Why using Live2D for modelling?

Fit with 2D game scenes

Although rendering a 3D model with special effects can achieve a 2.5D effect, like what miHoYo made for Genshin Impact with A cartoon-style light and shadow rendering method, it is still essentially three-dimensional modelling by this mean. The technology of Live2D, on the other hand, achieves a 2.5D effect by visualising the image in stereo, which best maintains the character of the 2D presentation.

Fig 1 Genshin Impact

For instance, Revue Starlight Re LIVE developed by Ateam Inc. which has been presented as musical and animation, and won audience favourites in every form. In developing the game, the operating and development team was keen to maintain the initial impression of animation, so that the player would not be alienated from the game with sense of disharmony on the game. By introducing Live2D, the game is able to retain the aesthetics revealed in 2D. Compared to the static screens, the abundance of dynamic frames broaden the boundaries of character performance and provide more realistic communication between characters and players.

Fig 1 Revue Starlight Re LIVE

Simple and easy

In addition to the fact that Live2D model fits better into 2D graphic designs like Japanese animation, modelling with Live2D is more efficient. Paul’s work (Modeling a Character in 3DS Max, 2nd Edition, 2010) about 3D character modelling demonstrate the common method of it. The head, hair, torso, legs, arms, back, neck and shoulders of the figure need to be modelled separately, which can be achieved by adjusting the basic geometry. In this process, it is also necessary to create accessories associated with these parts, e.g. facial features, chest, hips etc. These parts of body are then merged and optimised one by one, and finally mapped the textures. Although a certain amount of work can be reduced by mirroring the components, the overall process is still complex and cumbersome, with at least six to ten steps per part, not to mention numerous modifications of the model to conform original design. The process is the same in Gary's work (Maya 8.0 Character Modeling, 2007), but with some differences in the order of priority.

However, Live2D is capable for modelling 3D shapes simply by defining the portrait in two-dimensional image that maximise the exploitation of illustrations. By separating parts such as hair, facial features (to be separated into eyes, eyebrows, ears, nose, mouth), torso, arms and etc. on the image and setting them as separate parts. Then adds surfaces for deformation as parent, the individual parts of a 2D image are set up as children and can be deformed and moved the same with the parent component, but can also be adjusted autonomously. Adjusts the parts and sets keyframes when they move, the Live2D software will automatically generate a sequence of images that make the character look as if it is 'moving'.

The process is so simple that parameters can be entered with a single click even when setting up individual parts, although the parts built with manual adjustments could be better matched. Modelling operations such as texture painting, mesh editing and skinning in a 2D environment are considerably simpler with Live2D, easier to use while less time and effort are required than 3D models, directly decreasing game development time and bringing economic benefits to game companies by reducing the amount of development costs.

Increasing interactivity

What’s more, research on user experience in 3D stereoscopic games (Takatalo, Kawai, Kaistinen, Nyman and Häkkinen, 2011) reveals that it is not the more stereoscopic the game, the better the experience, and that a moderate level of stereoscopic separation can increase the user's presence for an optimal experience. And it has been experimentally proven that the stereoscopic 3D presentation causes a decrease in presence and an increase in arousal among weekly gamers, but the opposite in non-weekly gamers (Limperos, Waddell, H.Ivory & Ivory, 2014). Thus the 2.5D model enhances the dynamics of 2D games and improves the stereo feeling of the characters, meanwhile increasing gaming experience of players.

Reference:

- Steed, P. (2010) Modeling a Character in 3DS Max, 2nd Edition. 2nd ed. Plano, Tex: Jones & Bartlett Learning.
- Gary O. (2007) Maya 8.0 Character Modeling. Plano, Tex: Jones & Bartlett Learning.
- Takatalo, J. (2011) User Experience in 3D Stereoscopic Games. Media psychology. [Online] 14 (4), 387–414.
- Limperos, W. (2014) Psychological and Physiological Responses to Stereoscopic 3D Presentation in Handheld Digital Gaming: Comparing the Experiences of Frequent and Infrequent Game Players. Presence (Cambridge, Mass.). [Online] 23 (4), 341–353.