画像圧縮プログラムのSSIM比較
こちらのMサイズを圧縮します
元のサイズ: 1968495バイト
Photoshop 23.1.0 Release
| 圧縮設定 | 圧縮後サイズ | 圧縮率[%] | SSIM |
|---|---|---|---|
| Great | 1857313 | 94 | 0.949639 |
| Excellent | 808601 | 41 | 0.932082 |
| Very Good | 479340 | 24 | 0.90924 |
| Good | 271353 | 14 | 0.875281 |
| Fair | 198991 | 10 | 0.849169 |
| Poor | 181823 | 9.2 | 0.841396 |
| Very Poor | 127910 | 6.4 | 0.794932 |
TinyJpeg
| 圧縮設定 | 圧縮後サイズ | 圧縮率[%] | SSIM |
|---|---|---|---|
| 標準 | 405648 | 20 | 0.902383 |
WEBSITE PLANET
https://www.websiteplanet.com/ja/webtools/imagecompressor/
| 圧縮設定 | 圧縮後サイズ | 圧縮率[%] | SSIM |
|---|---|---|---|
| 標準 | 290346 | 15 | 0.921095 |
mozjpeg
brew install mozjpeg djpeg original.jpg | cjpeg -quality 80 > mozjpeg_80.jpg
| 圧縮設定 | 圧縮後サイズ | 圧縮率[%] | SSIM |
|---|---|---|---|
| 100 | 1590213 | 81 | 0.981104 |
| 95 | 822661 | 42 | 0.963509 |
| 90 | 575467 | 29 | 0.949991 |
| 85 | 459889 | 23 | 0.939699 |
| 80 | 393243 | 20 | 0.931684 |
| 75 | 345098 | 18 | 0.92489 |
| 70 | 314537 | 16 | 0.919387 |
| 65 | 287868 | 15 | 0.913465 |
| 60 | 266308 | 14 | 0.908341 |
| 55 | 248811 | 13 | 0.903638 |
| 50 | 234643 | 12 | 0.899497 |
| 45 | 221200 | 11 | 0.894545 |
| 40 | 205655 | 10 | 0.88877 |
| 35 | 191658 | 9.7 | 0.882185 |
| 30 | 174679 | 8.9 | 0.873319 |
| 25 | 157005 | 8.0 | 0.861803 |
| 20 | 137270 | 7.0 | 0.845599 |
| 15 | 115902 | 5.9 | 0.821904 |
| 10 | 90365 | 4.6 | 0.781339 |
guetzli
brew install guetzli guetzli --quality 100 original.jpg guetzli_100.jpg
| 圧縮設定 | 圧縮後サイズ | 圧縮率[%] | SSIM |
|---|---|---|---|
| 100 | 1692329 | 86 | 0.994849 |
| 98 | 1112058 | 56 | 0.982535 |
| 96 | 827689 | 42 | 0.975307 |
| 94 | 672226 | 34 | 0.968616 |
| 92 | 576593 | 29 | 0.961595 |
| 90 | 514508 | 26 | 0.956176 |
| 88 | 459795 | 23 | 0.951042 |
| 86 | 422522 | 21 | 0.945941 |
| 84 | 394728 | 20 | 0.941691 |
グラフに起こすと
見た感じではTinyJpegはPhotoshopのVery GoodとGoodに近しい性能でmozjpegはそれよりも良い成績で、WEBSITE PLANETとguetzliが最も良い成績となりました。 ただしguetzliは圧縮にだいぶ時間がかかるので、どこでも使えるというものでもなさそうです。
Macbook Pro '14で動画エンコード
環境
- MacBook Pro '14
- CPU: Apple M1Pro 10コア
- GPU: 16コア
- OS: macOS 12.1
素材
- sony_xavcs_30p.MP4
- 1920 × 1080 29.97fps
- 50Mbps
- Sony RX100M4 X-AVCS
- 2865F
- 1分36秒
- canon_uhd.MP4
- gh5_422_uhd.MP4
- 3840x2160 29.97fps
- 150Mbps
- LUMIX GH5
- 525F
- 17秒
FFmpeg編
FFmpegはバージョン4.4.1を使用します。
| Input | Codec | FPS |
|---|---|---|
| sony_xavcs_30p.MP4 | libx264 | 88 |
| ↑ | h264_videotoolbox | 61 |
| ↑ | libx265 | 9.01 |
| ↑ | hevc_videotoolbox | 57 |
| ↑ | vp9 | 6.7 |
| canon_uhd.MP4 | libx264 | 30 |
| ↑ | h264_videotoolbox | 51 |
| ↑ | libx265 | 2.71 |
| ↑ | hevc_videotoolbox | 15 |
| ↑ | vp9 | 5.6 |
| gh5_422_uhd.MP4 | libx264 | 14 |
| ↑ | h264_videotoolbox | 16 |
| ↑ | libx265 | 3.26 |
| ↑ | hevc_videotoolbox | 15 |
| ↑ | vp9 | 3.5 |
ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx264 -b:v 5000k fhd2fhd_x264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v h264_videotoolbox -b:v 5000k fhd2fhd_toolbox_h264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx265 -b:v 5000k fhd2fhd_x265_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v hevc_videotoolbox -b:v 5000k fhd2fhd_toolbox_h265_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v vp9 -b:v 5000k fhd2fhd_vp9_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v libx264 -b:v 5000k uhd2fhd_x264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v h264_videotoolbox -b:v 5000k uhd2fhd_toolbox_h264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v libx265 -b:v 5000k uhd2fhd_h265_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v hevc_videotoolbox -b:v 5000k uhd2fhd_toolbox_h265_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v vp9 -b:v 5000k uhd2fhd_vp9_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v libx264 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_x264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v h264_videotoolbox -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_toolbox_h264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v libx265 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v hevc_videotoolbox -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_toolbox_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v vp9 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_vp9_1.mp4
Adobe Media Encoder編
| Input | Codec | Min:Sec | FPS |
|---|---|---|---|
| sony_xavcs_30p.MP4 | H.264 | 0:15 | 191 |
| ↑ | HEVC(H.265) | 0:17 | 168 |
| canon_uhd.MP4 | H.264 | 0:27 | 52 |
| ↑ | HEVC(H.265) | 0:30 | 47 |
| gh5_422_uhd.MP4 | H.264 | 0:10 | 52 |
| ↑ | HEVC(H.265) | 0:11 | 47 |
nginxを使ってローカルにミニマム開発環境を整える
まず表示するファイルを用意します。
mkdir html echo "OK" > html/index.html
その次に nginx-min.conf というファイルを以下の内容でカレントディレクトリに作成します。
events { worker_connections 10; } http { server { listen 8080; listen [::]:8080; server_name localhost; location / { root html; index index.html index.htm; } } }
以下のコマンドで構文チェックします。ファイル名をしている時は絶対パスでないと、nginx標準のディレクトリ配下を見に行って今します。
nginx -t -p `pwd` -c nginx-min.conf
syntax is ok と test is successful と表示されていたらOKです。
nginx -p `pwd` -c nginx-min.conf -g "pid nginx.pid;"
で起動して、停止する時は
kill $(cat nginx.pid)
です。
次にローカル用の証明書を作ります。 https://github.com/FiloSottile/mkcert を使用します。
brew install mkcert
このコマンドでCAをインストールします
mkcert --install
こんなふうに言われます
Created a new local CA 💥 Sudo password: The local CA is now installed in the system trust store! ⚡️ Warning: "certutil" is not available, so the CA can't be automatically installed in Firefox! ⚠️ Install "certutil" with "brew install nss" and re-run "mkcert -install" 👈
証明書を作成します。
mkcert localhost
ローカルに秘密鍵と、証明書のペアが生成されます。
localhost-key.pem localhost.pem
nginxの設定を以下の様に書き換えて再度、起動します。
events { worker_connections 10; } http { server { listen 8080; listen [::]:8080; server_name localhost; root html; location / { index index.html index.htm; } } server { listen 8081 ssl; server_name localhost; ssl_certificate localhost.pem; ssl_certificate_key localhost-key.pem; location / { root html; index index.html index.htm; } } }
起動すると https://localhost:8081/ のURLで証明書のエラーが出ることもなく接続できるのが確認できると思います。
PHPにプロキシすることを考えます。
以下の様な app.php を用意します。
<?php phpinfo();
これを以下のコマンドで起動しておき
php -S localhost:8082 app.php
nginxの設定をプロキシするよう変えます。
http {
...
