There are not many options when it comes to making meaningful reductions in storage and streaming costs. Adopting a new, more efficient codec can deliver significant bandwidth savings but increases costs in the short-term. According to FirstLight Media, AI encoding can deliver up to 50% savings with your existing codec.
In this interview with FirstLight Media co-founders’ CEO André Christensen and CTO Juan Martin, they explain how their company is using AI and machine learning to improve the efficiency of the encode cycle. They also explain the additional business benefits the approach can enable.
Chapter 1: About FirstLight Media (00:41)
FirstLight Media is an end-to-end cloud video delivery platform. It started life as QuickPlay and was acquired by AT&T in 2016. In March of 2020, the technology was spun out of AT&T and became part of FirstLight Media. Machine learning and AI is a foundational part of the FirstLight platform.
Chapter 2: Save storage and streaming bandwidth with AI-assisted encoding (2:34)
The traditional way to encode video is to pick an appropriate profile for the type of content and encode the content with that fixed profile. Encoding in this way can result in video streams and video files much larger than they need to be. For example, consider a courtroom drama that has one car chase scene. Delivering the car chase in excellent quality requires relatively light compression. However, most of the movie is relatively static content, which can be compressed heavily with no loss in picture quality.
Mr. Martin says the company aims to minimize the time taken to encode content and to optimize the quality of content relative to the cost to deliver it. To achieve that, FirstLight analyzes each of the scenes of a movie or show. It then uses AI to set the optimal encoding profile for each scene based on the expected consumer viewing environment. The approach can yield as much as a 50% saving on streaming bandwidth and video storage space.
Chapter 3: AI encoding adds value beyond saving storage and streaming costs (4:49)
The scene-based encoding technique is being applied on small chunks – just a few seconds – of the video. The method has a couple of additional benefits beyond bandwidth and streaming cost savings.
The first benefit is the speed of encoding the video—an hour of content to be compressed and available in less than a minute. Getting new content into a video service more quickly can help boost engagement. As well, it is easier to assemble the smaller chunks of video into shorter versions and excerpts of the original. Such different formats can help drive engagement on smartphones and tablets.
 Encoding – a process to convert and compress video data into a form appropriate for streaming and storage.