Hasil Pencarian  ::  Simpan CSV :: Kembali

"Internet-based speech recognition applications prefer using TCP to ensure reliable speech data delivery. TCP-based speech recognition can be designed to push recognition updates on the fly, without waiting for all speech data to fully arrive at the server. We propose an ns-3-based emulation platform to evaluate the performance of TCP-based speech recognition. The server and the client are connected to the simulated network using a tap bridge. The real-time performance of full-duplex speech recognition is measured based on data size, loss rate, and propagation delay. For all our data samples, the application exhibits good performance when the propagation delay is 120 seconds and the loss rate is less than 0.3%, as well as when the propagation delay is 50 seconds and the loss rate is less than 0.5%."

"ABSTRAKImplementasi sistem pengenalan wicara berbasis jaringan, seperti: Internet, akan mengalami degradasi yang disebabkan oleh kehilangan dan keterlambatan data. Sebagian aplikasi pengenalan wicara lebih memilih keterlambatan data demi ketersediaan seluruh data wicara secara kalimat-per-kalimat. Ketersediaan seluruh data akan membantu sistem pengenalan wicara menjaga kinerja akurasi yang semestinya. Akan tetapi, pengguna biasanya lebih menghendaki batas keterlambatan yang wajar sebagai syarat dari kinerja memuaskan aplikasi.Dalam disertasi ini, sebuah model analitik dikembangkan untuk menginvestigasi batas waktu-tunda wajar sebuah skema aplikasi pengenal wicara berbasis TCP yang menempatkan sebuah pemenggal data wicara di klien. Batas waktu-tunda wajar didefinisikan sebagai keterlambatan maksimal yang diperkenankan dalam pengiriman seluruh data setiap kalimat wicara via TCP. Pengembangan model dilakukan melalui analisis transien berdasarkan kajian model discrete-time Markov dari multi-media streaming via TCP. Selanjutnya, sebuah metode perhitungan dari model distribusi keterlambatan paket aliran TCP pada kondisi steady-state diuji dengan membandingkan hasil-hasil perhitungannya dengan hasil investigasi dari model berbasis analisis transien. Hasil perbandingan menunjukan bahwa analisis transien adalah metode investigasi yang lebih tepat.Pada target penelitian berikutnya, sebuah kerangka kerja menggunakan protokol HTTP/2 plus Server Sent Event SSE diajukan sebagai solusi ketepatan waktu aplikasi pengenal wicara berbasis TCP. Kerangka kerja ini dibangun berdasarkan pada kerangka kerja pengenal wicara full-duplex yang dikembangkan dengan menggunakan teknologi WebSocket. Berdasarkan pada hasil percobaan, aplikasi menggunakan HTTP/2 plus SSE memiliki angka perbandingan kinerja latensi sebesar 3,6 lebih baik daripada aplikasi menggunakan WebSocket. Walaupun angka ini masih lebih kecil daripada angka kualitatif perbandingan ketepatan waktu yang lebih baik, yakni sebesar 5 , ada beberapa alasan dikemukakan yang berasal dari keunggulan-keunggulan fitur-fitur HTTP/2 dalam mengurangi latensi aplikasi dan juga dari kelemahan WebSocket bila ditempatkan dalam jaringan dengan proxy server, untuk menyimpulkan bahwa kerangka kerja aplikasi menggunakan HTTP/2 plus SSE dapat menjadi alternatif lebih baik daripada kerangka kerja aplikasi dengan WebSocket.

---

ABSTRACTImplementation of network based speech recognition, such as Internet, will suffer degradation due to packet loss and delays. Most of network speech recognition applications prefer to tolerate delay in order to receive all speech data completely that is delivered sentence by sentence. The availability of all speech data helps the application to save the expected acuraccy of recognition in case of no packet loss. However, users practically require an acceptable delay to have satisfactory performance of the application.In this research, an analytical model is developed to investigate the acceptable delay of TCP based speech recognition that employs a speech segmenter at the client. The acceptable delay is defined as a maximum allowable delay in sending all data for each speech sentence via TCP. For the purpose of model development, there are two analytical methods, i.e., transient analysis and steady state analysis. In the transient analysis, the investigation model is developed based on the discrete time Markov model of multimedia streaming via TCP, whereas in the steady state analysis, the investigation uses a calculation method of packet delay distribution model. Furthermore, the results of transient analysis experiment are compared with the calculation of packet delay distribution model at the steady state. The comparison shows that the transient analysis is more appropriate method of investigation.Next work, a framework using HTTP 2 protocol plus Server Sent Event SSE is proposed as a real time solution for TCP based speech recognition applications. This framework is developed on the basis of a full duplex speech recognition framework using WebSocket. Based on the experimentation results, the application of HTTP 2 plus SSE has a comparison factor of latency performance in amount of 3.6 better than the application of WebSocket. Although this factor is still smaller than a qualitative factor 5 that can state a better latency performance, there are some reason from the advantages of HTTP 2 features in reducing latency as well as from the limitation of WebSocket in a network with proxy server, to conclude that the framework of HTTP 2 plus SSE is a better alternative than the framework using WebSocket."

"In the current study, two experiments are reported that investigated the effects of simple white noise and mixture of white noise and other sounds on perception of speech. In both experiments, university students were recruited to listen to short sentences under various sound masking conditions. Experiment 1, where standard sets of speakers were used for both speech and masking stimuli, has shown that, compared to baseline where there was no masking sound, the participants had significantly greater difficulties in understanding the sentences where the average level of understanding was 28% for the white noise condition and 20% for the mixed noise condition in which white noise was mixed with pink noise and sounds of running water. In Experiment 2, a test model of the specially designed sound masking speaker was used to present the masking noise. Further, sounds of tweeting birds and healing music were added to the mixed noise from Experiment 1 to create the three masking noise conditions. The average level of understanding for the mixed noise condition was 14%, while that for the bird and music conditions were 24% and 30% respectively. The higher understanding rates for the latter conditions were due to lower volume of the mixed white noise in order to keep the overall volume including the birds and music at 55dB. There were also significant effects of sentence type and reading voice gender, suggesting that auditory legibility does not solely depend on the speech-to-noise sound level ratio, but also on other variables, such as, predictability of the sentences, and clarity of the speech. Feedback at the end of the sessions revealed that the participants found mixed noise less irritating than pure white noise, and they preferred mixed noise with bird tweeting or music even better. Thus, it was concluded that mixed noise with occasional sounds of tweeting birds, was the most suitable masking sound for commercial use, being efficient and not unpleasant."

"This updated book expands upon prosody for recognition applications of speech processing. It includes importance of prosody for speech processing applications; builds on why prosody needs to be incorporated in speech processing applications; and presents methods for extraction and representation of prosody for applications such as speaker recognition, language recognition and speech recognition. The updated book also includes information on the significance of prosody for emotion recognition and various prosody-based approaches for automatic emotion recognition from speech."

"Emosi atau perasaan manusia adalah salah satu faktor yang tidak dapat dikendalikan dalam aktivitas apapun. Tidak sedikit juga pekerjaan yang seringkali berkaitan dengan emosi manusia terutama di industri hiburan dan juga kesehatan. Oleh karena itu, 1 dekade kebelakang banyak riset yang dilakukan untuk mempelajari emosi manusia secara langsung maupun menggunakan teknologi. Pengembangan model speech emotion recognition berbahasa Indonesia masih sangat sedikit dan oleh karena itu dibutuhkan perbandingan secara spesifik pada penelitian ini diantara dua model classifier yaitu Convolutional Neural Network (CNN) dan juga Multilayer Perceptron (MLP) untuk menentukan model yang menghasilkan akurasi terbaik dalam memprediksi emosi dari suara manusia. Dalam speech recognition secara umum, salah satu faktor penting dalam mendapatkan model dengan akurasi terbaik adalah metode ekstraksi fiturnya. Oleh karena itu, penelitian ini menggunakan 3 fitur untuk melakukan pelatihan terhadap model yaitu Mel-frequency Cepstral Coefficients (MFCC), Mel-Spectrogram dan chroma. Dari 3 fitur ini, divariasikan dan menghasilkan 7 metode ekstraksi yang berbeda untuk digunakan sebagai input pelatihan model. Terakhir, untuk memastikan bahwa model sudah menggunakan parameter terbaik, dilakukan eksperimen dengan membandingkan model yang menggunakan batch size serta activation function yang berbeda. Ditemukan bahwa dengan menggunakan CNN dan fitur gabungan antara MFCC, mel-spectrogram dan juga chroma menghasilkan model dengan skor akurasi 50.6% sedangkan menggunakan MLP dengan fitur yang sama menghasilkan model dengan skor akurasi 58.47%.

