The current increasing need for privacy-preserving voice communications is leading to new ideas for securing voice transmission. This paper refers to a relatively new concept of sending encrypted speech as pseudo-speech in the audio domain over digital voice communication infrastructures, like 3G cellular network and VoIP. This work presents a novel distortion-tolerant speech encryption scheme for secure voice communications over voice channels that combines the robustness of analog speech scrambling and elevated security offered by digital ciphers like AES-CTR. The system scrambles vocal parameters of a speech signal (loudness, pitch, timbre) using distance-preserving pseudo-random translations and rotations on a hypersphere of parameters. Next, scrambled parameters are encoded to a pseudo-speech signal adapted to transmission over digital voice channels equipped with voice activity detection. Upon reception of this pseudo-speech signal, the legitimate receiver restores distorted copies of the initial vocal parameters. Despite some deciphering errors, an integrated neural-based vocoder based on the LPCNet architecture reconstructs an intelligible speech. The experimental implementation of this speech encryption scheme has been tested by simulations and sending an encrypted signal over FaceTime between two iPhones 6 connected to the same WiFi network. Moreover, speech excerpts restored from encrypted signals were evaluated by a speech quality assessment on a group of about 40 participants. The experiments demonstrated that the proposed scheme produces intelligible speech with a gracefully progressive quality degradation depending on the channel noise. Finally, the preliminary computational analysis suggested that the presented setting may operate on high-end portable devices in nearly real-time.

International audience

Continuous phase modulation (CPM) has gained increasing attention due to its favorable trade-off between power and bandwidth efficiency. Multi-h CPM appeared as a generalization of single-h schemes so as to decrease further the need for bandwidth expansion over the wireless channel. Despite the interesting characteristics of CPM, the decoding of the received signal is particularly difficult in a multi-path wireless environment with no diversity. To provide some level of diversity, several authors have proposed to combine CPM with Space-Time Block Coding (STBC). A new family of codes for CPM, based on L2-orthogonality was recently introduced in [8]. These full rate codes achieve full diversity and a low decoding complexity. In this paper, we detail a non trivial extension of these L2-orthogonal Space-Time codes using multi-h signaling schemes. These new codes achieve full diversity and a better spectral compactness by utilizing the available communication bandwidth more efficiently. Also, the decoding complexity is greatly decreased by using only one correlation filter bank for the detection of all transmitted signals.

Non-linear modulations, like the Continuous phase modulation (CPM) with its constant envelope property, have been appealing for energy efficient communication systems. In parallel, linear orthogonal Space-Time block codes (STBC) have emerged as a simple way of achieving spectral efficiency by full diversity and simple decoupled maximum-likelihood decoding. However, linear codes rely on pointwise orthogonality which leads to a well-known degradation of data rate for more than two antennas. In this paper, we introduce the concept of L2-orthogonality for non-linear Space-Time codes (STC). Our approach generalizes code design based on pointwise orthogonality. Namely, we are able to derive new codes with the same advantages as pointwise orthogonal STBC, i.e. low decoding complexity and diversity gain. At the same time, we are no longer limited by the restrictions of pointwise orthogonal codes, i.e. the reduction in data rate. Actually, we show how to construct full rate codes for any arbitrary number of transmit antennas. More precisely, a family of codes for continuous phase modulation (CPM) is detailed. The L2-orthogonality of these codes is ensured by a bank of phase correction functions which maintains the phase continuity but also introduces frequency offsets. We prove that these codes achieve full diversity and have full rate. Moreover, these codes don't put any restriction on the CPM parameters.

Low-frequency modulation of sound carry essential information for speech and music. They must be preserved for compression. The complex modulation spectrum has already been used for audio compression and is commonly obtained by spectral analysis of the sole temporal envelopes of the subbands out of a time/frequency analysis (modified discrete cosine transform combined with a modified discrete sine transform). However, amplitudes and tones of speech or music tend to vary slowly over time thus the temporal envelopes are often smooth and mostly of polynomial type. Processing in this domain usually creates undesirable distortions because only the magnitudes are taken into account and the phase data is often neglected. We remedy this problem with the use of a complex wavelet transform as a more appropriate envelope and phase processing tool. Complex wavelets carry both magnitude and phase explicitly with great sparsity and preserve well polynomials. Moreover an analytic Hilbert-like transform is possible with complex wavelets implemented as an orthogonal filter bank.

