Build Transcoding Templates to Control Video Output Quality - ApsaraVideo Media Processing - Alibaba Cloud -

A transcoding template defines the output parameters for audio and video transcoding jobs. It includes basic, video, and audio parameters. You can create custom templates or use system presets. This document explains how to create and use transcoding templates.

Template overview

Transcoding

Video transcoding converts a compressed bitstream to adapt content for different user devices and network bandwidths. Transcoding is a process of decoding and then re-encoding, so the source and output bitstreams can use the same or different video codecs.

VOD Media Processing currently supports the following five types of standard transcoding:

Type	Description	Use cases
Regular transcoding	Provides a basic transcoding service with comprehensive video transcoding capabilities. It is suitable for general playback and distribution scenarios and meets basic business needs without guaranteeing a specific video bitrate.	Suitable for standardizing long-form video content with complex formats and low distribution volumes.
Narrowband HD 1.0 transcoding	Narrowband HD 1.0 delivers video quality comparable to the source video at a relatively low bitrate. It can reduce distribution costs by 20% to 40% compared to regular transcoding, effectively balancing visual quality and transmission costs.	Suitable for business scenarios with high playback volume and a strong focus on visual quality, bitrate, and distribution costs.
Narrowband HD 2.0 transcoding	Narrowband HD 2.0 incorporates AI-powered image quality enhancement technology. Based on a human visual model, it reconstructs the encoding optimization goal, shifting the focus from "fidelity first" to "optimal subjective experience," and supports scene-based parameter tuning. At the same visual quality, it further reduces distribution costs by over 20% compared to Narrowband HD 1.0, significantly improving overall business performance.	Suitable for scenarios where source quality is poor or requires local adaptive optimization, playback volume is high, and there is a strong focus on cost control, premium user experience, and high-quality content presentation.
Audio transcoding	Audio transcoding can convert an audio file's format, extract audio streams from video files, and apply effects and enhancements.	Suitable for various scenarios that require converting audio file formats, adjusting audio quality parameters, or extracting audio from video to meet playback compatibility, storage optimization, and content creation needs.
Remuxing	Changes the video file's container format while preserving the original resolution and bitrate.	Suitable for business scenarios where the video size or bitrate does not need to be adjusted.

Scenarios and strategies

VOD Media Processing provides various scenario-based transcoding templates with five strategies: Quality First, Bitrate First, Extreme Compression, Balanced, and Vibrant Color. They cover scenarios such as short dramas, education, User-Generated Content (UGC), e-commerce, HD film and TV, and general-purpose video.

Transcoding scenarios

Short dramas: Suitable for new forms of web series, including vertical, micro, and serialized short dramas. This strategy optimizes fast-paced content, balances quality with bitrate, and adapts to vertical viewing on mobile devices.
Education: Designed for online courses, instructional videos, knowledge sharing, and training materials. Transcoding prioritizes the clarity and readability of text, charts, and presentation content.
User-Generated Content (UGC): Suitable for scenarios involving UGC, such as short videos, social sharing, and live stream replays.
E-commerce: Designed for e-commerce videos like live streams, product displays, and talent shows. Transcoding improves the clarity and detail of key areas, such as faces, enhancing the visual quality of Regions of Interest (ROI).
HD film and TV: Suitable for content with extremely high quality requirements, such as TV series, movies, and documentaries.
General: Suitable for standard on-demand or feed videos with no special requirements.

Transcoding strategies

Type	Strategy	Description
Narrowband HD 1.0	Quality First	Applies less compression to preserve more visual detail, prioritizing the viewing experience.
	Balanced	Balances bitrate and quality by compressing the file as much as possible without a noticeable decrease in subjective visual quality.
	Bitrate First	Uses high compression to prioritize bitrate reduction, which may result in some quality loss.
Narrowband HD 2.0	Vibrant Color	Aims for vivid colors and a strong visual impact to enhance the user's sensory experience and provide a premium visual experience.
	Extreme Compression	Maintains an acceptable level of quality while aggressively pursuing a low bitrate.
	Quality First	Saves more bandwidth than the 1.0 version while meeting high-quality requirements.
	Balanced	Achieves the best balance between quality and bitrate with current technology, making it the top choice for most scenarios.

