Project executed by: Anastasiia Kucherenko (UX / research); Genadijus Paleckis (Technical).
Brief: Create an optimised chat experience for online Bingo players. Chat has to be accessible without leaving a game.
Current solution: Players access chat through bottom navigation.
User surveys: to understand more about the audience, I conducted a few surveys during the live chat as well as on forums. Participants of the surveys were mostly females aged 26 – 46, living in the UK, spending a lot of the day time in their homes. Asking players why they like online Bingo, what they do while playing the game, and what activities they are generaly into, the surveys helped me to understand the players’ lifestyle.
Based on the results of the surveys, I created an online Bingo player persona.
Persona: Jessica, 38 y.o. female, Brighton, housewife, has 2 kids of age of 4 and 8. Jessica’s day is running around kids: while the oldest one is going to school, the youngest needs looking after at home. Jessica likes playing Bingo online because she feels more social, even when staying at home; and it doesn’t require her to spend too much time on her screen. However, when it comes to chatting, she can be too busy to type out messages.
Having taken Jessica’s experience into consideration, we decided to go with the option to have a speech recognition chat during the game experience.
Voice recognition is the technology by which sounds, words or phrases spoken by humans are converted into signals that are transformed into coding patterns to which meaning has been assigned.
The most common approaches to voice recognition can be divided into two classes: “template matching” and “feature analysis“.
Template matching is the simplest technique and has the highest accuracy when used properly, but it also suffers from the most limitations. It often is speaker dependent (each person’s voice is different, so the program must first be “trained” with a new user’s voice input before that user’s voice can be recognized by the program. During a training session, the program displays a printed word or phrase, and the user speaks that word or phrase several times into a microphone, 98% accuracy).
Feature analysis normally leads to speaker independent voice recognition (this method first processes the voice input using “linear predictive coding (LPC)”, then attempts to find characteristic similarities between the expected inputs and the actual digitized voice input, 90-95% accuracy).
For our Bingo players experience, we should use the combination of template matching and feature analysis recognition methods.
Template matching to activate commands:
1. “Send” – to activate type by voice system;
2. “Message to” + [username] – to send personal message* (what’s username input method?);
3. “Send emoji” + [emoji name] – to pick emoji.
Thus each message would go like one of the following patterns:
1. “Send” + [message (feature analysis)];
2. “Send emoji” + [Emoji name];
3. “Send” + “Message to” + [user name (feature analysis)] + [message (feature analysis)] .
The voice activation must be easily interchangable with typing, so the user has to be able to chose either option intuitively.
When the message has been typed (or voiced), or emoji selected to be sent, the emoji bar gets replaced with the Send and Cancel buttons:
Looking through Bingo players chat communication, I came across an observation that in the unhosted games, there are 43-64 chat comments per each game. All those comments can be divided into 3 categories:
1. #TG category. It takes about one half of all comments, and are an update of how many numbers a player has got left to win. More specifically those are 2tg (34.5%), 3tg (32%), 4tg (18%), 1tg (15%).
2. Abbreviations: wd/wdw(s) (39%), gl (20.3%), ty/tyvm (19%), and xx, pls/plzz, lol, syl/brb, omg that are less than 6% of the comments in this group each.
3. Without a host in the game, only 1-2 messages per game are anything else than the two above categories.
This brings me to a conclusion that for faster and more convenient chat input, there can be used emoji shortcuts. The 7 most used emoji can be shown on the swiping chat bar:
Emojis will also need to have a text shortcuts to be able to activate them by voice.
The bar with most popular emoji can be swiped onto the chat bar. The “more / plus” icon in the end of the bar will open the list of all icons with their text shortcuts.
Upon Loggin in to the game, a user sees a coachmark intoducing new chat experience.
User can also be asked to calibrate their voice to the speech recognition system. There should also be given a list of emoji and their verbal shortcuts.
Try the prototype [here].
Click around to see the chat works:
Testing the technical possibilities of the voice activated chat on JPJ games (turn CC on):
March 13, 2017
AI Design, UX Design