Transcribing could about to be a lot easier to do, thanks to an app in the form of MacWhisper, available on macOS.
Transcription services such as Otter and Transcribe enable you to convert audio files to text, so you can add it to a project or an interview to read back on.
Developed by Jordi Bruin, it's free to download, but there's also a Pro version available for $13 / £11 which enables faster transcribing. You can drop an MP3, MP4, WAV, or M4A file into the app, and by using OpenAI, you’re shown a window that displays the whole transcription, and you can edit parts of it if the app has spelled certain things wrong.
Bruin recently brought out version 2 which reduces the size of the app from 4GB to 8MB, and lets you drag and drop files directly from Apple’s Voice Memos app. So if you've used this on an iPhone to record an interview, for example, you can easily obtain a transcription on your Mac soon after.
I've been podcasting for years, and trying to transcribe episodes has always been time-consuming in making sure everything was correct. However, it's something that's always been important to me, as it can help someone with a hearing disability.
With this in mind, I used MacWhisper 2.0 for a recent episode of the iMore Show to see how well it transcribed what, Karen, Stephen, and I spoke about for an hour. I also took some time to talk to Bruin about how AI could be used as a force for good, as MacWhisper has.
Transcribing with ease
The episode that aired on February 19 was 62 minutes long, but it took MacWhisper just 10 minutes to transcribe. I was able to edit certain parts to replace 'IMoar' with 'iMore', while my name had an additional 'r' which was easily fixable, and I could then export it as a subtitles file, or a document.
In the grand scope of this, it was impressive, and a far cry from me manually transcribing podcasts and interviews back in 2020. I was able to scroll down to the 42-minute mark to find where we were giving our impressions on the Tetris trailer that debuted earlier in the week, for example, so I could go to another topic we were chatting about directly without scrubbing the timeline to aimlessly find it.
Speaking to Bruin, he's hopeful that apps like MacWhisper showcase how AI can be used for good. "I don’t think most people realize that something such as Whisper is also based on similar technology that allows things such as GPT to work," Bruin explains. "While Whisper and Large Language Models are different, they both build on the advancements in AI over the last few years. To me, Whisper really shows that all these advancements can be used in a number of ways we’ve never even considered."
However, accessibility could be the big winner here. AI could allow someone with visual or hearing impairments to help them enjoy podcasts and videos on YouTube for example. We asked Bruin if he's hopeful too that other apps like MacWhisper could take advantage of AI for needs like these. "I hope that AI will make it easier for developers to come up with innovative ways to solve accessibility challenges. Transcriptions for video and audio content are a very obvious improvement, but I’m also looking forward to seeing how AI can make complex computer interactions simpler for people with limited motor skills," Bruin explains.
The next frontier for accessibility?
AI could reach a point where it can generate a person providing sign language for any video for example, or it could work with a Braille Embosser to convert text, podcasts, and video into creating tactile dots for users.
"Having an AI that is trained on the specific movements that a person can do comfortably, to then translate them to complex (sets of) interactions would have a huge impact for a lot of people," Bruin continues. "My main takeaway is that as these complex technologies become more accessible to more developers and users, more solutions can be thought up together with people who need it most."
Bruin has other apps that take advantage of AI, such as TextAssistant which lets you use the technology for certain prompts and requests. From language translation, and simple explainers, to converting code to other languages and more.
However, MacWhisper seems like it could benefit users in ways other apps and AI services can't, and Bruin's not done. "While my main focus is on adding small quality-of-life improvements and features over the next few weeks. MacWhisper 3.0 will probably focus heavily on speaker detection and improved export options that are more customizable," Bruin reveals. "I want to release an iOS app at a later point but I'm going to have to think about how people would use it in that context. I just added a roadmap to the app where users can vote on their favorite features, so that should help me narrow down what I’ll be adding next!"
While MacWhisper is relatively new, it opens up a bunch of opportunities - not just for accessibility, but for students when creating reports, or when you want to watch something with subtitles on. There’s so much potential for AI to be a tool for everyone, and it looks like developers like Bruin are just getting started.