Subtitle burn-in

A subtitle burn-in template is a type of transcoding template used to embed subtitles into a video. This process embeds subtitles directly into the video frames, which improves playback compatibility and user experience by eliminating the need for an external file.

Audio-visual enhancement

Audio-visual enhancement improves source video quality through operations such as denoising, color and contrast enhancement, super-resolution, and SDR-to-HDR conversion. It enhances the viewing experience by improving color, brightness, and clarity. You can choose processing modules and their intensity based on the input video's characteristics.

To enable the audio and video enhancement feature, submit a ticket.

VOD Media Processing currently supports the following six types of audio-visual enhancement:

Type	Description	Use cases
Deinterlacing	Removes interlaced fields (by dropping either odd or even fields) and then doubles the frame rate. This process converts interlaced video to progressive video.	Suitable for converting interlaced videos (such as broadcast television content and old videos) to progressive-scan format.
Multi-frame denoising	Removes temporal noise from the video, resulting in a cleaner and more stable picture over time.	Suitable for most videos. It is generally recommended to enable this feature and adjust the denoising strength using parameters, except for very high-definition videos.
Compression artifact reduction	Removes compression noise caused by encoding (such as ringing and blocking artifacts) while enhancing edges and texture details. This improves image clarity while reducing noise.	Suitable for most videos. For high-definition videos, it primarily serves to enhance texture details.
Color and contrast enhancement	Adjusts the local and global contrast of the image while enhancing color vibrancy.	Suitable for most videos. The level of saturation enhancement can be adjusted through parameters.
Super-resolution	Improves video resolution and edge textures, significantly enhancing the overall clarity of the video. It currently supports 2x and 3x super-resolution scaling.	Suitable for most videos that require a resolution increase. We recommend using it in conjunction with compression artifact reduction.
SDR-to-HDR conversion	Converts Standard Dynamic Range (SDR) videos to High Dynamic Range (HDR) videos with a wide color gamut, significantly improving contrast, brightness, and color performance. It currently supports conversion to HLG and PQ HDR formats.	Suitable for converting SDR videos to HDR videos.

Watermark

Adding a watermark to a video involves embedding information such as images or text into the video stream during the encoding process. This creates a new video file with the watermark burned in. Typically, watermarks such as a company logo, a TV station icon, a user ID, or a nickname are added to videos to declare copyright ownership or promote a brand.

VOD Media Processing currently supports the following two types of watermarks:

Type	Description
Image watermark	Supports images or videos in PNG, GIF, APNG, and MOV formats. The watermark can be displayed at a specific position for the entire duration of the video, or for a specified time period (from the start time to the end time).
Text watermark	Burns one or more text strings onto the video as a watermark. You can configure the text's font, size, color, opacity, and outline. You can also apply different text content to different videos.

Note

Dynamic image watermarks require that the file extensions for GIF, APNG, and MOV formats be in lowercase. Static image watermarks do not have this restriction.
The watermark image must be stored in the same region as the video. For example, for a video stored in the China (Shanghai) region, the watermark image must also be stored in the China (Shanghai) region. Cross-region usage is not supported. To add or configure a storage location for a region, see Configure Storage Locations.

Create a transcoding template

Console

Log on to the Intelligent Media Service console.
In the left-side navigation pane, choose to go to the Template Management page.

On the Transcoding tab, create a template based on your requirements.

Create a transcoding template

Narrowband HD 1.0

You cannot create custom templates for Narrowband HD 1.0 transcoding. The system provides 216 preset templates that cover six scenarios (short dramas, education, UGC, HD film and TV, e-commerce, and general) and three transcoding strategies (Balanced, Bitrate First, and Quality First). For more information, see Preset Template Details.

