Mid-GOP I-Frame Generation for Low-Latency Video Playback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Temporal encoding schemes in video streams, such as H.264, require decoding frames leading up to the desired start time when playing mid-GOP, leading to resource waste and complexity due to burst decoding, especially for large GOP durations.
Innovation Solution
A video recorder transcoding system that generates new I-frames mid-GOP by decoding and replacing original I-frames with new I-frames, updating delta frames to reference the new I-frames, reducing the need for burst decoding and resource usage at playback devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If temporal encoding schemes (H.264) are used to compress video streams, then bandwidth efficiency is improved, but burst decoding complexity and resource usage increase at playback devices
Solution Approach 1:
The video recorder acts as an intermediary between the video source and playback devices. It performs transcoding to generate new I-frames at mid-GOP positions, converting the video stream into a format that reduces burst decoding requirements. This intermediary processing eliminates the need for playback devices to perform complex burst decoding operations.
Solution Approach 2:
The invention changes the encoding parameters by inserting new I-frames at specific positions within the GOP structure. This modifies the temporal encoding scheme to create shorter decoding dependencies, transforming the video stream from requiring burst decoding to allowing sequential decoding with reduced resource usage at playback devices.
2Productivity
If large GOP durations are used to reduce key frame frequency, then bandwidth efficiency is improved, but the need for burst decoding increases when playing mid-GOP
Solution Approach 1:
The invention segments the long GOP structure by inserting new I-frames at mid-GOP positions. This divides the original large GOP into smaller segments with independent decoding starts, allowing playback to begin at any point without requiring burst decoding of the entire original GOP sequence.
Solution Approach 2:
The video recorder performs preliminary transcoding actions to generate new I-frames at anticipated mid-GOP positions before playback occurs. This advance preparation eliminates the need for time-consuming burst decoding during actual playback operations.
3Ease of operation
If new I-frames are generated mid-GOP through transcoding, then client playback complexity is reduced, but server processing load increases
Solution Approach 1:
The video recorder creates copies of I-frames at mid-GOP positions through transcoding. These copied frames serve as new reference points for subsequent delta frames, allowing clients to decode from any point without requiring the original full-GOP structure, thus simplifying client playback while distributing the processing load.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods are disclosed for providing a requested video frame to a device. A group of pictures (GOP) for an encoded video stream captured by a camera can be obtained by a video recorder, where the GOP includes a key frame and, for each frame of multiple frames following the key frame, a delta frame that describes a delta from a previous frame or from a next frame to render the frame. The video recorder can generate, for at least one frame in the multiple frames, a new key frame corresponding to the at least one frame by obtaining the key frame and applying, to the key frame, a portion of the delta frames between the key frame and the at least one frame. The video recorder can provide the new key frame to the device.