Relationship between frequency shifting of vehicle sounds and acceleration impression

Abstract

Pioneer Corporation,25–1 Aza-Nishimachi, Yamada, Kawagoe, 350–8555 Japan(Received 22 May 2014, Accepted for publication 7 August 2014)Keywords: Vehicle sound, Acceleration impression, Frequency shifting, Paired comparison, Subjective evaluationPACS number: 43.50.Lj, 43.66.Hg, 43.66.Lj [doi:10.1250/ast.36.135]1. IntroductionThe use of vehicles that are propelled in whole or in partby electric motors, such as electric vehicles (EV) or hybridelectric vehicles (HEV), is becoming increasingly common inurban fleets. Many types of EV/HEVs emit less sound thanconventional internal combustion engine vehicles (ICEVs),particularly when they are driven at a low speed. It has been amatter of concern that the quietness of EV/HEVs may createdangerous situations for pedestrians, because they sometimescannot hear the sound of such vehicles in an urban soundenvironment [1]. Hence, regulations and recommendationsthat will mandate or recommend the installation of additionalsound-emitting devices in quiet vehicles are being consideredby some governments/organizations, such as the UnitedStates, Japan, and the World Forum for Harmonization ofVehicle Regulations (UNECE/WP.29). According to a guide-line [2], a warning sound should be automatically generatedwhen the vehicle is driven at a speed of less than 20km/h.This is because the difference between the sound level of anEV/HEV and an ICEV is significant in this speed range andthe tire-road noise becomes dominant at higher speeds.When the sound is designed, it should not be a simpleimitation of a conventional gasoline engine sound. Accordingto a questionnaire on the sound of EV/HEVs [3], conven-tional engine sounds are not the best means to identify avehicle’s presence and behavior in an urban sound environ-ment. In the survey, more than 50% of the respondentswho drove their cars daily replied that, as drivers, they hadexperienced a sense of danger owing to pedestrians’ lackof awareness of their cars. In contrast, only 1.6% of therespondents owned and drove HEVs.It is crucial to reveal the appropriate design of the sounds.For pedestrians, it is important to know the necessary soundlevel to give sufficient warning of an approaching vehicle inan urban environment. The authors have conducted a series ofstudies [4–6] to discuss the feasible sound level of the sound.There has also been some research concerning the temporaland spectral properties of such a sound. (e.g., [7,8]).Another concern is the frequency shifting of the sound.The regulations and recommendations on additional soundsfor quiet vehicles may specify requirements concerning thevariation in the frequency of the sounds as a function of thevehicle speed. The goal is to make it easier for pedestriansto recognize vehicle behavior such as acceleration [2]. Therelationship between frequency shifting and a pedestrian’simpression of a vehicle’s acceleration (hereafter referred toas ‘‘acceleration impression’’) needs to be clarified in orderto design the additional sounds and to develop appropriatelegislation. Although the frequency shifting of the sound of aconventional vehicle is primarily caused by the change in theengine revolutions per minute, the additional sounds can bedesigned independently of the vehicle’s propulsion mecha-nism. As mentioned above, merely a simple imitation of aconventional engine sound is not the best method of designinga sound that can be easily detected in an urban soundenvironment. Additionally, imitation of the engine soundmay not be the best means for pedestrians to recognize vehicleacceleration. Moreover, designing a new sound with achanging frequency that is independent of the vehiclepropulsion mechanism could also result in a novel drivingexperience for drivers and passengers in vehicles.To determine the relationship between frequency shiftingand acceleration impression, the authors conducted Scheffe’spaired comparison test using audio and visual stimuli in alaboratory. The visual stimulus was a vehicle traveling on aroad, monotonically accelerated from its stopped state. Thiswas recorded from two perspectives; from the cabin as theview of a passenger, and from the road as the view of apedestrian. The audio stimuli were a combination of three1/3-octave band noises. The central frequencies of the bandnoises were shifted independently. The effect of frequencyshifting on acceleration impression was analyzed usingrelative impression scores, which were defined as the dif-ferences relative to the stimulus without frequency shifting.2. MethodThe acceleration impression of the vehicle and theharmonization of the audio and visual stimuli were assessedthrough a paired comparison experiment.2.1. StimuliCombinations of audio and visual stimuli were used inthe experiment. The visual stimuli were videos of a vehicletraveling on a road, monotonically accelerated from itsstopped state. Two scenes were recorded from the interior