As some of you might have heard I’m currently working on a new service with a little team called voicenext which is an audio messaging service.
voicenext is a new way to communicate with your friends by using audio instead of text. For one this solves the most obvious problem: smartphone keyboards suck. It also solves a second problem which is getting the message across. Did you know there’s like 5 different ways to say “Hello” in German and depending on the tone it can mean different things? Now you know. And I bet this isn’t an exclusive problem to the German language either.
Now there’s a few problems to solve as well with voice notes. For example some people tend to send voice notes that are longer than a couple of minutes. I don’t know, but if you want to tell me so much wouldn’t it be easier to just call me? That’s why we limit voice notes to 30 seconds. We found this is a tolerable length for voice notes without compromising on content. Just skip your “uh”s and come straight to the point, done, we’re under 30 seconds.
Another issue we found is that voice notes don’t sound good, they’re either too low volume or there’s just a bunch of noise in the background. Next time your friend tries to fire up a nuke while sending a voice note tell them to use voicenext, their voice will be crystal clear.
Then there’s compression. As you might know compression is really awful. So how do we manage that your voice notes sound crystal clear while also not eating up too much data. Easy: use a compression algorithm that doesn’t make your voice sound absolutely awful. For that we use an industry standard OPUS algorithm with a bitrate that is the sweet spot between good quality and compression. Problem solved.
I hope I could give you a small insight on what we’re currently cooking up. Until next time 👋
Leave a Reply