms-variance, and percent-fluct: %f %f %f \n",
- rc->avg_st_encoding_bitrate,
- sqrt(rc->variance_st_encoding_bitrate),
- perc_fluctuation);
- if (frame_cnt != tot_num_frames)
- die("Error: Number of input frames not equal to output encoded frames != "
- "%d tot_num_frames = %d\n", frame_cnt, tot_num_frames);
-}
-
-vpx_codec_err_t parse_superframe_index(const uint8_t *data,
- size_t data_sz,
- uint32_t sizes[8], int *count) {
- // A chunk ending with a byte matching 0xc0 is an invalid chunk unless
- // it is a super frame index. If the last byte of real video compression
- // data is 0xc0 the encoder must add a 0 byte. If we have the marker but
- // not the associated matching marker byte at the front of the index we have
- // an invalid bitstream and need to return an error.
-
- uint8_t marker;
-
- marker = *(data + data_sz - 1);
- *count = 0;
-
-
- if ((marker & 0xe0) == 0xc0) {
- const uint32_t frames = (marker & 0x7) + 1;
- const uint32_t mag = ((marker >> 3) & 0x3) + 1;
- const size_t index_sz = 2 + mag * frames;
-
- // This chunk is marked as having a superframe index but doesn't have
- // enough data for it, thus it's an invalid superframe index.
- if (data_sz < index_sz)
- return VPX_CODEC_CORRUPT_FRAME;
-
- {
- const uint8_t marker2 = *(data + data_sz - index_sz);
-
- // This chunk is marked as having a superframe index but doesn't have
- // the matching marker byte at the front of the index therefore it's an
- // invalid chunk.
- if (marker != marker2)
- return VPX_CODEC_CORRUPT_FRAME;
- }
-
- {
- // Found a valid superframe index.
- uint32_t i, j;
- const uint8_t *x = &data[data_sz - index_sz + 1];
-
- for (i = 0; i < frames; ++i) {
- uint32_t this_sz = 0;
-
- for (j = 0; j < mag; ++j)
- this_sz |= (*x++) << (j * 8);
- sizes[i] = this_sz;
- }
- *count = frames;
- }
- }
- return VPX_CODEC_OK;
-}
-#endif
-
-// Example pattern for spatial layers and 2 temporal layers used in the
-// bypass/flexible mode. The pattern corresponds to the pattern
-// VP9E_TEMPORAL_LAYERING_MODE_0101 (temporal_layering_mode == 2) used in
-// non-flexible mode.
-void set_frame_flags_bypass_mode(int sl, int tl, int num_spatial_layers,
- int is_key_frame,
- vpx_svc_ref_frame_config_t *ref_frame_config) {
- for (sl = 0; sl < num_spatial_layers; ++sl) {
- if (!tl) {
- if (!sl) {
- ref_frame_config->frame_flags[sl] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_REF_ARF |
- VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF;
- } else {
- if (is_key_frame) {
- ref_frame_config->frame_flags[sl] = VP8_EFLAG_NO_REF_LAST |
- VP8_EFLAG_NO_REF_ARF |
- VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF;
- } else {
- ref_frame_config->frame_flags[sl] = VP8_EFLAG_NO_REF_ARF |
- VP8_EFLAG_NO_UPD_GF |
- VP8_EFLAG_NO_UPD_ARF;
- }
- }
- } else if (tl == 1) {
- if (!sl) {
- ref_frame_config->frame_flags[sl] = VP8_EFLAG_NO_REF_GF |
- VP8_EFLAG_NO_REF_ARF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_GF;
- } else {
- ref_frame_config->frame_flags[sl] = VP8_EFLAG_NO_REF_ARF |
- VP8_EFLAG_NO_UPD_LAST |
- VP8_EFLAG_NO_UPD_GF;
- }
- }
- if (tl == 0) {
- ref_frame_config->lst_fb_idx[sl] = sl;
- if (sl)
- ref_frame_config->gld_fb_idx[sl] = sl - 1;
- else
- ref_frame_config->gld_fb_idx[sl] = 0;
- ref_frame_config->alt_fb_idx[sl] = 0;
- } else if (tl == 1) {
- ref_frame_config->lst_fb_idx[sl] = sl;
- ref_frame_config->gld_fb_idx[sl] = num_spatial_layers + sl - 1;
- ref_frame_config->alt_fb_idx[sl] = num_spatial_layers + sl;
- }
- }
-}
-
-int main(int argc, const char **argv) {
- AppInput app_input = {0};
- VpxVideoWriter *writer = NULL;
- VpxVideoInfo info = {0};
- vpx_codec_ctx_t codec;
