Terms

Audio and video-specific terms

Video transcoding

Video transcoding converts an encoded video stream into another format to meet different network bandwidth requirements, device capabilities, and playback environments. The process involves decoding the source stream and re-encoding it in the target format. The source and output streams may use the same or different video encoding standards.

Live stream transcoding

Live stream transcoding converts a live video stream from one format to another in real time. Parameters such as encoding format, resolution, and bitrate can be adjusted to deliver transcoded streams across different devices and network conditions — supporting users with varying bandwidths, device capabilities, and latency requirements.

Container format

A container format is a file format that bundles compressed video streams, audio streams, and metadata (such as titles and subtitles) into a single file according to a defined specification.

Container formats fall into two categories: storage and streaming.

• Storage formats: AVI, ASF (WMA or WMV), MP4, MKV, and RMVB (RM or RA).

• Streaming formats: Flash Video (FLV), Transport Stream (TS), and MP4. The TS format is used with streaming protocols such as HTTP Live Streaming (HLS) and Real-Time Messaging Protocol (RTMP). The MP4 format is used with HTTP.

The following describes commonly used streaming container formats:

• MP4: a widely supported container format compatible with mobile devices (iOS and Android) and PC browsers. The MP4 file header stores all media metadata, and longer videos produce a larger header — increasing the time required to load the video. MP4 is best suited for short videos.

  An MP4 file consists of boxes (formerly known as atoms) that hold all media metadata, including media arrangement and timing information. The metadata references video frames and describes data layout. The longer the video, the larger the header, and the longer the load time.

• HLS: an HTTP-based media streaming protocol developed by Apple Inc. HLS uses the TS container format by default, splitting a stream into multiple TS segments. An M3U8 index file controls playback order. HLS requires minimal header-buffering time and is well suited for on-demand video. HLS is supported on mobile devices (iOS and Android) but is not natively supported in Internet Explorer on PCs — a custom player is required for PC playback. We recommend ApsaraVideo Player for Web.

• FLV: a format developed by Adobe. FLV is supported by Flash Player on PCs but not on mobile devices unless a dedicated player app is installed. Most mobile browsers, including those on iPhones, do not support FLV. We recommend ApsaraVideo Player.

• Dynamic Adaptive Streaming over HTTP (DASH): a streaming standard that uses fragmented MP4 (fMP4) to split a video into segments that can have different codec settings, such as resolution and bitrate. Players switch between segments seamlessly to implement adaptive bitrate streaming. The media presentation description (MPD) file in DASH is analogous to an HLS M3U8 file. YouTube and Netflix both use DASH for video delivery.

• HLS with fMP4: an updated HLS standard announced by Apple Inc. at WWDC 2016, which adds support for the fMP4 format alongside the traditional TS format. This lets a single transcoding job produce output compatible with both HLS and DASH simultaneously.

  HLS (including HLS with fMP4) and DASH are the most widely used adaptive streaming technologies. We recommend using HLS and DASH.

Codec

A codec is software or hardware that compresses and decompresses digital video. Most video compression is lossy. The following describes the common codec families:

• H.26X: developed by the International Telecommunication Union (ITU). This family includes H.261, H.262, H.263, H.264, and H.265.

  • H.261: used in early video conferencing and video calls.

  • H.263: used in video conferencing, video calls, and online videos.

  • H.264: also known as MPEG-4 Part 10 or MPEG-4 Advanced Video Coding (MPEG-4 AVC). It is the most widely deployed video compression standard for recording, compressing, and publishing high-definition video.

  • H.265: also known as High Efficiency Video Coding (HEVC). H.265 succeeds H.264 with twice the compression ratio at the same visual quality — reducing bitrate by 50% without quality loss. H.265 supports resolutions up to 8192 × 4320.

• MPEG: developed by Moving Picture Experts Group (MPEG), a joint working group of the International Organization for Standardization (ISO) and International Electrotechnical Commission (IEC). MPEG defines the following video compression standards:

  • MPEG-1 Part 2: used in VCD and some online videos, delivering video quality comparable to VHS.

  • MPEG-2 Part 2: equivalent to H.262, used in DVD, SVCD, and most digital video broadcasting and cable distribution systems.

  • MPEG-4 Part 2: used in network transmission, broadcasting, and media storage. Offers better compression than MPEG-2 and H.263 V1.

  • MPEG-4 Part 10: technically identical to ITU-T H.264. ITU-T calls it H.264; ISO and IEC call it MPEG-4 AVC. Both organizations developed the standard jointly.

• Audio Video Coding Standard (AVS): a series of digital audio and video coding standards developed by the Audio Video Coding Standards Workgroup of China. Two generations have been released:

  • AVS1 and AVS+: the first-generation standards. AVS+ achieves compression efficiency equivalent to H.264 or MPEG-4 AVC High Profile.

