WellSaid Labs Text-to-speech

TTS for eLearning: What Makes a Good Text-To-Speech Service?


Text-to-speech is the instructional designer’s new ally.

By increasing flexibility and accelerating time to market, TTS enables eLearning course creators to produce modules effectively. If you create training or eLearning content at scale, you can much more easily achieve your goal—educating an audience—by using the right text-to-speech. 

But, what makes one text-to-speech service better than the next? How do we, as content creators, make sure we are selecting the right TTS technology for our projects? What are the costs and constraints associated with producing content this way?

These are questions that we get from our Studio users and that we ask ourselves when thinking about our own productions.

To help answer these questions, evaluate your TTS based on the following criteria.

A Good TTS for eLearning Is:

Intelligible

It transmits the educational value the student or trainee needs.

Usable and reliable

Strong TTS for eLearning is intuitive, modern, and doesn’t require a complex technical configuration.

Cost effective

Look for options that significantly reduce production costs.

Several TTS options are available today. Each provides a different balance of pros and cons that will influence your selection process. WellSaid, for example, focuses on producing the most human-like synthetic voice. Others might optimize for mass processing of text with minimal attention to how it sounds, as long as “car” sounds sort of like /kär/

Can TTS for eLearning measure up to these criteria? We put it to the test and ran the following passage on multiple text-to-speech services:

That’s one small step for a man, one giant leap for mankind. Houston, Tranquillity Base here. The Eagle has landed.

Neil Armstrong

Here is how it sounds when one of our voices says it: 

This is with no text manipulation and no post-production editing. That’s it. No tweaking of pitch, cadence, emphasis, or any other sound engineering.

Pretty good so far.

Let’s up the ante a bit. Remember, we’re looking for the best text-to-speech service for eLearning: intelligible, usable and reliable, and cost effective. Let’s hear what text-to-speech can do against a training script excerpt used by real voice actors over at Edge Studio.

When we ran the same quote on other providers, we found they offered faster rendering and less realistic voices. None provided that life-like feeling we’ve worked so hard to replicate. This is of extreme importance for eLeaning courses; a monotonous, robotic-sounding voice will put students to sleep faster than chamomile tea after a hot shower.

TTS for eLearning Voice Training Script 1

 A Haunting Title

In this world, there is real evil. In the darkest shadows and in the most ordinary places. These are the true stories of the innocent and the unimaginable. Between the world we see and the things we fear.

eLearning Voice Training Script 2

A story

Every moment has a story. And every story matters. The first men painted stories on stone walls, the ancient Egyptians chose the chisel instead. The Incans told story with dances and fire, aborigines told it with star and spear.

eLearning Voice Training Script 3

Abiogenic Theory

The abiogenic theory holds that hydrocarbons were a component of the material that formed the earth, through accretion of solids, some 4.5 billion years ago. 

These are very tough scripts for a synthetic voice to master. I will leave it up to you to assess if the quality criteria was met.

One more thing: I couldn’t let this opportunity pass. Let’s try a tongue twister, just because 😃.

Bettie Botter bought some butter
But she said this butter’s bitter
If I put it in my batter
It will make my batter bitter
So she bought some better butter
Put it in her bitter batter
And it made her bitter batter better

Cool stuff, right?

When evaluating a text-to-speech service for eLearning (or TTS for eLearning), you want to know it can help you achieve your main goal: to clearly communicate your core content. In addition, it must at least meet your learner’s expectations. While some audiences of an eLearning module are required to complete the course, some are in ecosystems more sensitive to user feedback. If the completion rate of a module is substantially lower because of the quality of the narration, this can have a significant impact on your business.

You have something important to tell, and your students have an opportunity to learn. Let’s make learning engaging and consider the ultimate experience you are providing to them. 

That’s the ultimate goal: to succeed in educating.

READ MORE: Text-to-Speech, or TTS for Customer Service

Credits

Photo by Luca Bravo on Unsplash
Music by bensound

Tags: 

TOPICS