---

Emotions or human feelings are one of the factors that cannot be controlled in any activity. There are also many jobs that are often related to human emotions, especially in the entertainment and health industries. The development of speech emotion recognition models in Indonesian is still very little and therefore a specific comparison is needed in this study between two classifier models, namely Convolutional Neural Network (CNN) and Multilayer Perceptron (MLP) to determine the model that produces the best accuracy in predicting the emotion of the human voice.

In speech recognition in general, one of the important factors in acquiring a model with the best accuracy is the feature extraction method. Therefore, this study uses 3 features to train the model, namely Mel-frequency Cepstral Coefficients (MFCC), Mel-Spectrogram and chroma. From these 3 features, they were varied and resulted in 7 different extraction methods to be used as model training inputs.

Finally, to ensure that the model has used the best parameters, an experiment was conducted by comparing models using different batch sizes and activation functions. It was found that using CNN and the combined features of MFCC, mel-spectrogram and also chroma resulted in a model with an accuracy score of 50.6% while using MLP with the same features resulted in a model with an accuracy score of 58.47%."

"Cross-word modeling for Arabic speech recognition utilizes phonological rules in order to model the cross-word problem, a merging of adjacent words in speech caused by continuous speech, to enhance the performance of continuous speech recognition systems. The author aims to provide an understanding of the cross-word problem and how it can be avoided, specifically focusing on Arabic phonology using an HHM-based classifier."

"Laporan magang ini disusun untuk mengevaluasi prosedur review pengakuan pendapatan yang dilakukan oleh KAP HSB Advisory terhadap PT SB Indonesia Tbk. Lingkup pembahasan terfokus pada Kertas Kerja (KK) Contract Review PSAK 72 yang menggunakan pendekatan five-step process. Penulis melakukan evaluasi dengan membandingkan kerangka evaluasi yaitu, materi mengenai pendapatan yang telah dipelajari di mata kuliah Teori Akuntansi Keuangan. Melalui analisis dan evaluasi yang dilakukan, ditemukan bahwa prosedur pengakuan pendapatan sudah sesuai PSAK 72 Pendapatan dari Kontrak dengan Pelanggan dan sudah diimplementasikan dengan efektif. Penulis juga melakukan evaluasi diri terhadap kegiatan magang yang telah penulis jalankan sebagai advisory intern pada KAP HSB Advisory.

---

This internship report was prepared to evaluate the revenue recognition review procedures performed by KAP HSB Advisory on PT SB Indonesia Tbk. The scope of the discussion focuses on the Contract Review Working Paper PSAK 72 which uses a five-step process approach. The evaluation is conducted by comparing the evaluation framework, namely the theory on income that has been studied in the Financial Accounting Theory course. Through the analysis and evaluation conducted, it was found that the revenue recognition procedure was in accordance with PSAK 72 Revenue from Contracts with Customers and has been implemented effectively. The author also conducts a self-evaluation of the internship activities as an advisory intern at KAP HSB Advisory."

"Laporan magang ini bertujuan untuk mengevaluasi proses pengakuan pendapatan dari kontrak dengan pelanggan pada PT AAA yang diterapkan melalui pengisian kertas kerja Analisis Kontrak (AK). PT AAA merupakan perusahaan telekomunikasi yang memiliki beragam jasa antara lain Konektivitas, Internet, serta Teknologi Informasi dan Komunikasi (TIK). Kerangka yang digunakan untuk evaluasi adalah 5 Step Model dari PSAK 72: Pengakuan Pendapatan dari Kontrak dengan Pelanggan. Evaluasi dilakukan dengan membandingkan ketentuan menurut PSAK 72 dengan implementasi pengerjaan Analisis Kontrak (AK) yang dilakukan oleh Konsultan X untuk setiap tahapan dalam 5-Step Model. Hasil dari evaluasi ini menunjukkan bahwa proses pengakuan pendapatan dari kontrak dengan  pelanggan  dari PT AAA telah sesuai dengan PSAK 72. style="text-align: justify;"This internship report aims to evaluate PT AAA’s revenue recognition from contracts with customers which is implemented through fulfillment of working paper Analisis Kontrak (Contract Analysis). PT AAA is a telecommunication company which provides services such as Connectivity, Internet, and Information and Communication Technology (ICT). The framework used in this evaluation is 5-Step Model from PSAK 72: Revenue Recognition from Contracts with Customers. Evaluation is carried out by comparing each of steps according to PSAK 72 with the fulfillment of Analisis Kontrak (AK) by Konsultan X for each step in the 5-Step Model. The result of this evaluation indicates that the revenue recognition process from PT AAA is in accordance with PSAK 72."