Chiebot-Mirror
/
FFmpeg
mirror of https://github.com/FFmpeg/FFmpeg.git
8
0
Fork 0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
116915 Commits
37 Branches
400 Tags
474 MiB
Tag: Branch: Tree: 8a314e3f11
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '8a314e3f11'
${ noResults }
FFmpeg/libavcodec/nvenc_av1.c

213 lines
23 KiB
Raw Normal View History Unescape

avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/internal.h"
#include "avcodec.h"
#include "codec_internal.h"
#include "nvenc.h"
#define OFFSET(x) offsetof(NvencContext, x)
#define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
static const AVOption options[] = {
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
{ "preset", "Set the encoding preset", OFFSET(preset), AV_OPT_TYPE_INT, { .i64 = PRESET_P4 }, PRESET_DEFAULT, PRESET_P7, VE, .unit = "preset" },
{ "default", "", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_DEFAULT }, 0, 0, VE, .unit = "preset" },
{ "slow", "hq 2 passes", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_SLOW }, 0, 0, VE, .unit = "preset" },
{ "medium", "hq 1 pass", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_MEDIUM }, 0, 0, VE, .unit = "preset" },
{ "fast", "hp 1 pass", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_FAST }, 0, 0, VE, .unit = "preset" },
{ "p1", "fastest (lowest quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P1 }, 0, 0, VE, .unit = "preset" },
{ "p2", "faster (lower quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P2 }, 0, 0, VE, .unit = "preset" },
{ "p3", "fast (low quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P3 }, 0, 0, VE, .unit = "preset" },
{ "p4", "medium (default)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P4 }, 0, 0, VE, .unit = "preset" },
{ "p5", "slow (good quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P5 }, 0, 0, VE, .unit = "preset" },
{ "p6", "slower (better quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P6 }, 0, 0, VE, .unit = "preset" },
{ "p7", "slowest (best quality)", 0, AV_OPT_TYPE_CONST, { .i64 = PRESET_P7 }, 0, 0, VE, .unit = "preset" },
{ "tune", "Set the encoding tuning info", OFFSET(tuning_info), AV_OPT_TYPE_INT, { .i64 = NV_ENC_TUNING_INFO_HIGH_QUALITY }, NV_ENC_TUNING_INFO_HIGH_QUALITY, NV_ENC_TUNING_INFO_LOSSLESS, VE, .unit = "tune" },
{ "hq", "High quality", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TUNING_INFO_HIGH_QUALITY }, 0, 0, VE, .unit = "tune" },
{ "ll", "Low latency", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TUNING_INFO_LOW_LATENCY }, 0, 0, VE, .unit = "tune" },
{ "ull", "Ultra low latency", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TUNING_INFO_ULTRA_LOW_LATENCY }, 0, 0, VE, .unit = "tune" },
{ "lossless", "Lossless", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TUNING_INFO_LOSSLESS }, 0, 0, VE, .unit = "tune" },
{ "level", "Set the encoding level restriction", OFFSET(level), AV_OPT_TYPE_INT, { .i64 = NV_ENC_LEVEL_AV1_AUTOSELECT }, NV_ENC_LEVEL_AV1_2, NV_ENC_LEVEL_AV1_AUTOSELECT, VE, .unit = "level" },
{ "auto", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_AUTOSELECT }, 0, 0, VE, .unit = "level" },
{ "2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_2 }, 0, 0, VE, .unit = "level" },
{ "2.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_2 }, 0, 0, VE, .unit = "level" },
{ "2.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_21 }, 0, 0, VE, .unit = "level" },
{ "2.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_22 }, 0, 0, VE, .unit = "level" },
{ "2.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_23 }, 0, 0, VE, .unit = "level" },
{ "3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_3 }, 0, 0, VE, .unit = "level" },
{ "3.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_3 }, 0, 0, VE, .unit = "level" },
{ "3.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_31 }, 0, 0, VE, .unit = "level" },
{ "3.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_32 }, 0, 0, VE, .unit = "level" },
