Erfan Dastournejad
7 min readMar 18, 2022


Actual Conversations for the virtual world

In the second phase of my thesis project (Winter 2022) for background information and initial research please refer to part one of this article Medium Post (Part 1).

Amid the fairly good rate of COVID 19 vaccination, the University of Washington decided to start the winter quarter with remote learning, which gave me more time to spend on rapid prototyping and solidifying different aspects of my design. Based on the findings of my initial research and further analysis of the rapid prototyping data, I concluded that one reliable way of starting conversations in social VR is to evaluate how much both parties want to start that communication. For the purpose of this article, I call this “want” factor, “communication potential”. And since one side of the communication in my scenarios already indicates their desire for communication, I will only address the communication potential for the receiver side. This communication potential factor will be used to determine how much the receiver of the communication would accept it. The ultimate goal of my project is to help users, evaluate and rise this communication potential to make the initiation of conversation, easier and smoother.
In fact, in this document, I want to demonstrate three phases of communication potential implementation, for conversations in VR.

  1. Calculation
  2. Visualization
  3. Feedback
  4. Calculation
    To calculate the communication potential for a group or individual in VR, we can always borrow hints from real-life experiences. The caveat is that some interactions in VR are literally not possible in the real world, as much as some real-world pieces of information that are not directly transferable into VR. Below I’ll list some of such differences.
    Real-life pros
  • Intricate facial signals (gaze, head/neck orientation, …).
  • Socio-economic ques on who is who.
  • Physical attributes (attraction).
  • Shared environmental affordances (being in a club/ bar or stadium)

VR pros

  • Fabricated appearances
  • Audiovisual signals (icons, UI, etc)
  • Timeline of virtual places (how long someone is in/out of a conversation or place).
  • Safe space (facilitates more adventurous culture)

Given that there are inherent differences between VR interactions and real-world ones (Despite the similarities); I identified four contributing factors to the communication potential calculation. These four categories have affordances both in the real-world and VR which makes them more susceptible for examination and manipulation. The consequent of these individual factors then will be used to calculate the communication potential.

  • Personal Preferences
  • Topic
  • Comfort level
  • Anticipation

Personal Preferences
Everyone has preferences when it comes to picking who to talk to. The nature of these preferences and where they came from are not the subject of this topic, but their influence is a defining factor on the conversation potential. VR has the ability to mask many of these preference indicators (ie. how tall, short people are, abilities/disabilities, age, etc …) or give the user the ability to modify their physical attributes which in turn renders our perception of the looks, useless. However, in a virtual world where avatars can look as they desire the meaning of good-looking, handsomeness, sexiness, etc… will change (although the subject might persist as long as humans are). This will fundamentally change our ability to perceive and evaluate our personal preferences in interactions and potentially will lead to higher acceptance among avatars for communication.

Every conversation revolves around a topic. Mostly shared interests or recent events, these topics act as catalyzers to form and maintain conversations. The subject of the topic is highly dependent on the environment, for example, a conversation about “who is the killer” can mean different things at a funeral or at the cinema. Current social VR platforms try to give context and direction to conversations by putting people in different virtual environments/settings. But most times it lacks contextual information about the environment, hence any perception driven from that environment could be inaccurate. For instance, a virtual bar in VR lacks the (among many things) location information about that bar so people lose their context of mind about who they might encounter in that bar and what are the topics of their interest (ie. a revolving penthouse bar on top of a skyscraper in uptown Tokyo may not have the same visitors when replicated virtually in VR).
This lack of context for the environment makes environment-dependent topics a bit bizarre and unreliable. On the other hand, there might be no need for an environment where people can set up their own environments based on their topics of interest. In reality, people go to sports bars to talk about sport, romantic movies to flirt, networking conventions to talk business, and rallies for politics. Without going too much into the details of a VR solution for topics, for now, I just conclude that knowing/showing the topic of a conversation can raise the communication potential for people who share the same interest around that topic.

