From patchwork Mon Jun 29 03:58:03 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: James Almer X-Patchwork-Id: 20678 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id E75D444A2A5 for ; Mon, 29 Jun 2020 06:58:34 +0300 (EEST) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id B718568B752; Mon, 29 Jun 2020 06:58:34 +0300 (EEST) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mail-qk1-f175.google.com (mail-qk1-f175.google.com [209.85.222.175]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id 19DFF68B69F for ; Mon, 29 Jun 2020 06:58:28 +0300 (EEST) Received: by mail-qk1-f175.google.com with SMTP id q198so14158263qka.2 for ; Sun, 28 Jun 2020 20:58:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:subject:date:message-id:mime-version :content-transfer-encoding; bh=7MOSq24vlkdrVUGTZONqJ+ah1uhX70nOdfYecQo5uBU=; b=VcM2xSuSA637Oed8AstuNfXnVAOvaViLff32At6XDcVsvAi5ybC6KpS/Nk0JmGZ+AP 7tNVEE2FtQKLA2sgPPKsHtT3WdncuT6Mj23p/ZJGf4lYOhfPYtddqQacyIRG43SjnUt4 pFD9mogC/vId7Azd73+vg6KvXSSAhhHYG0/z7xWjbx58TxGBI6L+5zjejpXOivZLWzaP Vp+UKFfHlctbAJDth/aCLs5YB/y567JegjoXRjCOmyM+btO/mBeGiH7ziTg6G7nDXq3U Y3NE4pPMJQIq/SMHOVNnnBJKXOSdT09dV5K91NMIM2M8nD/Z5wc2/we/w6zp4Jd/t/II ZYyg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:subject:date:message-id:mime-version :content-transfer-encoding; bh=7MOSq24vlkdrVUGTZONqJ+ah1uhX70nOdfYecQo5uBU=; b=L4Dr3c4So/UqkS99F5fq7rtWGTiHO/cv+sEUxh2ShGc59c2cRcDIuzUlPUJFO+R2Z5 boUObpl7c80exaDom8NVz9f5r3B7gMsO1br6aefUs6ZWpgyu1ff0uO2iSIjqK1/vCm9Q 3Oczto4Ay60J339J14Tf28GCFKXyeo6SZMvNlKPph0xhwk5mlOue0IujpML06jQ23EW7 BoJhz7QW0BqF0ov6Rh+DX0yd3qcf2DEs7W7AqCYswKXdm1g6GqFLna3dYVGQbFDFvl8J pFoTWieSzXkOHoz9JUKIn+MBfgJXsb58OTMNQD+s6v9KOm5ybrtajQPypMc9QmYbJjAy jFmA== X-Gm-Message-State: AOAM530UjPYczUxBYAJKo6i/11tzEBd03CzhFp32QE63fSOH1cuRCf7+ jq1NJlFFXCqbzV25HMocKwQG7o78 X-Google-Smtp-Source: ABdhPJxdRpaTvbvaUWtq39XcdVEz+lUm0oJK4acs1QAASrpqt42rur/J9y7rhsqM3cS1t1LH0uoadQ== X-Received: by 2002:a05:620a:410e:: with SMTP id j14mr13409056qko.216.1593403105882; Sun, 28 Jun 2020 20:58:25 -0700 (PDT) Received: from localhost.localdomain ([181.23.95.153]) by smtp.gmail.com with ESMTPSA id c83sm15683912qkb.75.2020.06.28.20.58.24 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Sun, 28 Jun 2020 20:58:24 -0700 (PDT) From: James Almer To: ffmpeg-devel@ffmpeg.org Date: Mon, 29 Jun 2020 00:58:03 -0300 Message-Id: <20200629035803.35409-1-jamrial@gmail.com> X-Mailer: git-send-email 2.27.0 MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH] avcodec: Add an SVT-AV1 encoder wrapper X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" From: Daryl Seah Signed-off-by: Daryl Seah Signed-off-by: Jing SUN Signed-off-by: ZhiZhen Tang Signed-off-by: Zhong Li Signed-off-by: Xu Guangxin Signed-off-by: James Almer --- configure | 4 + doc/encoders.texi | 36 +++ doc/general.texi | 10 +- libavcodec/Makefile | 1 + libavcodec/allcodecs.c | 1 + libavcodec/libsvt_av1.c | 552 ++++++++++++++++++++++++++++++++++++++++ 6 files changed, 603 insertions(+), 1 deletion(-) create mode 100644 libavcodec/libsvt_av1.c diff --git a/configure b/configure index 