Narrowband HD 2.0

You cannot create custom templates for Narrowband HD 2.0 transcoding. The system provides 288 preset templates that cover six scenarios (short dramas, education, UGC, HD film and TV, e-commerce, and general) and four transcoding strategies (Extreme Compression, Balanced, Quality First, and Vibrant Color). For more information, see Preset Template Details.

Regular transcoding

Basic parameters

Parameter	Description
Template Name	Enter a custom name for the regular transcoding template.
Transcoding Format	Regular transcoding.
Container Format	Options: HLS (.m3u8+.ts), DASH (.mpd+.fmp4), CMAF (HLS), CMAF (DASH), 3GP (.3gp), AVI (.avi), FLV (.flv), F4V (.f4v), fMP4 (.mp4), MKV (.mkv), MOV (.mov), MP4 (.mp4), MPEG-2 TS (.ts), MXF (.mxf), and WebM (.mkv/webm).

Video parameters

Note

The parameters you need to configure vary depending on the selected container format. The actual parameters displayed on the page are the ones that apply.

Parameter		Description
Encoding Format		Select a codec. H.264, H.265, and AV1 are supported.
Bitrate Control		Bitrate control determines how the encoder sets the bitrate for the output file. Select a control mode from the drop-down list. The following modes are supported: Source Bitrate: The output video retains the same bitrate as the source video. Constant bitrate (CBR): Applies a uniform bitrate throughout the file, regardless of scene complexity, which typically results in larger file sizes. Average bitrate (ABR): The output file has a specified average bitrate. The bitrate is allocated based on scene complexity, with more complex scenes receiving a higher bitrate and simpler scenes a lower one. This mode ensures the output bitrate is within the expected range while allocating bitrate efficiently. CRF (Constant Rate Factor): Video quality is quantified on a scale where 0 is lossless and 51 is the worst. By setting a CRF value, you can achieve a consistent subjective quality throughout the video. However, the bitrate fluctuates based on scene complexity, making the final output bitrate unpredictable. Enter the relevant bitrate parameters.
Peak Bitrate		Enter the peak bitrate. Valid values: 10 to 50000. Unit: Kbps.
Resolution		Note If your input videos include both landscape and portrait orientations, we recommend using Set by Long/Short Side to prevent stretching or distortion. Resolution determines the dimensions of the output video. You can choose from the following settings based on your needs and the characteristics of the source video: Source Resolution: Maintains the original resolution of the source video. Set by Long/Short Side: Sets the resolution based on the long or short side of the video. This method automatically adapts to landscape (width > height) and portrait (height > width) videos, ensuring they are not stretched or distorted. Set by Height/Width: Directly specifies the output video's width and height. The horizontal dimension is fixed as width and the vertical as height, regardless of whether the input video is in landscape or portrait orientation.
Resolution Check		This parameter is required when Resolution is set to Set by Long and Short Sides or Set by Height/Width. It determines how transcoding is handled when the source video's resolution is smaller than the specified resolution. You have three options: Transcode at Source Resolution: The video is transcoded at its original resolution. Transcode at Specified Resolution: The video is upscaled to the specified resolution for transcoding. Do Not Transcode: The source video is used directly without any processing.
Frame Rate		The number of frames displayed per second. Source Frame Rate: Maintains the original frame rate of the input video, capped at 60 fps. Custom Frame Rate: You can specify an integer value between 0 and 60 fps. Common choices include 15, 25, and 30 fps.
Advanced Parameters	Segment Length	Enter the segment duration. The value must be an integer from 1 to 60. Note Each segment must contain at least one keyframe. Set the segment duration to a multiple of the GOP duration. If the segment duration is less than the GOP duration or not a multiple of it, the transcoding service adjusts the keyframe interval.
	GOP	The maximum interval between keyframes, specified by time or number of frames. The value must be an integer from 1 to 100,000. Note A larger GOP value increases compression efficiency but reduces encoding speed. It also increases the duration of individual streaming segments and the seek response time in the player.
	Encoding Profile	This parameter only applies when the codec is H.264. It supports profiles for high-resolution devices, standard-resolution devices, and mobile devices. Different hardware supports different profiles. When a video is output in multiple resolutions, we recommend setting the lowest resolution to a profile for mobile devices (baseline) to ensure playback on low-end devices. Set other resolutions to a profile for standard-resolution devices (main) or high-resolution devices (high).
	Buffer	Enter a value from 1000 to 128000. Unit: Kbps. Note The buffer size controls bitrate fluctuations. A larger buffer size allows for greater bitrate fluctuation and results in higher video quality.
	Scan Mode	Select a scan mode. Options: Source Mode, Auto Deinterlace, Interlaced, and Progressive. Note If you set the scan mode to Sequential Scan or Interlaced Scan and it does not match the source video's mode, the transcoding will fail. We recommend setting the scan mode to Scan Mode of Input Video or Automatic Deinterlacing to improve compatibility.
	HDR of Source Video	When the input is an HDR source, you must set the dynamic range for the transcoded output. To ensure correct playback of HDR video, your display device and player must support HDR. Otherwise, issues such as overexposure and color casts may occur during playback. Map to SDR: Converts the HDR video to an SDR format for standard devices, aiming to preserve as much color richness as possible. Note This feature is free during the public preview. After the preview period, it will be billed based on processing duration. Keep HDR: Maintains the dynamic range of the source video.
	Color Format	Select the color format you need. By default, the source video format is used.

