Square wave excitation was used for a number of reasons: (1) this made it easy to see any artifacts (deviation from square wave) in the excitation source output at various output powers and frequencies, (2) laser diodes make stable square waves, not sine waves, and (3) it was compatible with the square wave nature of the detector response (i.e. Improvements, observed using square wave excitation, are documented in time domain and frequency domain. One of the most useful test signals in electronics is a humble square wave. The “ideal” square wave is a superposition of an infinite number of sine waves, each contributing it’s required amplitude and phase.
The ABS Speech Information Hiding Algorithm Based on Filter Similarity
WuZhijun , in Information Hiding in Speech Signal for Secure Communication, 2015
3.3.1 LPC Substitution Algorithm
Analysis of the excitation signal in the ABS coding scheme indicates that the similarity between two speech waveform frames is decided by two factors: synthesis filters with greater similarity, and the excitation signal they share. For two synthesis filters with great similarity, if the excitation signal is different, their speech waveforms will be different. Therefore, when the filter similarity of S is greater than So, the LPC coefficients of the current frame cannot be transmitted, but still use the signal of the current frame. This guarantees to keep the speech quality of both the current frame and synthetic speech.
The flow chart of the LPC coefficient substitution algorithm based on filter similarity is shown in Figure 3.4[24].
If the filter similarity S between synthesis filters for the ith and (i–1)th frames exceeds the threshold So, the LPC coefficients for i th frame are replaced by the of (i–1)th frame; that is, substituting with . is the similarity of the neighboring ith and (i–1)th frames, and means substitution of the LPC parameters of the ith frame with of the (i–1)th frame.
The threshold So is a dynamic value, which depends on speech quality, filter similarity, and other factors. The selection of similarity threshold So should guarantee that the LPC substitution algorithm reaches a high hiding capacity (i.e., data rate) and maintains good speech quality.