- vpx_codec_enc_cfg_t enc_cfg;
- SvcContext svc_ctx;
- uint32_t i;
- uint32_t frame_cnt = 0;
- vpx_image_t raw;
- vpx_codec_err_t res;
- int pts = 0; /* PTS starts at 0 */
- int frame_duration = 1; /* 1 timebase tick per frame */
- FILE *infile = NULL;
- int end_of_stream = 0;
- int frames_received = 0;
-#if OUTPUT_RC_STATS
- VpxVideoWriter *outfile[VPX_TS_MAX_LAYERS] = {NULL};
- struct RateControlStats rc;
- vpx_svc_layer_id_t layer_id;
- vpx_svc_ref_frame_config_t ref_frame_config;
- int sl, tl;
- double sum_bitrate = 0.0;
- double sum_bitrate2 = 0.0;
- double framerate = 30.0;
-#endif
- struct vpx_usec_timer timer;
- int64_t cx_time = 0;
- memset(&svc_ctx, 0, sizeof(svc_ctx));
- svc_ctx.log_print = 1;
- exec_name = argv[0];
- parse_command_line(argc, argv, &app_input, &svc_ctx, &enc_cfg);
-
- // Allocate image buffer
-#if CONFIG_VPX_HIGHBITDEPTH
- if (!vpx_img_alloc(&raw, enc_cfg.g_input_bit_depth == 8 ?
- VPX_IMG_FMT_I420 : VPX_IMG_FMT_I42016,
- enc_cfg.g_w, enc_cfg.g_h, 32)) {
- die("Failed to allocate image %dx%d\n", enc_cfg.g_w, enc_cfg.g_h);
- }
-#else
- if (!vpx_img_alloc(&raw, VPX_IMG_FMT_I420, enc_cfg.g_w, enc_cfg.g_h, 32)) {
- die("Failed to allocate image %dx%d\n", enc_cfg.g_w, enc_cfg.g_h);
- }
-#endif // CONFIG_VPX_HIGHBITDEPTH
-
- if (!(infile = fopen(app_input.input_filename, "rb")))
- die("Failed to open %s for reading\n", app_input.input_filename);
-
- // Initialize codec
- if (vpx_svc_init(&svc_ctx, &codec, vpx_codec_vp9_cx(), &enc_cfg) !=
- VPX_CODEC_OK)
- die("Failed to initialize encoder\n");
-
-#if OUTPUT_RC_STATS
- if (svc_ctx.output_rc_stat) {
- set_rate_control_stats(&rc, &enc_cfg);
- framerate = enc_cfg.g_timebase.den / enc_cfg.g_timebase.num;
- }
-#endif
-
- info.codec_fourcc = VP9_FOURCC;
- info.time_base.numerator = enc_cfg.g_timebase.num;
- info.time_base.denominator = enc_cfg.g_timebase.den;
-
- if (!(app_input.passes == 2 && app_input.pass == 1)) {
- // We don't save the bitstream for the 1st pass on two pass rate control
- writer = vpx_video_writer_open(app_input.output_filename, kContainerIVF,
- &info);
- if (!writer)
- die("Failed to open %s for writing\n", app_input.output_filename);
- }
-#if OUTPUT_RC_STATS
- // For now, just write temporal layer streams.
- // TODO(wonkap): do spatial by re-writing superframe.
- if (svc_ctx.output_rc_stat) {
- for (tl = 0; tl < enc_cfg.ts_number_layers; ++tl) {
- char file_name[PATH_MAX];
-
- snprintf(file_name, sizeof(file_name), "%s_t%d.ivf",
- app_input.output_filename, tl);
- outfile[tl] = vpx_video_writer_open(file_name, kContainerIVF, &info);
- if (!outfile[tl])
- die("Failed to open %s for writing", file_name);
- }
- }
-#endif
-
- // skip initial frames
- for (i = 0; i < app_input.frames_to_skip; ++i)
- vpx_img_read(&raw, infile);
-
- if (svc_ctx.speed != -1)
- vpx_codec_control(&codec, VP8E_SET_CPUUSED, svc_ctx.speed);
- if (svc_ctx.threads)
- vpx_codec_control(&codec, VP9E_SET_TILE_COLUMNS, (svc_ctx.threads >> 1));
- if (svc_ctx.speed >= 5 && svc_ctx.aqmode == 1)
- vpx_codec_control(&codec, VP9E_SET_AQ_MODE, 3);
-
-
- // Encode frames
- while (!end_of_stream) {
- vpx_codec_iter_t iter = NULL;
- const vpx_codec_cx_pkt_t *cx_pkt;
- if (frame_cnt >= app_input.frames_to_code || !vpx_img_read(&raw, infile)) {
- // We need one extra vpx_svc_encode call at end of stream to flush
- // encoder and get remaining data
- end_of_stream = 1;
- }
-
- // For BYPASS/FLEXIBLE mode, set the frame flags (reference and updates)
- // and the buffer indices for each spatial layer of the current
- // (super)frame to be encoded. The temporal layer_id for the current frame
- // also needs to be set.
- // TODO(marpan): Should rename the "VP9E_TEMPORAL_LAYERING_MODE_BYPASS"
- // mode to "VP9E_LAYERING_MODE_BYPASS".
- if (svc_ctx.temporal_layering_mode == VP9E_TEMPORAL_LAYERING_MODE_BYPASS) {
- layer_id.spatial_layer_id = 0;
- // Example for 2 temporal layers.
- if (frame_cnt % 2 == 0)
- layer_id.temporal_layer_id = 0;
- else
- layer_id.temporal_layer_id = 1;
- // Note that we only set the temporal layer_id, since we are calling
- // the encode for the whole superframe. The encoder will internally loop
- // over all the spatial layers for the current superframe.
- vpx_codec_control(&codec, VP9E_SET_SVC_LAYER_ID, &layer_id);
- set_frame_flags_bypass_mode(sl, layer_id.temporal_layer_id,
- svc_ctx.spatial_layers,
- frame_cnt == 0,
- &ref_frame_config);
- vpx_codec_control(&codec, VP9E_SET_SVC_REF_FRAME_CONFIG,
- &ref_frame_config);
- // Keep track of input frames, to account for frame drops in rate control
- // stats/metrics.
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- ++rc.layer_input_frames[sl * enc_cfg.ts_number_layers +
- layer_id.temporal_layer_id];
- }
- }
-
- vpx_usec_timer_start(&timer);
- res = vpx_svc_encode(&svc_ctx, &codec, (end_of_stream ? NULL : &raw),
- pts, frame_duration, svc_ctx.speed >= 5 ?
- VPX_DL_REALTIME : VPX_DL_GOOD_QUALITY);
- vpx_usec_timer_mark(&timer);
- cx_time += vpx_usec_timer_elapsed(&timer);
-
- printf("%s", vpx_svc_get_message(&svc_ctx));
- if (res != VPX_CODEC_OK) {
- die_codec(&codec, "Failed to encode frame");
- }
-
- while ((cx_pkt = vpx_codec_get_cx_data(&codec, &iter)) != NULL) {
- switch (cx_pkt->kind) {
- case VPX_CODEC_CX_FRAME_PKT: {
- if (cx_pkt->data.frame.sz > 0) {
-#if OUTPUT_RC_STATS
- uint32_t sizes[8];
- int count = 0;
-#endif
- vpx_video_writer_write_frame(writer,
- cx_pkt->data.frame.buf,
- cx_pkt->data.frame.sz,
- cx_pkt->data.frame.pts);
-#if OUTPUT_RC_STATS
- // TODO(marpan/wonkap): Put this (to line728) in separate function.
- if (svc_ctx.output_rc_stat) {
- vpx_codec_control(&codec, VP9E_GET_SVC_LAYER_ID, &layer_id);
- parse_superframe_index(cx_pkt->data.frame.buf,
- cx_pkt->data.frame.sz, sizes, &count);
- // Note computing input_layer_frames here won't account for frame
- // drops in rate control stats.
- // TODO(marpan): Fix this for non-bypass mode so we can get stats
- // for dropped frames.
- if (svc_ctx.temporal_layering_mode !=
- VP9E_TEMPORAL_LAYERING_MODE_BYPASS) {
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- ++rc.layer_input_frames[sl * enc_cfg.ts_number_layers +
- layer_id.temporal_layer_id];
- }
- }
- for (tl = layer_id.temporal_layer_id;
- tl < enc_cfg.ts_number_layers; ++tl) {
- vpx_video_writer_write_frame(outfile[tl],
- cx_pkt->data.frame.buf,
- cx_pkt->data.frame.sz,
- cx_pkt->data.frame.pts);
- }
-
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- for (tl = layer_id.temporal_layer_id;
- tl < enc_cfg.ts_number_layers; ++tl) {
- const int layer = sl * enc_cfg.ts_number_layers + tl;
- ++rc.layer_tot_enc_frames[layer];
- rc.layer_encoding_bitrate[layer] += 8.0 * sizes[sl];
- // Keep count of rate control stats per layer, for non-key
- // frames.
