@eyevinn/autovmaf

Index

Namespaces

Enumerations

QualityAnalysisModel

Classes

Type aliases

Variables

logger

Functions

Type aliases

AWSPipelineConfiguration

AWSPipelineConfiguration: { ecsCluster: string; ecsContainerName: string; ecsSecurityGroup: string; ecsSubnet: string; ecsTaskDefinition: string; inputBucket: string; mediaConvertEndpoint: string; mediaConvertRole: string; mediaConvertSettings: any; outputBucket: string }

Type declaration

ecsCluster: string
ecsContainerName: string
ecsSecurityGroup: string
ecsSubnet: string
ecsTaskDefinition: string
inputBucket: string
mediaConvertEndpoint: string
mediaConvertRole: string
mediaConvertSettings: any
outputBucket: string

BitrateRange

BitrateRange: { max?: number; min?: number }

Type declaration

Optional max?: number
Optional min?: number

BitrateResolutionVMAF

BitrateResolutionVMAF: { bitrate: number; resolution: Resolution; vmaf: number }

Type declaration

bitrate: number
resolution: Resolution
vmaf: number

JobDescription

JobDescription: { bitrates?: number[]; encodingProfile: string | Record<string, string>; method?: "bruteForce" | "walkTheHull"; models?: ("HD" | "PhoneHD" | "UHD")[]; name: string; pipeline: string | LocalPipelineConfiguration; pipelineVariables?: {}; reference: string; resolutions?: Resolution[]; skipExisting?: boolean; skipTranscode?: boolean }

Describes a ABR-analysis job and can be used to create jobs using the createJob()-function.

Type declaration

Optional bitrates?: number[]

A list of bitrates to test. By default a list of bitrates between 150 kbit/s to 9000 kbit/s.
encodingProfile: string | Record<string, string>

Path to a JSON-file that defines how the reference should be encoded. When using AWS, this is a MediaConvert configuration. For local pipelines, this is key-value pairs that will be passed as command line arguments to FFmpeg. For inline pipeline definition, this should be key-value pairs See an example for AWS at examples/encoding-profile.json.
Optional method?: "bruteForce" | "walkTheHull"

The method to use when analyzing the videos. Either bruteForce or walkTheHull. By default bruteForce. NOTE: walkTheHull is not implemented at the moment.
Optional models?: ("HD" | "PhoneHD" | "UHD")[]

A list of VMAF-models to use in evaluation. This can be HD, MobileHD and UHD. HD by default.
name: string

This will name the folder in which to put the files.
pipeline: string | LocalPipelineConfiguration

Path to a YAML-file that defines the pipeline, or an inline local pipeline configuration. See examples/pipeline.yml for an example AWS-pipeline.
Optional pipelineVariables?: {}

Values that will be substituted into the encoding options. Currently only supported for local pipeline
- [key: string]: string[]
reference: string

Path to the reference video to analyze. Normally a local path, but when using AWS, this can also be an S3-URI.
Optional resolutions?: Resolution[]

A list of resolutions to test. By default it will test all resolutions in the example ABR-ladder provided by Apple in the HLS Authoring Spec.
Optional skipExisting?: boolean

Skip transcode if outfile allready exists
Optional skipTranscode?: boolean

Skip transcode and run analysis only, files are assumed to be allready present

LocalPipelineConfiguration

LocalPipelineConfiguration: { easyVmafExtraArgs?: {}; easyVmafPath: string; ffmpegEncoder: "libx264" | "h264_videotoolbox"; ffmpegOptions: {}; ffmpegPath: string; pythonPath: string; singlePass?: boolean; skipDefaultOptions?: boolean }

Type declaration

Optional easyVmafExtraArgs?: {}
- [key: string]: string
easyVmafPath: string
ffmpegEncoder: "libx264" | "h264_videotoolbox"
ffmpegOptions: {}
- [key: string]: string
ffmpegPath: string
pythonPath: string
Optional singlePass?: boolean
Optional skipDefaultOptions?: boolean

Resolution

Resolution: { height: number; range?: BitrateRange; width: number }

Type declaration

height: number
Optional range?: BitrateRange
width: number

Variables

logger

logger: Logger

Functions

createJob

createJob(description: JobDescription, pipelineData?: any, encodingProfileData?: any, concurrency?: boolean): Promise<void>

Creates a analysis job according to a job description.

example

Example of creating a job from a job description.

const { createJob, qualityAnalysisModelToString } = require('@eyevinn/autovmaf');

const abrLadder = await = createJob({
  name: "hello",
  pipeline: "pipeline.yml",
  encodingProfile: "profile.json",
  reference: "reference.mp4",
  models: ["HD", "PhoneHD"],
  resolutions: [{ width: 1280, height: 720, range: undefined }],
  bitrates: [600000],
  method: "bruteForce"
});

example

Example of creating a job from a job description with range set.

const { createJob, qualityAnalysisModelToString } = require('@eyevinn/autovmaf');

const abrLadder = await = createJob({
  name: "hello",
  pipeline: "pipeline.yml",
  encodingProfile: "profile.json",
  reference: "reference.mp4",
  models: ["HD", "PhoneHD"],
  resolutions: [{ width: 1280, height: 720, range: { min: 400000, max: 600000} }],
  bitrates: [400000, 600000, 800000],
  method: "bruteForce"
});

example

Example of creating a job from a YAML-file.

const { createJob, JobDescription } = 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);

Parameters

description: JobDescription

An object that describes the job to create.
Optional pipelineData: any
Optional encodingProfileData: any
concurrency: boolean = true

Returns Promise<void>

getVmaf

getVmaf(filename: string, onProgress?: (index: number, filename: string, total: number) => void): Promise<{ filename: string; vmaf: number }[]>

