The advances happening with AI in audio are just as exciting as AI image creation. At least, that is true for me. Studio Sound “is a regenerative speech effect that uses AI to remove noise and echo, then restore and elevate speech.” It is designed to make flawed recordings sound like they were made in a studio.
The first audio file below was recorded in my office without attempting to mask the background dryer noise. My office here at home is right next to our laundry room. It’s always been very difficult for me to do voice recordings. I’d have to wait for certain times of the day or record in my closet. For this test, I left my office door open so the sound could filter in. It’s probably a fairly extreme test of studio sound, though also likely a very common noise condition for those working in a home office.
The audio below was enhanced using the Studio Sound feature in Descript. You can apply the effect at any percentage you wish. The sound you’re hearing below was set to 100%. There was so much dryer noise in the background that anything other than 100% introduced the sound back in.
Is it perfect? No, of course not. Recording this in a studio would certainly produce a richer sound. But it is most certainly useable audio for 99% of the situations I might need my voice to be recorded. Below is a video of me playing with the Studio Sound filter percentage so you can experience it in real-time.