{ "3.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_33 }, 0, 0, VE, .unit = "level" },
{ "4", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_4 }, 0, 0, VE, .unit = "level" },
{ "4.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_4 }, 0, 0, VE, .unit = "level" },
{ "4.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_41 }, 0, 0, VE, .unit = "level" },
{ "4.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_42 }, 0, 0, VE, .unit = "level" },
{ "4.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_43 }, 0, 0, VE, .unit = "level" },
{ "5", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_5 }, 0, 0, VE, .unit = "level" },
{ "5.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_5 }, 0, 0, VE, .unit = "level" },
{ "5.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_51 }, 0, 0, VE, .unit = "level" },
{ "5.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_52 }, 0, 0, VE, .unit = "level" },
{ "5.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_53 }, 0, 0, VE, .unit = "level" },
{ "6", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_6 }, 0, 0, VE, .unit = "level" },
{ "6.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_6 }, 0, 0, VE, .unit = "level" },
{ "6.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_61 }, 0, 0, VE, .unit = "level" },
{ "6.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_62 }, 0, 0, VE, .unit = "level" },
{ "6.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_63 }, 0, 0, VE, .unit = "level" },
{ "7", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_7 }, 0, 0, VE, .unit = "level" },
{ "7.0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_7 }, 0, 0, VE, .unit = "level" },
{ "7.1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_71 }, 0, 0, VE, .unit = "level" },
{ "7.2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_72 }, 0, 0, VE, .unit = "level" },
{ "7.3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LEVEL_AV1_73 }, 0, 0, VE, .unit = "level" },
{ "tier", "Set the encoding tier", OFFSET(tier), AV_OPT_TYPE_INT, { .i64 = NV_ENC_TIER_AV1_0 }, NV_ENC_TIER_AV1_0, NV_ENC_TIER_AV1_1, VE, .unit = "tier"},
{ "0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TIER_AV1_0 }, 0, 0, VE, .unit = "tier" },
{ "1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TIER_AV1_1 }, 0, 0, VE, .unit = "tier" },
{ "rc", "Override the preset rate-control", OFFSET(rc), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, INT_MAX, VE, .unit = "rc" },
{ "constqp", "Constant QP mode", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_PARAMS_RC_CONSTQP }, 0, 0, VE, .unit = "rc" },
{ "vbr", "Variable bitrate mode", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_PARAMS_RC_VBR }, 0, 0, VE, .unit = "rc" },
{ "cbr", "Constant bitrate mode", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_PARAMS_RC_CBR }, 0, 0, VE, .unit = "rc" },
{ "multipass", "Set the multipass encoding", OFFSET(multipass), AV_OPT_TYPE_INT, { .i64 = NV_ENC_MULTI_PASS_DISABLED }, NV_ENC_MULTI_PASS_DISABLED, NV_ENC_TWO_PASS_FULL_RESOLUTION, VE, .unit = "multipass" },
{ "disabled", "Single Pass", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_MULTI_PASS_DISABLED }, 0, 0, VE, .unit = "multipass" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "qres", "Two Pass encoding is enabled where first Pass is quarter resolution",
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TWO_PASS_QUARTER_RESOLUTION }, 0, 0, VE, .unit = "multipass" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "fullres", "Two Pass encoding is enabled where first Pass is full resolution",
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_TWO_PASS_FULL_RESOLUTION }, 0, 0, VE, .unit = "multipass" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "highbitdepth", "Enable 10 bit encode for 8 bit input",OFFSET(highbitdepth),AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "tile-rows", "Number of tile rows to encode with", OFFSET(tile_rows), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, NV_MAX_TILE_ROWS_AV1, VE },
