Reliable verbatim Speech to Text Reporting (STTR) for live events

TLDR: If you are looking for accurate and quality live captioning for events and conferences then you NEED to request NRCPD registered 'Speech to Text Reporters (STTR)' formerly known as palantypists.

Please do not use 'respeaking' it is not quick enough, accurate enough or fit for live captioning work.

(This blog originally written in 2018, updated in December 2020)

Skip to how to book genuine STTR.

My focus on captioning for deaf people

While I know captioning is useful for many groups, I speak from a deaf perspective. Deaf people may not have had access to the audio, or only partial access. If there are errors, deaf people may not notice or realise they are there and will come away with an incorrect understanding of what was said.

What is the difference between Speech to Text Reporting (STTR) and Respeaking for captions

Speech to Text Reporting (STTR)

Speech to Text Reporters known as STTR operators use a specialist keyboard (palantype or stenograph) and software so they can input words phonetically and have them show up as actual words on a screen. STTR ops can produce 200-300 words per minute at 98% accuracy. Slightly lower accuracy rates may seem reasonable, but at more than 2-3% errors the intelligibility of the captions decreases and the risk of errors changing meaning increases a lot. Errors are usually very obvious and occur most in people's names and specialist words. Providing STTR operators with preparation materials in advance which include people's names and specialist words reduces the error rate and increases their speed considerably.

STTR ops are registered with organisations like NRCPD (National Registers of Communication Professionals working with Deaf and Deafblind people) who require members to provide proof of their qualifications, undergo training including in ethics of communication support for deaf people and have a have a code of conduct/ethics as well as other things.

Respeaking

Respeakers work in pairs with one respeaker listening to what is said and re-speaking it into specialist software which they have trained to their speech, and the second respeaker going back to correct errors. Every 15 minutes they switch roles to rest their voices.

The accuracy rate of respeaking is lower and errors are less obvious as they tend to just miss out words entirely or substitute the correct word with another word which may not obviously look out of place. Respeakers can't go as fast as 200 words per minute. The slowness means they lag more and more behind the speaker which makes it harder for deaf users to follow. Respeaking often ends up (without indicating they're doing it) skipping entire sentences or paragraphs when they get too far behind the speaker and try to catch up.

Respeakers have no professional registration and no guarantee of being trained about the importance of their role to represent speech correctly for those who otherwise do not have access to the audio. Respeaking is better used for captioning non-live content where someone can listen and correct the captions better afterwards and speed is less important.

Why am I writing about live captioning?

Today I tried to access a livestream of a disability conference about ableism (disablism) in academia. I saw the caption provider and my heart sank...

Respeaking...

This article about respeaking covers many of the issues

Respeaking errors

I listened carefully to 5 minutes of audio and even I as a partially deaf person picked up multiple errors including:

• Entire sentences which I half-heard didn't show up at all in the captions.


• There were several occasions where single words were just missed out.


• At one point the word "sensitive" was captioned as "intensive". That's a massive change to the meaning of what the speaker was saying.


• Another error was an academic's 'principle' "Frank's remission society" was mis-captioned as "part of the remission society" removing the academic's name and work which was effectively an inline citation which was a critical shortcut to the meaning.

Complaining to the conference organisers

I tweeted some complaints about the quality of the captioning and it turns out the conference organisers weren't happy either! They had provided full scripts in advance which would definitely have helped STTR operators and should have helped respeakers.

Of course by the time issues are detected on the day, it's too late. The conference organisers are unhappy. The people relying on captions are unhappy.

Conference organisers don't know the field of "live captioning" support" so are at risk of being scammed or ripped off by agencies which only actually provide poor quality captions. Cost isn't even an issue here as respeaking costs the same or often more than genuine STTR.

How can I book reliable quality STTR?

STTR can be provided in-person or remotely. In person is best for live conferences and often not more expensive for half or whole day events. For anything lasting more than an hour or so, two STTR operators will be needed "co working" because of the demands of the work.

Remote can work well for webinars and very short events but is less good for on the ground events as they lose visual context. There are risks around managing technology to transmit the event audio to the captioner and maintain suitable internet connectivity.

There are two main ways of booking STTR (in person, or remote).

Go directly to STTR Operators

The Association of Verbatim Speech to Text Reporters (AVSTTR) (pronounced Av-Ster) website has options to
1) "Book a reporter" which puts out a general enquiry to everyone in their organisation or
2) "Our reporters" which shows individuals and where they are based so you can contact them directly and see if they are available for your event.

AVSTTR members are all registered with NRCPD and reliable. STTR operators prefer to be contacted directly and you may be able to get better rates.

AVSTTR website contains lots of useful information about STTR as a service.

Go to an agency for STTR operators

Often organisations have procurement rules or it can be easier to use an agency.

However BEWARE some agencies may appear professional, experienced and competent, but in practice frequently provide a very poor captioning service like happened at the conference above. One agency I will not use is AiMedia - who wrecked the above conference and at least 3 other events by sneaking in respeaking instead of STTR; sometimes even when STTR was actually requested.

If I use an agency, I ALWAYS go to either:

121Captions who are deaf-led by Tina Lannin who uses STTR herself and have served me well for over 7 years now.

or

MyClearText who are STTR-operator "writer" led by people who used to do decent captioning for the BBC.

121Captions and MyClearText websites contain lots of useful information about STTR/live captioning as a service.

There is a great film by some American STTRs (called CART) about the technology and STTR writer expertise behind everything they do. Audio description due soon and will be linked when available.

Go back to top of this blogpost.