Created By: UnclGhost on July 16, 2011 Last Edited By: UnclGhost on July 18, 2011

Sliding Scale Of Robot Voices

A classification of how artificial-sounding robots in fiction are.

Name Space:
Page Type:
This scale classifies how "robotic" or "human" an artificial voice in fiction sounds.

Text to Speech

Robots sound like a consumer-grade voice synthesis program such as Microsoft Sam. Ironically, this is the most "realistic" depiction since that's what computer voices sound like today, but it's almost never seen (or heard) in fiction. When it is, it's often Played for Laughs, and may include references to real-life programs, such as the infamous SWAH bug.
  • Auto in Wall E is voiced by an actual text-to-speech program.
  • Justified when Stephen Hawking appears in fiction; he uses a TTS program to talk in real life.

Monotone Human

Robots sound like a human saying their lines with very little emotion or inflection. Expect to hear phrases like "does-not-compute" a lot. This was common in early sci-fi movies and TV shows, but nowadays is a Discredited Trope and is mostly Played for Laughs. This is what people often mean when they describe someone as "robotic"-sounding.
  • Battle droids in Star Wars sound like this in the first two prequels.

Borderline Human

Robots sound like humans saying their lines in an idiosyncratic or automated-sounding way, like a prerecorded message.
  • GLaDOS in Portal is voiced by a voice actress who based her delivery of each line on an actual text-to-speech output, after the dev team discovered the comedic potential of text-to-speech voices calmly saying malicious things. She sounds more human near the end and in the sequel.

Filtered Human

Robots sound exactly like humans speaking through a voice filter.
  • C-3PO in Star Wars has a recognizable, human-sounding voice run through a filter. Presumably, this is important for his protocol and etiquette-related duties.

Perfectly Human Sounding

Robots sound just like humans and there are no audio cues to indicate that they're robots. This includes both emotional and emotionless voices, provided the voice itself sounds human.
Community Feedback Replies: 15
  • July 16, 2011
    Bender from Futurama
  • July 17, 2011
    Title is waaaay too long. How about Perfected Text To Speech?
  • July 17, 2011
    How about Sliding Scale Of Robotic Voices and show a range of robots who sound like machines to robots that sound more "human"?
  • July 17, 2011
    In Durarara, whenever Celty 'speeks' through a laptop or phone, shes somehow able to convey a vast range of emotions. Including Sadness, Anger, and Irritation

    Although its not in the future...

  • July 17, 2011
    ^^ Concur with making this a sliding scale, since the range of representations is so wide.

    Should also talk a bit more about what makes the job so difficult in Real Life and where the state of the art is going.
  • July 17, 2011
    Sliding Scale makes sense, although I'm still going to note how practically nothing is over at the end of "actually voiced by a computer". If anyone knows anything about the Real Life difficulties, please feel free to add it.

    Pjnick300, which category would you say Celty fits in best?
  • July 17, 2011
    Daleks are not robots. And they are not "without inflection" - their Catchphrase is frantically spoken, with added venom. They can even scream, reflect, or plead. They do not belong in "Monotone Human."
  • July 17, 2011
    To add to monotone human:
    • HAL, from 2001ASpaceOdyssey
    • Whatever the robot's name from Moon is

    Perfectly sounding human:

    I would also add another category. I'm not sure where it would fit, but something like Pre-recorded Snippets to include robots that technically "talk", but don't actually create any new arrangements of words or sounds, they just replay pre-recorded segments at the appropriate time. Wall E (and Eve), would count under this.
  • July 17, 2011
    Sorry about the Daleks, fixed.

    I'm not sure about the pre-recorded snippets because it seems like Wall-E and Eve change the pitch and duration of their sounds and can even learn new words, although Wall-E has a little bit of pronunciation trouble, although at the same time they don't really fit into any of the categories here. I dunno, how abstract should this scale get? Should we have R2-D2 at one end and Bender on the other?

    I think the "monotone human" criteria should be a little more on the side of stereotypical "ERR-OR, ERR-OR" robot voices than Gertie (the robot from Moon) or Hal.
  • July 17, 2011
    I don't think Celty counts, people just read the text like you would with a notepad, the voiceover is just for viewer convenience, nobody else can hear it ...yet.

    Shinra, Shizuo and Izaya are just really good at reading her emotions, which is why they don't always need to look at her when talking.

    How about hatsune miku?
  • July 17, 2011
    • Aigis from Persona 3 travels from "Borderline Human" to "Perfectly Human" in the English dub. In this case, it is used to mark her character development throughout the game.
  • July 18, 2011
    In Douglas Adams' H 2 G 2, The Company Choir of the Sirius Robotics Corporation, programmed to sing Share and Enjoy in voices that are exactly a flattened semi-fifth out of tune.
  • July 18, 2011
    You could also add another category of robots who are The Incomprehensible to the viewer, e.g. R2D2 and H.E.L.P.eR.
  • July 18, 2011
    Cybermen are an interesting example in that they've changed over the years. The original cloth-faced ones had a Borderline Human, singsong delivery. This evolved through a Filtered Human period where they seem to show emotions including anger and satisfaction, to the current incarnations which are Monotone Human but very much NOT played for laughs (apart from the line "You would destroy five million Cybermen with four Daleks??").
  • July 18, 2011
    See also Creepy Monotone.