User guide overview

更新时间:
复制 MD 格式

Alibaba Cloud ApsaraVideo Media Processing (MPS), formerly known as ApsaraVideo for Media Transcoding (MTS), provides a cost-effective, elastic, and highly scalable method for transcoding multimedia data into formats suitable for playback on all platforms. MPS also performs multimodal analysis on media content to enable automated review, smart production, and copyright protection.

Usage notes

To quickly understand the integration process, familiarize yourself with the core concepts of this product, such as jobs, tasks, MPS queues, workflows, and templates. For more information, see Terms.

Process overview

MPS supports various media processing tasks, such as transcoding, snapshots, media information analysis, and video AI. The core integration process is the same for all task types and consists of two invocation patterns: synchronous and asynchronous.

Most tasks require file downloads and take time to analyze and process. For this reason, only tasks that can be completed quickly, such as single snapshots and media information analysis, support both synchronous and asynchronous invocations. Other tasks, such as transcoding and video AI, support only asynchronous invocations.

Sync task flow

image

Applicable tasks

A job for taking a single screenshot and analyzing media information.

Process description

You can submit sync tasks by using an SDK or an API. After submission, the API immediately returns the task result. You can also poll the query API to retrieve the task result, but receiving results through Message Service (MNS) messages is not supported.

  1. Upload the video to Object Storage Service (OSS).

  2. Submit a sync task. In the API call, specify the sync invocation pattern as described in the documentation.

  3. Receive the task result returned by the submission API.

    Note

    If the input file is too large, the task might time out and fail. You can add a retry mechanism as needed.

  4. (Optional) Poll the query API.

  5. (Optional) Receive the task result returned by the query API.

Asynchronous task flow

image

Applicable tasks

All MPS task types, including media transcoding (standard transcoding, Narrowband HD 1.0, Narrowband HD 2.0, and audio/video enhancement), snapshots, media information analysis, and video AI (smart production, smart tagging, automated review, and media fingerprint).

Process description

You can submit asynchronous tasks from the MPS console, by calling job submission SDKs or APIs, or by triggering a configured workflow. After submission, the task enters an MPS queue and is scheduled based on priority and submission time, within the concurrency limit. The task may not be complete when the API call returns a response. You must periodically poll the query API or configure an MNS queue for the MPS queue or workflow to obtain the task results.

  1. Upload the video to OSS.

  2. Submit an asynchronous task.

  3. Poll the query API.

  4. Receive the task result returned by the query API.

  5. (Optional) An MNS message notification is sent.

  6. (Optional) Receive and parse the MNS message.

Global configurations

Configuration

Description

Reference

Activate services

Activate MPS, OSS, and other related products, and grant access permissions to your cloud resources.

Activate services

Accounts and authorization

ApsaraVideo Media Processing supports Resource Access Management (RAM) and Security Token Service (STS). You can grant permissions in the following ways:

  • Grant permissions to a RAM user by attaching a system authorization policy.

  • Grant permissions to a RAM user by creating a custom policy.

  • Grant temporary access permissions to a role using STS.

Overview of identity and permission management

Pipeline Management

MPS uses pipelines to manage your production queues.

Create a pipeline

(Optional) Message notifications

ApsaraVideo Media Processing supports receiving message notifications through Simple Message Queue (formerly MNS).

Set message notifications

(Optional) Monitoring service

MPS supports using Cloud Monitor to monitor metrics such as request success rate and the number of submitted transcoding jobs.

Configuring monitoring and alerts

Feature usage

Feature

Description

Reference

Media asset management

MPS lets you upload media files to OSS and manage the resulting media assets, including basic information, source file information, playback information, and AI data. You can retrieve, update, search, download, and delete assets.

Media asset management

Get media information

MPS lets you retrieve key information about audio and video files stored in OSS, including resolution, bitrate, frame rate, encoding format, and file format.

Media transcoding

MPS supports video transcoding, audio transcoding, container format conversion, Narrowband HD, audio/video enhancement, and accelerated transcoding. You can use transcoding templates to simplify development.

Simple video editing

MPS supports simple post-production tasks for videos, such as audio extraction, video extraction, concatenation, clipping, adding opening and ending credits, and audio mixing.

Add watermarks

MPS supports adding visible image or text watermarks and invisible digital watermarks to videos.

Add captions

MPS supports adding captions to videos. This is useful for multilingual source videos, silent playback, noisy environments, or making content accessible to hearing-impaired viewers. Hard subtitles and packaged subtitles are supported.

None

Video snapshots

MPS lets you take snapshots of a video at specified times and sizes to create video thumbnails, sprites, and progress bar thumbnails for players.

Video snapshots

Video AI

MPS supports smart production, automated review, smart tagging, and media fingerprint.

Workflows

MPS media workflows process uploaded files by performing tasks such as transcoding, review, packaging, and analysis to generate media files.

You can create workflows from the console or by using the API to build cloud-based audio and video processing flows.