AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

What Makes a Good Speech Tokenizer for LLM-Centric Speech Generation? A Systematic Study

Xiaoran Fan, Zhichao Sun, Yangfan Gao, Jingfei Xiong, Hang Yan, Yifei Cao, Jiajun Sun, Shuo Li, Zhihao Zhang, Zhiheng Xi, Yuhao Zhou, Senjie Jin and 10 more

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use a decoupled tokenizer (FACodec) instead of coupled tokenizers inside the SLM (ablation)LLM-centric SLM · LibriTTS test-cleanWER: 13.63 -> 6.07Improved
use decoupled FACodec instead of coupled BigCodec despite BigCodec's better reconstruction (ablation)LLM-centric SLM · RoleTriviaQA in-domainSIM (all speakers): 0.21 -> 0.61Improved
use semi-decoupled SpeechTokenizer instead of coupled tokenizers (ablation)LLM-centric SLM · LibriTTS test-cleanUTMOS: 8 codebooks: 4.05Improved
add multi-token prediction with 12 heads to decoupled SLM (ablation)FACodec SLM · LibriTTS test-cleanWER: 6.07 -> 3.01Improved
increase MTP compression from 3 to 12 tokens per hidden state (ablation)FACodec SLM · LibriTTS test-cleanWER: 5.84 -> 3.01Improved
increase MTP compression from 1 to 12 tokens per hidden state (ablation)FACodec SLM · LibriTTS test-cleanUTMOS: 3.93 -> 3.67Hurt
increase MTP compression from 1 to 12 tokens per hidden state (ablation)FACodec SLM · LibriTTS test-cleanspeech synthesis success rate: 0.87 -> 1.00Improved
increase MTP compression from 1 to 12 tokens per hidden state (ablation)FACodec SLM · LibriTTS test-cleanSIM (seen speakers): 0.50 -> 0.47Hurt
add speaker embedding to NTP decoupled SLM (ablation)FACodec NTP SLM · LibriTTS test-cleanWER: 6.07 -> 5.51Improved
add speaker embedding to MTP-3H decoupled SLM (ablation)FACodec MTP-3H SLM · LibriTTS test-cleanWER: 5.84 -> 7.04Hurt
add speaker embedding to MTP-12H decoupled SLM (ablation)FACodec MTP-12H SLM · LibriTTS test-cleanSIM: 0.47 -> 0.60Improved
add speaker embedding to MTP-12H decoupled SLM (ablation)FACodec MTP-12H SLM · LibriTTS test-cleanWER: 3.01 -> 2.94Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy