FREE – Auto-Transcribing an Mp3 using Youtube

In my previous blog post I discussed how to create a Mp3 recording from a VoIP call over the internet. This blog will discuss the concept I’ve been asked the most questions about:

Wait. Auto-transcribe… for free. How did you manage that?

Well, gentle reader, It’s not the easiest thing to imagine doing.

There are a lot of pay-services for transcription who will do a much better job. However, if you’re like me: Strapped for cash and required to do you own transcription, you can use this trick.

Stage 1: Setup

First you need to make sure you have acquired/done a few things. Click the hyperlinks to find out more.

  1. Made an Mp3 recording of your interview.
  2. Decided on the transcription style you want to use. – As a note this method wont work for verbatim style transcription as YouTube edits ‘Um’s’ etc. out.
  3. Registered and Activated a YouTube account – activation is not necessary for transcription’s under 15 mins but I had an hour long interview so…
  4. Familiarised yourself with YouTube’s privacy settings lest you broadcast a private interview to the masses.
  5. Worked out how to turn your Mp3 into an Mp4 – I used this handy converter but I’ve seen anything from using Windows Movie maker to coding the FFMPEG facility on your desktop. In short; you take an image and merge it with your Mp3 to make an Mp4. The format that YouTube uses.
  6. My post-YouTube method uses Word and Excel but I imagine you can do some similar tricks with open office.

Stage 2: YouTube

If, like me, you are a novice to YouTube here’s some pointers on how to upload your video. This is also a good point to check your privacy settings are the correct ones. Avoiding the aforementioned broadcasting issue.

Once you’re in the creator studio things look a bit like this:

Screencapture of YouTube creator studio

You’ll notice my Mp3 interviews uploaded in the centre square, if you select one, YouTube takes you to a screen that look like this:

Screencapture of YouTube creator studio
Select the Subtitles/CC option

Note my handy red arrow, select the Subtitles/CC option from this screen and it will open the captions.

NOTE: If your auto-captions aren’t there. Don’t panic! Go make a cup of tea and come back and check again, sometimes it takes a little while to crunch your captions.

You are looking for the ‘English (Automatic)’ option (below) click on it.

Just to give you an idea, this auto-caption took about 20 mins to turn up and it’s a short video.

Select the ‘Edit’ button  and your screen will allow you to edit the Auto-transcribed captions.

Select the ‘edit’ option to activate the editing aspect.

I personally like to select the ‘pause when editing’ function at the bottom as it makes things easier.

Caption proofreading and edit
Make sure the ‘Pause video while typing’ option is selected

Once you’re set up you can select the dialog on the left and edit (remember to save regularly in case of internet issues – learn from my mistake).

YouTube’s auto-transcriptions are pretty good (thanks to our Google overlords) but they’re not perfect. They lack any punctuation/grammar and sometimes create hilarious mistakes. So it’s important that you edit.

TIP: I put an ! to note a change in speaker. This is a VERY good idea for later when you need to separate things out in stage 3.

Once you’ve edited. Congratulations you have completed stage two!

Stage 3: Extracting and formatting

So you’ve created your Mp4 and edited the auto-captions on YouTube. Now for the best bit, turning your work into an interview transcript that works. There is no way to do this without proofreading. If you want to trust the computer you can but, I think it’s a bit reckless.

Download your captions.

Screencapture of Downloading captions
The format doesn’t matter as the next bit gets rid of this anyway.

This file will not open! You need to open notepad and drag and drop the file in. At which point you get something that looks like this:

Image of notepad with captions
Please excuse my nonsense speech. 🙂

Open excel and copy this in once you get that you can do some quick editing hacks. Now you can just edit this by hand but if you have a huge document like I did this will speed things up.