7495f35faa..ac69e78581 100755 --- a/configure +++ b/configure @@ -267,6 +267,7 @@ External library support: --enable-libspeex enable Speex de/encoding via libspeex [no] --enable-libsrt enable Haivision SRT protocol via libsrt [no] --enable-libssh enable SFTP protocol via libssh [no] + --enable-libsvtav1 enable AV1 encoding via SVT [no] --enable-libtensorflow enable TensorFlow as a DNN module backend for DNN based filters like sr [no] --enable-libtesseract enable Tesseract, needed for ocr filter [no] @@ -1803,6 +1804,7 @@ EXTERNAL_LIBRARY_LIST=" libspeex libsrt libssh + libsvtav1 libtensorflow libtesseract libtheora @@ -3223,6 +3225,7 @@ libopencore_amrnb_decoder_deps="libopencore_amrnb" libopencore_amrnb_encoder_deps="libopencore_amrnb" libopencore_amrnb_encoder_select="audio_frame_queue" libopencore_amrwb_decoder_deps="libopencore_amrwb" +libsvt_av1_encoder_deps="libsvtav1" libopenh264_decoder_deps="libopenh264" libopenh264_decoder_select="h264_mp4toannexb_bsf" libopenh264_encoder_deps="libopenh264" @@ -6372,6 +6375,7 @@ enabled libsoxr && require libsoxr soxr.h soxr_create -lsoxr enabled libssh && require_pkg_config libssh libssh libssh/sftp.h sftp_init enabled libspeex && require_pkg_config libspeex speex speex/speex.h speex_decoder_init enabled libsrt && require_pkg_config libsrt "srt >= 1.3.0" srt/srt.h srt_socket +enabled libsvtav1 && require_pkg_config libsvtav1 "SvtAv1Enc >= 0.8.4" EbSvtAv1Enc.h svt_av1_enc_init_handle enabled libtensorflow && require libtensorflow tensorflow/c/c_api.h TF_Version -ltensorflow enabled libtesseract && require_pkg_config libtesseract tesseract tesseract/capi.h TessBaseAPICreate enabled libtheora && require libtheora theora/theoraenc.h th_info_init -ltheoraenc -ltheoradec -logg diff --git a/doc/encoders.texi b/doc/encoders.texi index 17a0f4c821..4c946cf92d 100644 --- a/doc/encoders.texi +++ b/doc/encoders.texi @@ -1601,6 +1601,42 @@ useful for screen content. Default is true. @end table +@section libsvt-av1 + +SVT-AV1 encoder wrapper. + +Requires the presence of the SVT-AV1 headers and library during configuration. +You need to explicitly configure the build with @code{--enable-libsvtav1}. + +@subsection Options + +@table @option +@item qmax +Sets the maximum quantizer to use when using bitrate mode. + +@item qmin +Sets the minimum quantizer to use when using bitrate mode. + +@item qp +Uses quantizer mode to encode at the given quantizer (0-63). + +@item hielevel +Enable scene change detection. + +@item la_depth +Set number of frames to look ahead (0-120). + +@item preset +Set encoding preset. + +@item tier + +@item tile-rows +Selects how many rows of tiles to encode with (0-6). + +@item tile-columns +Selects how many columns of tiles to encode with (0-4). + @section libkvazaar Kvazaar H.265/HEVC encoder. diff --git a/doc/general.texi b/doc/general.texi index ea34b963b5..3c0d547eb4 100644 --- a/doc/general.texi +++ b/doc/general.texi @@ -263,6 +263,14 @@ Go to @url{https://github.com/xiph/rav1e/} and follow the instructions to build the C library. To enable using rav1e in FFmpeg, pass @code{--enable-librav1e} to @file{./configure}. +@section SVT-AV1 + +FFmpeg can make use of the Scalable Video Technology for AV1 library for AV1 encoding. + +Go to @url{https://github.com/OpenVisualCloud/SVT-AV1/} and follow the instructions +for installing the library. Then pass @code{--enable-libsvtav1} to configure to +enable it. + @section TwoLAME FFmpeg can make use of the TwoLAME library for MP2 encoding. @@ -820,7 +828,7 @@ following image formats are supported: @item Autodesk RLE @tab @tab X @tab fourcc: AASC @item AV1 @tab E @tab E - @tab Supported through external libraries libaom, libdav1d and librav1e + @tab Supported through external libraries libaom, libdav1d, librav1e and libsvtav1 @item Avid 1:1 10-bit RGB Packer @tab X @tab X @tab fourcc: AVrp @item AVS (Audio Video Standard) video @tab @tab X diff --git a/libavcodec/Makefile b/libavcodec/Makefile index 5a6ea59715..67d5da04d6 100644 --- a/libavcodec/Makefile +++ b/libavcodec/Makefile @@ -1024,6 +1024,7 @@ OBJS-$(CONFIG_LIBRAV1E_ENCODER) += librav1e.o OBJS-$(CONFIG_LIBSHINE_ENCODER) += libshine.o OBJS-$(CONFIG_LIBSPEEX_DECODER) += libspeexdec.o OBJS-$(CONFIG_LIBSPEEX_ENCODER) += libspeexenc.o +OBJS-$(CONFIG_LIBSVT_AV1_ENCODER) += libsvt_av1.o OBJS-$(CONFIG_LIBTHEORA_ENCODER) += libtheoraenc.o OBJS-$(CONFIG_LIBTWOLAME_ENCODER) += libtwolame.o OBJS-$(CONFIG_LIBVO_AMRWBENC_ENCODER) += libvo-amrwbenc.o diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c index fa0c08d42e..82ac50b00b 100644 --- a/libavcodec/allcodecs.c +++ b/libavcodec/allcodecs.c @@ -725,6 +725,7 @@ extern AVCodec ff_librsvg_decoder; extern AVCodec ff_libshine_encoder; extern AVCodec ff_libspeex_encoder; extern AVCodec ff_libspeex_decoder; +extern AVCodec ff_libsvt_av1_encoder; extern AVCodec ff_libtheora_encoder; extern AVCodec ff_libtwolame_encoder; extern AVCodec ff_libvo_amrwbenc_encoder; diff --git a/libavcodec/libsvt_av1.c b/libavcodec/libsvt_av1.c new file mode 100644 index 0000000000..090b4c66de --- /dev/null +++ b/libavcodec/libsvt_av1.c @@ -0,0 +1,552 @@ +/* + * Scalable Video Technology for AV1 encoder library plugin + * + * Copyright (c) 2018 Intel Corporation + * + * 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 this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include +#include "EbSvtAv1ErrorCodes.h" +#include "EbSvtAv1Enc.h" + +#include "libavutil/common.h" +#include "libavutil/frame.h" +#include "libavutil/imgutils.h" +#include "libavutil/opt.h" +#include "libavutil/pixdesc.h" +#include "libavutil/avassert.h" + +#include "internal.h" +#include "encode.h" +#include "packet_internal.h" +#include "avcodec.h" +#include "profiles.h" + +typedef enum eos_status { + EOS_NOT_REACHED = 0, + EOS_SENT, + EOS_RECEIVED +}EOS_STATUS; + +typedef struct SvtContext { + AVClass *class; + + EbSvtAv1EncConfiguration enc_params; + EbComponentType *svt_handle; + + EbBufferHeaderType *in_buf; + int raw_size; + int max_tu_size; + + AVFrame *frame; + + AVBufferPool *pool; + + EOS_STATUS eos_flag; + + // User options. + int hierarchical_level; + int la_depth; + int enc_mode; + int rc_mode; + int scd; + int qp; + + int