A 2.4 kbit/s MELP coder candidate for the new U.S. Federal Standard

By Alan McCree, K. Truong, E. George, T. Barnwell, V. Viswanathan. 1996

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Speech & Audio

Citations in the corpus

High Fidelity Neural Audio Compression (2022, background): “Despite some advances (Griﬃn & Lim, 1985; McCree et al., 1996), modeling the excitation signal has remained a challenging task.”
Speech Resynthesis from Discrete Disentangled Self-Supervised Representations (2021, background): “Despite some advances [28, 29], modeling the excitation signal has remained a challenge.”
Low Bit-rate Speech Coding with VQ-VAE and a WaveNet Decoder (2019, methodology): “For some evaluations we also include other speech codecs for comparison: Codec2 [11] at 2.4kbps, MELP [17] at 2.4kbps, AMR-WB [10] at 23.05kbps and the original signal after companding and quantization using 8-bit µ -law [18] (giving a bit-rate of 128kbps). µ -law is the companding function used by…”
LPCNET: Improving Neural Speech Synthesis through Linear Prediction (2018, background): “Despite some advances [10, 11], modeling the excitation signal has remained a challenge.”
Wavenet Based Low Rate Speech Coding (2017, methodology): “Quality was evaluated using POLQA and listening tests against a number of unmodiﬁed reference coders: Codec 2 (2.4 kb/s), MELP (2.4 kb/s) [29], Speex wideband (2.4 kb/s) [30], ITU-T G.711 µ -law at 16 kHz (128 kb/s), and ITU-T G.722.2 AMR-WB (23.05 kb/s).”
Improving Opus Low Bit Rate Quality with Neural Speech Synthesis (2019, background): “Examples are linear predictive vocoders [1, 2] or sinusoidal coders [3, 4].”
A Real-Time Wideband Neural Vocoder at 1.6 kb/s Using LPCNet (2019, background): “The quality obtained exceeds what is achievable with existing low-bitrate vocoders such as MELP.”
Noise power spectral density estimation based on optimal smoothing and minimum statistics (2001, methodology): “The noise estimator was tested in conjunction with a multiplicatively modified minimum mean square error log spectral amplitude (MM-MMSE-LSA) estimator [2], [6] and the 2400 bps MELP [ 16 ] speech coder.”
Speaker recognition from coded speech and the effects of score normalization (2001, unknown): “Finally, the Mixed Excitation Linear Prediction[8] (MELP) coder which is the new U.”
Robust Low Rate Speech Coding Based on Cloned Networks and Wavenet (2020, methodology): “However, compared to parametric coding, waveform coding requires a rate that typically is an order of magnitude larger, e.g., compare [7,8] to the adaptive multi-rate wideband (AMR-WB) coder [12], standardized as ITU-T G722.”