How AI can help you with your audiobook productions: text and audio formats

Artificial Intelligence (AI) is a complicated topic with many applications across a diverse set of industries. From your phone’s automated personal assistant (Siri or Alexa), the suggestions algorithm on Netflix, to your Roomba vacuum, AI tools are all around us. However, when people hear the term AI, they often think of Skynet from the Terminator series or Ultron from the Avengers, robots or human-like machines intent on destroying the Earth and the human race. However, AI is defined as a computer system (or algorithm) that has been designed by humans for the purpose of executing a task to benefit humans. 

Computers and software processes can do two things with AI and machine learning – synthesis or analysis. Let’s just talk about analysis – about how using the power of data gathering and computational thinking can make audiobook production better, helping all elements of audio production including how books are planned, written and formatted. This is a huge and fascinating topic, and the more prevalent that AI tools become in the near future, the better we can all embrace the possibilities!

Uncovering invisible patterns

People can recognise patterns to a certain extent. Modern computers take it to a whole other level. AI allows us to view patterns in new and thoughtful ways, allowing us to analyse and act on new data. There’s two ways to think of this as well. 

You can run pattern recognition algorithms over a book once it’s completed, or you can run that process while writing is in progress. As a quick example, you can find out at the end of a book-writing project that you overused certain words or phrases. 

Or, if the AI algorithm is running in real-time, it can tell you that this is the third time you’ve used the word ‘all’ in one paragraph, enabling the author to fix it in editing or to adjust on the fly. Compared with AI tools, there’s a good chance that a human may not catch this as they are focusing on the story as a whole. Even an experienced editor is going to miss most patterns because that’s too much to think about in a single project. AI analysis can work with multiple projects in entire databases and continuously update what types of patterns it’s looking for.

Making hard or repetitive things easier

One of the things AI excels at is making quick work out of time-consuming and frustrating tasks. Audiobook engineering is viciously repetitive, and there are many tasks that take an enormous amount of brainpower during editing, proofing, mastering, organisation, labelling, transfer and distribution.

If any part of those processes can be automated, AI is the best candidate for the job. It should always require a human to run the processes and filter the results, but almost all of the worst parts of audiobook production are the parts that don’t require thinking, they just require rote actions. Let AI do the menial stuff! An example on the authorship side of things might be book formatting. If a software tool is set up the right way, specs can be set by the author, and then the tool will automatically and instantly reformat the book as desired by the author. For audiobook production, tools like Pozotron allow for audiobook preparation, proofing and QC to be completed more accurately in a fraction of the time it takes a human to do alone. 

Use the machine to analyse training techniques

It’s possible to set up AI to help train authors, editors and technicians that are part of book publishing and production processes. And this isn’t about squeezing blood from a stone and pushing the ‘harder faster better NOW!’ factory worker mentality. This is just about letting a machine track you and your progress doing various tasks, and giving you a readout at the end to help you improve. For example, based on the right data sets and general setup, an AI analysis program could easily figure out 

  • best times of day for training engineers, or 
  • how many mistakes proofers find when they’re hungry or angry, or 
  • how long of a break people should take to refresh creative and technical abilities.

No human could gather all that data about 10 people simultaneously training around the world, but the right AI process could handle that in a flash.

Better results with less effort

Have you specifically established your goals or KPIs for a project? For example, these could include more money, quicker turnarounds, greater self-reported happiness, shorter milestones, more people satisfied.

Once you know those goals for a book project, you can let an AI measure them either in progress or at the end of the project, or as a part of a continuous improvement process or cycle. Through comprehensive analysis, AI can help you figure out where the hurdles are, and take steps to improve the system as a whole. Notice that no part of this is taking any humanity out of the creative process, and no part is reducing the talent required to make a great book production! 

Audiobook proofing as an example of an AI-based solution

All of this theory is great, but what’s an example in practice? At Pozotron, we created an AI tool to help proofreaders with a simple but highly frustrating task – finding misreads between a script and its matched audiobook narration. Using an algorithm that gets updated every few months, the software scans every word of narrated audio, compares it to the uploaded script, and gives you a readout of potential discrepancies. This is using AI for the power of good in the world. What used to be a horrible task with a high incidence of failure now is made much easier, and proofers can concentrate on the decision-making parts of their job instead of the parts that a machine can find more easily and consistently.

Contribute your ideas

This isn’t all that AI can do for publishers: in fact, at Pozotron we have a ton of creative ideas about how AI could help book professionals in their search for the next great novel and audiobook. We are actively planning how Pozotron can make writing and publishing novels and adapting them for audio easier and more efficient, so feel free to contribute your ideas in the comments below or to contact me personally. I encourage you to take away from this article that AI tools can get you to where you need to go faster than ever before, enabling you to focus on the human elements of creativity that can’t be replicated by AI!

Ryan Hicks is the Director of Outreach at Pozotron, a software company that focuses on helping narrators and production houses use technology to save time at every step of the audio production process. Pozotron’s AI-powered software suite will help you find efficiencies in project preparation, editing, proofing and QC to get as close to the 1:1 ratio as possible.

Responses

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  1. Thanks to Ryan from Pozotron for contributing this piece on AI and automation for your audiobook processes – a super interesting read!