SWET Toolbox: Two Apps to Help You Transcribe Audio Files

Reviewed by Winifred A. Bird

As a journalist, I love having transcribed versions of important interviews on file. I find scanning a Word document for key information much easier than delving back into an old audio file, and I like being able to mark up the text and return to it months or even years later. I’m much less enthusiastic about the time-consuming process of producing these transcripts. Fortunately, while computers are not yet able to take a recording of a natural conversation and turn it into an accurate typed document, many tools exist to make the job go faster. In this column I review two: oTranscribe, a simple, free transcription app, and the similarly named Transcribe, a fancier piece of software that costs 20 USD per year.

To use this program, simply go to the oTranscribe website, upload your audio file, and start typing into the space provided (you may need to convert your audio file to a type of file that oTranscribe can handle). There’s no need to log in or buy any software, and after you have used the program once, you can access it even when you are not connected to the internet—just open up your browser and type in the Web address. Once you’ve finished transcribing, download the file to your computer as a plain text file (you can convert it to an MS Word file later, although having to perform this extra step is slightly annoying). The main benefit of using oTranscribe instead of the default audio player on your computer, such as iTunes, is that you can stop, start, rewind, fast-forward, and change the speed of the audio file by using “hot keys” on your keyboard such as F1 and the escape key. In other words, instead of pausing to click over to the audio player with your mouse in order to stop or rewind, you simply shoot a finger up to the hot key and keep typing virtually uninterrupted. This saves a lot of time. You can also insert timestamps in the transcript that tell you where the audio and typed versions match up. That’s about all oTranscribe offers: the basics, done well, with no frills.

If you’re willing to pay a modest sum, you can get a few more features, some of which are truly wonderful. Transcribe provides everything that oTranscribe does, plus the ability to transcribe with a foot pedal, and the option to dictate your transcription instead of typing it.

I’ll explain dictation later—first a note about foot pedals. These are external pieces of equipment that cost between 50 and 100 USD and plug into your computer, allowing you to stop, start, rewind, and fast forward using your foot instead of the hot keys. I bought one of these several years back for a large transcription project but never got hooked; the hot keys seemed just as easy to use, without the fuss of the extra equipment. This seems to be a question of personal preference, however, since lots of people swear by foot pedals.

The more exciting feature of Transcribe is its dictation capability. It turns out that while computers do a pitiful job of automatically transcribing a real conversation, they’re perfectly capable of transcribing a clearly spoken dictation. Here’s what I mean:

Typed transcription:
Me: Thank you, great, and about how long do you have to talk this morning?

Interviewee: I can talk for about a half hour this morning and then I can certainly do things over email ad nauseam later this week.

Automatic transcription:
great and about how long do you have to talk this morning I can watch her this morning and then I can put my Ed… [at this point, the computer appears to have given up]

Dictated transcription:
thank you great and about how long do you have to talk this morning I can talk for about a half hour this morning and then I can certainly do things over email at nauseum and talk again later this week

As you can see, the dictated transcription is perfect except for the spelling of “ad nauseam” (which I admit I had to look up myself) and some missing punctuation. To use this feature, you simply plug a pair of headphones into the computer and listen to the original file, repeating back what you hear into the computer’s microphone as you go. The program automatically types whatever you say. In order to save significant time compared to conventional transcribing, you need to speak as you listen, rather than pausing the recording to repeat back a phrase at a time. This is easier than it sounds. I slowed down the speed of the recording to 80 percent and after about ten minutes of practice, I was sailing along, transcribing my audio with hardly a pause. I did have to go back to fix some punctuation and add spaces here and there, but overall, the process was far faster than conventional transcription.

Amazingly, the program provides dictation capability in Japanese as well as several other languages. The drawback is that the software doesn’t always choose the correct kanji, and going back to fix these mistakes takes so long it renders the program almost useless. Here’s an example:

Original sentence

Dictated sentence
日本人はキリスト教やイスラム教のように述べたいの紙ではなく 自然万物のあらゆるものに神を見出してきました

Those of you with a native or nearly native Japanese accent may have better luck than me (don’t ask me how 唯一絶対の神 became 述べたいの紙!!). I could imagine using the dictation feature to listen to a Japanese recording and take down notes or a rough translation in English – something I often do to avoid the particularly onerous task of transcribing in a foreign language. Using the time-stamp feature would allow me to go back later and get a precise transcription of the quotes I planned to use.

My verdict: If you transcribe regularly, sign up for Transcribe’s free seven-day trial to see if you like the dictation process better than conventional transcription. The program’s small cost could end up saving you many, many hours of work. On the other hand, if you only transcribe occasionally, oTranscribe is probably adequate. 

Please contact Winnie at winifredabird[at]yahoo.co.jp to suggest items for review, comment on the reviews published here, or volunteer to be a guest reviewer.