How do you optimize the latency and quality of live broadcast with DASH?

1 What is DASH?

DASH is a protocol that allows the server to divide a video into segments of different bitrates and resolutions, and the client to request the most suitable segment based on the network conditions and device capabilities. This way, the video can adapt to changing bandwidth and avoid buffering or degradation. DASH is not tied to any specific codec, container, or encryption method, and it supports multiple audio and video tracks, subtitles, and metadata.

2 How does DASH affect latency?

DASH introduces some latency due to the segmentation and adaptation process. The server needs to encode and package the video into segments, and the client needs to download and buffer the segments before playing them. The segment duration, which is typically between 2 and 10 seconds, determines the minimum latency that can be achieved. The longer the segment, the more data can be delivered in one request, but also the more delay is introduced.

3 How can you reduce latency with DASH?

To reduce latency with DASH, you can employ several techniques, such as decreasing the segment duration to shorten the delay between encoding and playback (although this may increase the number of requests and overhead, and reduce adaptation efficiency). Additionally, you can use chunked transfer encoding (CTE) to deliver segments as they are being encoded instead of waiting for the entire segment to be ready; this can decrease latency by up to 50%, but requires support from both the server and the client. Finally, low-latency DASH (LL-DASH) can enable clients to request partial segments that are not yet completed, and HTTP/2 or HTTP/3 can improve connection and transmission speed; this can reduce latency to less than two seconds, however, it also requires support from both the server and the client.

4 How does DASH affect quality?

DASH affects quality by allowing the client to select the best segment for the current network and device conditions. The quality is measured by metrics such as resolution, bitrate, frame rate, and quality of experience (QoE). The higher the resolution and bitrate, the better the image quality, but also the more bandwidth and processing power are needed. The frame rate affects the smoothness and motion of the video, and the QoE reflects the user's perception and satisfaction.

5 How can you improve quality with DASH?

To improve quality with DASH, you can employ a variety of techniques, such as adaptive bitrate (ABR) algorithms that adjust the segment selection based on network bandwidth, buffer occupancy, and user preferences. Content-aware encoding (CAE) is another option to optimize encoding parameters according to the complexity of the video content. Additionally, multi-bitrate tiles (MBT) can divide the video into spatial regions that can be streamed at different bitrates and resolutions. This can enhance the quality for certain regions, such as the center or edges of the video.