Support SSML

Follow

Roger Kibbe

• January 14, 2019 22:32

Support SSML - it is widely used in voice apps to enhance output and the user experience.

The SSML spec is a W3C standard (see https://www.w3.org/TR/speech-synthesis11/)and quite comprehensive. Incrementally adding SSML features makes sense.

Note that Alexa, Google and Cortana support SSML

I suggest the following tags have high priority

<audio> = embed audio clips. support for background audio nice addition - see Google's implementation of the <par>, <seq> and <media> tags

<p> <s> <break> = break elements

<say-as> = control how to say an element

<prosody> = rate, pitch and volume

<emphasis> = emphasize

<lang> = language

<sub> = substitute

Lower priority

<phoneme> - Support IPA and X-Sampa

<voice> - Change voice

23

• 

  Ahmed Bouzid

  • February 17, 2019 19:48

  Agreed.  

  Do you know if they support playing audio (MP3) files?

  AB

   

  PS: Good to see you here, Roger. :-)

  1

  Comment actions Permalink
• 

  Roger Kibbe

  • February 19, 2019 22:17
  • Edited

  Thanks Ahmed. Just sent you an e-mail vis-a-vis this.

  0

  Comment actions Permalink
• 

  Jennie Stenhouse

  • March 07, 2019 11:44

  I am also interested in the ability to play MP3 files! 

  0

  Comment actions Permalink
• 

  Stephen Keefer

  • March 19, 2019 14:28

  Is Bixby able to speak a response?

  0

  Comment actions Permalink
• 

  Roger Kibbe

  • March 19, 2019 17:44

  Yes Bixby speaks responses. By default the dialog template text is spoken but you can add a "speech" key to template to have spoken output be different than the template text.

  See https://bixbydevelopers.com/dev/docs/dev-guide/developers/refining-dialog.intro-dialog for more info on dialog and please use our StackOverflow community - https://stackoverflow.com/questions/tagged/bixby if you have further questions or need additional info.

  0

  Comment actions Permalink
• 

  Malgorzata Misiaszek

  • June 12, 2019 16:06

  hello :)
  I have a question - if Bixby is able to speak a response, is it also able to speak a response in a different language without chnging Bixby language nor system language? If so, could you please provide me with some examples of capsules that do it for me to be able to inset that feature in my own capsule?
  
  I would be grateful :)

  0

  Comment actions Permalink
• 

  Roger Kibbe

  • August 02, 2019 16:37

  The ability to speak in a different language/use a different voice for that language is the first SSML feature we have enabled - see https://bixbydevelopers.com/dev/docs/reference/type/template.speech

  0

  Comment actions Permalink
• 

  Mark Tucker

  • August 17, 2019 18:12

  Please add support for the <break> tag next

  2

  Comment actions Permalink
• 

  Daniel Mittendorf

  • September 09, 2019 08:49

  I'm currently developing a capsule for a customer who really wants to use SSML <audio> for a more immersive game-play. As we're on a clock regarding this capsule, how does the timeline look regarding the audio-SSML feature?

  1

  Comment actions Permalink

Please sign in to leave a comment.