Audio parameters

Parameter	Description
Disable Audio	Specifies whether to disable the audio. If you disable audio, the output file will not contain an audio stream or related media information. This feature is suitable for scenarios where you need to extract only the video stream from a file.
Encoding Format	Select a codec. The supported codecs vary depending on the container format. The actual codecs displayed on the page prevail.
Encoding Profile	You can set the profile when the codec is AAC. For multi-channel audio sources like 5.1 or 7.1, we recommend setting the profile to `aac_low`. For general playback devices, we recommend `aac_he`. At the same audio quality, `aac_he` can save half the audio bitrate compared to `aac_low`, with a common low bitrate value of 64 Kbps. For high-end playback devices, we recommend `aac_he_v2`. This format not only has a smaller file size and better sound quality, but also has a common low bitrate range of 32–48 Kbps.
Sample Rate	Select a sample rate from the drop-down list. Unit: Hz. The supported sample rates vary by container and codec format. For a complete list, see Supported Sample Rates.
Audio Bitrate	Options: Use Source Audio Bitrate and Average Bitrate. Use Source Audio Bitrate: Maintains the original bitrate of the input audio. Average Bitrate: Sets a fixed average bitrate for the output audio.
Bitrate Check	This parameter is required when Audio Bitrate is set to Average Bitrate. It determines how transcoding is handled when the source audio bitrate is lower than the specified bitrate. The following three options are available: Transcode at Source Audio Bitrate: Transcodes the audio at its original bitrate. Transcode at Specified Bitrate: Re-encodes the audio to the specified average bitrate. Do Not Transcode: Uses the original audio directly without transcoding.
Sound Channels	Select the number of channels from the drop-down list. You can choose to keep the original number of audio channels. The default value is 2.
Volume Normalization	If you enable Volume Normalization, the system automatically adjusts the loudness of audio files to make them consistent. This helps prevent large loudness differences between files. This setting is supported only for single-stream audio output and not for multi-stream audio. Note The professional edition of loudness normalization provides more natural and precise loudness control and is billed based on processing duration. It is suitable for applications with higher audio quality requirements. For more information, see Audio-visual enhancement pricing.