  • AVS2: the second-generation standard, designed for efficient compression of UHD (4K and higher) and HDR video. AVS2 achieves twice the compression efficiency of AVS+ and H.264 or MPEG-4 AVC, surpassing H.265 or HEVC.

• Other codecs — such as Google's VP8 and VP9, and RealNetworks' RealVideo — are rarely used for online video.

Before selecting a codec, verify that it is supported by your target playback clients, such as apps and web browsers. Use the most widely supported codecs available. ApsaraVideo VOD supports H.264 or MPEG-4 AVC and H.265 or HEVC as video codecs, with H.264 or MPEG-4 AVC as the default. For audio codecs, ApsaraVideo VOD supports MP3, AAC, VORBIS, and FLAC, with MP3 as the default.

Bit rate

The bit rate is the number of bits transmitted per second, measured in bits per second (bit/s). In the video field, the bit rate is equivalent to the bitrate. A higher bit rate means more data transmitted per second, resulting in better audio or video quality but a larger file size.

Bitrate

Bitrate is the amount of data a video file uses per unit of time, and is the primary control parameter for image quality in video encoding. It is measured in bits per second (bit/s), typically expressed in Kbit/s or Mbit/s. For videos at the same resolution, a higher bitrate means lower compression and higher quality.

File size scales directly with bitrate: File size = Duration × Bitrate ÷ 8. For example, a 60-minute 720p video at 1 Mbit/s has a file size of 3,600 s × 1 Mbit/s ÷ 8 = 450 MB.

Resolution

Resolution is the number of pixels in each dimension of a video frame — for example, 1,280 × 720 means 1,280 pixels wide and 720 pixels tall. Higher resolution means more pixels per frame and sharper images.

Resolution and bitrate are closely linked. Higher resolutions require higher bitrates to maintain quality. Each resolution has a recommended bitrate range:

• Below the lower limit: image quality is poor.

• Above the upper limit: file size and bandwidth consumption increase, but visual quality improves only marginally.

Frame rate

Frame rate is the number of frames displayed per second, measured in frames per second (FPS) or Hz. A higher frame rate produces smoother, more lifelike motion.

• 25–30 FPS is sufficient for most content.

• 60 FPS delivers an immersive, high-motion experience.

• Beyond 75 FPS, improvements in perceived smoothness diminish significantly.

Setting a frame rate higher than your monitor's refresh rate wastes GPU processing power, as the extra frames cannot be displayed. At the same resolution, higher frame rates require proportionally more GPU capability.

GOP

A Group of Pictures (GOP) is a sequence of consecutive frames in an MPEG-encoded video. Each GOP starts with an I-frame and ends just before the next I-frame. A GOP contains three frame types:

• Intra coded picture (I-frame): a complete, self-contained keyframe that can be decoded independently. Every GOP begins with an I-frame, which is always the first frame in a video sequence.

• Predictive coded picture (P-frame): encodes only the difference relative to the previous I-frame or P-frame. P-frames use fewer bits than I-frames, but are sensitive to transmission errors because they depend on earlier frames.

• Bidirectionally predictive coded picture (B-frame): encodes differences relative to both the preceding and the following frames. B-frames achieve the highest compression ratio among the three types, but require the most decoding resources.

The GOP value defines the keyframe interval — the number of frames between two Instantaneous Decoding Refresh (IDR) frames. At least one keyframe is required per second. The keyframe interval in seconds = GOP value ÷ frame rate. For example, with a default GOP value of 250 and a frame rate of 25 FPS, the interval is 250 ÷ 25 = 10 seconds.

Choosing the right GOP value balances video quality, file size (bandwidth), and seeking performance:

• Larger GOP values reduce file size. However, if the value is too large, frames near the end of a GOP become distorted and video quality degrades.

• A larger GOP value also slows seeking. When a user seeks to a position, the player jumps to the nearest preceding keyframe. A larger GOP means a greater distance from the target position to that keyframe — more P-frames and B-frames to decode, and longer load time.

• P-frames and B-frames take longer to encode than I-frames. A large GOP value produces many of them, reducing encoding efficiency.

• A small GOP value requires a higher bitrate to maintain image quality, increasing bandwidth consumption.

Sampling rate

Sampling rate (or sampling frequency) is the number of audio samples captured per second from a continuous signal, measured in Hz. A higher sampling rate captures more detail and produces more natural-sounding audio.

Sound channel

A sound channel is an independent audio signal recorded or played back from a specific spatial position. The channel count refers to the number of audio sources during recording or the number of speakers during playback.

IDR frame alignment

An IDR (Instantaneous Decoding Refresh) frame is a special type of I-frame. Unlike a regular I-frame — where subsequent P-frames and B-frames can reference I-frames that appeared earlier in the stream — no frame after an IDR frame can reference any frame before it. The first I-frame in a sequence is always an IDR frame.

