Adaptive Podcasting

In 2022, the BBC presented its Adaptive Podcasting project, which it released as Open Source at the end of last year. Although usage is currently limited, adaptive Podcasting could be of interest for a variety of formats.

Traditionally, podcast production has been linear. Typically, a script is created or a conversation is recorded. The audio files are then cut, mixed, and possibly further processed before they are made available via an RSS feed.

Some publishers already deviate from this norm, especially in the area of advertising. Advertisements are inserted more or less dynamically (DAI) into the respective podcast episode. This method works quite well and can be observed, for example, with the Vergecast, which personally addresses me with German-language advertising.

Adaptive Podcasting goes beyond this idea and focuses on the content level. The podcast player accesses certain user information, such as the time of day or available listening duration.

Producers can insert so-called cues into their podcasts on the other hand. Based on these cues, a decision can then be made as to which audio segment should be played, taking into account the available data.

This technology could be particularly relevant for news programs that can play appropriate regional news at certain points based on user data. But also story-driven podcasts could vary their content depending on the time of day and offer an even more immersive experience, as everything fits the current situation of the listener.

The BBC’s technology is based on SMIL and is essentially a slimmed-down version of it. Cooperation is hoped for, as this is crucial so that the idea does not disappear into oblivion. Podcasters and audio players in particular must keep up with this and support the corresponding functions.

It will become critical when listeners want to replay or recommend episodes and find out that the content has changed. How do listeners deal with this? How can producers use this technology without losing trust? And how is it ensured that such a technology does not become a disaster for data protection? These are questions that still have to be answered, and which will certainly contribute to the fact that the distribution will probably extend itself a bit.

If this idea prevails, in addition to classical podcasts, many interesting formats could emerge that sound individually for each listener and make the experience even more personal.

Further information is available directly at the BBC or in the GitHub-Repo.️