forced_idr; + + int tier; + + int tile_columns; + int tile_rows; +} SvtContext; + +static const struct { + EbErrorType eb_err; + int av_err; + const char *desc; +} svt_errors[] = { + { EB_ErrorNone, 0, "success" }, + { EB_ErrorInsufficientResources, AVERROR(ENOMEM), "insufficient resources" }, + { EB_ErrorUndefined, AVERROR(EINVAL), "undefined error" }, + { EB_ErrorInvalidComponent, AVERROR(EINVAL), "invalid component" }, + { EB_ErrorBadParameter, AVERROR(EINVAL), "bad parameter" }, + { EB_ErrorDestroyThreadFailed, AVERROR_EXTERNAL, "failed to destory thread" }, + { EB_ErrorSemaphoreUnresponsive, AVERROR_EXTERNAL, "semaphore unresponsive" }, + { EB_ErrorDestroySemaphoreFailed, AVERROR_EXTERNAL, "semaphore unresponsive" }, + { EB_ErrorCreateMutexFailed, AVERROR_EXTERNAL, "failed to creat mutex" }, + { EB_ErrorMutexUnresponsive, AVERROR_EXTERNAL, "mutex unresponsive" }, + { EB_ErrorDestroyMutexFailed, AVERROR_EXTERNAL, "failed to destory muxtex" }, + { EB_NoErrorEmptyQueue, AVERROR(EAGAIN), "empty queue" }, +}; + +static int svt_map_error(EbErrorType eb_err, const char **desc) +{ + int i; + + av_assert0(desc); + for (i = 0; i < FF_ARRAY_ELEMS(svt_errors); i++) { + if (svt_errors[i].eb_err == eb_err) { + *desc = svt_errors[i].desc; + return svt_errors[i].av_err; + } + } + *desc = "unknown error"; + return AVERROR_UNKNOWN; +} + +static int svt_print_error(void *log_ctx, EbErrorType err, + const char *error_string) +{ + const char *desc; + int ret = svt_map_error(err, &desc); + + av_log(log_ctx, AV_LOG_ERROR, "%s: %s (0x%x)\n", error_string, desc, err); + + return ret; +} + +static int alloc_buffer(EbSvtAv1EncConfiguration *config, SvtContext *svt_enc) +{ + const int pack_mode_10bit = + (config->encoder_bit_depth > 8) && (config->compressed_ten_bit_format == 0) ? 1 : 0; + const size_t luma_size_8bit = + config->source_width * config->source_height * (1 << pack_mode_10bit); + const size_t luma_size_10bit = + (config->encoder_bit_depth > 8 && pack_mode_10bit == 0) ? luma_size_8bit : 0; + + EbSvtIOFormat *in_data; + + svt_enc->raw_size = (luma_size_8bit + luma_size_10bit) * 3 / 2; + + // allocate buffer for in and out + svt_enc->in_buf = av_mallocz(sizeof(*svt_enc->in_buf)); + if (!svt_enc->in_buf) + return AVERROR(ENOMEM); + + svt_enc->in_buf->p_buffer = (unsigned char *)av_mallocz(sizeof(*in_data)); + if (!svt_enc->in_buf->p_buffer) + return AVERROR(ENOMEM); + + svt_enc->in_buf->size = sizeof(*svt_enc->in_buf); + svt_enc->in_buf->p_app_private = NULL; + + return 0; + +} + +static int config_enc_params(EbSvtAv1EncConfiguration *param, + AVCodecContext *avctx) +{ + SvtContext *svt_enc = avctx->priv_data; + const AVPixFmtDescriptor *desc; + + param->source_width = avctx->width; + param->source_height = avctx->height; + + desc = av_pix_fmt_desc_get(avctx->pix_fmt); + param->encoder_bit_depth = desc->comp[0].depth; + + if (desc->log2_chroma_w == 1 && desc->log2_chroma_h == 1) + param->encoder_color_format = EB_YUV420; + else if (desc->log2_chroma_w == 1 && desc->log2_chroma_h == 0) + param->encoder_color_format = EB_YUV422; + else if (!desc->log2_chroma_w && !desc->log2_chroma_h) + param->encoder_color_format = EB_YUV444; + else { + av_log(avctx, AV_LOG_ERROR , "Unsupported pixel format\n"); + return AVERROR(EINVAL); + } + + if (avctx->profile != FF_PROFILE_UNKNOWN) + param->profile = avctx->profile; + + if (avctx->level != FF_LEVEL_UNKNOWN) + param->level = avctx->level; + + if ((param->encoder_color_format == EB_YUV422 || param->encoder_bit_depth > 10) + && param->profile != FF_PROFILE_AV1_PROFESSIONAL ) { + av_log(avctx, AV_LOG_WARNING, "Foprcing Professional profile\n"); + param->profile = FF_PROFILE_AV1_PROFESSIONAL; + } else if (param->encoder_color_format == EB_YUV444 && param->profile != FF_PROFILE_AV1_HIGH) { + av_log(avctx, AV_LOG_WARNING, "Forcing High profile\n"); + param->profile = FF_PROFILE_AV1_HIGH; + } + + // Update param from options + param->hierarchical_levels = svt_enc->hierarchical_level; + param->enc_mode = svt_enc->enc_mode; + param->tier = svt_enc->tier; + param->rate_control_mode = svt_enc->rc_mode; + param->scene_change_detection = svt_enc->scd; + param->qp = svt_enc->qp; + + param->target_bit_rate = avctx->bit_rate; + + if (avctx->gop_size > 0) + param->intra_period_length = avctx->gop_size - 1; + + if (avctx->framerate.num > 0 && avctx->framerate.den > 0) { + param->frame_rate_numerator = avctx->framerate.num; + param->frame_rate_denominator = avctx->framerate.den; + } else { + param->frame_rate_numerator = avctx->time_base.den; + param->frame_rate_denominator = avctx->time_base.num * avctx->ticks_per_frame; + } + + if (param->rate_control_mode) { + param->max_qp_allowed = avctx->qmax; + param->min_qp_allowed = avctx->qmin; + } + + param->intra_refresh_type = svt_enc->forced_idr + 1; + + if (svt_enc->la_depth >= 0) + param->look_ahead_distance = svt_enc->la_depth; + + param->tile_columns = svt_enc->tile_columns; + param->tile_rows = svt_enc->tile_rows; + + return 0; +} + +static void read_in_data(EbSvtAv1EncConfiguration *param, const AVFrame *frame, + EbBufferHeaderType *header_ptr) +{ + EbSvtIOFormat *in_data = (EbSvtIOFormat *)header_ptr->p_buffer; + int bytes_shift = param->encoder_bit_depth > 8 ? 1 : 0; + int frame_size = av_image_get_buffer_size(frame->format, frame->width, frame->height, 1); + + in_data->luma = frame->data[0]; + in_data->cb = frame->data[1]; + in_data->cr = frame->data[2]; + + in_data->y_stride = AV_CEIL_RSHIFT(frame->linesize[0], bytes_shift); + in_data->cb_stride = AV_CEIL_RSHIFT(frame->linesize[1], bytes_shift); + in_data->cr_stride = AV_CEIL_RSHIFT(frame->linesize[2], bytes_shift); + + header_ptr->n_filled_len = frame_size; +} + +static av_cold int eb_enc_init(AVCodecContext *avctx) +{ + SvtContext *svt_enc = avctx->priv_data; + EbErrorType svt_ret; + int ret; + + svt_enc->eos_flag = EOS_NOT_REACHED; + + svt_ret = svt_av1_enc_init_handle(&svt_enc->svt_handle, svt_enc, &svt_enc->enc_params); + if (svt_ret != EB_ErrorNone) { + return svt_print_error(avctx, svt_ret, "Error initializing encoder handle"); + } + + ret = config_enc_params(&svt_enc->enc_params, avctx); + if (ret < 0) { + svt_av1_enc_deinit_handle(svt_enc->svt_handle); + svt_enc->svt_handle = NULL; + av_log(avctx, AV_LOG_ERROR, "Error configuring encoder parameters\n"); + return ret; + } + + svt_ret = svt_av1_enc_set_parameter(svt_enc->svt_handle, &svt_enc->enc_params); + if (svt_ret != EB_ErrorNone) { + svt_av1_enc_deinit_handle(svt_enc->svt_handle); + svt_enc->svt_handle = NULL; + return svt_print_error(avctx, svt_ret, "Error setting encoder parameters"); + } + + svt_ret = svt_av1_enc_init(svt_enc->svt_handle); + if (svt_ret != EB_ErrorNone) { + svt_av1_enc_deinit_handle(svt_enc->svt_handle); + svt_enc->svt_handle = NULL; + return svt_print_error(avctx, svt_ret, "Error initializing encoder"); + } + + if (avctx->flags & AV_CODEC_FLAG_GLOBAL_HEADER) { + EbBufferHeaderType *headerPtr = NULL; + + svt_ret = svt_av1_enc_stream_header(svt_enc->svt_handle, &headerPtr); + if (svt_ret != EB_ErrorNone) { + return svt_print_error(avctx, svt_ret, "Error building stream header"); + } + + avctx->extradata_size = headerPtr->n_filled_len; + avctx->extradata = av_mallocz(avctx->extradata_size + AV_INPUT_BUFFER_PADDING_SIZE); + if (!avctx->extradata) { + av_log(avctx, AV_LOG_ERROR, + "Cannot allocate AV1 header of size %d.\n", avctx->extradata_size); + return AVERROR(ENOMEM); + } + + memcpy(avctx->extradata, headerPtr->p_buffer, avctx->extradata_size); + + svt_ret = svt_av1_enc_stream_header_release(headerPtr); + if (svt_ret != EB_ErrorNone) { + return svt_print_error(avctx, svt_ret, "Error freeing stream header"); + } + } + + svt_enc->frame = av_frame_alloc(); + if (!svt_enc->frame) + return AVERROR(ENOMEM); + + return alloc_buffer(&svt_enc->enc_params, svt_enc); +} + +static int eb_send_frame(AVCodecContext *avctx, const AVFrame *frame) +{ + SvtContext *svt_enc = avctx->priv_data; + EbBufferHeaderType *headerPtr = svt_enc->in_buf; + + if (!frame) { + EbBufferHeaderType headerPtrLast; + + if (svt_enc->eos_flag == EOS_SENT) + return 0; + + headerPtrLast.n_alloc_len = 0; + headerPtrLast.n_filled_len = 0; + headerPtrLast.n_tick_count = 0; + headerPtrLast.p_app_private = NULL; + headerPtrLast.p_buffer = NULL; + headerPtrLast.flags = EB_BUFFERFLAG_EOS; + + svt_av1_enc_send_picture(svt_enc->svt_handle, &headerPtrLast); + svt_enc->eos_flag = EOS_SENT; + return 0; + } + + read_in_data(&svt_enc->enc_params, frame, headerPtr); + + headerPtr->flags = 0; + headerPtr->p_app_private = NULL; + headerPtr->pts = frame->pts; + + svt_av1_enc_send_picture(svt_enc->svt_handle, headerPtr); + + return 0; +} + +static AVBufferRef* get_output_ref(AVCodecContext *avctx, SvtContext *svt_enc, int filled_len) +{ + if (filled_len > svt_enc->max_tu_size) { + const int max_frames = 8; + int max_tu_size; + + if (filled_len > svt_enc->raw_size * max_frames) { + av_log(avctx, AV_LOG_ERROR, "TU size > %d raw frame size.