Audio Response

An audio response allows participants to record their answer. Participants can always replay their recording to examine its content. If rerecording is allowed, participants can rerecord themselves an unlimited number of times until they are satisfied with their recording.

Recommended Practices

• This feature works most reliably in Chrome and Firefox.

Required Properties

type

UsageExample code

• Definition: The type of response
• Possible values: For an audio response, type must be specified as "audio".

{
    "type": "audio"
}

Optional Properties

auto_start

UsageExample code

• Definition: Whether the audio response will start recording automatically.
• Possible values: true or false
• Default: false

{
  "type": "audio",
  "auto_start": true
}

delay

UsageExample code

• Definition: Delay the displaying of a response, relative to the latest onset of all stimuli on a given trial, by a certain number of seconds. This property is useful for forcing participants to read some text stimulus before responding.
• Possible values: A number in seconds, decimals supported (e.g., 1.5)
• Default: 0

{
  "type": "audio",
  "delay": 1.5
}

duration_timer

UsageExample code

• Definition: Whether to show a duration timer that indicates the time elapsed since the reponse becomes visible.
• Possible values: true,false
• Default: false

{
  "type": "audio",
  "duration_timer": true
}

duration_timer_onset

Dependencies

Requires duration_timer to be set to "true".

UsageExample code

• Definition: When to show the duration timer. When set to "start", the timer will appear as soon as the response is displayed and animate a stopwatch effect. When set to "end", the duration timer will only appear once the participant has made a reponse with no stopwatch effect.
• Possible values: "start", "end"
• Default: start

{
  "type": "audio",
  "duration_timer": true,
  "duration_timer_onset": "start"
}

instruction

UsageExample code

• Definition: A short text providing the necessary instructions for participants about the current response. Use empty quotes ("") to remove the instruction text.
• Possible values: Any string
• Default: "Please speak into the microphone clearly."

{
  "type": "audio",
  "instruction": "The text in these quotation marks will be displayed."
}

key_stop

UsageExample code

• Definition: The key a participant should press to stop the recording.
• Possible values: Any character
• Default: " " (i.e., the space bar)

{
  "type": "audio",
  "key_stop": "j"
}

onset_detection

UsageExample code

• Definition: Whether to perform onset detection in this recording based on on real-time analysis of recording volume. If set to true, the audio response will record the time (in milliseconds) at which the participant first starts speaking (or making a loud noise).
• Possible values: true or false
• Default: false
• See also: onset_sensitivity

{
  "type": "audio",
  "onset_detection": true
}

onset_sensitivity

Dependencies

Requires onset_detection to be set to true.

UsageExample code

• Definition: The sensitivity of the onset detection feature. We don't recommend changing this setting unless there are hypothesis-driven reasons for your study to use a lower or higher sensitivity setting.
• Possible values: an integer number between 1 and 99, with 1 being the least sensitive and 99 being the most sensitive.
• Default: 50
• See also: onset_detection

{
  "type": "audio",
  "onset_detection": true,
  "onset_sensitivity": 60
}

padding

UsageExample code

• Definition: The number of milliseconds the response will continue to record after the participant hits stop. This parameter prevents participants from cutting off their response too quickly (i.e., hitting stop while still speaking).
• Possible values: a number in milliseconds (e.g. 600)
• Default: 500

{
  "type": "audio",
  "padding": 600
}

parent

UsageExample code

• Definition: If specified, this response will inherit all properties from another defined response (aka, the parent), unless the same property is explicitly defined in this response.
• Possible values: The name of another defined response.

Given an audio response named "aud1" defined as follows:

{
  "type": "audio",
  "rerecording_allowed": false,
  "delay": 1.3,
  "padding": 600
}

We can create a child response, "aud2", that inherits the rerecording_allowed and padding properties of its parent, but has its own specified delay property.

{
  "type": "audio",
  "parent": "aud1",
  "delay": 2.5,
}

rerecording_allowed

UsageExample code

• Definition: Whether participants can rerecord themselves.

• Possible values: true, false

  Note

  When set to false, participants can still replay their recording but have no way of overwriting it and rerecording it.

• Default: true

{
  "type": "audio",
  "rerecording_allowed": false
}

speech_to_text

Experimental Feature

We are in the process of updating this feature given the recent advancements in language modeling. For the time being, the quality of speech recognition works best on Google Chrome, and is noticeably lower if participants use the Safari browser on Apple devices.

UsageExample code

• Definition: Whether to automatically transcribe the recording to English text.
• Possible values: true, false
• Default: false

{
  "type": "audio",
  "speech_to_text": true
}

stop_on_silence

UsageExample code

• Definition: Whether to automatically stop the recording when FindingFive detects that the participant has stopped speaking. The detection sensitivity can be tuned by the onset_sensitivity property (a larger value means more sensitive).

• Possible values: true or false

  Property in tandem

  When this property is set to true, it automatically sets onset_detection to true as well and cannot be overridden.

• Default: false

{
  "type": "audio",
  "stop_on_silence": true
}

trim_leading_silence

UsageExample code

• Definition: Whether to automatically trim the leading silence of a recording. The sensitivity at which "silence" is determined can be tuned by the onset_sensitivity property (a larger value means more sensitive - that is, a tiny bit of noise will be considered as valid audio input and included in the output).

• Possible values: true or false

  Property in tandem

  When this property is set to true, it automatically sets onset_detection to true as well and cannot be overridden.

• Default: false

{
  "type": "audio",
  "trim_leading_silence": true
}

Recorded Data

• value: The name of the sound file that is saved; this sound file is included in the downloadable zip archive containing your participant data.
• rt: the onset time if onset_detection is turned on; empty otherwise