Using transcription to add value to audio podcasts
Unlock keyword-rich content to improve searchability using speech technology.
Podcasts are big business in the USA, with 1 in 3 Americans tuning in—that’s over 21 million hours of audio every day.
They have seen a huge surge in popularity in recent years—for producers they are an affordable medium and a convenient way for stories to find an audience, while listeners appreciate the wealth of free content available to them. Portable and easy to access, popular podcasts quickly gather a loyal and regular listener base.
Broadcasters are increasingly turning to transcription services to add extra value to their audio podcasts. For those seeking to expand their audience, search engines can index text better than audio, and transcripts provide keyword-rich content, improving searchability. Listeners wishing to quote or share comments on social media would have to first transcribe them, so providing a transcript increases ease of shareability, driving more traffic to the website and boosting audience figures. Traditional transcription services can be slow and expensive, which is why Speechmatics’ Automatic Speech Recognition (ASR) software is a popular solution, providing full, accurate transcripts of audio blogs within minutes, rather than the hours a human needs to listen to the files in real-time. Public Interest Podcast, based in the Washington, D.C., metropolitan area, offers a positive and entertaining window into the lives of public servants, with a collection of conversations with politicians, activists and community figures on issues such as healthcare, the environment, education and business. The podcast’s archives give a fascinating insight into community and public service, which is why it was decided that the collection should be transcribed to create a book.
Manual transcription was considered as an option but was too slow and expensive. Instead, Public Service Podcast turned to Speechmatics’ automated cloud transcription service for pre-recorded audio batches to speed up the process and get the best value for money.
Accessed via a web console, the pay-as-you-go service uses speech recognition to transcribe hours of podcast audio in minutes, accurately identifying and labelling the changing speakers and quickly returning the results by email, text or JSON, complete with time-stamping and word confidence scores.
“”"We want our podcasts to reach the largest possible audience, By turning them into a book, we can reach people who don’t connect with audio content—we want to maximize accessibility."
