Live Captioning Cat Power's Sea of Love

Today I hit a new milestone with steno – I was able to do a live transcription of a song for the first time!

Check it out!


Live Captioning 'Sea of Love' with Steno (using Plover)

Yes, but how does steno work?

It guarantees speeds of up to 225 wpm (and higher!), it’s been around for what seems like ages, the courts have been using it to record trials verbatim in their entirety… how could something like that possibly work? It sounds really crazy when you think about it.

If you’ve only worked with qwerty (or even if you’ve taken up something like Dvorak or Colemak), prepare for some mind-benders; steno redefines a lot of what you might assume to be true about a keyboard. 😄

Steno keyboards come in different shapes and sizes, but the keyboard – ever since 1911 – has had this basic layout:

STPH * FPLTD  <----- top row
SKWR * RBGSZ  <----- bottom row
  AO   EU     <----- thumbs

So. How do you use this beast?

1. Your hands rest between the top and bottom row

Wait, what!? That’s right, your hands sit between the two rows, literally on the cracks of the keys.

(and, not to get ahead of ourselves too much, but one amazing implication of this is that on a steno keyboard your hands rarely move more than a half key away. The one exception to this being numbers.)

2. Multiple keys are pressed at the same time to create words

  • On a traditional keyboard, you type letters – one. at. a. time.
  • On a steno machine (as they’re called), you “chord” an entire word by pressing multiple keys down at the same time (much like you would press a chord on a piano)

Fun fact for this point: stenographers are said to “write” words, not type them.

3. The steno keyboard has layers

On a Qwerty keyboard, a given letter only has one meaning; the letter ‘t’ means one thing – the letter ‘t’.

On a steno machine, though, any given letter – when combined with other keys – has the potential to represent a number of different things.

If we look at the left side of the board, for example:

  • H can represent ‘H’ by itself
  • but if you happen to be pressing P and H together, then you actually get ‘M’
  • and if you press T, P, and H all at the same time, then you get ‘N’

So, the individual keys are really just the beginning. The combination of keys is the next step, and those combinations open up all kinds of possibilities.

As an exercise for the reader, consider how many combinations of letters are possible on a steno keyboard. 😉

Want to see the different layers actually used in steno? Check out the graphic at the top of this page. (And, yes, those all get used. Regularly, in fact.)

4. Steno is based on the way words sound (not the way they are written)

This is another really wild thing about steno; when you’re “writing” words in steno, you’re actually sounding them out, instead of attempting to write them the way they look.

The word “friend”, for example, is written phonetically as “FREND”, because that’s the way we say it. (I’m simplifying things a bit because you have to actually hit several keys to get the ‘f’ and the ‘n’ – per point three – but that’s how the word ends up getting sounded out after you take care of that part.)

5. While steno is mostly phonetic, it can bend the rules quite a bit when representing words

In general, steno represents a word by having:

  • the left side of the board give the opening consonant
  • the thumbs give the middle vowel
  • and the right side of the board give the final consonant

But many words don’t fit the consonant-vowel-consonant pattern. There are generally two ways to handle this. Either the word can be formed by making additional strokes or by getting a little clever and squeezing the word into a single stroke somehow.

“How can you squeeze a word into a single stroke”, you ask? Here are some tricks that I’ve seen so far:

  • Allow the word to be written out of order on the board. Normally, the word is “written across the board” left-to-right. This trick makes an exception, letting the word be written across the board out of order. (This is called an inversion).
  • Skip parts of a word that are pronounced more softly. Or, sometimes, drop a syllable altogether, even if it is a full fledged syllable.
  • Get downright dirty and map a word to just a handful of keys that are more of a nod to the actual word than any attempt to represent how it sounds. (SG being used for ‘something’, for example). This trick is especially common with the most frequently used words in English.

To summarize:

  • steno uses a condensed board where you can press multiple keys at the same time
  • groups of keys map to different sounds. All the sounds you will ever need are already tucked somewhere in the board
  • and, finally, there are all kinds of tricks that let you smash even the longest of words into a single stroke

There’s just one last thing that makes this all work.

You.

Well, okay, more specifically you, your brain, and frankly, a lot of practice. The board sets the stage for you to be able to drop your hands on the board and have an arbitrarily long word pop up on the screen. But, that’s only if you know the word and feel comfortable enough with the board to chord it at a reasonable speed.

What kind of practice is needed to come up to speed with steno? That, my friend, is a topic for another day.

STAEU TAOUPBD

Gearing up for steno

Okay, so, to recap… I went to PyCon, got to hang out with some of the realtime captioners, and got hooked on steno.

The conference was wrapping up. Where would I go from here?

Well, before leaving the conference, I asked Mirabai and some others if they thought this was something that people could actually pick up and use for coding. To be honest, I was a little scared about what the answer would be. Much to my amazement, the answer was an emphatic “Yes!” It was another unbelievable moment. It’s not every day that a pro tells you a new skill is just waiting for you.

If you’re interested, you can actually see for yourself what it looks like for a developer to code in steno – check it out: coding in steno.

Before I headed home, I was already looking into getting a steno machine that I could practice on. But that’s the thing: where do you go if you want to get started with steno?

Well, one fantastic resource is… the steno community! There’s a pretty active Discord server for Plover that hosts a rich community of makers, tinkerers, and hobbyists that have come up with all kinds of steno boards that you can use.

Here’s the board that I bought while still at the conference (that’s right; I couldn’t even wait until I got back home to buy a board 😄) – it’s called the Georgi, one of many killer boards by Jeremy Bernhardt:

stenoing at Coava

After a few months of practicing steno, I bought a second board that looked like something I could use while traveling – the TinyMod, a stellar board by Charley Shattuck:

TinyMod

Both boards have been awesome. Love ‘em both. 😎 Here are some things that I enjoy about each board:

Georgi:

  • keys are super light
  • split keyboard – gives you a lot of flexibility when trying to find an ergonomic setup
  • all around beautiful board (has an amazing silk screen on the back)
  • has a Qwerty mode if you want to switch back to a traditional board at any point!

TinyMod:

  • board is a single unit – makes it a lot easier to tote around when traveling
  • keys require more force to activate (makes the board a bit more forgiving if you like to rest your hands on the board)
  • another solid board with a beautiful design

Both boards have been vetted by a pro and received high marks! Here are links to both reviews:

For more information on boards that are supported by Plover check out this page.

Its pretty incredible to have sturdy, well-designed boards at reasonable prices come out of the maker community. I’m extremely grateful to have these available (Honestly, not sure where I’d even start looking to get a board if these weren’t around!) If you’ve been thinking about getting started with steno, I definitely recommend checking them out!