server {
...
location / {
proxy_pass http://localhost:8082/;
}
https://localhost:8081/ を開くとPHPの詳細が確認できると思います。
[追記]
nginxの最新バージョンではlogの出力先が固定されているので -e オプションで出力先を変更する必要がある
nginx -e /dev/null -p `pwd` -c nginx-min.conf -g "pid nginx.pid;"
macOS CatalinaでTensorflow 1.15.3をビルド
ビルド環境
conda create -n tf_38_2 python=3.8 conda activate tf_38_2
pipモジュールのインストール
pip install -U --user pip six 'numpy<1.19.0' wheel setuptools mock 'future>=0.17.1' 'gast==0.2.2' typing_extensions pip install -U --user keras_applications --no-deps pip install -U --user keras_preprocessing --no-deps
git clone https://github.com/tensorflow/tensorflow tensorflow_38_2 cd tensorflow_38_2 git checkout v1.15.3
Bazelのインストール
export BAZEL_VERSION=0.26.1 curl -fLO "https://github.com/bazelbuild/bazel/releases/download/${BAZEL_VERSION}/bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh" chmod +x "bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh" ./bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh --user bazel --version
この２つのパッチを参考にファイルを修正します。
Fix GCC 10.1 compile error. by cbalint13 · Pull Request #40654 · tensorflow/tensorflow · GitHub
py-tensorflow1: fix Python 3.8 build. · macports/macports-ports@f63da02 · GitHub
configureの内容
% ./configure WARNING: --batch mode is deprecated. Please instead explicitly shut down your Bazel server using the command "bazel shutdown". You have bazel 0.26.1 installed. Please specify the location of python. [Default is /Users/tak/opt/anaconda3/envs/tf_38_2/bin/python]: Found possible Python library paths: /Users/tak/opt/anaconda3/envs/tf_38_2/lib/python3.8/site-packages Please input the desired Python library path to use. Default is [/Users/tak/opt/anaconda3/envs/tf_38_2/lib/python3.8/site-packages] Do you wish to build TensorFlow with XLA JIT support? [Y/n]: y XLA JIT support will be enabled for TensorFlow. Do you wish to build TensorFlow with OpenCL SYCL support? [y/N]: n No OpenCL SYCL support will be enabled for TensorFlow. Do you wish to build TensorFlow with ROCm support? [y/N]: n No ROCm support will be enabled for TensorFlow. Do you wish to build TensorFlow with CUDA support? [y/N]: n No CUDA support will be enabled for TensorFlow. Do you wish to download a fresh release of clang? (Experimental) [y/N]: n Clang will not be downloaded. Do you wish to build TensorFlow with MPI support? [y/N]: n No MPI support will be enabled for TensorFlow. Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native -Wno-sign-compare]: Would you like to interactively configure ./WORKSPACE for Android builds? [y/N]: n Not configuring the WORKSPACE for Android builds. Do you wish to build TensorFlow with iOS support? [y/N]: n No iOS support will be enabled for TensorFlow. Preconfigured Bazel build configs. You can use any of the below by adding "--config=<>" to your build command. See .bazelrc for more details. --config=mkl # Build with MKL support. --config=monolithic # Config for mostly static monolithic build. --config=gdr # Build with GDR support. --config=verbs # Build with libverbs support. --config=ngraph # Build with Intel nGraph support. --config=numa # Build with NUMA support. --config=dynamic_kernels # (Experimental) Build kernels into separate shared objects. --config=v2 # Build TensorFlow 2.x instead of 1.x. Preconfigured Bazel build configs to DISABLE default on features: --config=noaws # Disable AWS S3 filesystem support. --config=nogcp # Disable GCP support. --config=nohdfs # Disable HDFS support. --config=noignite # Disable Apache Ignite support. --config=nokafka # Disable Apache Kafka support. --config=nonccl # Disable NVIDIA NCCL support. Configuration finished
