Perceptual Long-Term Variable-Rate Sinusoidal Modeling of Speech

Girin, Laurent; Firouzmand, Mohammad; Marchand, Sylvain

doi:10.1109/tasl.2006.885928

Cited by 14 publications

(18 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This model was first used for cepstral modeling in [28], [35]. Then it was applied to the LT modeling of harmonic parameters in [18], [19] and LP parameters in [20]. The DCM is defined as follows:X…”

Section: B Discrete Cosine Model (Dcm) For the Lt-modeling Of The Stmentioning

confidence: 99%

“…At the same time, it was proposed in [18] to model the LT trajectory of sinusoidal speech parameters (the phase and the amplitude of each harmonic) with a Discrete Cosine Model (DCM). In contrast to [17], where the length of parameter trajectories and the order of the model were fixed, in [18] the long-term frames are continuously voiced sections of speech, which exhibit very variable size and "shape": such a section can contain several phonemes or syllables. Therefore, the LT-DCM is adjusted to F. Ben Ali and S. Djaziri-Larbi are with Université Tunis El Manar, Ecole Nationale d'Ingénieurs de Tunis, Signals and Systems Lab, Tunis, Tunisia.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Low Bit-Rate Speech Codec Based on a Long-Term Harmonic Plus Noise Model

Ali¹,

Djaziri-Larbi²,

Girin³

2016

J. Audio Eng. Soc.

Self Cite

View full text Add to dashboard Cite

The long-term harmonic plus noise model (LT-HNM) for speech shows an interesting data compression, since it exploits the smooth evolution of the time trajectories of the short-term harmonic plus noise model parameters, by applying a discrete cosine model (DCM). In this paper, we extend the LT-HNM to a complete low bit-rate speech coder. A Normalized Split Vector Quantization (NSVQ) is proposed to quantize the variable dimension LT-DCM vectors. The NSVQ is designed according to the properties of the DCM vectors obtained from a standard speech database. The obtained LT-HNM coder reaches an average bit-rate of 2.7kbps for wideband speech. The proposed coder is evaluated in terms of modeling and coding errors, bit-rate, listening quality and intelligibility.

show abstract

Section: B Discrete Cosine Model (Dcm) For the Lt-modeling Of The Stmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Low Bit-Rate Speech Codec Based on a Long-Term Harmonic Plus Noise Model

Ali¹,

Djaziri-Larbi²,

Girin³

2016

J. Audio Eng. Soc.

Self Cite

View full text Add to dashboard Cite

show abstract

“…The secret key is generated using the indices of VQ corresponding to the neighboring frames derived from the natural speech's characters. (Girin, 2007) presents an encryption algorithm based on time-trajectory model of the sinusoidal components corresponding to voiced speech signals. This method uses the amplitude and phase parameters of the discrete cosine functions which are applied for each voiced segment of the speech.…”

Section: Related Workmentioning

confidence: 99%

Performance Analysis of Real Time Implementations of Voice Encryption Algorithms using Blackfin Processors

Duta

Gheorghe

Țăpuș

2016

Proceedings of the 2nd International Conference on Information Systems Security and Privacy

View full text Add to dashboard Cite

A large part of the latest research in speech coding and speech encryption algorithms is motivated by the need of obtaining secure military communications, to allow effective operation in a hostile environment. Since the bandwidth of the communication channel is a sensitive problem in military applications, low bitrate speech compression methods and high throughput encryption algorithms are mostly used. Several speech encryption methods are characterized by very strict requirements in power consumption, size, and voltage supply. These requirements are difficult to fulfill, given the complexity and number of functions to be implemented, together with the real time requirement and large dynamic range of the input signals. To meet these constraints, careful optimization should be done at all levels, ranging from algorithmic level, through system and circuit architecture, to layout and design of the cell library. The key points of this optimization are among others, the choice of the algorithms, the modification of the algorithms to reduce computational complexity, the choice of a fixed-point arithmetic unit, the minimization of the number of bits required at every node of the algorithm, and a careful match between algorithms and architecture. This paper describes the performance analysis on Digital Signal Processor (DSP) platform of some of the recently proposed voice encryption algorithms, as well as the performance of stream ciphers such as Grain v1, Trivium and Mickey 2.0 (which are suited for real time voice encryption). The algorithms were ported onto a fixed point DSP, Blackfin 537, and stage by stage optimization was performed to meet the real time requirements. Memory optimization techniques such as data placement and caching were also used to reduce the processing time. The goal was to determine which of the evaluated encryption algorithms is best suited for real time secure communications.

show abstract

“…This is because in the proposed dictionary, only voiced speech frames can be sparsely represented and thus the deviation between the V/NV metric computed for voiced and nonvoiced frames is generally large. In order to have fast implementation, DCT can be preferred, but in this case the information of the speech source (sparse vector) and speech production system (dictionary) is coupled (Girin et al, 2007). This coupling produces ambiguous results for characterizing voiced sounds in certain cases, which results in a lower V/NV accuracy.…”

Section: Robustness Of Warped-lp Dictionary For V/nv Detectionmentioning

confidence: 99%