- if (tl == layer_id.temporal_layer_id &&
- !(cx_pkt->data.frame.flags & VPX_FRAME_IS_KEY)) {
- rc.layer_avg_frame_size[layer] += 8.0 * sizes[sl];
- rc.layer_avg_rate_mismatch[layer] +=
- fabs(8.0 * sizes[sl] - rc.layer_pfb[layer]) /
- rc.layer_pfb[layer];
- ++rc.layer_enc_frames[layer];
- }
- }
- }
-
- // Update for short-time encoding bitrate states, for moving
- // window of size rc->window, shifted by rc->window / 2.
- // Ignore first window segment, due to key frame.
- if (frame_cnt > rc.window_size) {
- tl = layer_id.temporal_layer_id;
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- sum_bitrate += 0.001 * 8.0 * sizes[sl] * framerate;
- }
- if (frame_cnt % rc.window_size == 0) {
- rc.window_count += 1;
- rc.avg_st_encoding_bitrate += sum_bitrate / rc.window_size;
- rc.variance_st_encoding_bitrate +=
- (sum_bitrate / rc.window_size) *
- (sum_bitrate / rc.window_size);
- sum_bitrate = 0.0;
- }
- }
-
- // Second shifted window.
- if (frame_cnt > rc.window_size + rc.window_size / 2) {
- tl = layer_id.temporal_layer_id;
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- sum_bitrate2 += 0.001 * 8.0 * sizes[sl] * framerate;
- }
-
- if (frame_cnt > 2 * rc.window_size &&
- frame_cnt % rc.window_size == 0) {
- rc.window_count += 1;
- rc.avg_st_encoding_bitrate += sum_bitrate2 / rc.window_size;
- rc.variance_st_encoding_bitrate +=
- (sum_bitrate2 / rc.window_size) *
- (sum_bitrate2 / rc.window_size);
- sum_bitrate2 = 0.0;
- }
- }
- }
-#endif
- }
-
- printf("SVC frame: %d, kf: %d, size: %d, pts: %d\n", frames_received,
- !!(cx_pkt->data.frame.flags & VPX_FRAME_IS_KEY),
- (int)cx_pkt->data.frame.sz, (int)cx_pkt->data.frame.pts);
- ++frames_received;
- break;
- }
- case VPX_CODEC_STATS_PKT: {
- stats_write(&app_input.rc_stats,
- cx_pkt->data.twopass_stats.buf,
- cx_pkt->data.twopass_stats.sz);
- break;
- }
- default: {
- break;
- }
- }
- }
-
- if (!end_of_stream) {
- ++frame_cnt;
- pts += frame_duration;
- }
- }
-
- // Compensate for the extra frame count for the bypass mode.
- if (svc_ctx.temporal_layering_mode == VP9E_TEMPORAL_LAYERING_MODE_BYPASS) {
- for (sl = 0; sl < enc_cfg.ss_number_layers; ++sl) {
- const int layer = sl * enc_cfg.ts_number_layers +
- layer_id.temporal_layer_id;
- --rc.layer_input_frames[layer];
- }
- }
-
- printf("Processed %d frames\n", frame_cnt);
- fclose(infile);
-#if OUTPUT_RC_STATS
- if (svc_ctx.output_rc_stat) {
- printout_rate_control_summary(&rc, &enc_cfg, frame_cnt);
- printf("\n");
- }
-#endif
- if (vpx_codec_destroy(&codec)) die_codec(&codec, "Failed to destroy codec");
- if (app_input.passes == 2)
- stats_close(&app_input.rc_stats, 1);
- if (writer) {
- vpx_video_writer_close(writer);
- }
-#if OUTPUT_RC_STATS
- if (svc_ctx.output_rc_stat) {
- for (tl = 0; tl < enc_cfg.ts_number_layers; ++tl) {
- vpx_video_writer_close(outfile[tl]);
- }
- }
-#endif
- printf("Frame cnt and encoding time/FPS stats for encoding: %d %f %f \n",
- frame_cnt,
- 1000 * (float)cx_time / (double)(frame_cnt * 1000000),
- 1000000 * (double)frame_cnt / (double)cx_time);
- vpx_img_free(&raw);
- // display average size, psnr
- printf("%s", vpx_svc_dump_statistics(&svc_ctx));
- vpx_svc_release(&svc_ctx);
- return EXIT_SUCCESS;
-}