- Defined in get-vmaf.ts:83
Returns the VMAF-values from a JSON-file or a directory of JSON-files. Can be used on both local paths as well as S3-URIs.
example
Example of retrieving a list of VMAF-scores from S3.
```
const vmafFiles = await getVmaf('s3://path/to/vmaf/');
vmafFiles.forEach(file => {
  console.log(file.filename + ': ' + file.vmaf);
});
```
Parameters
- filename: string
  
  The path to the file or directory. Can be a local path or a S3-URI.
- onProgress: (index: number, filename: string, total: number) => void = ...
  - - (index: number, filename: string, total: number): void
    - Parameters
      - index: number
      - filename: string
      - total: number
      Returns void
Returns Promise<{ filename: string; vmaf: number }[]>
A list of objects with filename and VMAF-scores.

loadPipeline

loadPipeline(pipelineFilename: string, encodingProfile?: string): Promise<Pipeline>

- Defined in load-pipeline.ts:41
Loads a pipeline from a YAML file and an encoding profile from a JSON file.

Parameters
- pipelineFilename: string
  
  The local path to the pipeline YAML.
- Optional encodingProfile: string
  
  The local path to the encoding profile JSON. If left undefined, the encoding profile will be set to an empty object.
Returns Promise<Pipeline>
A pipeline that can be used to transcode or analyze videos with.

loadPipelineFromObjects

loadPipelineFromObjects(pipelineData: any, encodingProfileData?: any): Promise<Pipeline>

- Defined in load-pipeline.ts:67
Loads a pipeline and an encoding profile from a JSON object.

Parameters
- pipelineData: any
  
  The object containing the pipeline data.
- Optional encodingProfileData: any
Returns Promise<Pipeline>
A pipeline that can be used to transcode or analyze videos with.

qualityAnalysisModelToString

qualityAnalysisModelToString(model: QualityAnalysisModel): string

- Defined in models/quality-analysis-model.ts:15
Converts from a QualityAnalysisModel-enum to a readable string.

Parameters
- model: QualityAnalysisModel
  
  The model to convert.
Returns string
A readable string of the model name.

stringToQualityAnalysisModel

stringToQualityAnalysisModel(model: "HD" | "PhoneHD" | "UHD"): QualityAnalysisModel

- Defined in models/quality-analysis-model.ts:32
Converts from a string to a QualityAnalysisModel-enum.

Parameters
- model: "HD" | "PhoneHD" | "UHD"
  
  The string to convert. Can be either "HD", "PhoneHD", or "UHD".
Returns QualityAnalysisModel
A QualityAnalysisModel-enum depending on the input string.

suggestLadder

suggestLadder(directoryWithVmafFiles: string, filterFunction?: (bitrate: number, resolution: Resolution, vmaf: number) => boolean, includeAllBitrates?: boolean, onProgress?: (index: number, filename: string, total: number) => void): Promise<LadderAndVmafPairs>

- Defined in suggest-ladder.ts:18
Suggests an optimal ABR-ladder from a directory of VMAF-files. Only supports loading from S3 at the moment.

Parameters
- directoryWithVmafFiles: string
  
  URI to the directory with VMAF-files.
- filterFunction: (bitrate: number, resolution: Resolution, vmaf: number) => boolean = ...
  
  Optional function to filter values from the analysis.
  - - (bitrate: number, resolution: Resolution, vmaf: number): boolean
    - Parameters
      - bitrate: number
      - resolution: Resolution
      - vmaf: number
      Returns boolean
- includeAllBitrates: boolean = false
  
  If true, return optimal resolutions for all bitrates.
- onProgress: (index: number, filename: string, total: number) => void = ...
  - - (index: number, filename: string, total: number): void
    - Parameters
      - index: number
      - filename: string
      - total: number
      Returns void
Returns Promise<LadderAndVmafPairs>
Returns the optimal ladder.

Index

Namespaces

Enumerations

Classes

Type aliases

Variables

Functions

Type aliases

AWSPipelineConfiguration

Type declaration

ecsCluster: string

ecsContainerName: string

ecsSecurityGroup: string

ecsSubnet: string

ecsTaskDefinition: string

inputBucket: string

mediaConvertEndpoint: string

mediaConvertRole: string

mediaConvertSettings: any

outputBucket: string

BitrateRange

Type declaration

Optional max?: number

Optional min?: number

BitrateResolutionVMAF

Type declaration

bitrate: number

resolution: Resolution

vmaf: number

JobDescription

Type declaration

Optional bitrates?: number[]

encodingProfile: string | Record<string, string>

Optional method?: "bruteForce" | "walkTheHull"

Optional models?: ("HD" | "PhoneHD" | "UHD")[]

name: string

pipeline: string | LocalPipelineConfiguration

Optional pipelineVariables?: {}

[key: string]: string[]

reference: string

Optional resolutions?: Resolution[]

Optional skipExisting?: boolean

Optional skipTranscode?: boolean

LocalPipelineConfiguration

Type declaration

Optional easyVmafExtraArgs?: {}

[key: string]: string

easyVmafPath: string

ffmpegEncoder: "libx264" | "h264_videotoolbox"

ffmpegOptions: {}

[key: string]: string

ffmpegPath: string

pythonPath: string

Optional singlePass?: boolean

Optional skipDefaultOptions?: boolean

Resolution

Type declaration

height: number

Optional range?: BitrateRange

width: number

Variables

logger

Functions

createJob

Parameters

description: JobDescription

Optional pipelineData: any

Optional encodingProfileData: any

concurrency: boolean = true

Returns Promise<void>

getVmaf

Parameters

filename: string

onProgress: (index: number, filename: string, total: number) => void = ...

Parameters

index: number

filename: string

total: number

Returns void

Returns Promise<{ filename: string; vmaf: number }[]>