When the decoder encounters an IDR frame, it flushes the reference frame buffer immediately. This prevents decoding errors from propagating forward in the stream. All frames after the IDR frame are coded as a new sequence. IDR frames also enable random access: most players seek to the IDR frame nearest to the target position, avoiding the need to reverse-parse earlier frames.

When transcoding a video into multiple versions at different bitrates, enabling IDR frame alignment ensures the IDR frames in all output files are aligned precisely by time and frame content. This allows a player to switch between bitrate versions without visible lag.

Profile

A profile defines the set of encoding capabilities available for a specific class of applications. H.264 provides three profiles:

• Baseline Profile: uses I-frames and P-frames, supports only progressive video and context-adaptive variable-length coding (CAVLC). Designed for low-complexity devices and applications that require fault tolerance, such as mobile video calls.

• Main Profile: adds B-frame support and interlaced video, along with both CAVLC and context-adaptive binary arithmetic coding (CABAC). Used in mainstream consumer electronics such as MP4 players, portable video players, PSPs, and iPods.

• High Profile: extends the Main Profile with 8 × 8 inter-prediction, custom quantization, lossless video coding, and additional YUV formats including 4:4:4. Used for broadcast and disc storage, particularly in high-definition television applications. Blu-ray Disc uses this profile.

Bitrate control method

Bitrate control methods determine how the encoder allocates bits across a video stream. The three main methods are:

• Variable bitrate (VBR): allocates bits based on scene complexity. Simple scenes use fewer bits; complex scenes use more. VBR balances quality and file size efficiently.

• Constant bitrate (CBR): maintains a fixed bitrate throughout the stream. CBR files are larger than VBR or ABR files at comparable quality levels.

• Average bitrate (ABR): targets a specified average bitrate while allowing the instantaneous bitrate to vary. LAME introduced ABR to address the quality-size trade-off between CBR and VBR. ABR divides the stream into segments of 50 frames, encoding simpler segments at lower bitrates and complex segments at higher bitrates.

  The peak bitrate in some segments may exceed the target, but the overall average stays within the specified range. ABR is a modified form of VBR that keeps the average output within a defined bound while adapting to content complexity. Alibaba Cloud uses ABR by default.

Container format conversion

Container format conversion rewraps an audio or video file from one container format to another — for example, converting an AVI file to MP4 — without re-encoding the audio or video streams. The compressed streams are extracted from the source container and written into the target container as-is.

Compared with transcoding, container format conversion has two main advantages:

• Fast processing: encoding and decoding account for most transcoding time. Since container format conversion skips both steps, processing is significantly faster.

• Lossless quality: no re-encoding means no quality loss.

The output file retains the same resolution and bitrate as the input, making it effectively equivalent to the source quality.

Narrowband HD™

Narrowband HD™ is a proprietary encoding optimization technology built on human vision models. It applies different encoding strategies to different content types to reduce bitrate by 20 to 40% while maintaining the same perceived visual quality.

IMS-specific terms

Local file

A local file is a media file stored on your device that has not yet been uploaded to Object Storage Service (OSS).

Input file

An input file is a media file stored in OSS. A local file uploaded to OSS can be used as an input file.

Output file

An output file is a media file or file set that MPS generates after transcoding and stores in OSS.

Custom template

A custom template defines a set of transcoding parameters — including audio, video, and container settings — and is identified by a unique ID. Custom templates can be created in any region and apply to all transcoding jobs in that region.

Preset template

A preset template is an intelligent transcoding template built into MPS. Preset templates dynamically adjust transcoding settings based on the characteristics of the input file, generating the optimal output under specific bandwidth conditions. For details, see Preset template details.

Watermark

MPS supports adding up to four static watermarks to an output file. Create a watermark template to define reusable attributes such as position, offset, and size. When submitting a transcoding job, specify the watermark template and the watermark content to apply.

Common terms

File format

A file format is identified by a filename extension (such as .doc, .jpg, or .avi) that tells the operating system which application to use to open the file. Common video filename extensions include .avi, .mpg, and .mp4. Video files with recognized extensions can be opened directly by video players installed on your PC.

UTC (ISO 8601 standard time format)

Coordinated Universal Time (UTC) is the global time standard based on atomic seconds. UTC serves as the reference for all time zones. The acronym UTC is a compromise between the English abbreviation CUT and the French abbreviation TUC.

All time fields returned by ApsaraVideo VOD APIs and all time parameters in API requests use UTC by default. Times follow the ISO 8601 format: YYYY-MM-DDThh:mm:ssZ. For example, 2017-01-11T12:00:00Z represents 20:00:00 on January 11, 2017 in UTC+8 (China Standard Time), which is 8 hours ahead of UTC.

Region

A region is the physical location where an Alibaba Cloud service is deployed. Select a region close to your business to reduce access latency and improve user experience.

OSS

OSS stands for Object Storage Service, Alibaba Cloud's cloud object storage. MPS reads input media files from OSS and writes output files back to OSS.