\n", max_frames); + return NULL; + } + + max_tu_size = 1 << av_ceil_log2(filled_len); + av_buffer_pool_uninit(&svt_enc->pool); + svt_enc->pool = av_buffer_pool_init(max_tu_size + AV_INPUT_BUFFER_PADDING_SIZE, NULL); + if (!svt_enc->pool) + return NULL; + + svt_enc->max_tu_size = max_tu_size; + } + av_assert0(svt_enc->pool); + + return av_buffer_pool_get(svt_enc->pool); +} + +static int eb_receive_packet(AVCodecContext *avctx, AVPacket *pkt) +{ + SvtContext *svt_enc = avctx->priv_data; + EbBufferHeaderType *headerPtr; + AVFrame *frame = svt_enc->frame; + EbErrorType svt_ret; + AVBufferRef *ref; + int ret = 0, pict_type; + + if (svt_enc->eos_flag == EOS_RECEIVED) + return AVERROR_EOF; + + ret = ff_encode_get_frame(avctx, frame); + if (ret < 0 && ret != AVERROR_EOF) + return ret; + if (ret == AVERROR_EOF) + frame = NULL; + + eb_send_frame(avctx, frame); + av_frame_unref(svt_enc->frame); + + svt_ret = svt_av1_enc_get_packet(svt_enc->svt_handle, &headerPtr, svt_enc->eos_flag); + if (svt_ret == EB_NoErrorEmptyQueue) + return AVERROR(EAGAIN); + + ref = get_output_ref(avctx, svt_enc, headerPtr->n_filled_len); + if (!ref) { + av_log(avctx, AV_LOG_ERROR, "Failed to allocate output packet.\n"); + svt_av1_enc_release_out_buffer(&headerPtr); + return AVERROR(ENOMEM); + } + pkt->buf = ref; + pkt->data = ref->data; + + memcpy(pkt->data, headerPtr->p_buffer, headerPtr->n_filled_len); + memset(pkt->data + headerPtr->n_filled_len, 0, AV_INPUT_BUFFER_PADDING_SIZE); + pkt->size = headerPtr->n_filled_len; + pkt->pts = headerPtr->pts; + pkt->dts = headerPtr->dts; + if (headerPtr->pic_type == EB_AV1_KEY_PICTURE) { + pkt->flags |= AV_PKT_FLAG_KEY; + pict_type = AV_PICTURE_TYPE_I; + } else if (headerPtr->pic_type == EB_AV1_INTRA_ONLY_PICTURE) { + pict_type = AV_PICTURE_TYPE_I; + } else if (headerPtr->pic_type == EB_AV1_INVALID_PICTURE) { + pict_type = AV_PICTURE_TYPE_NONE; + } else + pict_type = AV_PICTURE_TYPE_P; + + if (headerPtr->pic_type == EB_AV1_NON_REF_PICTURE) + pkt->flags |= AV_PKT_FLAG_DISPOSABLE; + + if (headerPtr->flags & EB_BUFFERFLAG_EOS) + svt_enc->eos_flag = EOS_RECEIVED; + + ff_side_data_set_encoder_stats(pkt, headerPtr->qp * FF_QP2LAMBDA, NULL, 0, pict_type); + + svt_av1_enc_release_out_buffer(&headerPtr); + + return 0; +} + +static av_cold int eb_enc_close(AVCodecContext *avctx) +{ + SvtContext *svt_enc = avctx->priv_data; + + if (svt_enc->svt_handle) { + svt_av1_enc_deinit(svt_enc->svt_handle); + svt_av1_enc_deinit_handle(svt_enc->svt_handle); + } + if (svt_enc->in_buf) { + av_free(svt_enc->in_buf->p_buffer); + av_freep(&svt_enc->in_buf); + } + + av_buffer_pool_uninit(&svt_enc->pool); + av_frame_free(&svt_enc->frame); + + return 0; +} + +#define OFFSET(x) offsetof(SvtContext, x) +#define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM +static const AVOption options[] = { + { "hielevel", "Hierarchical prediction levels setting", OFFSET(hierarchical_level), + AV_OPT_TYPE_INT, { .i64 = 4 }, 3, 4, VE , "hielevel"}, + { "3level", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 3 }, INT_MIN, INT_MAX, VE, "hielevel" }, + { "4level", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 4 }, INT_MIN, INT_MAX, VE, "hielevel" }, + + { "la_depth", "Look ahead distance [0, 120]", OFFSET(la_depth), + AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 120, VE }, + + { "preset", "Encoding preset [0, 8]", + OFFSET(enc_mode), AV_OPT_TYPE_INT, { .i64 = MAX_ENC_PRESET }, 