I have been asking this question for years at conferences – what do we need to know for 1| great text to speech auditions and 2| how to give the best performance if we do book them? No one could give me a good answer. The best I got was – be consistent. Mike and I have both been in the finals for some big text to speech jobs and I booked a small one overseas. Still I was unsure how to give a great performance…until NOW!
I’ve been working with a text to speech company and the program manager I’ve been working with was so gracious to share with me WHAT TO KNOW FOR A TEXT TO SPEECH JOB! These tips will help you in auditions and once you book one.
How the Text to Speech Process Works
First of all, we need to understand how it works. The modern way companies create synthetic voices is by using neural networksmodeled on the human brain. You may have heard of neural pathways. These are like a road map for our brain to make sense of things.
When you record audio for a synthetic voice, a machine creates a map based on your audio and the script you read. It begins to learn how the sounds and words connect to create a pathway for future input.
Why is this important? Because it leads to the first thing to understand about recording a text to speech job…
WHAT TO KNOW:
1| Whatever style you record the script in, will be the style the machine creates. Did you do it all super upbeat? That’s how the machine will map your voice.
2| Because of #1, it would be a good idea to determine the purpose/goal of the text to speech/synthetic voice project you’re doing and choose a specific delivery based on that. Will it be used for e-learning? A brand voice like Alexa or Siri? Deliver the lines not on the script but on the purpose/goal.
3| Pacing – for a TTS project, you want to aim for an audiobook narration-type pace. Not too slow and not too fast. You need to be heard clearly.
4| Consistency is key (like I was told at conferences) but let me go a little deeper. This is what allows the machine to find the map between the audio and text easier. That’s WHY it’s important. A further tip for helping with consistency is – TTS scripts are long and time-consuming. Do not try and get it done quickly. You need to space the recording out over days so your voice doesn’t fatigue. When it’s fatigued, you can hear it in the recording and that affects the final outcome. Keeping your voice healthy helps with consistency. I will also add that TTS scripts get very monotonous. You’re doing the same thing over and over again. Just know that it’s part of it and do what you can to keep the same energy and pace throughout.
5| Clearly enunciate the words. This is key in the machine being able to create a map.
6| TTS is interested in unique, authentic voices…not voices mimicking robots. Many people send in TTS auditions and think they need to sound like robots. Give a natural delivery that allows your authentic unique voice to come through.
7| Audio quality is EXTREMELY important for TTS jobs. The audio needs to be clean, free of mouth clicks and outside noise. The machines can often pick up things we can barely hear. If you’re recording a TTS job and don’t have a quality set-up at home, it’s best to book time at a studio, like Atlanta Voiceover Studio.
8| Pronunciations – understand the type of pronunciations wanted – are they general American? More global? Make sure you’re pronouncing the words correctly.
One more tip would be to ask the company how they prefer the files to be delivered: individually (each prompt)? in one file labeled? any specific audio settings?
I think TTS jobs can be a great thing for you as a voice talent BUT you MUST be informed beforehand. It’s important to make sure the rate is correct (check Gravy for the Brain for a great TTS rate guide), the company is ethical (are they part of Open Voice Network?) and that you are protected (check Nava’s AI Voice contract).