ビルド実行
bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package
ビルド完了したら以下のパッチを参考にソース修正
./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
pip install /tmp/tensorflow_pkg/tensorflow-1.15.3-cp38-cp38-macosx_10_15_x86_64.whl
これでビルドは完了です。
macOS CatalinaでTensorflow 2.3.0をビルド
ビルド環境
macOS用のAnacondaをインストール
conda create -n tf_38 python=3.8 conda activate tf_38
以下がセットアップ内容
Collecting package metadata (current_repodata.json): done Solving environment: done ==> WARNING: A newer version of conda exists. <== current version: 4.8.3 latest version: 4.8.4 Please update conda by running $ conda update -n base -c defaults conda ## Package Plan ## environment location: /Users/tak/opt/anaconda3/envs/tf_38 added / updated specs: - python=3.8 The following packages will be downloaded: package | build ---------------------------|----------------- pip-20.2.2 | py38_0 1.7 MB python-3.8.5 | h26836e1_0 20.7 MB setuptools-49.6.0 | py38_0 747 KB ------------------------------------------------------------ Total: 23.1 MB The following NEW packages will be INSTALLED: ca-certificates pkgs/main/osx-64::ca-certificates-2020.7.22-0 certifi pkgs/main/osx-64::certifi-2020.6.20-py38_0 libcxx pkgs/main/osx-64::libcxx-10.0.0-1 libedit pkgs/main/osx-64::libedit-3.1.20191231-h1de35cc_1 libffi pkgs/main/osx-64::libffi-3.3-hb1e8313_2 ncurses pkgs/main/osx-64::ncurses-6.2-h0a44026_1 openssl pkgs/main/osx-64::openssl-1.1.1g-h1de35cc_0 pip pkgs/main/osx-64::pip-20.2.2-py38_0 python pkgs/main/osx-64::python-3.8.5-h26836e1_0 readline pkgs/main/osx-64::readline-8.0-h1de35cc_0 setuptools pkgs/main/osx-64::setuptools-49.6.0-py38_0 sqlite pkgs/main/osx-64::sqlite-3.33.0-hffcf06c_0 tk pkgs/main/osx-64::tk-8.6.10-hb0a8c7a_0 wheel pkgs/main/noarch::wheel-0.35.1-py_0 xz pkgs/main/osx-64::xz-5.2.5-h1de35cc_0 zlib pkgs/main/osx-64::zlib-1.2.11-h1de35cc_3
pipモジュールのインストール
pip install -U --user pip six 'numpy<1.19.0' wheel setuptools mock 'future>=0.17.1' 'gast==0.3.3' typing_extensions pip install -U --user keras_applications --no-deps pip install -U --user keras_preprocessing --no-deps
git clone https://github.com/tensorflow/tensorflow tensorflow_38 cd tensorflow_38 git checkout v2.3.0
Bazelのインストール
export BAZEL_VERSION=3.1.0 curl -fLO "https://github.com/bazelbuild/bazel/releases/download/${BAZEL_VERSION}/bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh" chmod +x "bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh" ./bazel-${BAZEL_VERSION}-installer-darwin-x86_64.sh --user export PATH="$PATH:$HOME/bin" bazel --version
Xcodeの設定
sudo xcode-select -s /Applications/Xcode.app/Contents/Developer sudo xcodebuild -license bazel clean --expunge
このパッチを参考にファイルを修正します。
Fix GCC 10.1 compile error. by cbalint13 · Pull Request #40654 · tensorflow/tensorflow · GitHub
configureの内容
