About a year ago, Adobe released a web-based tool for recording and editing spoken audio clips called Project Shasta. This project integrated artificial intelligence (AI) to improve the recording quality of podcasts.
Project Shasta has since been renamed Adobe Podcast and is part of a new collection of (still?) free online audio tools that can now be found under the Adobe Podcast umbrella.
The latest tool, Enhanced Speech, now closes the loop again. The online tool denoises audio files and tries to remove all irrelevant background noise fully automatically. In addition, Enhanced Speech is supposed to significantly improve the speech quality so that the recording is hardly inferior to a clip created in the studio. Of course, the term studio is to be seen relatively in this context. The declared purpose of the tool is to enable anyone to create professional audio recordings in an "unprofessional" environment at home.
Adobe Enhanced Speech uses AI to improve the speech quality of an audio clip
As we understand it, you don't need a Creative Cloud subscription to try out the beta version for free, but you do need to register with Adobe.
Afterwards, an audio file can be uploaded via drag & drop in the browser. The tool supports MP3 and WAV formats, whereby the audio file can be a maximum of 1 GB in size or one hour long. The processing time should then be in the range of seconds - afterwards the improved audio version is available for download.
We ourselves have not yet tried the tool, but are of course pleased when large companies make such small, but certainly useful tools available to a wider circle of users free of charge. Especially if you don't have the adequate GPU power available yourself.
However, one should also be aware that the uploaded data is usually used by the companies to further optimize their AI models. With consequences that are perhaps not even assessable today.