Laporkan Masalah

Sintesis Suara Polifoni dengan Algoritma Karplus-Strong Menggunakan STM32F401CCU6

Shofi Na'ila Haniefah, Enas Dhuhri Kusuma, S.T., M.Eng.; Ir. Addin Suwastono, S.T., M.Eng., IPM.

2026 | Skripsi | TEKNIK ELEKTRO

Sintesis suara polifoni pada mikrokontroler memerlukan pengelolaan sumber daya komputasi dan memori agar beberapa nada dapat dihasilkan secara bersamaan dan real time. Penelitian ini bertujuan mengimplementasikan dan mengevaluasi sintesis suara petikan gitar polifoni berbasis algoritma Karplus-Strong pada mikrokontroler STM32F401CCU6.

Sistem dirancang untuk menghasilkan lima nada gitar yang dikendalikan menggunakan lima joystick analog. Masukan dibaca melalui ADC, setiap nada diproses secara independen menggunakan Karplus-Strong, kemudian sampel seluruh kanal aktif dijumlahkan dan dikirim melalui I2S dengan DMA menuju DAC PCM5102A. Pengujian meliputi akurasi frekuensi fundamental, waktu peluruhan, distribusi harmonik, waktu pemrosesan, beban CPU, penggunaan memori, latensi respons, dan pemetaan pluck velocity.

Hasil pengujian menunjukkan bahwa sistem mampu menghasilkan lima nada secara simultan dan real time. Galat frekuensi fundamental maksimum sebesar 0,4592%. Waktu eksekusi DSP maksimum berada pada kisaran 10-11 us, lebih kecil daripada periode sampling 44,1 kHz sebesar 22,68 us. Pengisian blok audio menghasilkan beban CPU sebesar 23?ngan headroom 77%. Penggunaan SRAM dan Flash masing-masing sebesar 38,53?n 13,81%, sedangkan latensi respons rata-rata sebesar 9 ms. Peningkatan pluck velocity menghasilkan kecenderungan peningkatan RMS, amplitudo puncak, dan amplitudo peak-to-peak. Pemetaan empat level dipilih karena setiap tingkat lebih mudah dicapai dan dibedakan dibandingkan pemetaan delapan level. Berdasarkan hasil tersebut, STM32F401CCU6 memiliki sumber daya pemrosesan dan memori yang memadai untuk menjalankan sintesis Karplus-Strong polifoni pada konfigurasi penelitian ini.

Polyphonic sound synthesis on a microcontroller requires the management of computational and memory resources to generate multiple notes simultaneously and in real time. This study aims to implement and evaluate a polyphonic plucked-string sound synthesis system based on the Karplus-Strong algorithm using an STM32F401CCU6 microcontroller.

The system is designed to generate five guitar notes controlled by five analog joysticks. The input signals are acquired through the ADC, each note is processed independently using the Karplus-Strong algorithm, and the samples from all active channels are summed and transmitted to a PCM5102A DAC through I2S using DMA. The evaluation covers fundamental-frequency accuracy, decay time, harmonic distribution, processing time, CPU load, memory usage, response latency, and pluck velocity mapping.

The results demonstrate that the system can generate five notes simultaneously in real time. The maximum fundamental-frequency error is 0.4592%. The maximum DSP execution time is approximately 10-11 us, which is shorter than the 22.68 us sampling period at 44.1 kHz. Audio block processing produces a CPU load of 23% with 77% headroom. SRAM and Flash utilization are 38.53% and 13.81%, respectively, while the average response latency is 9 ms. Increasing the pluck velocity tends to increase the RMS value, peak amplitude, and peak-to-peak amplitude. A four level mapping is selected because each level is easier to reach and distinguish than those in the eight level mapping. These results indicate that the STM32F401CCU6 provides sufficient processing and memory resources to perform polyphonic Karplus-Strong synthesis under the configuration used in this study.

Kata Kunci : STM32F401CCU6, Karplus-Strong, sintesis suara, petikan senar, polifoni.

  1. S1-2026-502927-abstract.pdf  
  2. S1-2026-502927-bibliography.pdf  
  3. S1-2026-502927-tableofcontent.pdf  
  4. S1-2026-502927-title.pdf