The need for fast turnaround time is what drives most to use an automatic transcription service over other alternatives. People know that a speech-to-text service is not only faster than DIY but also more affordable than hiring a professional. However, what a lot of people don’t consider is how audio quality affects transcript accuracy. Unlike humans, machines are unable to pick up on things like strong accents and overtalk. With this in mind, it’s important to make sure the audio you intend to convert into a transcript is recorded as clearly as possible. Especially if you’re using a speech-to-text transcription machine, like Temi.

Difference between good audio and bad audio

There are some simple recording tips to improve audio quality, which will, in turn, increase the accuracy of your transcripts.

Here are a few things to keep in mind for future recordings:

1. Record in a quiet setting — Avoid lots of background noise. Things like music, traffic, and the buzz of a busy coffee shop will make it difficult to sort the words you speak from what’s going on behind you.
2. Speak clearly, loudly, and slowly — If it’s accuracy you’re looking for, enunciate your words a little more than you would in a normal conversation.
3. Avoid talking over the other person — automated transcription does best with one to two speakers. If there do happen to be multiple people involved in a recorded conversation, make sure that people are not talking over one another.
4. Invest in quality recording equipment — It doesn’t have to be expensive, but a dedicated mic or recording device will go a long way in improving audio quality. You could also opt to use our app for iOS or Android.

If you aren’t sure if the audio you have is a good fit for automated transcription, ask yourself:
  • Is the Is the background noise-free?
  • Are speakers close to the mic?
  • Is conversation free of overtalk?
  • Do speakers have clear accents?

If you can answer yes to these four questions, chances are you’re in good shape. Also, keep in mind that your first file is free to transcribe. We recommend running a test file through our speech recognition system to see what type of tweaks, if any, your recording needs in order to produce the most accurate transcript possible.


Transcript quality depends heavily on your audio quality. Follow these simple steps to ensure that your next transcript a usable one. While automatic transcription is not perfect, it can get pretty close to capturing everything that is said in a clearly recorded conversation. And for any errors that do come through, our editing tool is an easy way to make corrections, name speakers, and customize your format.

Get a Free Transcript