شناسايي خودكار سبك موسيقي

Automatic Recognition of Music Genre

در دهه اخير به دليل رشد سريع داده‌هاي موسيقي كه به صورت فايل‌هاي صوتي در اينترنت و از طريق پايگاه داده‌هاي بسيار بزرگ قابل دسترس هستند، توجه محققان به روش‌هاي پردازش خودكار سيگنال هاي موسيقي بيشتر از گذشته شده است. شناسايي خودكار سبك موسيقي نيز به عنوان يكي از زمينه‌هاي تحقيقي جذاب در اين عرصه مي‌باشد. در اين مقاله روش‌هاي مختلفي براي استخراج ويژگي، انتخاب ويژگي و مدل سازي براي شناسايي خودكار هشت سبك موسيقي شامل سبك‌هاي Celtic، Classic، Classic Piano، Jazz، Metal، Persian Classic، Relaxing و Dance كه از آلبوم‌هايي با نوازنده‌هاي مختلف جمع‌آوري شدند، پياده‌سازي و مورد ارزيابي قرار گرفت. از ميان روش‌هاي رايج استخراج ويژگي كه در حوزه تشخيص گفتار، شناسايي آلت موسيقي و هم چنين شناسايي سبك موسيقي مورد استفاده قرار مي‌گيرند، انواع ويژگي‌هاي كوتاه مدّت، ميان مدّت و بلند مدّت بررسي شده و از ويژگي‌هاي بلند مدّت به دليل اين كه در تعيين سبك موسيقي كارآيي چنداني نداشتند، صرف نظر گرديد. با استفاده از يك روش انتخاب ويژگي و مطالعه تركيب‌هاي مختلف ويژگي‌ها، بردار ويژگي مناسب به دست آمد كه درنهايت با استفاده از حدود سي ويژگي برتر انتخاب شده از بين 122 ويژگي، به كارآيي 90% در تشخيص سبك موسيقي دست يافتيم. مدل سازي سبك‌هاي موسيقي با و بدون در نظر گرفتن توالي زماني، مورد ارزيابي قرار گرفت. بقيه مطالب به دليل زياد بودن درج نگرديد.

Abstract Nowadays, automatic analysis of music signals has gained a considerable importance due to the growing amount of music data found on the Web. Music genre classification is one of the interesting research areas in music information retrieval systems. In this paper several techniques were implemented and evaluated for music genre classification including feature extraction, feature sclcction and music genre modeling on a database of 8 different music genres containing Celtic, Classic, Classic Piano, Jazz, Metal, Persian Classic, Relaxing and Dance music. This database was gathered from several albums composed by different musicians. Short, middle and long term features were studied and finally only short and middle term features were used in our experiments. The long term features were discarded due to their low performance in music genre classification. Two modeling types of the music genres were evaluated. In the first type, only distribution of the feature vcctors was used and in the sccond type, the ordering of the feature vectors was taken into account. Some modeling techniques such as ANN, GMM, Decision Tree and SVM were used individually and in a hierarchical approach. We proposed a taxonomy which classifies the music genres in a hierarchy where there are a small number of classes in the root and large number of classes in leaves. In fact, each class at the root of taxonomy contains one or more music genres and each genre is represented as a leaf at the bottom of the taxonomy. In addition, several classifiers were used simultaneously, in a way that each of them classifies the music genres individually. The decision is finally made using a voting algorithm. Besides, several short-term feature extraction techniques which have successfully been applied in speech recognition, music instrument classification and also music genre classification were studied and after analysis of the experimental results using statistical measures and different combinations of features, a near optimal feature vector was selected.