In this paper we present an alternative separable implementation of L2-orthogonal space-time codes (STC) for continuous phase modulation (CPM). In this approach, we split the STC CPM transmitter into a single conventional CPM modulator and a correction filter bank. While the CPM modulator is common to all transmit antennas, the correction filter bank applies different correction units to each antenna. Thereby desirable code properties as orthogonality and full diversity are achievable with just a slightly larger bandwidth demand. This new representation has three main advantages. First, it allows to easily generalize the orthogonality condition to any arbitrary number of transmit antennas. Second, for a quite general set of correction functions that we detail, it can be proved that full diversity is achieved. Third, by separating the modulation and correction steps inside the receiver, a simpler receiver can be designed as a bank of data independent inverse correction filters followed by a single CPM demodulator. Therefore, in this implementation, only one correlation filter bank for the detection of all transmitted signals is necessary. The decoding effort grows only linearly with the number of transmit antennas.

To combine the power efficiency of continuous phase modulation (CPM) with enhanced performance in fading environments, some authors have suggested to use CPM in combination with space-time codes (STC). Recently, we have proposed a CPM ST-coding scheme based on L2-orthogonality for two transmitting antennas. In this paper we extend this approach to the three antennas case. We analytically derive a family of coding schemes which we call parallel code (PC). This code family has full rate and we prove that the proposed coding scheme achieves full diversity as confirmed by accompanying simulations. We detail an example of the proposed ST codes that can be interpreted as a conventional CPM scheme with different alphabet sets for the different transmit antennas which results in a simplified implementation. Thanks to L2-orthogonality, the decoding complexity, usually exponentially proportional to the number of transmitting antennas, is reduced to linear complexity.

The technical realization of modern wireless receivers yields significant interfering IQ-imbalances, which have to be compensated digitally. To cancel these IQ-imbalances, we propose an algorithm using iterative blind source separation (IBSS) as well as information about the modulation scheme used (hence the term semi-blind). The novelty of our approach lies in the fact that we match the nonlinearity involved in the IBSS algorithm to the probability density function of the source signals. Moreover, we use approximations of the ideal non-linearity to achieve low computational complexity. For severe IQ-mismatch, the algorithm leads to 0.2 dB insertion loss in an AWGN channel and with 16-QAM modulation.

We study the issue of regularity for multiwavelets. We generalize here the concept of balancing for higher degree discrete-time polynomial signals and link it to a very natural factorization of the lowpass refinement mask that is the counterpart of the well-known zeros at /spl pi/ condition for wavelets. This enables us to clarify the subtle relations between approximation power, smoothness and balancing order. Using these new results, we are also able to construct a family of orthogonal multiwavelets with symmetries and compact support that is indexed by the order of balancing.

Biorthogonal wavelets have been used with great success in most of the recent transform image coders. By using the new balanced multiwavelets, one can now easily design fully orthogonal linear phase FIR transform schemes. The aim of our work is to assess whether the added orthogonality yields a performance gain compared to traditional biorthogonal transforms. As comparison platform we use the well-known SPIHT codec, which is based on the significance tree quantization (STQ) principle. Without any particular fine-tuning the multiwavelet codec performs within 0.5 dB of SPIHT. A closer inspection shows however that it is hard to improve on this, therefore re-establishing the rule of thumb that strict orthogonality is not a key factor in image transform coding. More details can be obtained on the [WEB] at http://lcavwww.epfl.ch/~weidmann/mwcoder.

This paper deals with multiwavelets which are a generalization of wavelets in the context of multirate filter banks and their applications to signal processing and especially compression. By their inherent structure, multiwavelets are fit for processing multi-channel signals. First, we recall some general results on multiwavelets and the convergence of the iterated matrix product. Then, we define under what conditions we can apply systems based on multiwavelets to one-dimensional signals in a simple way. That means we give some natural and simple conditions that should help in the design of new multiwavelets for signal processing. Finally, we provide some tools in order to construct multiwavelets with the required properties, the so-called 'balanced multiwavelets'.