Wyniki wyszukiwania

In the area of signal processing the transforms are typically used because they offer specific point of view on the signal, revealing its properties, suitable for example for its better decomposition [4]-[6], [15]. An obvious example is the Fourier transform so much suited for acoustic signals treatment and for many other physical signals [5]. However, the digital signal may not have any ref[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2007/4

» Kodowanie dźwięku w cyfrowych systemach transmisyjnych

TOMASZ MARCINIAK

Jakość sygnałów (kodeków) audio można zdefiniować za pomocą trzech głównych parametrów [1]: szerokości pasma sygnału, wierności odtwarzania oraz realizmu przestrzennemu. W przypadku sygnałów audio wysokiej jakości wymagane jest pasmo o szerokości, którego górna wartość jest maksymalną częstotliwością słyszaną przez człowieka i osiąga 20 kHz. Standard płyty kompaktowej CD, który jeszcze dziś [...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2007/4

» Algorithms and optimization methods for parallel processing in one-thread microcontroller based systems

PAWEŁ PAWŁOWSKI

Optimization of various, important, well predefined computation tasks is an important aspect of the art of digital signal processing. Unfortunately, because of a strong demand to put solutions as quickly as possible into market, many commercially available algorithms are not fully optimized. Short time to market (TTM) and desirable performance of the design are often reached much more by mea[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2007/4

» Speech recognition by means of feature extraction method based on slope transformation assisted

ADAM DĄBROWSKI

SZYMON DRGAS

Current automatic speech recognition (ASR) systems achieve already acceptable recognition accuracy for clean speech. However their performance degrades, if they are subjected to speech corrupted with noise present in practical environments. It has been observed that additive white noise severely degrades the performance of mel-cepstra based recognition systems [8]. This problem becomes espec[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2007/4

» Rozpoznawanie samogłosek na podstawie sekwencji wideo

ADAM KONIECZKA

Rozpoznawanie ludzkiej mowy zazwyczaj odbywa się na podstawie analizy dźwięków wypowiadanych przez daną osobę. Ogół definicji tego pojęcia odnosi się jedynie do analizy strumienia dźwiękowego. W literaturze znajdujemy określenie: "Rozpoznawanie mowy to technologia pozwalająca komputerowi wyposażonemu w urządzenie do pobierania dźwięku, np. takie jak mikrofon, interpretować mowę ludzką, na pr[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2008/4

» Digital filtering by discrete trigonometric transforms realized with generalized convolution

In this paper we present considerations, which form an appropriate background for justification a hypothesis that digital filtering based on any transform belonging to the family of discrete trigonometric transforms (DTTs) may be efficiently realized possibly at the cost of some slight and usually fully acceptable approximation with the so-called fast generalized convolution instead of less [...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2008/4

» ACCINO 2 - a microprocessor measurement system with IMEMS accelerometers and gyroscope for testing road restraint systems

PAWEŁ PAWŁOWSKI

aCCino 2 is a microcontroller based measurement arrangement for testing road restraint systems with measurement and computation of: ASI (acceleration severity index), THIV (theoretical head impact velocity) and PHD (post-impact head deceleration). These parameters are needed to test and certify new road restraint systems in Europe according to PN-EN 1317- 1: 1998 [1] standard. Nowadays in Eu[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2008/6

» Influence of subband signal denoising for voice activity detection

TOMASZ MARCINIAK

Exact detection of voice activity endpoints is crucial in many speech processing procedures like, e.g., in: - speech coding in telephone communication, - speech enhancement, - automatic speech recognition (ASR). Determination of endpoints can significantly improve efficiency of these tasks. For example, in case of speech coding, the transmission speed can be reduced, because, as it can be observed, even a duplex conversation signal includes up to 60% of silence [1]. Transmission of only active speech signal parts reduces the channel occupation and saves the battery power. In case of automatic speech recognition, precise detection of word boundaries is a crucial step, which decides about the recognition effectiveness. VAD/EPD can be realized in time and/or in frequency domain [...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2009/3

» Efficient image watermarking in the transform domain with the discrete trigonometric transforms

In the area of the still image watermarking, the approach where the watermark is introduced in the transform domain is considered in numerous descriptions. Apart from the Fourier and wavelet transform the discrete cosine transform (DCT) is mentioned quite often [1,3-5,10-12]. However, the commonly used DCT, is in fact one member of a larger set of 16 discrete cosine and sine transforms [2, 6] also referred to as the discrete trigonometric transforms (DTTs). In this context, the DCT commonly proposed for watermarking is the two dimensional DCT of type II-even (in short: DCT IIe), while there are all together four even DCTs, four odd DCTs, and similarly eight respective discrete sine transforms (DSTs) - four even and four odd [2,6], thus there are 16 DTTs in total. In the presen[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2009/3

» Discrete trigonometric transform (DTT) fi lters

Finite impulse response (FIR) fi ltering, based on fi lter characteristics defi ned in the Fourier transform domain (frequency domain) is one of the most basic and fundamental procedures within the area of digital signal processing [8]. On the other hand the discrete trigonometric transforms (DTTs) seem to attract some attention in the context of fi ltering similar to the well known convolution, which is the result of the convolutionmultiplication property described for the case of the symmetric convolution proposed by Martucci [1]. Some examples of such applications may be found in [4, 7, 9]. In this paper novel approach based on the generalized convolution matrix defi ned for the DTTs [2-4] is proposed and documented with a set of computational examples. The concept exploits[...] więcej»
w zeszycie ELEKTRONIKA - KONSTRUKCJE, TECHNOLOGIE, ZASTOSOWANIA 2010/3