Transcription from audio data. The more professional it is, the more time and effort it takes. The traditional method is to input text little by little while listening to audio recorded with an IC recorder , but it takes a lot of time.
But humans are evolving creatures. A method for automatically transcribing audio data using Google Docs has been devised! So in this article, we will explain how to do that.
How to do automatic transcription in Windows 10
What you need to prepare is audio data, a PC (both Windows and Mac), and Google Docs. Just a few more settings. With just this, you will be able to automatically transcribe from audio data.
First, I will explain how to auto-transcribe in Windows 10.
Change sound input and output settings
　Open the “Sound” control panel and open the “Recording” tab. Right-click on “Stereo Mixer” and set it to “Default Device “. This is the only setting in Windows 10.
If you don’t see Sound Mixer
Download and install the application from the “Virtual Audio Cable” site .
Depending on the PC you are using, “Sound Mixer” may not be displayed. Then install a third party app . The app to use is ” Virtual Audio Cable “. Install this in advance. Note that you need administrator privileges during installation, so right-click the installation file and select “Run as administrator”.
Change the input device to “CABLE Output (VB-Audio Virtual Cable)”.
Then, open “Sound” from “Windows Settings” and change “Input device ” to “CABLE Output (VB-Audio Virtual Cable)” in the “Input” column. Preparations are now complete.
Voice Typing in Google Docs
Enable voice input in Google Docs and click the speaker icon.
Next, open Google Docs and select Tools -> Dictation. Then a speaker icon will appear on the screen, so click it to start voice input.
Play an audio file and Google Docs automatically recognizes and transcribes the audio.
If you play the audio file in this state, Google Docs will automatically start voice input and the voice will be converted to text .
If the recording condition is good, it can be converted into text with high precision , but manual correction is essential.
Here is the result of transcribing the sentence I read alone in the room. I understand the general meaning, but the details need to be corrected. Also, for audio files with loud surrounding sounds, such as seminars, the accuracy is considerably lower, so the results tend to be greatly influenced by the recording conditions .
