Pith. sign in

REVIEW

TridentSE: Guiding Speech Enhancement with 32 Global Tokens

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2210.12995 v1 pith:RVY4EU5F submitted 2022-10-24 eess.AS cs.SD

TridentSE: Guiding Speech Enhancement with 32 Global Tokens

classification eess.AS cs.SD
keywords globalinformationtokenstridentseenhancementspeechcapturedetails
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

In this paper, we present TridentSE, a novel architecture for speech enhancement, which is capable of efficiently capturing both global information and local details. TridentSE maintains T-F bin level representation to capture details, and uses a small number of global tokens to process the global information. Information is propagated between the local and the global representations through cross attention modules. To capture both inter- and intra-frame information, the global tokens are divided into two groups to process along the time and the frequency axis respectively. A metric discriminator is further employed to guide our model to achieve higher perceptual quality. Even with significantly lower computational cost, TridentSE outperforms a variety of previous speech enhancement methods, achieving a PESQ of 3.47 on VoiceBank+DEMAND dataset and a PESQ of 3.44 on DNS no-reverb test set. Visualization shows that the global tokens learn diverse and interpretable global patterns.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.