How to set up your ElevenLabs voice in Predictable

Setting up your ElevenLabs voice in Predictable is quick and easy - it should take you less than a minute! To start, you'll need an ElevenLabs account, where you can either create custom voices or use pre-made ones available for download.

Step 1: Get Your ElevenLabs API Key

Log into your ElevenLabs account and head to the My Account section in the bottom left corner. From there, you'll find an option for API keys. Clicking on this will open a modal where you can create a new API key. Name it whatever you like and click Create API Key. Remember to copy the key, as it will only be shown once.

Step 2: Connect ElevenLabs to Predictable

Next, open Predictable and go to Settings > Speech, then select ElevenLabs. (If you don’t see this option, make sure you have the latest version of Predictable installed.) Paste in your API key from ElevenLabs and click Connect My Voices. Your ElevenLabs voices should now appear under My Online Voices, ready to be used in Predictable.

Using ElevenLabs voices offline in Predictable

ElevenLabs voices work online, meaning an internet connection is required to generate new speech. However, Predictable has features to enhance offline use:

1. Stored Phrases: Once you've spoken a phrase using ElevenLabs, it will be stored, allowing you to play it back later without an internet connection and with reduced latency.

2. Pre-download Phrases: You can pre-download all the phrases in your Grid by navigating to ElevenLabs settings within Predictable. This saves the phrases using the currently selected voice for offline use. Be mindful that this could use a large portion of your ElevenLabs quota, depending on the size of your Grid.

   
   

This setup allows you to use your custom voices effectively in Predictable, both online and offline.

Literate client with slow degenerative condition

High tech AAC and Low tech AAC outcome

Background

Dave has Parkinson’s disease, and he lives at home with his family. Dave reported that he struggles to communicate verbally and this can be frustrating. Dave’s speech is quiet and slurred, which requires careful listening by his conversation partners.

He does not have any form of AAC to help him communicate.

A scenario was given whereby Dave fell and he was trying to say something but he could not be understood.

Dave uses skype and telephone but his speech can be difficult to understand by others. Dave said that during group conversations he misses the opportunity to interject and join in, so he feels left out.

Assessment

Physical abilities

Dave was observed mobilising with a 4 wheeled walker, this had

a laser beam guide to reduce ‘freezing’/improve stride length when walking.

Dave is able to sit in midline unsupported. However, he cannot maintain this posture and tends to gradually fall forward.

Dave has some generalised weakness of his upper limbs. His movements are slowed with some reduced co-ordination. Dave has difficulties with activities that require fine motor skills. Dave is waiting for a powered wheelchair.

Dave is computer literate and knowledgeable in technology.

Dave no longer works due to his condition.

Sensory

No visual or hearing deficits observed or reported. Dave is able to read standard size text approx. 12 –14 points, at a near distance (approx. 30cm).

Cognition

Dave was able to attend in tasks and he showed some insight into his condition.

However, there were some minor cognitive difficulties noted during the intervention.

Language and Literacy

Dave is literate and has no language difficulties.

Dave tried 3 text based communication devices during the assessment. Dave

was able to use the Allora and Lightwriter, he benefited from the use of the keyguard for accurate targeting.

Dave was able to access the iPad by using a stylus, as when using his finger he would make mistakes (mis-hits). Dave expressed that he preferred using the iPad as he liked the easy selection from prediction, and the easy editing of the phrase bank on the AAC app Predictable.

Dave also quite liked the Allora.

Dave was given the opportunity to try the iPad with the Predictable App and the Allora during

a trial period.

Following the trial, Dave said that he preferred using the iPad. Dave was also shown another text-based app, Proloquo4text as it has other features to speed up communication. Paul had time to explore the app but preferred Predictable.

Outcome/ Recommendations

Plan and Recommendations: In the first instance the KM CAT team advised Dave’s speech therapist that Dave would benefit from a paper-based alphabet chart which could be used

via partner assisted scanning, as on some days, Dave would find it difficult to point to the letters.

Also for Dave to have a communication passport for when he goes into hospital etc.

Dave was issued with an iPad with Predictable app. Dave was also introduced to a FAB for days when his access is not so good and he struggles to use his iPad. Dave finds that the FAB helps him get his message across when he is feeling poorly and less able. The inbuilt keyguard on the FAB helps his targeting. Dave is able to use his AAC. However, he needs prompts to use them in some instances when his speech is unclear.

Head tracking vs. Eye gaze technology

Predictable now offers 3 new features that rely on Apple Face ID advanced technology:

• Head Tracking

• Scan & Track

• Gesture Speak

These new features aim to offer alternative ways of interacting with Predictable if you cannot rely on direct touch.

What are the advantages of using Head tracking instead of eye gaze technology?

- Predictable Head Tracking doesn't require any additional hardware or calibration. It is already in-built in your iPad / iPhone

- As Apple Face ID technology recognises any human face with two eyes, a nose, and a mouth, it allows many other ways of selecting keys on the screen (such as Mouth open, Tongue out, Cheek puff, etc.)

- It allows you to keep using your iPad as your assistive device, which is more socially accepted but also easier to carry compared to any other big hardware.

- Gesture Speak allows you to attach messages or actions to your different face gestures (e.g. you could have the message "Help me please" to be triggered every time you frown and the Speak button to be triggered every time you smile)

- Scan & Track allows you to combine different scanning and selective methods. If you cannot rely on your head movements to point a key on the screen, you can turn the auto-scan on and keep using "Blink" to trigger a selection.

Find out more about the selection triggers available in Predictable 6: