<< Chapter < Page Chapter >> Page >
Spectrograms visually represent the speach signal, and the calculation of the Spectrogram is briefly explained.

We know how to acquire analog signals for digital processing ( pre-filtering , sampling , and A/D conversion ) and to compute spectra of discrete-time signals (using the FFT algorithm ), let's put these various components together to learn how the spectrogram shown in [link] , which is used to analyze speech , is calculated. The speech was sampled at a rate of 11.025 kHzand passed through a 16-bit A/D converter.

Music compact discs (CDs) encode their signals at a sampling rate of 44.1 kHz. We'll learn the rationale for thisnumber later. The 11.025 kHz sampling rate for the speech is 1/4 of the CD sampling rate, and was the lowest availablesampling rate commensurate with speech signal bandwidths available on my computer.

Looking at [link] the signal lasted a little over 1.2 seconds. How long was thesampled signal (in terms of samples)? What was the datarate during the sampling process in bps (bits per second)?Assuming the computer storage is organized in terms of bytes (8-bit quantities), how many bytes of computer memory doesthe speech consume?

Number of samples equals 1.2 11025 13230 . The datarate is 11025 16 176.4 kbps. The storage required would be 26460 bytes.

Got questions? Get instant answers now!

Speech spectrogram

The resulting discrete-time signal, shown in the bottom of [link] , clearly changes its character with time. To display these spectral changes, thelong signal was sectioned into frames : comparatively short, contiguous groups of samples.Conceptually, a Fourier transform of each frame is calculated using the FFT. Each frame is not so long that significantsignal variations are retained within a frame, but not so short that we lose the signal's spectral character. Roughly speaking, the speech signal's spectrum is evaluated over successive time segments and stacked side by side so that the x -axis corresponds to time and the y -axis frequency, with color indicating the spectral amplitude.

An important detail emerges when we examine each framed signal ( [link] ).

Spectrogram hanning vs. rectangular

The top waveform is a segment 1024 samples long taken from the beginning of the "Rice University" phrase. Computing [link] involved creating frames, here demarked by the vertical lines, that were 256 sampleslong and finding the spectrum of each. If a rectangular window is applied (corresponding to extracting a frame fromthe signal), oscillations appear in the spectrum (middle of bottom row). Applying a Hanning window gracefully tapers thesignal toward frame edges, thereby yielding a more accurate computation of the signal's spectrum at that moment of time.
At the frame's edges, the signal may change very abruptly, a feature not present in theoriginal signal. A transform of such a segment reveals a curious oscillation in the spectrum, an artifact directlyrelated to this sharp amplitude change. A better way to frame signals for spectrograms is to apply a window : Shape the signal values within a frame so that the signal decaysgracefully as it nears the edges. This shaping is accomplished by multiplying the framed signal by the sequence w n . In sectioning the signal, we essentially applied a rectangular window: w n 1 , 0 n N 1 . A much more graceful window is the Hanning window ; it has the cosine shape w n 1 2 1 2 n N . As shown in [link] , this shaping greatly reduces spurious oscillations in each frame'sspectrum. Considering the spectrum of the Hanning windowed frame, we find that the oscillations resulting from applying therectangular window obscured a formant (the one located at a little more than half the Nyquist frequency).

What might be the source of these oscillations? To gain some insight, what is thelength- 2 N discrete Fourier transform of a length- N pulse? The pulse emulates the rectangular window, and certainly has edges.Compare your answer with the length- 2 N transform of alength- N Hanning window.

The oscillations are due to the boxcar window's Fourier transform, which equals the sinc function.

Got questions? Get instant answers now!

Non-overlapping windows

In comparison with the original speech segment shown in the upper plot, the non-overlapped Hanning windowed version shownbelow it is very ragged. Clearly, spectral information extracted from the bottom plot could well miss importantfeatures present in the original.

If you examine the windowed signal sections in sequence to examine windowing's effect on signal amplitude, we see that wehave managed to amplitude-modulate the signal with the periodically repeated window ( [link] ). To alleviate this problem, frames are overlapped (typically by half a frame duration). This solutionrequires more Fourier transform calculations than needed by rectangular windowing, but the spectra are much better behavedand spectral changes are much better captured.

The speech signal, such as shown in the speech spectrogram , is sectioned into overlapping, equal-length frames, with a Hanning window appliedto each frame. The spectra of each of these is calculated, and displayed in spectrograms with frequency extending vertically,window time location running horizontally, and spectral magnitude color-coded. [link] illustrates these computations.

Overlapping windows for computing spectrograms

The original speech segment and the sequence of overlapping Hanning windows applied to it are shown in the upper portion.Frames were 256 samples long and a Hanning window was applied with a half-frame overlap. A length-512 FFT of each frame wascomputed, with the magnitude of the first 257 FFT values displayed vertically, with spectral amplitude valuescolor-coded.

Why the specific values of 256 for N and 512 for K ? Another issue is how was the length-512 transform of each length-256 windowed framecomputed?

These numbers are powers-of-two, and the FFT algorithm can be exploited with these lengths. To compute a longertransform than the input signal's duration, we simply zero-pad the signal.

Got questions? Get instant answers now!

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Fundamentals of signal processing. OpenStax CNX. Nov 26, 2012 Download for free at http://cnx.org/content/col10360/1.4
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Fundamentals of signal processing' conversation and receive update notifications?

Ask