{ "tile-columns", "Number of tile columns to encode with", OFFSET(tile_cols), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, NV_MAX_TILE_COLS_AV1, VE },
{ "surfaces", "Number of concurrent surfaces", OFFSET(nb_surfaces), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, MAX_REGISTERED_FRAMES, VE },
{ "gpu", "Selects which NVENC capable GPU to use. First GPU is 0, second is 1, and so on.",
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
OFFSET(device), AV_OPT_TYPE_INT, { .i64 = ANY_DEVICE }, -2, INT_MAX, VE, .unit = "gpu" },
{ "any", "Pick the first device available", 0, AV_OPT_TYPE_CONST, { .i64 = ANY_DEVICE }, 0, 0, VE, .unit = "gpu" },
{ "list", "List the available devices", 0, AV_OPT_TYPE_CONST, { .i64 = LIST_DEVICES }, 0, 0, VE, .unit = "gpu" },
avcodec/nvenc: add option to control subsampling of packed rgb input
1 year ago
{ "rgb_mode", "Configure how nvenc handles packed RGB input.",
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
OFFSET(rgb_mode), AV_OPT_TYPE_INT, { .i64 = NVENC_RGB_MODE_420 }, 0, INT_MAX, VE, .unit = "rgb_mode" },
{ "yuv420", "Convert to yuv420", 0, AV_OPT_TYPE_CONST, { .i64 = NVENC_RGB_MODE_420 }, 0, 0, VE, .unit = "rgb_mode" },
{ "yuv444", "Convert to yuv444", 0, AV_OPT_TYPE_CONST, { .i64 = NVENC_RGB_MODE_444 }, 0, 0, VE, .unit = "rgb_mode" },
{ "disabled", "Disables support, throws an error.", 0, AV_OPT_TYPE_CONST, { .i64 = NVENC_RGB_MODE_DISABLED }, 0, 0, VE, .unit = "rgb_mode" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "delay", "Delay frame output by the given amount of frames",
OFFSET(async_depth), AV_OPT_TYPE_INT, { .i64 = INT_MAX }, 0, INT_MAX, VE },
{ "rc-lookahead", "Number of frames to look ahead for rate-control",
OFFSET(rc_lookahead), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, VE },
avcodec/nvenc_av1: Correct CQ range for AV1 The Constant Quality (CQ) range for the AV1 codec is actually 0 to 63, contrary to what is stated in the header and documentation. Signed-off-by: Diego Felix de Souza <ddesouza@nvidia.com> Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
7 months ago
{ "cq", "Set target quality level (0 to 63, 0 means automatic) for constant quality mode in VBR rate control",
OFFSET(quality), AV_OPT_TYPE_FLOAT, { .dbl = 0.}, 0., 63., VE },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "init_qpP", "Initial QP value for P frame", OFFSET(init_qp_p), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
{ "init_qpB", "Initial QP value for B frame", OFFSET(init_qp_b), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
{ "init_qpI", "Initial QP value for I frame", OFFSET(init_qp_i), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
{ "qp", "Constant quantization parameter rate control method",
OFFSET(cqp), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
{ "qp_cb_offset", "Quantization parameter offset for cb channel",
OFFSET(qp_cb_offset), AV_OPT_TYPE_INT, { .i64 = 0 }, -12, 12, VE },
{ "qp_cr_offset", "Quantization parameter offset for cr channel",
OFFSET(qp_cr_offset), AV_OPT_TYPE_INT, { .i64 = 0 }, -12, 12, VE },
{ "no-scenecut", "When lookahead is enabled, set this to 1 to disable adaptive I-frame insertion at scene cuts",
OFFSET(no_scenecut), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "forced-idr", "If forcing keyframes, force them as IDR frames.",
OFFSET(forced_idr), AV_OPT_TYPE_BOOL, { .i64 = 0 }, -1, 1, VE },
{ "b_adapt", "When lookahead is enabled, set this to 0 to disable adaptive B-frame decision",
OFFSET(b_adapt), AV_OPT_TYPE_BOOL, { .i64 = 1 }, 0, 1, VE },
{ "spatial-aq", "set to 1 to enable Spatial AQ", OFFSET(aq), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "temporal-aq", "set to 1 to enable Temporal AQ", OFFSET(temporal_aq), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "zerolatency", "Set 1 to indicate zero latency operation (no reordering delay)",
OFFSET(zerolatency), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "nonref_p", "Set this to 1 to enable automatic insertion of non-reference P-frames",
