This study explores whether hours listening to music of young adults with self-reported normal hearing is associated with auditory working memory and if hours listening to music and auditory working memory can predict speech-in-noise perception. Thirty native Swedish speaking university students with self-reported normal hearing in the ages ranging from 21 to 29 years old (M= 23.2) participated in filling out a self-reporting questionnaire concerning their listening habits, a listening span test and a speech-in-noise test. A hierarchical multiple linear regression analysis was performed. The results did not suggest a significant correlation between hours listening to music and auditory working memory nor did it indicate that hours listening to music and auditory working memory could significantly predict speech-in-noise perception. These insignificant findings might be due to several reasons including methodological issues such as the sample size, communication difficulties due to poor internet connection and/or the use of self-reported answers. These results and the arguments presented in the discussion indicate that further research is needed to better answer the research questions of the current study.

In a world where information, entertainment and e-commerce are growing rapidly in terms of volume and options, it can be challenging for individuals to find what they want. Search engines and recommendation systems have emerged as solutions, guiding the users. A typical example of this is Spotify, a music streaming company that utilises users listening data and other derived metrics to provide personalised music recommendation. Spotify has a hypothesis that external factors affect users listening preferences and that some of these external factors routinely affect the users, such as workout routines and commuting to work. This work aims to find time- based listening habits in users’ music listening history to decrease the entropy in the data, resulting in a better understanding of the users. While this work primarily targets listening habits, the method can, in theory, be applied on any time series-based dataset. Listening histories were split into hour vectors, vectors where each element represents the distribution of a label/genre played during an hour. The hour vectors allowed for a good representation of the data independent of the volume. In addition, it allowed for clustering, making it possible to find hours where similar music was played. Hour slots that routinely appeared in the same cluster became a profile, highlighting a habit. In the final implementation, a user is represented by a profile vector allowing different profiles each hour of a week. Several users were profiled with the proposed approach and evaluated in terms of decrease in Shannon entropy when profiled compared to when not profiled. On average, user entropy dropped by 9% with highs in the 50% and a small portion of users not experiencing any decrease. In addition, the profiling was evaluated by measuring cosine similarity across users listening history, resulting in a correlation between gain in cosine similarity and decrease in entropy. In conclusion, users become more predictable and interpretable when profiled. This knowledge can be used to understand users better or as a feature for recommender systems and other analysis. / I en värld där information, underhållning och e-handel har vuxit kraftig i form av volym och alternativ, har individer fått det svårare att hitta det som de vill ha. Sökmotorer och rekommendationssystem har vuxit fram som lösningar till detta problem och hjälpt individer att hitta rätt. Ett typexempel på detta är Spotify, en musikströmningstjänst som använder sig av användares lyssningsdata för att rekommendera musik och annan personalisering. Spotify har en hypotes att externa faktorer påverkar användares lyssningspreferenser, samt att vissa av dessa faktorer påverkar användaren rutinmässigt som till exempel träningsrutiner och pendlade till jobbet. Målet med detta arbete är att hitta tidsbaserade lyssningsvanor i användares musiklyssningshistorik för att sänka Shannon entropin i data, resulterande i en bättre förståelse av användarna. Arbetet är primärt gjort för att hitta lyssningsvanor, men metoden kan i teorin appliceras på valfri godtycklig tidsserie dataset. Lyssningshistoriken delades in i timvektorer, radvektorer med längden x där varje element representerar fördelningen av en etikett/ genre som spelas under en timme. Timvektorerna skapade möjligheten till att använda klusteranalys som användes för att hitta timmar där liknande musik spelats. Timvektorer som rutinmässigt hamnade i samma kluster blev profiler, som användes för att markera vanor. I den slutgiltiga produkten representeras en användare av en profilvektor som tillåter en användare att ha en profil för varje timme i veckan. Ett flertal användare blev profilerade med den föreslagna metoden och utvärderade i form av sänkning i entropi när de blev profilerade gentemot när de inte blev profilerade. I genomsnitt sänktes användarnas entropi med 9%, med några över användare 50%, samt ett fåtal som inte fick någon sänknings alls. Profilering blev även utvärderad genom att mäta cosinuslikhet över en användares lyssningshistorik. Detta resulterade i en korrelation mellan ökning i cosinuslikhet och sänkning i entropi vid användandet av profilering. Slutsatsen som kan dras är att användare blir mera förutsägbara och tolkbara när de har blivit profilerade. Denna kunskap kan användas till att förstå användare bättre eller användas som en del av ett rekommendationssystem eller annan analys.

The efficacy of aural modeling in music education at the primary and secondary levels is well documented, and anecdotal evidence among university studio voice teachers abounds. However, this topic has not previously been explored with undergraduate level singers using acoustic analysis of the singing voice. This investigation utilized a survey on listening behaviors to examine undergraduate voice students’ use of recorded aural models. In addition, an empirical study measured the effect of repeated exposure to recorded aural models on participants’ vocal production. Research was conducted at Southwestern University, a private liberal arts institution in Georgetown, Texas. Study participants were divided into two groups. The control group performed a newly-composed melody after a recorded aural model of the melody was played a single time. The experimental group completed 10-minute listening assignments once a day for a five-day period before performing the same melody. Data between the non-listening and listening groups was compared. Using a second newly composed melody, the control group then completed a five-day listening assignment and performed the second melody. Pre- and post-listening data from this group of subjects was compared. Listening assignments were adapted from a speech pathology remediation technique known as auditory bombardment. They involved listening to multiple repetitions of the recorded aural model without attempting to practice singing the melody. The study measured four acoustic parameters: musical accuracy (pitch and rhythm), vowel/consonant articulation, use of vibrato, and ratio of power between overtones above and below 2 kHz. The listening behavior survey revealed that most students use recorded aural models in their practice time. However, results indicated that students would benefit from professional quality aural models and specific information about appropriate time parameters for listening activities. Results of the empirical study revealed a statistically significant 20-30% improvement in vocal production in both the experimental listening group and the control group post-listening. These data demonstrate that focused periods of listening to an aural model are effective in improving vocal production, even within a short period of time. The results of this study support the inclusion of aural modeling in the applied voice studio. / text