Audio transcoding

Basic parameters

Parameter	Description
Template Name	Enter a custom name for the audio transcoding template.
Transcoding Format	Audio transcoding.
Container Format	Options: AAC (.aac), M4A (.m4a), MP3 (.mp3), MP4 (.mp4), WAV (.wav).

Audio parameters

Parameter	Description
Disable Audio	Specifies whether to disable the audio. If you disable audio, the output file will not contain an audio stream or related media information.
Encoding Format	Select a codec. The supported codecs vary depending on the container format. The actual codecs displayed on the page prevail.
Encoding Profile	You can set the profile when the codec is AAC. For multi-channel audio sources like 5.1 or 7.1, we recommend setting the profile to `aac_low`. For general playback devices, we recommend `aac_he`. At the same audio quality, `aac_he` can save half the audio bitrate compared to `aac_low`, with a common low bitrate value of 64 Kbps. For high-end playback devices, we recommend `aac_he_v2`. This format not only has a smaller file size and better sound quality, but also has a common low bitrate range of 32-48 Kbps.
Sample Rate	Select a sample rate from the drop-down list. Unit: Hz. The supported sample rates vary by container and codec format. For a complete list, see Supported Sample Rates.
Audio Bitrate	Options: Use Source Audio Bitrate and Average Bitrate. Use Source Audio Bitrate: Maintains the original bitrate of the input audio. Average Bitrate: Sets a fixed average bitrate for the output audio.
Bitrate Check	This parameter is required when Audio Bitrate is set to Average Bitrate. It determines how transcoding is handled when the source audio bitrate is lower than the specified bitrate. The following three options are available: Transcode at Source Audio Bitrate: Transcodes the audio at its original bitrate. Transcode at Specified Bitrate: Re-encodes the audio to the specified average bitrate. Do Not Transcode: Uses the original audio directly without transcoding.
Sound Channels	Select the number of channels from the drop-down list. You can choose to keep the original number of audio channels. The default value is 2.
Volume Normalization	If you enable Volume Normalization, the system automatically adjusts the loudness of audio files to make them consistent. This helps prevent large loudness differences between files. This setting is supported only for single-stream audio output and not for multi-stream audio. Note The professional edition of loudness normalization provides more natural and precise loudness control and is billed based on processing duration. It is suitable for applications with higher audio quality requirements. For more information, see Audio-visual enhancement pricing.

Remuxing

Keeps the original resolution and bitrate of the video, updating only the container format. It supports output in mp4, flv, m3u8 (with ts segments), and m3u8 (with fmp4 segments) formats.

Subtitle burn-in

Create subtitle burn-in template

Parameter	Description
Template Name	Enter a name for the subtitle burn-in template.
External Subtitle File Format	Select the external subtitle file format: srt or ass.
External Subtitle Encoding Format	Select the external subtitle encoding format. If you set it to auto, detection may be inaccurate. We recommend specifying a specific encoding format.

Audio-visual enhancement

Parameter	Description
Template Name	Enter a name for the audio-visual enhancement template.
Container Format	Select a container format.
Deinterlace	Select whether to enable deinterlacing if the video contains interlaced segments.
Remove Frame Type	When deinterlacing is enabled, select whether to remove the even or odd fields.
Multi-frame Denoising	Select whether to enable multi-frame denoising if the video contains temporal noise.
Denoising Intensity	When multi-frame denoising is enabled, enter the denoising strength. Valid values: [0.5, 5]. A lower value applies stronger denoising.
Remove Compression Distortion	Specifies whether to enable the removal of video compression artifacts.
Color and Contrast Enhancement	Specifies whether to enable color and contrast enhancement.
Degree of Saturation Enhancement	When color and contrast enhancement is enabled, enter the saturation enhancement level. Valid values: [0, 1]. A lower value applies stronger saturation enhancement.
Super Resolution	Specifies whether to enable super-resolution.
Super-resolution Ratio	When super-resolution is enabled, select the magnification ratio for the resolution. The width and height are scaled proportionally.
Output video dimensions	Preset Resolution: Select a system-preset resolution. Custom Resolution: Enter a width and height. Valid values: [128, 4096]. Unit: px. Note If this parameter is not set, the actual output video dimensions will be the dimensions of the source video after being scaled by the magnification ratio.
SDR to HDR	Specifies whether to enable SDR-to-HDR conversion. This parameter only applies when the output video codec is H.265.

