Input logic
tl;dr
Visit corner.builders and try to navigate the room. You can set the viewpoint from the settings panel in the bottom. This step defines the input to the projection function that will come soon!
Goal of the project
Please visit the inception post of the project, and the other editions of this newsletter for more info on this project
Where we are:
We have a website with basic room parameters. We have a floating camera that we can navigate. For the complete history, please visit the other editions of this newsletter.
Goal of this session:
Setting the basic logic of how to set up the input to the projection function. We need a way for the user to specify a viewpoint and a direction, moving, going back to the stored viewpoint, modifying it. It is a certainty that the user-flow will not be definitive. However, what we design now should be simple enough, and workable for an MVP. What is out of scope of this session: the actual projection of the image, and also image selection. Instead we will use our rectangle image, and make it float in front of the viewpoint. So looking from the viewpoint, there is no difference, but looking from the side you can see that the image is floating, and not on the walls. This will come later.
Requirements:
Simplicity. Try to make it clear and unconfusing. Already failed this one I feel :'D
Have a Set Viewpoint mode or a FreeMove mode. In "Set Viewpoint", all user actions like camera orientation, camera position will be stored and defining the "Viewpoint": the point where the projected image on the walls will be seen undistorted. In Free Move the user can navigate the room without changing the viewpoint. They can then observe the projection on a different angle than the viewpoint an see the distorted image.
领英推荐
In Set Viewpoint mode, the user cannot make the camera move with zoom-in/zoom-out (or mouse scroll). Instead this action will change the scale of the image. This is a good limitation as the camera position xyz position should be quite precise and require the user to enter numbers in the input fields. The orientation/scale will be done by fingers on the screen (or mouse) to adjust on the environment.
When switching from FreeMove to Set Viewpoint modes, the user should "teleport back" to the Viewpoint, preventing the mistake of instantly erasing the previous viewpoint, and modify from there.
There should be a "Go to viewpoint" button to also go to viewpoint without change the mode.
Hopefully as a user this should feel natural and without need for explanations (reading this). If it's not, then it means I should do better.
You can find here a short video about the new functionality:
Conclusion:
We have a way to define our viewpoint position and view orientation, and scale (do we also need image rotation?).
Next step:
For now the image is floating. The next logical step is to write the function projection, and actually project this image on the walls, instead of just making it float in the air.