  1. Using wildcards (in Excel this is *) like this: *:*:*.*,*:*:*.* find and replace (Ctrl H) the time signatures with a blank cell.
  2. Remove any blank cells using this handy trick.
  3. Highlight the cells remaining and transpose them into another worksheet allowing you to paste into word without applying a strange format.
  4. Paste into word using the ‘keep text only‘ option to remove any lingering formatting.
  5. Find and Replace the ! with a paragraph mark (copy this: ^p into your ‘replace’ field).
  6. Copy newly paragraphed document BACK into excel.
  7. At this point you may need to TRIM your speech section as there will be a space at the front.
  8. Add another column and assign a speaker to each section – In my case it was an interview between two people so I just dragged it down using excels Autofill function.
  9. Paste this back into a word table.
  10. Proofread the result to check each section makes sense and the correct people are highlighted as speaking.

Voila. You have a transcription that you didn’t have to type up one word at a time.

Now I’m not going to claim that this technique is easy, but once you get the hang of it you can get long transcriptions done very fast. I did an hour long interview in two hours from interview end to final transcription, including compiling time for YouTube and converting from Mp3/Mp4.

Got a quicker method? Found a spelling/grammar/general mistake in this post? Please comment and let me know. I also respond well to gratuitous praise about how this made your dissertation/assignment easier (I know it made mine MUCH easier).

Next up: Library lists on Twitter

Sarah out.



TUTORIAL: Interviewing and Recording people over the internet

Greetings reader! Welcome to my blog. This is my first post addressing some of the prickly issues I turned up when doing my dissertation.

First question I kept getting was

How did you manage to interview people in Orkney? Especially when they couldn’t use Skype?

Well gentle reader. I used this marvellous tool called the internet! I joke, I joke (or at least I hope so, if you don’t know what the internet is then what are you doing here?)

Here’s a few quick pointers for how I went about interviewing and recording people online. Rather than write a complete Tutorial I thought I’d link to the people I found while doing my research as they have more tech skills than I do!

1. Get the hardware

I have a pretty impressive desktop (no I don’t know the specs, I’m a librarian, I had a friend build it for me) so this technique may kill low-end laptops. For which I apologise.


Get a headset with mic. I went to my local Claus Olson and grappled a cheap headset after two weeks worth of failed tests. It made everything so much easier.

2. Can’t use Skype? Find another tool!

Many of the libraries I contacted straight-out told me they couldn’t get Skype through their firewall. So I had to think outside the traditional contact box.

After a brief session of GoogleFu, I found AppearIn.

AppearIn Homepage

It had several advantages:

  1. Free
  2. Easy to share with the interviewee
  3. Private (you can ‘lock’ rooms once you’ve claimed them)
  4. Easy to use (no download as it runs through the browser or even your mobile!)
  5. Limited ‘chat’ function (handy for when the voice bit goes wrong!)

I used it with little-to-no issues.

3. That’s great, but how do I record it?

Many readers will know that you can record Skype calls. However, I had to create some tech workarounds for my interviews. I discovered the technical term for what I was doing was Voice Over Internet Protocol (“VoIP”) calls. Which helped make my GoogleFu stronger when researching.

If you want to follow what I did you’ll need these two tools:



(Download at: you’ll also need the digital  cable too)

A donation-ware programme

Rather than try to explain what I did I’ll link to the Youtube video (warning: he occasionally swears) I used. In short it links the mic back round to the headphones. This DOES mean you can hear yourself (note my comments about a headset – it just plain doesn’t work with speakers) but it means you can also record yourself.


(Download at:

Image of Audacity
An open source sound editing tool

Once you’ve done all the finangling with Soundmeeter all you have to do is hit the ‘record’ button on Audacity and it records. If it’s not working check that the input is the correct thing (I found it sometimes reverted and ignored the audio coming through Soundmeeter)

I used their MP3 function to export and save my audio as an MP3 which then allowed me to auto-transcribe it – the subject for the next blogpost.

There we have it, this is my Frankenstein’s monster version of online recording, I’m sure there’s slicker ways to do it, but this worked for non-tech me so…

Hope all these links help out, do contact me if you notice any issues/have any more tips!

Sarah Out