Podcast Transcription Software Review
When we started our SaaS Marketing Insights podcast we didn’t really think about the value of transcribing the audio into text. We assumed there was a service that could do it quickly and easily, and there is. In fact, there are dozens of them. But which service is the best for transcribing podcast audio? We decided to scratch our own itch and find out for ourselves.
Before you read on it’s worth mentioning that, as with all our other blog posts, there are no affiliate links here – our findings are based purely on our testing and own experience. We hope you find what we have learned useful. If you do, please feel free to share it or add a comment at the end of the article.
Podcast Transcriptions: why bother?
Transcribing audio podcasts benefits both visitors and search engines alike. Visitors can scan the text quickly, finding the information they’re looking for without having to listen. Search engines can index the transcription and make it accessible to searchers, bring new visitors to your site. You can also use excerpts of the transcription as part of your show promotion.
Podcast Transcription Testing Methodology
We took the first 10 minutes of audio from our podcast interview with Rand Fishkin and uploaded it to six leading online transcription services. We then exported the results and compared them to our own manually transcribed version. The excerpt featured both Rand’s voice (American English) and Paul, our CEO (British English) in a standard podcast interview format, originally recorded over a Zoom video call in MPEG4 audio format then edited and exported as an MP3 file. The audio quality wasn’t first class with one short section affected by a poor internet connection. As such, it is representative of real-life audio podcast quality.
Podcast Transcription Software Test Results
Based on feedback from podcasters, we shortlisted six solutions to test. Below is shown an example of our manually transcribed text for comparison with the example we provide for each solution:
Touted as an AI powered visual editor for audio transcription, Descript is the only solution featured here that requires you to download and install the software. In our tests Descript and Otter performed best out of all the automated software solutions.
Price: From 15c per minute, pay as you go.
Otter turns voice conversations into written notes. Featuring a web interface (Chrome preferred) Android and iOS apps. It also connects to your Zoom account to automatically transcribe video conversations. Industry leading performance and a freemium business model makes an irresistible combination.
Price: 600 minutes’ free transcription per month. Paid options from $8.33 per month (billed annually) for 100 hours per month.
Rev seems to be considered the gold standard for podcast transcription. Powered by real people rather than AI, the accuracy is excellent. However, this is reflected in the premium price. Turnaround is surprisingly reasonable though – 2 hours for our 10 minute sample file. In our tests Rev’s transcription was more accurate than any AI powered solution, with very few errors.
Price: From $1 per minute for human powered transcription.
Simon Says is aimed squarely at audio and video professionals requiring quick transcription. With support for transcribing multiple languages and a plethora of export options supporting many audio and video production software, this solution seems ready for every workflow. However, in our tests the end results were the least accurate, with some of the errors seeming weirdly comical!
Price: Starting from $15 per hour, pay as you go.
Temi is Rev’s AI powered transcription service, and uses the same interface to manage your transcriptions. It may be one tenth of the price of its human powered counterpart, Rev, however you pay for what you get and in our tests accuracy was not the best.
Price: 10c per minute. A free 45 minute trial is available.
Another AI based offering, Trint supports input in English and an amazing 27 other languages from a wide range of source file types. While Trint’s export options are second only to Simon Says, it’s Zapier integration opens up a world of possibilities. Which is just as well because it didn’t distinguish itself for accuracy in our tests. iPhone app also available.
Price: From $15 per hour, pay as you go.
Conclusions & Recommendations
If you want accuracy, and are happy to pay for it then currently Rev is the best of the solutions we tested. However, if you have little or no budget and are happy to spend time making minor corrections we found that both Otter and Descript deliver an equally high level of accuracy. However, they are still inferior to Rev.
|Cost per min||$0.15||$0.00||$1.00||$0.25||$0.10||$0.25|
(10 minute audio)
|2 mins||3 mins||120 mins||6 mins||5 mins||4 mins|
|Ease of Use|
So now we’ve scratched our own itch we have decided to transcribe the rest of our podcast episodes using Otter because of its Zoom integration and 10 hours of free audio transcription per month. Of course, your use case my vary from ours so it’s worthwhile researching a solution that fits your needs. Happy testing!
Last updated 1st August 2019.