Recorded footsteps VS synthesised footsteps walking on aggregated materials : How do experienced listeners evaluate audio quality and realism when comparing granularsynthesised footsteps to recorded footsteps

Sammanfattning

Foley artists are often tasked with the assignment to replace the footsteps in a movie scene. This is often done by walking on different materials that correspond with the materials on screen, or substitute materials that have similar textures and produce a similar sound. However, with modern technology, a person without experience with Foley can produce similar sounds using a synthesiser. This thesis investigates how subjects rate the performance, realism and audio quality when comparing recorded and granular synthesised footsteps walking on aggregated materials. A listening test was conducted with 16 participants comparing the two methods with 3 different aggregated materials, consisting of gravel, sand and snow. After each evaluation, participants assessed the performance and realism. This data was cross-analysed with qualitative responses and evidenced that recorded stimulus was preferred. The reasoning behind this preference differed for each material, but the main factors were weight, level of details, rhythm/tempo and variance between footsteps. However, it can be concluded that the elements that make a footstep sound realistic, are similar for materials who share common sonic attributes, but differ between materials that do not share the same attributes. Thus, it’s hard to define what makes a footstep sound realistic, it all depends on the material. The participants also had to evaluate the sound quality of each stimulus on a 5 point scale. A quantitative analysis, using ANOVA, revealed no significant difference of audio quality between the recorded and synthesised stimuli. This also makes the conclusion that higher sound quality does not equal higher realism.