The enhancement of speech which is corrupted by noise is commonly performed in the short-time discrete Fourier transform domain. In case only a single microphone signal is available, typically only the spectral amplitude is modified. However, it has recently been shown that an improved spectral phase can as well be utilized for speech enhancement, e.g., for phase-sensitive amplitude estimation. In this paper, we therefore present a method to reconstruct the spectral phase of voiced speech from only the fundamental frequency and the noisy observation. The importance of the spectral phase is highlighted and we elaborate on the reason why noise reduction can be achieved by modifications of the spectral phase. We show that, when the noisy phase is enhanced using the proposed phase reconstruction, instrumental measures predict an increase of speech quality over a range of signal to noise ratios, even without explicit amplitude enhancement.

STFT phase reconstruction in voiced speech for an improved single-channel speech enhancement

An algorithm for estimating signals from short-time magnitude spectra is introduced offering a significant improvement in quality and efficiency over current methods. The key issue is how to invert a sequence of overlapping magnitude spectra (a ldquospectrogramrdquo) containing no phase information to generate a real-valued signal free of audible artifacts. Also important is that the algorithm performs in real-time, both structurally and computationally. In the context of spectrogram inversion, structurally real-time means that the audio signal at any given point in time only depends on transform frames at local or prior points in time. Computationally, real-time means that the algorithm is efficient enough to run in less time than the reconstructed audio takes to play on the available hardware. The spectrogram inversion algorithm is parameterized to allow tradeoffs between computational demands and the quality of the signal reconstruction. The algorithm is applied to audio time-scale and pitch modification and compared to classical algorithms for these tasks on a variety of signal types including both monophonic and polyphonic audio signals such as speech and music.

Real-Time Signal Estimation From Modified Short-Time Fourier Transform Magnitude Spectra

Recent advances in text-to-speech (TTS) led to the development of flexible multi-speaker end-to-end TTS systems. We extend state-of-the-art attention-based automatic speech recognition (ASR) systems with synthetic audio generated by a TTS system trained only on the ASR corpora itself. ASR and TTS systems are built separately to show that text-only data can be used to enhance existing end-to-end ASR systems without the necessity of parameter or architecture changes. We compare our method with language model integration of the same text data and with simple data augmentation methods like SpecAugment and show that performance improvements are mostly independent. We achieve improvements of up to 33% relative in word-error-rate (WER) over a strong baseline with data-augmentation in a low-resource environment (LibriSpeech-100h), closing the gap to a comparable oracle experiment by more than 50%. We also show improvements of up to 5% relative WER over our most recent ASR baseline on LibriSpeech-960h.

Generating Synthetic Audio Data for Attention-Based Speech Recognition Systems

A method and system is provided for encoding and decoding of speech signals at a low bit rate. The continuous input speech is divided into voiced and unvoiced time segments of a predetermined length. The encoder of the system uses a linear predictive coding model for the unvoiced speech segments and harmonic frequencies decomposition for the voiced speech segments. Only the magnitudes of the harmonic frequencies are determined using the discrete Fourier transform of the voiced speech segments. The decoder synthesizes voiced speech segments using the magnitudes of the transmitted harmonics and estimates the phase of each harmonic from the signal in the preceding speech segments. Unvoiced speech segments are synthesized using linear prediction coding (LPC) coefficients obtained from codebook entries for the poles of the LPC coefficient polynomial. Boundary conditions between voiced and unvoiced segments are established to insure amplitude and phase continuity for improved output speech quality.

Harmonic adaptive speech coding method and system

We present a computationally efficient real-time algorithm for constructing time-domain audio signals from spectrograms. The Single-Pass Spectrogram Inversion (SPSI) algorithm is similar to the synthesis step in phaselocked vocoders, but with phase rates at spectral peaks determined solely from the magnitude spectra using quadratic interpolation. The algorithm provides good quality results in a single fully-deterministic pass, and can also provide excellent initial phase estimates for improved results with iterative spectrogram inversion techniques such as the Griffin-Lim algorithm.

Single Pass Spectrogram Inversion

In this paper, speech synthesis directly from the processed Short-Time Fourier Transform Magnitude (STFTM) using the LSEE-MSTFTM algorithm [6,7] is compared to more conventional algorithms for several speech processing applications. For the applications considered, the most improvement occurs for time-scale modification of multiple speaker speech and noisy speech since these input signals are not well modeled by the analysis/synthesis system used for comparison. However, for the applications of speech synthesis from speech model parameters, time-scale modification of clean speech, speech enhancement by spectral subtraction, and helium speech enhancement, significant improvement is not gained by using the LSEE-MSTFTM algorithm. Significantly better results are not obtained since a good STFT phase estimate is available and employed in the conventional approaches to these applications.

Jae S Lim

Papers

Speech synthesis from short-time Fourier transform magnitude and its application to speech processing

Abschätzung von Anregungsparametern