Blogging with Speech to Text
Having random ideas
I often get a little bored while driving to and from work. Most days I either listen to music or podcasts. Then there are days when I don’t want to listen to either, and my mind comes up with some interesting, and for the most part, bizarre ideas. So on the drive to work this morning (11/11/2019), I was thinking to myself “I wonder if I could record myself talking about topics and get it transcribed into text to then use on the site as an article”. Which at first seemed a bit weird, but who hasn’t sat in a moving car talking to themselves 😆.
So on my return from work I decided to give it a go and see what would happen. Below is an unedited version of what happens when a South African born engineer tries talking in his best english, to then get transcribed back into United Kingdom English. I recorded myself using iOS Voice Memos and uploaded the file to AWS Transcribe to output the text below. To be fair its actually not that bad.
So what come out
So this’s gonna be interesting. I had this really random idea that seeing I spend quite a bit of time driving to and from the office, that I could potentially use that time to write drug articles. But how can you write and drive? You might see. Well, it appears, have this miraculous little thing. And this little tool called a ws transcribe, which allows you descend in voice or video any form off audio based file into their system, and they will generate text off the back of it. So how does this help me? So if I certainly come up with an idea while I’m driving home like I am right now, I can start a voice memo on my phone, for instance, and generating audio half. That’s I could then add End of is transcribed and generate my HTML and always use on the rock. Easy. No. So this is a very simple tutorial. Not even a tutorial should say It’s a very simple example of me using that This is unedited. I have used it directing from the voice recording straight to transcribe straight into world. So hopefully that’s a work, and we won’t get any characters or any form of Wade swearing in between, I hopefully were in the future. Utilise this a little bit more work correctly. If not, I will find another way to do it. Until then, happy days.
Where to go from here
Although the first attempt wasn’t perfect, it did get me an article fairly quickly. So if I was to use this as a proper method of “writing” future articles I would need to speak a little more clearly and then take the time to edit the output from AWS Transcribe. Maybe something to work on in the near future 🤓