0, MAX_ENC_PRESET, VE }, + + { "tier", "Set tier (general_tier_flag)", OFFSET(tier), + AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, VE, "tier" }, + { "main", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, VE, "tier" }, + { "high", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, VE, "tier" }, + + FF_AV1_PROFILE_OPTS + +#define LEVEL(name, value) name, NULL, 0, AV_OPT_TYPE_CONST, \ + { .i64 = value }, 0, 0, VE, "avctx.level" + { LEVEL("2.0", 20) }, + { LEVEL("2.1", 21) }, + { LEVEL("2.2", 22) }, + { LEVEL("2.3", 23) }, + { LEVEL("3.0", 30) }, + { LEVEL("3.1", 31) }, + { LEVEL("3.2", 32) }, + { LEVEL("3.3", 33) }, + { LEVEL("4.0", 40) }, + { LEVEL("4.1", 41) }, + { LEVEL("4.2", 42) }, + { LEVEL("4.3", 43) }, + { LEVEL("5.0", 50) }, + { LEVEL("5.1", 51) }, + { LEVEL("5.2", 52) }, + { LEVEL("5.3", 53) }, + { LEVEL("6.0", 60) }, + { LEVEL("6.1", 61) }, + { LEVEL("6.2", 62) }, + { LEVEL("6.3", 63) }, + { LEVEL("7.0", 70) }, + { LEVEL("7.1", 71) }, + { LEVEL("7.2", 72) }, + { LEVEL("7.3", 73) }, +#undef LEVEL + + { "rc", "Bit rate control mode", OFFSET(rc_mode), + AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 3, VE , "rc"}, + { "cqp", "Const Quantization Parameter", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, INT_MIN, INT_MAX, VE, "rc" }, + { "vbr", "Variable Bit Rate, use a target bitrate for the entire stream", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, INT_MIN, INT_MAX, VE, "rc" }, + { "cvbr", "Constrained Variable Bit Rate, use a target bitrate for each GOP", 0, AV_OPT_TYPE_CONST,{ .i64 = 2 }, INT_MIN, INT_MAX, VE, "rc" }, + + { "qp", "QP value for intra frames", OFFSET(qp), + AV_OPT_TYPE_INT, { .i64 = 50 }, 0, 63, VE }, + + { "sc_detection", "Scene change detection", OFFSET(scd), + 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 = 1 }, 0, 1, VE }, + + { "tile-columns", "Log2 of number of tile columns to use", OFFSET(tile_columns), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 4, VE}, + { "tile-rows", "Log2 of number of tile rows to use", OFFSET(tile_rows), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 6, VE}, + + {NULL}, +}; + +static const AVClass class = { + .class_name = "libsvt_av1", + .item_name = av_default_item_name, + .option = options, + .version = LIBAVUTIL_VERSION_INT, +}; + +static const AVCodecDefault eb_enc_defaults[] = { + { "b", "7M" }, + { "g", "-1" }, + { "qmin", "0" }, + { "qmax", "63" }, + { NULL }, +}; + +AVCodec ff_libsvt_av1_encoder = { + .name = "libsvt_av1", + .long_name = NULL_IF_CONFIG_SMALL("SVT-AV1(Scalable Video Technology for AV1) encoder"), + .priv_data_size = sizeof(SvtContext), + .type = AVMEDIA_TYPE_VIDEO, + .id = AV_CODEC_ID_AV1, + .init = eb_enc_init, + .receive_packet = eb_receive_packet, + .close = eb_enc_close, + .capabilities = AV_CODEC_CAP_DELAY | AV_CODEC_CAP_AUTO_THREADS, + .pix_fmts = (const enum AVPixelFormat[]){ AV_PIX_FMT_YUV420P, + AV_PIX_FMT_YUV420P10, + AV_PIX_FMT_NONE }, + .priv_class = &class, + .defaults = eb_enc_defaults, + .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, + .wrapper_name = "libsvt_av1", +};