EUVIMEDEUVIMED entdecken
Crossref

AKDQ: Adaptive Key-Delta Quantization for Low-Bitrate RVQ-Based Neural Speech Codecs

Xiao-Hang Jiang, Yang Ai, Zhen-Hua Ling, Ji Wu

IEEE Signal Processing Letters · 2026 · 33 · 4201-4205

DOI: 10.1109/lsp.2026.3744302 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Xiao-Hang Jiang und 3 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Gefördert von: National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Low-bitrate speech coding is essential for bandwidth-constrained channels and on-device applications, but conventional residual vector quantization (RVQ)-based codecs employ a fixed number of codebooks per frame regardless of content. Adjacent speech frames often exhibit strong similarity, and quantizing them independently with the same scheme results in redundant encoding, thereby wasting bitrate. Therefore, we propose Adaptive Key-Delta Quantization (AKDQ), which replaces the fixed per-frame VQ allocation of RVQ with an adaptive frame-wise key-or-delta policy. AKDQ is a plug-in strategy that requires no modification to the codec's encoder–decoder backbone. Specifically, in AKDQ, each frame is classified as either a key-frame (K-frame) quantized with all VQs, or a delta-frame (D-frame) quantized with only a few VQs on a delta correction from the previous reconstruction, thereby reducing bitrate. The K/D decision is made by a budget-aware causal scheduler through a bit-bank mechanism, keeping the system streamable with no additional lookahead. Experiments on both causal and non-causal codec families demonstrate that AKDQ effectively reduces bitrate redundancy. For instance, on DAC, using AKDQ at 1.5 kbps achieves reconstruction quality comparable to using the baseline RVQ at 2.0 kbps, saving 25% of the bitrate.

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Jiang, X., Ai, Y., Ling, Z., & Wu, J. (2026). AKDQ: Adaptive Key-Delta Quantization for Low-Bitrate RVQ-Based Neural Speech Codecs. IEEE Signal Processing Letters, 33, 4201-4205. https://doi.org/10.1109/lsp.2026.3744302

Vancouver

Jiang X, Ai Y, Ling Z, Wu J. AKDQ: Adaptive Key-Delta Quantization for Low-Bitrate RVQ-Based Neural Speech Codecs. IEEE Signal Processing Letters. 2026;33:4201-4205. doi:10.1109/lsp.2026.3744302

BibTeX
@article{Jiang2026,
  author = {Jiang, Xiao-Hang and Ai, Yang and Ling, Zhen-Hua and Wu, Ji},
  title = {AKDQ: Adaptive Key-Delta Quantization for Low-Bitrate RVQ-Based Neural Speech Codecs},
  journal = {IEEE Signal Processing Letters},
  year = {2026},
  volume = {33},
  pages = {4201--4205},
  doi = {10.1109/lsp.2026.3744302}
}
Kostenlos starten →