User Test – Sing Song


Questions I aim to test for:

  • Are the words too confusing?
  • Do I need to play recordings of the words to the user before asking them to participate?
  • Are the controls immediately understandable to a new user?
  • Is the user able to successfully operate the interface?
  • Is the user interested?
  • How long is the user interested in engaging in the activity?
  • Does the user repeat an action?



  • Most users wanted to hear the pronunciation of the words being played before they attempted to record themselves saying it.
  • Others, noting the thrill of learning something new, wanted to try to say the word themselves before being provided with a pronunciation.
  • From a design aspect, users preferred seeing the words on screen rather than on queue cards as they thought having the word on paper while speaking and looking for the design to appear on the screen was distracting.
  • How would I use this word? Provide some context for the usage of the word (e.g.) sentence usage, most common situations to use it.
  • Benedetta pointed out that she preferred the words on paper, minus the pronunciation hint. As most folks are not trained to use the dictionary’s pronunciation tool. She deemed it irrelevant.


  • Users want some type of feedback to determine if they are saying the word correctly.
  • Have he most recent attempt to say the word in the front layer. Keep previous versions in a back layer (i.e.) show drawing for each iteration.
  • Show what the perfect drawing of the word so users can compare. Want to see the similarity.
  • Draw immediately as the word is being said.
  • Provide the user with a copy of their drawing — this can be emailed or a printer connected to provide a copy. — or— a servo controlled arm that draws a physical rendition of what’s on screen. The printed version should include the word, it’s meaning and some brief context of what is being represented.
  • If the drawing is beautiful the user will try to say another word!

General design:

  • A mic entices users to say something – so having that component there would engage folks.
  • The system should be designed so that users find their way around without instructions.
  • Give the user more than one try, since I am prompting them to generate something unique.
  • Try building a mic to analyze the sound.

My take-aways

I have to make some decisions regarding the final design of the interface.

  • Do I give the user a sound byte to inform proper pronunciation?
  • How should I represent the perfect Trinidadian pronunciation of the word?
  • What should the input (mic) look like? Should I build one?
  • Buttons: should I create separate buttons for input and output? i.e recording the sound and playing it back.
  • Do I want to incorporate the ability to print a physical copy of the drawing? If so, how?


Leave a Reply

Your email address will not be published.