OFFSET(nonref_p), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "strict_gop", "Set 1 to minimize GOP-to-GOP rate fluctuations",
OFFSET(strict_gop), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "aq-strength", "When Spatial AQ is enabled, this field is used to specify AQ strength. AQ strength scale is from 1 (low) - 15 (aggressive)",
OFFSET(aq_strength), AV_OPT_TYPE_INT, { .i64 = 8 }, 1, 15, VE },
{ "weighted_pred","Enable weighted prediction", OFFSET(weighted_pred),AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
all: use designated initializers for AVOption.unit Makes it robust against adding fields before it, which will be useful in following commits. Majority of the patch generated by the following Coccinelle script: @@ typedef AVOption; identifier arr_name; initializer list il; initializer list[8] il1; expression tail; @@ AVOption arr_name[] = { il, { il1, - tail + .unit = tail }, ... }; with some manual changes, as the script: * has trouble with options defined inside macros * sometimes does not handle options under an #else branch * sometimes swallows whitespace
11 months ago
{ "b_ref_mode", "Use B frames as references", OFFSET(b_ref_mode), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, NV_ENC_BFRAME_REF_MODE_MIDDLE, VE, .unit = "b_ref_mode" },
{ "disabled", "B frames will not be used for reference", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_BFRAME_REF_MODE_DISABLED }, 0, 0, VE, .unit = "b_ref_mode" },
{ "each", "Each B frame will be used for reference", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_BFRAME_REF_MODE_EACH }, 0, 0, VE, .unit = "b_ref_mode" },
{ "middle", "Only (number of B frames)/2 will be used for reference", 0,AV_OPT_TYPE_CONST, { .i64 = NV_ENC_BFRAME_REF_MODE_MIDDLE }, 0, 0, VE, .unit = "b_ref_mode" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "dpb_size", "Specifies the DPB size used for encoding (0 means automatic)",
OFFSET(dpb_size), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, INT_MAX, VE },
{ "ldkfs", "Low delay key frame scale; Specifies the Scene Change frame size increase allowed in case of single frame VBV and CBR",
OFFSET(ldkfs), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, UCHAR_MAX, VE },
{ "intra-refresh","Use Periodic Intra Refresh instead of IDR frames",
OFFSET(intra_refresh),AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "timing-info", "Include timing info in sequence/frame headers",
OFFSET(timing_info), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE },
{ "extra_sei", "Pass on extra SEI data (e.g. a53 cc) to be included in the bitstream",
OFFSET(extra_sei), AV_OPT_TYPE_BOOL, { .i64 = 1 }, 0, 1, VE },
{ "a53cc", "Use A53 Closed Captions (if available)", OFFSET(a53_cc), AV_OPT_TYPE_BOOL, { .i64 = 1 }, 0, 1, VE },
{ "s12m_tc", "Use timecode (if available)", OFFSET(s12m_tc), AV_OPT_TYPE_BOOL, { .i64 = 1 }, 0, 1, VE },
avcodec/nvenc: add support for lookahead_level
9 months ago
#ifdef NVENC_HAVE_LOOKAHEAD_LEVEL
{ "lookahead_level", "Specifies the lookahead level. Higher level may improve quality at the expense of performance.",
OFFSET(lookahead_level), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, NV_ENC_LOOKAHEAD_LEVEL_AUTOSELECT, VE, .unit = "lookahead_level" },
{ "auto", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LOOKAHEAD_LEVEL_AUTOSELECT }, 0, 0, VE, .unit = "lookahead_level" },
{ "0", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LOOKAHEAD_LEVEL_0 }, 0, 0, VE, .unit = "lookahead_level" },
{ "1", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LOOKAHEAD_LEVEL_1 }, 0, 0, VE, .unit = "lookahead_level" },
{ "2", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LOOKAHEAD_LEVEL_2 }, 0, 0, VE, .unit = "lookahead_level" },
{ "3", "", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_LOOKAHEAD_LEVEL_3 }, 0, 0, VE, .unit = "lookahead_level" },
avcodec/nvenc: Multi NVENC Split Frame Encoding in HEVC and AV1 When Split frame encoding is enabled, each input frame is partitioned into horizontal strips which are encoded independently and simultaneously by separate NVENCs, usually resulting in increased encoding speed compared to single NVENC encoding. Signed-off-by: Diego Felix de Souza <ddesouza@nvidia.com> Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
9 months ago
#endif
#ifdef NVENC_HAVE_SPLIT_FRAME_ENCODING
{ "split_encode_mode", "Specifies the split encoding mode", OFFSET(split_encode_mode), AV_OPT_TYPE_INT, { .i64 = NV_ENC_SPLIT_AUTO_MODE }, 0, NV_ENC_SPLIT_DISABLE_MODE, VE, .unit = "split_encode_mode" },
{ "disabled", "Disabled for all configurations", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_SPLIT_DISABLE_MODE }, 0, 0, VE, .unit = "split_encode_mode" },
{ "auto", "Enabled or disabled depending on the preset and tuning info", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_SPLIT_AUTO_MODE }, 0, 0, VE, .unit = "split_encode_mode" },
{ "forced", "Enabled with number of horizontal strips selected by the driver", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_SPLIT_AUTO_FORCED_MODE }, 0, 0, VE, .unit = "split_encode_mode" },
{ "2", "Enabled with number of horizontal strips forced to 2 when number of NVENCs > 1", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_SPLIT_TWO_FORCED_MODE }, 0, 0, VE, .unit = "split_encode_mode" },
{ "3", "Enabled with number of horizontal strips forced to 3 when number of NVENCs > 2", 0, AV_OPT_TYPE_CONST, { .i64 = NV_ENC_SPLIT_THREE_FORCED_MODE }, 0, 0, VE, .unit = "split_encode_mode" },
avcodec/nvenc: add support for lookahead_level
9 months ago
#endif
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ NULL }
};
static const FFCodecDefault defaults[] = {
{ "b", "2M" },
{ "qmin", "-1" },
{ "qmax", "-1" },
{ "qdiff", "-1" },
{ "qblur", "-1" },
{ "qcomp", "-1" },
avcodec/nvenc: let preset decide default gop size Default GOP size is now set by preset and tuning info. This GOP size is only overwriten if -g value is provided. Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
2 years ago
{ "g", "-1" },
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
{ "bf", "-1" },
{ "refs", "0" },
{ NULL },
};
static const AVClass av1_nvenc_class = {
.class_name = "av1_nvenc",
Revert "all: Don't set AVClass.item_name to its default value" Some callers assume that item_name is always set, so this may be considered an API break. This reverts commit 0c6203c97a99f69dbaa6e4011d48c331e1111f5e.
11 months ago
.item_name = av_default_item_name,
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
const FFCodec ff_av1_nvenc_encoder = {
.p.name = "av1_nvenc",
CODEC_LONG_NAME("NVIDIA NVENC av1 encoder"),
.p.type = AVMEDIA_TYPE_VIDEO,
.p.id = AV_CODEC_ID_AV1,
.init = ff_nvenc_encode_init,
FF_CODEC_RECEIVE_PACKET_CB(ff_nvenc_receive_packet),
.close = ff_nvenc_encode_close,
.flush = ff_nvenc_encode_flush,
.priv_data_size = sizeof(NvencContext),
.p.priv_class = &av1_nvenc_class,
.defaults = defaults,
.p.pix_fmts = ff_nvenc_pix_fmts,
.p.capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_HARDWARE |
avcodec/nvenc: handle frame durations and AV_CODEC_CAP_ENCODER_REORDERED_OPAQUE
2 years ago
AV_CODEC_CAP_ENCODER_FLUSH | AV_CODEC_CAP_DR1 |
AV_CODEC_CAP_ENCODER_REORDERED_OPAQUE,
avcodec/nvenc: add AV1 encoding support The encoder seems to be trading blows with hevc_nvenc. In terms of quality at low bitrate cbr settings, it seems to outperform it even. It produces fewer artifacts and the ones it does produce are less jarring to my perception. At higher bitrates I had a hard time finding differences between the two encoders in terms of subjective visual quality. Using the 'slow' preset, av1_nvenc outperformed hevc_nvenc in terms of encoding speed by 75% to 100% while performing above tests. Needless to say, it always massively outperformed h264_nvenc in terms of quality for a given bitrate, while also being slightly faster.
2 years ago
.caps_internal = FF_CODEC_CAP_NOT_INIT_THREADSAFE |
FF_CODEC_CAP_INIT_CLEANUP,
.p.wrapper_name = "nvenc",
.hw_configs = ff_nvenc_hw_configs,
};