Over the past few years, I have found that converting speech to text has a lot of value.
Speech recorded or otherwise is not very searchable without a transcript.
In 2016, I created three other podcast series. As an experiment, I wanted to see which was more valuable. Audio alone or audio with transcripts. Guess which one was viewed more? Hint: the series with transcripts was viewed/listened to a lot more.
People value the transcripts. Enough to make an eBook out of them.
I have used a variety of tools for transcripts.
I tried Dragon Speak Naturally by Nuance a while back which is trained to your voice, but I kept forgetting I had it available to me. Now there is similar technology on my laptop (when connected to the web).
If you prefer to have your voice transcribed, most Apple, Chrome, and Windows-based computers now have the option of activating speech-to-text (aka dictation) where you turn on this feature (with a hotkey) and start talking as you watch your words appear on your page. If you can talk, your computer can write for you quite literally. Helps for brainstorming and writing any stream of consciousness. Some tools work better than others. Just don’t forget to edit the text later since we speak differently than we read.
When it comes to transcribing audio recordings, I used one online vendor for a successful Kickstarter project. This same vendor did a great job back then. However, after while they began to deliver slower (1 week+), less consistently and results were not as accurate as I had previously seen. This did not just happen once or twice. This caused me to look elsewhere within the free market for similar transcription services.
No transcripts are 100% accurate due to nuances in language and even audio recordings, so you need to check the transcripts received against the audio you sent. I do.
In 2016, I started to use Rev and now get transcripts for my audio recordings in less than 12 hours.
I also found a few contractors from Upwork who do transcription well but are much slower which is okay when it is not time-sensitive, but you have to hunt for them because they get a lot of work from around the world.
In 2018, I started using Temi to get transcripts from my audio recordings within a few minutes. Yes, that’s right…transcripts in minutes, not hours nor days. Faster and less expensive. Game changer! A 16-minute audio file gets transcribed in less than half that time thanks to machine-generated transcription which seems to improve every time I use it. Temi costs a fraction of the price of human-generated transcription. Accuracy is very close to human-generated transcription, however, I review every transcript I get back from humans or machines since there are specific words and context that most would miss if I did not.
Now back to creating more eBooks based on audio interviews recorded and transcribed.