@eyevinn/autovmaf
Advanced tools
By optimizing ABR-ladders for specific content, you will make sure to not have wasteful rungs and this has been shown to cut bandwidth usage in half.
Transcoding and VMAF analysis can either be run in AWS or locally. When running in aws, you will need a running ECS cluster with a task definition configured to run easyvmaf-s3.
npm install --save @eyevinn/autovmaf
A few environment variables can be set. These are:
LOAD_CREDENTIALS_FROM_ENV=true //Load AWS credentials from environment variables
AWS_REGION=eu-north-1
AWS_ACCESS_KEY_ID=ABCD...
AWS_SECRET_ACCESS_KEY=EFGH...
To generate VMAF measurements, you will need to define a job which can be created with the createJob()-function.
const { createJob } = require('@eyevinn/autovmaf');
const vmafScores = await createJob({
name: "MyVMAFmeasurements",
pipeline: "pipeline.yml",
encodingProfile: "profile.json",
reference: "reference.mp4",
models: ["HD", "PhoneHD"], // optional
resolutions: [{ // optional
width: 1280,
height: 720,
range: { // optional
min: 500000,
max: 600000
}
}],
bitrates: [ // optional
500000,
600000,
800000
],
method: "bruteForce" // optional
});
When creating a job, you can specify:
examples/pipeline.yml for an example AWS-pipeline.examples/aws/encoding-profile.json.
For local pipelines, this is key-value pairs that will be passed as command line arguments to FFmpeg. If pipeline
data is inlined in the job definition, encodingProfile can be omitted and key-value pairs can instead be set
in the ffmpegOptions property of the pipeline object.bruteForce or walkTheHull. By default bruteForce. NOTE: walkTheHull is not implemented at the moment.const { createJob } = require('@eyevinn/autovmaf');
const YAML = require('yaml');
const fs = require('fs');
const parseResolutions = resolutions => {
resolutions.map(resolutionStr => ({
width: parseInt(resolutionStr.split('x')[0]),
height: parseInt(resolutionStr.split('x')[1]),
}));
};
const jobFile = fs.readFileSync('job.yml', 'utf-8');
const jobData = YAML.parse(jobFile);
const job = {
...jobData,
resolutions: jobData['resolutions'] !== undefined ? parseResolutions(jobData['resolutions']) : undefined,
};
createJob(job);
An example of creating a job from a YAML-file can be seen in the examples-folder.
Using getVmaf(), you can read VMAF-scores from a JSON-file or a directory of JSON-files. This works on both local paths as well as S3-URIs with a "s3://"-prefix.
Example:
const vmafFiles = await getVmaf('s3://path/to/vmaf/');
vmafFiles.forEach(file => {
console.log(file.filename + ': ' + file.vmaf);
});
When running with the cli, all transcoding and vmaf analysis will be run locally.
Installing with npm -g will make the autovmaf command available in your path
npm install -g @eyevinn/autovmaf
EASYVMAF_PATH - needs to point to the file easyVmaf.py from your
easyVmaf installation.FFMPEG_PATH - only needs to be set if ffmpeg is not in your path.PYTHON_PATH - only needs to be set if python is not in yur path.Available command line options for the cli can be listed with the --help argument
autovmaf [source]
run transcode and vmaf analysis for videofile source
Commands:
autovmaf [source] run transcode and vmaf analysis for
videofile source [default]
autovmaf suggest-ladder <folder> Suggest bitrate ladder given vmaf results
autovmaf export-csv <folder> Export Vmaf results as csv
Positionals:
source SOURCEFILE [string]
Options:
--version Show version number [boolean]
--help Show help [boolean]
--resolutions List of resolutions, ie 1920x1080,1280x720... [string]
--bitrates List of bitrates, ie 800k,1000k,... [string]
--name Name for this autovmaf run [string]
--models List of VMAF Models to use [string]
--job File with job definition [string]
--saveAsCsv Save VMAF measurements as a .csv file in addition to a JSON
file [boolean] [default: false]
--skipTranscode Skip transcode and run vmaf on allready transcoded files
[boolean] [default: false]
--skipExisting Skip transcode for allready transcoded files
[boolean] [default: true]
--probeBitrate Read bitrate of transcoded file with ffprobe
[boolean] [default: false]
--ffmpeg-options List of options to pass to ffmpeg, on the form
key1=value1:key2=value2 [string]
Output files will be stored in a folder corresponding to the argument given to the --name option.
If resolutions and/or bitrates are not specified default values will be used, See above.
With the --job option, a path to a yaml or json file with a job definition can be passed to to the cli. The values
defined in the file can be overriden with other commandline options. For instance the reference video defined
in the job file can be overridden by passing a source file on the command line.
It is possible to iterate over other variables than bitrate and resolutions when running a local encode. For instance, to run transcode and vmaf analysis with x265 in CRF mode for a number of CRF values, a job definition like below can be used (also available in examples/local/local-job-crf.yaml)
models:
- HD
resolutions:
- width: 1920
height: 1080
bitrates:
- 0
pipeline:
ffmpegEncoder: libx265
singlePass: true
skipDefaultOptions: true
ffmpegOptions:
"-pix_fmt": "yuv420p"
"-preset": "veryslow"
"-x265-params": "crf=%CRF%:scenecut=0:keyint=50:min-keyint=50:open-gop=0"
easyVmafExtraArgs:
"-threads": 20
pipelineVariables:
CRF:
- 22
- 26
- 30
- 34
This will run transcode and vmaf analysis for CRF values 22,26,30, and 34. Variables are used in the ffmpeg options
by insterting %VARIABLENAME%. This string will then be subtituted with a value from the list of values from
pipelineVariables.VARIABLENAME. Note that when running CRF encode or other non-ABR mode, skipDefaultOptions must
be set to avoid injecting bitrate options to ffmpeg. Also note that the cli needs to be run with the --probe-bitrate
option to get the correct bitrate from the transcoded files.
autovmaf --resolutions 1920x1080,1280x720,960x540 --bitrates 500k,800k,1200k,1600k,2000k,3000k,4000k --name my-autovmaf-test1 my-source-video.mp4
With the above command, when the run is finished transcoded files will be available in the folder my-autovmaf-test1, and vmaf-data in the folder my-autovmaf-test1/HD.
npm test
Eyevinn Technology is an independent consultant firm specialized in video and streaming. Independent in a way that we are not commercially tied to any platform or technology vendor.
At Eyevinn, every software developer consultant has a dedicated budget reserved for open source development and contribution to the open source community. This give us room for innovation, team building and personal competence development. And also gives us as a company a way to contribute back to the open source community.
Want to know more about Eyevinn and how it is to work here. Contact us at work@eyevinn.se!
FAQs
Unknown package
The npm package @eyevinn/autovmaf receives a total of 10 weekly downloads. As such, @eyevinn/autovmaf popularity was classified as not popular.
We found that @eyevinn/autovmaf demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 0 open source maintainers collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
PEP 770 proposes adding SBOM support to Python packages to improve transparency and catch hidden non-Python dependencies that security tools often miss.
Security News
Socket CEO Feross Aboukhadijeh discusses open source security challenges, including zero-day attacks and supply chain risks, on the Cyber Security Council podcast.
Security News
Research
Socket researchers uncover how threat actors weaponize Out-of-Band Application Security Testing (OAST) techniques across the npm, PyPI, and RubyGems ecosystems to exfiltrate sensitive data.