In this paper, the contribution of visual and acoustic preference to multi-sensory landscape evaluation was quantitatively compared. The real landscapes were treated as dual-sensory ambiance and separated into visual landscape and soundscape. Both were evaluated by 63 respondents in laboratory conditions. The analysis of the relationship between respondent's visual and acoustic preference as well as their respective contribution to landscape preference showed that (1) some common attributes are universally identified in assessing visual, aural and audio-visual preference, such as naturalness or degree of human disturbance; (2) with acoustic and visual preferences as variables, a multi-variate linear regression model can satisfactorily predict landscape preference (R2 = 0.740), while the coefficients of determination for a unitary linear regression model were 0.345 and 0.720 for visual and acoustic preference as predicting factors, respectively; (3) acoustic preference played a much more important role in landscape evaluation than visual preference in this study (the former is about 4.5 times of the latter), which strongly suggests a rethinking of the role of soundscape in environment perception research and landscape planning practice.

Soundscape perception is increasingly recognized as an important part of landscape preference and environmental experience. However, few studies have juxtaposed visual landscape preference and soundscape preference to compare their contributions to overall landscape preference. This paper aims to quantify and compare the contribution of audiovisual perception to visitors’ overall park landscape preference. The landscape preferences of visitors at seven sample sites in a city park were investigated through field questionnaires in three dimensions: visual landscape, acoustic landscape, and audiovisual landscape. The results showed that visitors’ visual landscape preference (VLP = 7.53) was generally higher than soundscape preference (SP = 7.08), while the influence of auditory preference (57%) on overall landscape preference (OLP) was found to be greater than that of visual preference (43%). The ratio of audio/visual contribution to the overall landscape preference decreased as the average sound level of the sample sites increased. Of all the population characteristics, only the educational level (sig = 0.034) could be used as an effective predictor of OLP (Impact coefficient = −0.103). In addition, older visitors rated OLP lower than young visitors, and females rated OLP lower than males. It was found that visual harmony, color richness, color contrast, plant coverage, and plant diversity were the main visual landscape attributes that influenced visitors’ visual preferences, while acoustic harmony, quietness, sound vitality, and acoustic richness were the main soundscape attributes that impacts visitors’ auditory preference. The results of this study may be useful for park landscape design and regeneration.