Course materials for UCL module COMP0160
For the last few weeks we have been looking at the perceptual consequences of increasingly complex interacting stimuli, in the context of pseudo-haptic effects in a virtual reality environment. The core pseudo-haptic illusion under consideration, arising from a mismatch between the visually-perceived movement of a virtual object and the kinaesthetic perception of the corresponding limb movement, seems to be susceptible to modulation by other concurrent (visual and auditory) stimuli that the brain integrates into its understanding of the scene.
In today’s lab we add some further variations, because obviously the situation wasn’t complicated enough. But we also, more importantly, consider:
Once again we will implement these experiments in both screen-based and headset-based VR settings. One of the things to consider is how the experience of each environment differs, and what the perceptual consequences of that might be.
The screen-based Unity environments you have used so far give you more or less unlimited scope for adjusting the details of an experiment. You have access to the parameters. You control what is onscreen. You have the source code and can rewrite it at will.
The headset versions are much more prescriptive. This week the experiment app gives you more (any!) configuration options, and you might choose to design a sequence of experiments that varies only very specific stimuli at a time, or all at once. Nevertheless, you are still confined by the options I have provided for you, and you should think about what that means.
As in previous sessions, you should work in pairs or groups for these tasks.
In previous weeks you have used a Unity environment in which you manually altered object attributes in the inspector. This a straightforward and very versatile approach, but it is unwieldy to manage and there may be a risk of the experimenter’s expectations influencing the subject.
A useful alternative is to have software run the experiment automatically. You saw this approach in action on the VR headsets, for which manual experimentation is impractical. This week we will apply similar approaches onscreen as well.
We will begin with an example of setting up and coding such an experiment in Unity.
(Note: this demo was promised a few weeks ago but deferred due to illness.)
The experimental setup in this case is very simple, considering only variations in a single object attribute, mass, which you will recall determines the C/D ratio. However, the same general idea can be used for more elaborate experiments, and you’ll get to try a more elaborate version below.
A couple of things to note:
The main screen-based project this week is an automated experiment manager similar to that developed above, but with much more configurability.
Once again, you will need to set this up in a new Unity project. The process this time is a little more fiddly than before, mostly because this project uses multiple scenes. But the initial stages should be familiar by now:
Launch the Unity Hub application and sign into your Unity account if you have not already done so. Create a new project for this session:
Download lab9.zip and unzip it. Locate the lab9.unitypackage
file inside the resulting folder.
lab9.unitypackage fileMake sure the Game panel is open in the viewer area. You should see a form with some controls and text. However, the text display needs some supporting resources and also, somewhat exasperatingly, seems to need a couple of run ups before it will work properly.
In the Hierarchy window, select Canvas → Headline. This should bring up a dialog saying that you need to import starter assets for TMPro. Click the button to do so.
Now, run the scene. You will probably see that there are lot of blank spaces on the onscreen form.
Stop the scene running. You should now see that the form text is populated, at least with placeholders like Experiment Settings [XXXX]. If you run it at this point you should be able to configure an experiment, but it will fail when you try to start it. This is because the program can’t find the scenes which implement the experiments. To make them available, they need to be added to the scene list.
You should now be able to run this scene to configure an experiment, then press Start Experiment to run it.
As with the headset experiments from last week, this project assigns a random 4-letter
code to each experiment and saves the data as a CSV file with this code as the filename.
The CSVs are saved to the application’s persistent data folder, which is typically in
some non-obvious place like ~/Library/Application Support/comp0160/lab_9/
or C:\Users\Matt\AppData\LocalLow\DefaultCompany\. The full path to each
output file will be logged to the Unity console, so check that if you
have trouble finding your data.
The experiment project you set up above allows you to configure the following experimental parameters:
Given that these are the options at your disposal, you are tasked with designing an experiment that will tell you something about the subjects’ perceptual responses to this environment. Once again you should consider some or all of the following:
Having figured out your experiment, perform it on each other and gather the data. As in previous weeks, you may each do separate experiments or design one collectively that you then all participate in.
If you have time, do the data analysis. Does it show what you expect? Does it show anything at all?
Apps for exploring the effect of various object parameters and for running automated experiments are pre-installed on the Pico 4 VR headsets. You will use these for Tasks 3–5 below.
Before doing so, you may need to configure (or reconfigure) the headset’s play boundary.
If you are starting the headset from a powered down state, or if your present location is significantly different from its previous setting, it will invite you to reset the boundary automatically. If not — for example, if you are using a headset previously used in the seat next to you — you should manually reset the boundary from the settings screen.
Configure a Stationary Boundary to be Small and Sitting. Ensure that the floor level is appropriate for your surroundings — try setting this automatically at first, but if in doubt adjust it manually. Then Enter VR World.
If the dock is not visible in the lower part of your vision, open it by pressing the ○ button on one of your controllers. Choose the Library option, and then in the left hand sidebar of the window that opens, choose Unknown. Among the available options should be apps called CDratio and lab9_experiments — run these for the tasks below.
(The headset will warn you that the apps are from unknown developers and might damage your device. This is true, but run them anyway. Even if you don’t trust us, it’s our device and we’ll damage it if we want to.)
The CDratio app provides an exploratory environment with two blocks whose properties you can adjust with sliders, similar to those you used last week. The available parameters this time are:
In addition, there is a global control for audio Volume. Note that this is auxiliary to the system volume control — if your headset is muted outside the app, it will remain silent regardless of any changes you make to the in-app volume.
Play around with all the available properties and consider how they seem, subjectively, to affect your experience of lifting the blocks in the scene.
The lab9_experiments app, just like the screen-based project, allows you to configure an experiment, specifying which factors to vary and how many trials to perform, and it will then run that experiment and save the results. There are only so many things you can configure, but that’s still a lot of variation. How are you going to deal with it?
Go back to Task 1 above. Review the questions.
Design a set of experiments to perform on the headset.
You may wish to reuse your previous design unchanged. Or maybe you’d like to revise it in the light of your results the first time around. Perhaps the headset environment is different enough that you’d like to change your hypothesis? The choice is yours.
Do the experiments.
Analyse the data.
Like last week, results will need to be retrieved from the headsets. We will again, do this after the lab and post the files to Moodle, so you probably won’t have the chance to do any analysis during the session. But if you finish the experiments with a lot of time to spare, ask the tutors — we may be able to get the data for you sooner so you have a chance to look at it during the session.