Watermark

Parameter	Description
Template name	Enter a name for the watermark template.
Watermark type	Select the watermark type: Image Watermark or Text Watermark.
Image watermark
Watermark asset	Upload the watermark image. PNG and GIF formats are supported. The file size cannot exceed 20 MB.
Watermark position	Select the watermark position.
Image width	By Value: Sets the watermark image width based on the entered value. Valid values: [8, 4096]. Unit: px. By Percentage: Sets the watermark image width based on the entered percentage, accurate to two decimal places. Valid values: [0, 100]. Unit: %.
Image height	By Value: Sets the watermark image height based on the entered value. Valid values: [8, 4096]. Unit: px. By Percentage: Sets the watermark image height based on the entered percentage, accurate to two decimal places. Valid values: [0, 100]. Unit: %.
Horizontal offset	By Value: Sets the horizontal offset of the watermark image based on the entered value. Valid values: [8, 4096]. Unit: px. By Percentage: Sets the horizontal offset of the watermark image based on the entered percentage, accurate to two decimal places. Valid values: [0, 100]. Unit: %.
Vertical offset	By Value: Sets the vertical offset of the watermark image based on the entered value. Valid values: [8, 4096]. Unit: px. By Percentage: Sets the vertical offset of the watermark image based on the entered percentage, accurate to two decimal places. Valid values: [0, 100]. Unit: %.
Dynamic watermark	Specifies whether to enable dynamic watermarks. If enabled, the watermark appears only during the specified time.
Start time	Enter the watermark's start time. The value must be between 0 and the video's length, in seconds.
End mode	When dynamic watermark is enabled, select the end mode for the watermark: Last until end of video: The watermark is displayed until the end of the video. Last for a specific duration: Enter the duration for the watermark. The value must be from 0 to the video length. Unit: seconds.
Text watermark
Text watermark content	Enter the content for the text watermark.
Font	Select a font.
Font size	Enter the font size. Valid values: [5, 119]. Unit: px.
Font color	Select a font color.
Font opacity	Enter the font opacity. Valid values: [0, 100]. Unit: %.
Horizontal offset	By Value: Sets the horizontal offset of the watermark text based on the entered value. Valid values: [8, 4096]. Unit: px. By Percentage: Sets the horizontal offset of the watermark text based on the entered percentage, accurate to two decimal places. Valid values: [0, 100]. Unit: %.
Vertical offset	By Value: Sets the vertical offset of the watermark text based on the entered value. Valid values: [8, 4096]. Unit: px. By percentage: Set the vertical offset of the watermark text by entering a percentage from 0 to 100. The value is accurate to two decimal places. The unit is %.
Outline width	Enter the outline width for the watermark text. Valid values: [0, 4096]. Unit: px.
Outline color	Select an outline color for the watermark text.

FAQ: Are IMS and MPS transcoding templates interchangeable?

Intelligent Media Services (IMS) and Media Processing (MPS) are two independent products, and their transcoding templates are not interchangeable.

IMS transcoding templates: managed on the Template Management page of the IMS console. Templates are categorized by scenario (such as short dramas and HD film and TV) and include Narrowband HD 1.0/2.0, regular transcoding, audio transcoding, and remuxing types.
MPS transcoding templates: managed on the transcoding template page of the MPS console. Templates are named by container format and codec, and support both custom and system preset templates.

To achieve transcoding effects in MPS similar to those of IMS, refer to the resolutions of IMS short drama templates (480P, 720P) and select MPS preset templates with the corresponding resolutions.