Fifteen years after its release, MPEG-2 remains entrenched as the codec of choice for broadcast delivery. But a group of early adopters spanning industries from sports to education are exploring newer compression techniques in order to push beyond UHD into super resolutions.
That’s a key takeaway from the “Encoding & Transcoding Trends 2021” survey, produced by Streaming Media and sponsored by Bitmovin.
While MPEG-2 is embedded in the transmission chain (if it ain’t broke, why fix it) it often occurs side-by-side to more modern, live-streaming codecs such as H.264 (AVC) and H.265 (HEVC).
Another key finding was the fact that early adopters are more open to not just near-term codecs such as AV1 and EVC, but also show continued interest in VP9.
VP9 is tied tightly to mobile device usage on many Android devices, including Asia-Pacific markets where the device landscape already supports VP9.
Usage is more prevalent among a cross-industry group of early adopters. What’s interesting about this early adopter group is that they not only use higher content acquisition resolutions, but sometimes up to 8K and beyond.
This group also has a higher tendency to use cloud-based encoding. Even if they have hardware-based encoding on premise, the use of cloud-based encoding allows early adopters to try out newer resolutions, frame rates, and bit depth without committing to new on-prem equipment purchases, the report says.
“Organizations that haven’t moved their encoding to the cloud will make the move in the next 12 months. Having said that, hybrid approaches look to also trend upwards over the next twelve months.”
The industry buzz around per-title encoding continues to grow, as does interest in context-aware encoding (CAE) but the reality, based on both anecdotal evidence and these survey responses, is that CAE hasn’t yet caught on.
Having said that, the early adopter subgroup’s responses should be a wakeup call to anyone not using CAE. The survey showed that almost 30% of early adopters regularly use CAE and per-title encoding, and a full 79% of early adopters use it for at least a portion of their encoding workflow.
It’s not hard to see why: per-title encoding analyzes each single asset and optimizes it exactly for particular commercial requirements.
As Bitmovin’s Markus Hafellner notes, “Once you have experienced per-title encoding, it really helps in your workflow design and a lot of the additional commercial factors that influence the encoding decisions. The core philosophy is really that [CAE] is intelligent: If you have a big enough content catalog, there are no two pieces that are exactly the same.”
One other area to note is education. Over the last 18 months, educational institutions have used video like never before, and so interest in encoding techniques — for both live and on-demand VOD — has grown considerably.