Content-adaptive Encoder Preset Prediction for Adaptive Live Streamin...

In live streaming applications, a fixed set of bitrate-resolution pairs(known as bitrate ladder) is generally used to avoid additional pre-processingrun-time to analyze the complexity of every video content and determine theoptimized bitrate ladder. Furthermore, live encoders use the fastest availablepreset for encoding to ensure the minimum possible latency in streaming. Forlive encoders, it is expected that the encoding speed is equal to the videoframerate. An optimized encoding preset may result in (i) increased Quality ofExperience (QoE) and (ii) improved CPU utilization while encoding. In thislight, this paper introduces a Content-Adaptive encoder Preset predictionScheme (CAPS) for adaptive live video streaming applications. In this scheme,the encoder preset is determined using Discrete Cosine Transform(DCT)-energy-based low-complexity spatial and temporal features for every videosegment, the number of CPU threads allocated for each encoding instance, andthe target encoding speed. Experimental results show that CAPS yields anoverall quality improvement of 0.83 dB PSNR and 3.81 VMAF with the samebitrate, compared to the fastest preset encoding of the HTTP Live Streaming(HLS) bitrate ladder using x265 HEVC open-source encoder. This is achieved bymaintaining the desired encoding speed and reducing CPU idle time.