Procedural generated trees - How distribution and growth factors affect perceived realism

Jæger, Peter Hørdum

4. term

Medialogy, Master

2018

2018-05-31

121

Dette studie undersøger, hvordan placeringen af træer og de regler, der styrer deres vækst i et videospil, påvirker spillernes oplevede realisme. Oplevet realisme henviser her til, hvor troværdige objekter ser ud og opfører sig for spillere, og den formes af deres tidligere erfaringer og forventninger. Undersøgelsen fokuserede på proceduralt genererede træer—træer skabt af algoritmer frem for håndbyggede modeller. To uafhængige variable blev manipuleret: fordeling (tilfældig placering versus økologisk fordeling, der efterligner mønstre i naturen) og vækstmodel (en algoritme, der tager højde for begrænsende faktorer for trævækst, versus en der ikke gør). Samlet set var der ingen statistisk signifikante forskelle mellem testbetingelserne i vurderingen af scenens realisme. Der var dog en indikation af, at træernes form blev opfattet som mere realistisk, når træer var placeret med økologisk fordeling, og når væksten blev modelleret med begrænsende faktorer. Disse resultater peger på, at mere naturtro placerings- og vækstregler kan få vegetation til at fremstå mere troværdig for spillere, selv hvis den overordnede realismevurdering ikke ændrer sig betydeligt.

This study investigates how the placement of trees and the rules governing their growth in a video game influence players’ perceived realism. Perceived realism here refers to how believable objects look and behave to players, and it is shaped by their previous experience and expectations. The research focused on procedurally generated trees—trees created by algorithms rather than hand-built models. Two independent variables were manipulated: distribution (random placement versus ecological distribution that mimics patterns found in nature) and the growth model (an algorithm that accounts for limiting factors that constrain tree growth versus one that does not). Overall, there were no statistically significant differences between the test conditions in how realistic the scenes were judged. However, there was an indication that tree shape was perceived as more realistic when trees were placed using ecological distribution and when growth was modeled with limiting factors. These findings suggest that more natural placement and growth rules can make vegetation appear more believable to players, even if the overall realism rating does not change significantly.

[This abstract was generated with the help of AI]

Download
View record in AAU Student Projects