Comfort Level
The level of comfort is a bundle of factors determining how comfortable is an individual to start a conversation. Variables like the local time zone (is it 6 pm or 3 am where this person presides now), language (Can they talk the same language as us), physical location (are they in a quiet or busy environment is it summer or winter there, etc …), and cultural openness (is the person comfortable talking to specific people or about specific topics). This might seems like a Duh! at first but affordances in VR brought a slew of design opportunities here. And with great power comes great responsibility. Calculating comfort level might be as easy as showing the local hour and language over each Avatar, however, publicly displaying such information could cause severe privacy issues for some users. In this project, I intentionally avoid raising privacy issues and as a principle, I don’t want to access more data than the platforms are already accessing. Consequently, calculating comfort level as a sensitive topic, will not be included in this project and set aside for later examinations.

People most likely want to know what is going to happen as a result of an interaction. In real life, we identify approaches, gazes, subtle smiles, and inviting looks to anticipate an imminent conversation. The lack of all these signals in VR makes for a vague gray area where people need to send very distinct and abrupt signals before initiating a conversation. On the other hand, the conversation initiator wants to know (beforehand) how likely it is to get rejected. So anticipation makes for an important factor for the sender and receiver of a conversation initiation. With more diligent signaling of anticipation, we can raise the communication potential in both parties.

  1. Visualization
    By limiting the design scope to the four areas above, now it is important to determine the ways to signal/convey this information. Likely in the realm of VR audiovisual signals are compatible with the current tech, ( but the haptics are looming on the horizon). These design solutions are each based on a simple HMW question and the reasoning behind them will make it easier for later evaluation.

A. How might we create more approachable/desirable avatars for conversations?

  1. A good smile is key: No need for expensive facial trackers, just an automatic default natural smile for everyone.
  2. Avoid gender, race, class stereotypical signifiers: If we put the tools there people will use them to present themselves, by focusing on other non-stereotypical ques we will make it harder for people to perceive/judge certain groups of people.

B. How might we indicate conversation topics/interests for an avatar?

  1. Interest Selection: an obvious solution is to let each user select their topics of interest and then display those topics via visual or contextual means.
  2. Prompt Locations: Another way to approach this question would be to set up specific locations for different topics so people who are interested in the topic would be able to find and gather in those locations.
  3. Prompt Setup: Users could set up a location and select their interest as a prompt (ie. a conversation table about politics). Others will see these locations and decide if they want to join.
  4. Prompt Generation: Based on the previous conversations or group discussions, the system will generate and display prompts

C. How might we indicate their openness to a conversation?

  1. One idea is to use view-based orientation to show interest, for instance, if you are not interested in talking, as I approach you I will always see your back to me (without changing your actual orientation).
  2. To indicate presence we can decrease opacity gradually. When someone is inactive for a while their avatar turns transparent.

D. How might we prepare the conversation receiver, for a possible interaction and create anticipation?

  1. A time-sensitive halo could do just for this purpose. When people stand near each other for a short time (usually to start a conversation) helo forms around them individually and after a short time, both halo bubbles will merge into a larger cell to indicate presence and interest.
  2. Group hao bubbles can change color (or show text) the longer they are communicating, to indicate to others that this group is talking for a while.

A combined scenario:

  1. Three avatars are talking, a bubble connecting them shows some topics based on what is being discussed, a strong halo around them shows their group and the fact that they are talking for a while now.
  2. A newcomer is interested in the discussion and decides to join
  3. By moving closer and standing there the halo will encompass him to showing everyone in the group that someone joined.
  4. A group member decides to leave the conversation, as she waves to everyone and leaves, her halo detaches from the group and at the same time, her bubble of topics will follow her for a while, indicating her possible interest topics of her.

3. Feedback
My VR prototype at this point can re-create all the above scenarios. the feedback however requires more thorough testing of the said interactions. So my plan for the next few weeks is to put the prototype to test and see if the feedback piece works as I intend it to work. And ultimately does my VR experience make conversations easier or not?