% ./configure You have bazel 3.1.0 installed. Please specify the location of python. [Default is /Users/tak/opt/anaconda3/envs/tf_38/bin/python3]: Found possible Python library paths: /Users/tak/opt/anaconda3/envs/tf_38/lib/python3.8/site-packages Please input the desired Python library path to use. Default is [/Users/tak/opt/anaconda3/envs/tf_38/lib/python3.8/site-packages] Do you wish to build TensorFlow with OpenCL SYCL support? [y/N]: n No OpenCL SYCL support will be enabled for TensorFlow. Do you wish to build TensorFlow with ROCm support? [y/N]: n No ROCm support will be enabled for TensorFlow. Do you wish to build TensorFlow with CUDA support? [y/N]: n No CUDA support will be enabled for TensorFlow. Do you wish to download a fresh release of clang? (Experimental) [y/N]: n Clang will not be downloaded. Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native -Wno-sign-compare]: Would you like to interactively configure ./WORKSPACE for Android builds? [y/N]: n Not configuring the WORKSPACE for Android builds. Do you wish to build TensorFlow with iOS support? [y/N]: n No iOS support will be enabled for TensorFlow. Preconfigured Bazel build configs. You can use any of the below by adding "--config=<>" to your build command. See .bazelrc for more details. --config=mkl # Build with MKL support. --config=monolithic # Config for mostly static monolithic build. --config=ngraph # Build with Intel nGraph support. --config=numa # Build with NUMA support. --config=dynamic_kernels # (Experimental) Build kernels into separate shared objects. --config=v2 # Build TensorFlow 2.x instead of 1.x. Preconfigured Bazel build configs to DISABLE default on features: --config=noaws # Disable AWS S3 filesystem support. --config=nogcp # Disable GCP support. --config=nohdfs # Disable HDFS support. --config=nonccl # Disable NVIDIA NCCL support. Configuration finished
ビルド開始
bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package
インストール作業
./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
pip install /tmp/tensorflow_pkg/tensorflow-2.3.0-cp38-cp38-macosx_10_15_x86_64.whl
これでインストール完了です。
iMac proでエンコード
環境
素材
いつもの
- sony_xavcs_30p.MP4
- 1920 × 1080 29.97fps
- 50Mbps
- Sony RX100M4 X-AVCS
- 11175F
- 6分13秒
- canon_uhd.MP4
- gh5_422_uhd.MP4
- 3840x2160 29.97fps
- 150Mbps
- LUMIX GH5
- 1770F
- 59秒
Adobe Media Encoder編
| Input | Codec | Min:Sec | FPS |
|---|---|---|---|
| sony_xavcs_30p.MP4 | H.264 | 1:03 | 177 |
| ↑ | HEVC(H.265) | 1:12 | 155 |
| canon_uhd.MP4 | H.264 | 0:14 | 101 |
| ↑ | HEVC(H.265) | 0:13 | 109 |
| gh5_422_uhd.MP4 | H.264 | 0:20 | 88 |
| ↑ | HEVC(H.265) | 0:20 | 88 |
sony_xavcs_30p.MP4 のエンコードの場合、ほぼCPUは使われず、GPUだけとなりました。
GPUのエンコードが律速となっています。
canon_uhd.MP4の場合、先ほどよりはCPUの負荷が上がっていますが、それでも利用率は低いです。
あとH.265はGPUも少し利用率低めです。
gh5_422_uhd.MP4 では大分CPU使用率が高くなりました。主にデコード処理に使われているのだと思います。GPUも使用率高めです。なのでこの動画形式だとCPUもGPUも同等に使い切れる変換形式だと思います。
FFmpeg編
FFmpegはバージョン4.2.3を使用します。
| Input | Codec | FPS |
|---|---|---|
| sony_xavcs_30p.MP4 | libx264 | 81 |
| ↑ | h264_videotoolbox | 189 |
| ↑ | libx265 | 27 |
| ↑ | hevc_videotoolbox | 157 |
| canon_uhd.MP4 | libx264 | 26 |
| ↑ | h264_videotoolbox | 52 |
| ↑ | libx265 | 10.36 |
| ↑ | hevc_videotoolbox | 45 |
| gh5_422_uhd.MP4 | libx264 | 28 |
| ↑ | h264_videotoolbox | 52 |
| ↑ | libx265 | 12 |
| ↑ | hevc_videotoolbox | 45 |
HEVC Toolboxが使えるのでh.265は少し早くエンコード出来る様です。
ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx264 -b:v 5000k fhd2fhd_x264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v h264_videotoolbox -b:v 5000k fhd2fhd_toolbox_h264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx265 -b:v 5000k fhd2fhd_x265_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v hevc_videotoolbox -b:v 5000k fhd2fhd_toolbox_h265_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v libx264 -b:v 5000k uhd2fhd_x264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v h264_videotoolbox -b:v 5000k uhd2fhd_toolbox_h264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v libx265 -b:v 5000k uhd2fhd_h265_1.mp4 ffmpeg -y -i canon_uhd.MP4 -c:v hevc_videotoolbox -b:v 5000k uhd2fhd_toolbox_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v libx264 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_x264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v h264_videotoolbox -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_toolbox_h264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v libx265 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -c:v hevc_videotoolbox -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_toolbox_h265_1.mp4
Windows機で動画エンコード
環境
- CPU: Core i7 9750H
- GPU: GeForce GTX 1660Ti
- OS: Windows 10 Home 1903
素材
前回と同じように1080p 29.97Fに変換します。
- sony_xavcs_30p.MP4
- 1920 × 1080 29.97fps
- 50Mbps
- Sony RX100M4 X-AVCS
- 11175F
- 6分13秒
- canon_uhd.MP4
- gh5_422_uhd.MP4
- 3840x2160 29.97fps
- 150Mbps
- LUMIX GH5
- 1770F
- 59秒
Adobe Media Encoder編
| Input | Codec | Min:Sec | FPS |
|---|---|---|---|
| sony_xavcs_30p.MP4 | H.264 | 0:26 | 429 |
| ↑ | HEVC(H.265) | 0:41 | 272 |
| canon_uhd.MP4 | H.264 | 0:12 | 118 |
| ↑ | HEVC(H.265) | 0:13 | 109 |
| gh5_422_uhd.MP4 | H.264 | 0:27 | 65 |
| ↑ | HEVC(H.265) | 0:28 | 63 |
sony_xavcs_30p.MP4 と canon_uhd.MP4の場合はIntel内臓GPUでデコードを行い、GeForceでエンコードを行なっているため、かなり早いです。
gh5_422_uhd.MP4の場合は、デコードをCPUで行なっているためこちらが性能は下がり気味です。
FFmpeg編
FFmpegはChocolateyでインストールしたバージョン4.2.2を使用します。
| Input | Codec | FPS |
|---|---|---|
| sony_xavcs_30p.MP4 | libx264 | 44 |
| ↑ | h264_nvenc | 266 |
| ↑ | libx265 | 16 |
| ↑ | hevc_nvenc | 166 |
| canon_uhd.MP4 | libx264 | 40 |
| ↑ | h264_nvenc | 92 |
| ↑ | libx265 | 17.5 |
| ↑ | hevc_nvenc | 91 |
| gh5_422_uhd.MP4 | libx264 | 32 |
| ↑ | h264_nvenc | 80 |
| ↑ | libx265 | 21.01 |
| ↑ | hevc_nvenc | 78 |
ffmpegの場合は全てCPUでデコードを行います。Adobe Media Encoder程速度は出ていない印象です。ただしnvencの性能は高いです。
ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx264 -b:v 5000k fhd2fhd_x264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v h264_nvenc -b:v 5000k fhd2fhd_nvenc_h264_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v libx265 -b:v 5000k fhd2fhd_h265_1.mp4 ffmpeg -y -i sony_xavcs_30p.MP4 -c:v hevc_nvenc -b:v 5000k fhd2fhd_nvenc_hevc_1.mp4 ffmpeg -y -i canon_uhd.MP4 -vf "scale=1920:-1" -c:v libx264 -b:v 5000k uhd2fhd_x264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -vf "scale=1920:-1" -c:v h264_nvenc -b:v 5000k uhd2fhd_nvenc_h264_1.mp4 ffmpeg -y -i canon_uhd.MP4 -vf "scale=1920:-1" -c:v libx265 -b:v 5000k uhd2fhd_h265_1.mp4 ffmpeg -y -i canon_uhd.MP4 -vf "scale=1920:-1" -c:v hevc_nvenc -b:v 5000k uhd2fhd_nvenc_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -vf "scale=1920:-1" -c:v libx264 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_x264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -vf "scale=1920:-1" -c:v h264_nvenc -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_nvenc_h264_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -vf "scale=1920:-1" -c:v libx265 -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_h265_1.mp4 ffmpeg -y -i gh5_422_uhd.MP4 -vf "scale=1920:-1" -c:v hevc_nvenc -b:v 5000k -pix_fmt yuv420p uhd422_to_fhd_nvenc_h265_1.mp4
