Delay-less frequency domain packet-loss concealment for tonal audio signals

Abstract

A delay-less packet-loss concealment (PLC) method for stationary tonal signals is presented, that addresses audio codecs utilizing a modified discrete cosine transformation (MDCT). In the case of a frame loss, tonal components are detected using the last two received spectra and their accompanied pitch information. Phases of the tonal components are subsequently predicted, aiming for a continuous phase evolution between successive frames. Thus, accurate estimates of tonal components are achieved for stationary signals during frame losses, leading to an improved listening impression. This method is used in the recently standardized enhanced voice services (EVS) codec. Subjective quality evaluation using a MUSHRA test demonstrates its performance

ARMAS: Active Reconstruction of Missing Audio Segments

“…The signals provided in the audio gaps in the DNN structure were time-frequency coefficients (either complex-values or magnitude). In the work of Sperschneider et al [7], the authors presented a delay-less packet-loss concealment (PLC) method for stationary tonal signals, which addresses audio codecs that utilizes a modified discrete cosine transformation (MDCT). In the case of a frame loss, tonal components are identified using the last two obtained spectra and their pitch information.…”

EVS Channel Aware Mode Robustness to Frame Erasures

“…There is for example a specific transition mode that codes speech and audio onsets so that individual lost frames do not adversely affect audio quality [8]. For tonal music and stable speech vowels there are separate coding modes that handle these type of stable segments very well in case of frame loss [9]. Global gains are used independently from the previous frames in order to reduce error propagation [10].…”