Forum Discussion

ConsultingServi's avatar
ConsultingServi
Community Member
6 years ago

Text to Speech (TTS) question

My agency has traditionally used others to narrate content, but that has started to become prohibitive due to amount of time/cost of the employee recording, other pressing duties, and so on. I've been looking at TTS to supplement some or most of the narration points.  However, until Storyline's TTS becomes more natural and more adjustable, it is not an option for us.

I looked at over 20 other TTS so far. I have it down to four: Amazon Polly, Azure, IBM Watson, and Voicery. I haven't looked into Azure or IBM Watson much yet.

Do any of you have experience with one or some of these, or another suggested one...that is NOT one of the following:

Abodit, Acapela, AI Word Smith, Audiobookmaker, AX Semantics,  Cepstral, iSpeech, Ivona, Natural Reader, Panopretor, Quill, Read the Words, ReadSpeaker, SpeechKit, TextAloud, Textengine.io, TextSpeech Pro, TTS Reader

  • UlisesM's avatar
    UlisesM
    Community Member

    I find that some are better than others, but they will never replace live audio recording. In my experience (and this is just my experience), TTS never satisfies the audio needs I have for my courses.

    • WendyFarmer's avatar
      WendyFarmer
      Super Hero
      Ulises Musseb

      I find that some are better than others, but they will never replace live audio recording. In my experience (and this is just my experience), TTS never satisfies the audio needs I have for my courses.

      Agree not in the final production but I use it to allow reviewers to hear the narration and see the closed captions and find there is less rework than if they are just reading words on a storyboard

  • Thanks for the feedback. I currently use the Articulate TTS as that review part, and to gauge timing. So far it paces about 50-75% slower than an actual narrator, which benefits me in building courses.

    I completely get the points being made. I can go into a long explanation, but what it boils down to, is although TTS is not ideal, it is still primarily the only option on the table. My option is to abandon it and let someone else make the choice who may or may not do the same level of research, or provide a recommendation in hopes that I can choose the best one, and/or influence the decision further.

    Any answer to the original question would be greatly appreciated.

  • Bob_Mongiovi's avatar
    Bob_Mongiovi
    Community Member

    Not sure if this question is still relevant anymore, but if it is, have you looked into Google Text to Speech API?  You'll probably need someone who knows how to program, as there isn't a "tool" you can just download to create audio files.

    Here are the samples they provide.  The WaveNet ones are the more natural-sounding ones:
    https://cloud.google.com/text-to-speech/docs/voices