Electric vehicles can be utterly silent; thus manufacturers have started to artificially add alert sounds to cars to lower the risk for pedestrians in the streets. Yet, urban sound environments are often loud, chaotic, and uncontrolled, with pedestrians subjected to a multitude of cognitive stimuli, so that these environments become a challenge. On the example of Shenzhen City – a modern yet loud “high-tech” city with a comparatively high ratio of electric cars – this research studies how pedestrians perceive electric vehicles sonically as objects that are embedded in the urban soundscape. Five different street types served for data collection. Utilizing the collected data, a lab study involving 20 participants simulated urban street scenes with video and audio, followed by a questionnaire on audibility, comfort, attention, sense of control, and sense of safety. Findings indicate that the underground parking lot, with its specific soundscape, has the highest perceived audibility and attention. The food street provides more comfort, control and safety. Overall, for these three perception types, daytime is higher ranked than nighttime, while audibility and attention are less affected by the time of day. This study sheds light on urban soundscape characteristics and contributes future research directions including context-specific, dynamic sound design for enhancing electric vehicle alert sounds and pedestrian-vehicle interaction in dense urban environments.