ApsaraVideo VOD offers a comprehensive media moderation service for videos, audio, images, and text. This service helps you detect harmful content or elements, such as terrorism, political sensitivity, pornography, advertisements, abuse, and undesirable scenes in your media assets. It reduces manual review costs, improves content quality, and enhances platform governance and user experience. This topic describes the benefits, features, application scenarios, and usage of the media moderation service in ApsaraVideo VOD.
Benefits
The basic moderation capabilities are consistent with Alibaba Cloud Content Moderation. The service detects violations in images, videos, text, and audio. This includes risk detection for content related to terrorism, political sensitivity, pornography, advertisements, abuse, and undesirable scenes.
Media moderation in ApsaraVideo VOD aggregates individual reviews for all resources within a media asset. This includes video, audio, images (such as snapshots and thumbnails), and text (such as titles, descriptions, and tags). You only need to make a single call to process all media resources, which is more convenient and efficient.
ApsaraVideo VOD supports automated moderation through workflows and custom configuration templates. Moderation can be automatically triggered after a media file is uploaded, and the results are returned using real-time callbacks.
ApsaraVideo VOD moderation is deeply integrated with all stages of the video lifecycle, including media asset management, editing, production, distribution, and playback. This integration meets the requirements of complex scenarios. For example, you can choose to transcode a video only after it passes moderation, and discard or archive it otherwise. You can also automatically clip non-compliant segments before rapid distribution.
Features
Feature | Description | References |
Automated review | The automated review service is based on massive amounts of annotated data and deep learning algorithms. It accurately detects prohibited content in videos, thumbnails, and titles from multiple dimensions, including audio, text, and visuals. The service includes modules for pornography, terrorism, and political sensitivity. It can be used in scenarios such as short video platforms and media moderation to significantly improve video moderation efficiency. | |
Manual review | The manual review service lets you unpublish non-compliant videos in advance or promptly. This avoids or reduces the negative impact after publication. | |
Reviewing security IP addresses | The security IP service for moderation ensures that for videos in the Checking or Blocked state, only requests from specified security IPs can obtain the playback address. This is often used for internal review. |
Scenarios
For short videos and large volumes of user-generated content (UGC), you can use automated video moderation technology to quickly detect non-compliant content. Then, you can manually review videos with suspected violations. This process greatly improves moderation efficiency.
For video-on-demand platforms and the media industry, you can use automated video moderation technology to efficiently detect scenes that involve terrorism and political sensitivity in VOD videos. This helps you address the latest political sensitivities and strictly control non-compliant content.
Usage notes
The final moderation result of an audio or video file is determined by both the automated and manual moderation results. However, the manual moderation result always overrides the automated moderation result.
After media moderation, only audio and video files in the Normal state can be played. Audio and video files in the Flagged or Reviewing state can be played only in the ApsaraVideo VOD console or using a security IP for moderation. You cannot obtain playback addresses for these files using an API or SDK. However, any valid playback addresses that have already been exposed can still be used for playback.
How to use
Prerequisites
ApsaraVideo VOD is activated. For more information, see Activate ApsaraVideo VOD.
The audio or video file to be moderated must be uploaded to the ApsaraVideo VOD media asset library. For more information, see Upload media files.
Step 1: Configure the moderation process
When you upload an audio or video file to ApsaraVideo VOD, it is processed based on the configured moderation process. This process determines whether the uploaded file can be played. The two moderation processes are Review After Publish and Review Before Publish. Review After Publish (default) allows an audio or video file to be played immediately after processes such as transcoding and content delivery acceleration are complete. Review Before Publish prevents a file from being accessed after it is uploaded and processed. The file can be played only after it passes moderation.
Log on to the ApsaraVideo VOD console.
In the left-side navigation pane, choose Review Management > Settings.
On the moderation settings page, click Modify next to Review Process.
Select a moderation process.
Moderation process
Process description
Review After Publish (Default)
When an audio or video file is uploaded to ApsaraVideo VOD, its status is set to Normal by default after transcoding is successful. The file can be played directly.
Later, you can initiate a machine-assisted or manual moderation for confirmation in Video Moderation as needed.
During manual moderation, if you mark an audio or video file as Blocked, the file cannot be played using a playback address obtained through an API or SDK. It can only be played in the ApsaraVideo VOD console or using a security IP for moderation. However, any valid playback addresses that have already been exposed can still be used for playback.
Review Before Publish
When an audio or video file is uploaded to ApsaraVideo VOD, it enters the moderation process by default after transcoding is successful. The file status is marked as Reviewing.
You must then initiate a manual moderation in Video Moderation. Otherwise, the file will remain in the Reviewing state. A file in the Reviewing state cannot be played using a playback address obtained through an API or SDK. It can only be played in the ApsaraVideo VOD console or using a security IP for moderation. However, any valid playback addresses that have already been exposed can still be used for playback.
Click OK to complete the configuration.
Step 2: (Optional) Configure event notifications
Before you start media moderation, you can configure event notifications for ApsaraVideo VOD. This lets you receive callback messages that contain the media moderation status and results after moderation is complete. For more information, see Configure callbacks. The following event notifications must be enabled for media moderation:
Manual moderation: ManualAuditComplete
Automated moderation:
Enable Audio/Video Analysis Complete to receive the Audio/Video Analysis Complete event notification. You can call the SubmitAIMediaAuditJob operation to start a review job only after you receive this notification.
Enable Video AI Processing Complete. This switch is shared by all AI processing tasks. After you enable it, you can receive the Automated Review Complete event notification.
Step 3: Start manual or automated moderation
You can first perform an automated moderation for a preliminary review. Then, you can manually review the audio and video files with suspected violations. This process reduces the manual moderation workload and ensures the accuracy of the moderation results. For more information, see the following documents:
You cannot obtain playback addresses for videos in the Checking or Blocked state by calling API operations. You can configure security IPs for moderation to ensure that only requests from these IPs can obtain playback addresses. This configuration meets internal moderation requirements. For more information, see Security IP for moderation.