Manual transcripts, corrupted audio files, labor-intensive post-production \u2014 it\u2019s hard to believe how many aspects of our work used to consume so much time and energy just a few years ago. Thanks to AI-powered tools, but also to years of hands-on experience and learning, the way we work has fundamentally evolved and become far more efficient.<\/p>\n
A microphone kit that once cost well over a thousand euros and weighed at least ten kilograms now fits into a small handbag. There is already a wide range of compact wireless microphones on the market that can be controlled either via smartphone or directly through the included receiver. Wireless audio sets like the R\u00d8DE Wireless Pro (suitable for recording two people) are also designed so that recordings cannot clip. Background noise is heavily suppressed as well.<\/p>\n
Even if unwanted or distracting background noise still ends up in the recorded audio, there are now easy-to-use and highly effective solutions. AI tools such as Auphonic and Adobe Enhance can filter out hiss, hum, and other noise, level out differences in volume between speakers, and automatically cut filler words or overly long pauses. If the raw recordings are very poor, however, these tools can sometimes alter the voices of the speakers more than intended.
\nAfter any AI processing, we of course listen through everything again, make fine adjustments in the editing software, cut where needed, add the music bed, and so on.<\/p>\n
It\u2019s almost unimaginable that until a few years ago we still transcribed some episodes by hand. Early automated transcripts were often unusable, especially when stronger dialects were involved. Today, countless transcription options are available. We use the Riverside platform, originally designed for remote recording \u2014 it now also offers automatic transcripts, show notes, title suggestions, and social media snippets.<\/p>\n
Both in Riverside and on other platforms (and of course directly in ChatGPT), episode descriptions can be generated automatically. However, these texts all tend to sound alike. At most, we use them as a rough starting point and then write the show notes ourselves.<\/p>\n