Recently, I came across this interesting article on Business Insider that spoke extensively on how the ChatGPT voice-to-text feature was much more impressive than any existing voice-to-text feature out there.
This article details how the AI transcribed summaries from a podcast, which easily could pass as something written by a human being. This is because OpenAI uses a technology called Whisper to power its ChatGPT voice-to-text feature. This got me thinking, how then, can you best use this feature? Here is how:
- Download and install the Voice Control for ChatGPT extension from the Chrome Web Store. When you download this extension, it lets you record your voice and submit it to ChatGPT. It also allows you to listen to responses aloud.
- Once downloaded, then open the ChatGPT app on your browser or device.
- Then, select the microphone button below the input field. With this selection, you activate the voice control feature, and you will be asked to allow microphone access. Click Allow.
- Say something like ‘Hey ChatGPT’ and begin to converse with the bot. You can then ask it questions or give commands.
- To stop this feature, click on the microphone again and say, ‘Bye, ChatGPT’.
Transcribing with Whisper AI
Alternatively, you can also use OpenAI Whisper to make transcriptions. Downloading Whisper is a longer process, but once downloaded, you can transcribe interviews, lectures, and podcasts or create written records of spoken content. Whisper is the x-factor in ChatGPT’s impressive voice-to-text performance.
With Whisper AI installed, here is how to then go about transcribing audio files:
- Open a command prompt in the direction containing your audio file.
- Then type whisper <filename> (then replace the <filename> with your audio file’s name), then click Enter. Watch as the AI transcribes the text.
- You can transcribe speech in other languages beyond English. You will specify the language using the –language option (CSS code). So, if you want to transcribe an audio file in French, you will go to the CSS code and type in whisper French.wav –language French.
And there you have it. With ChatGPT and Whisper, you will be using the most advanced and fastest voice-to-text feature that is currently out in the market.