@inproceedings{mclaughlin-alm-2026-turnstyle,
title = "{T}urn{S}tyle: Suggesting Style Prompts for Instruction-Guided Text-to-Speech Using Bimodal Dialogue Contexts",
author = "McLaughlin, Jack and
Alm, Cecilia O",
editor = "T.Y.S.S., Santosh and
Ghosh, Sreyan and
Duraiswami, Ramani and
Watanabe, Shinji and
Arora, Siddhant and
Kumar, Sonal and
Adi, Yossi and
Manocha, Dinesh and
Garcia Perera, Leibny Paola",
booktitle = "The Proceedings of Speech and Audio Language Models Workshop (2nd Edition)",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.salma-1.1/",
pages = "1--10",
ISBN = "979-8-89176-496-5",
abstract = "Instruction-guided text-to-speech (TTS) systems are among the most intuitive for controllable speech synthesis, but writing descriptive instructions in the form of style prompts is too cumbersome for users to maintain natural conversation rhythm. We introduce TurnStyle, a prompt generation system, that produces style prompts for instruction-guided TTS, leveraging a bimodal dialogue history (speech and text transcripts) and a user script query. To address the unintuitive interface problem, TurnStyle creates a style prompt by conditioning the Llama-3.2-3B-Instruct model on a prefix vector generated from an embedded bimodal representation of a dialogue. Computational experiments with performance measured across standard metrics like BERTscore and ROUGE-L show that TurnStyle{'}s architecture improved on a comparative baseline model for the understudied task of style prompt generation. The introduced model also consumed over 15 times less energy than the baseline and required less memory, suggesting that the TurnStyle framework may feasibly run on edge devices."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="mclaughlin-alm-2026-turnstyle">
<titleInfo>
<title>TurnStyle: Suggesting Style Prompts for Instruction-Guided Text-to-Speech Using Bimodal Dialogue Contexts</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jack</namePart>
<namePart type="family">McLaughlin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cecilia</namePart>
<namePart type="given">O</namePart>
<namePart type="family">Alm</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>The Proceedings of Speech and Audio Language Models Workshop (2nd Edition)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Santosh</namePart>
<namePart type="family">T.Y.S.S.</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sreyan</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ramani</namePart>
<namePart type="family">Duraiswami</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shinji</namePart>
<namePart type="family">Watanabe</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Siddhant</namePart>
<namePart type="family">Arora</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sonal</namePart>
<namePart type="family">Kumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yossi</namePart>
<namePart type="family">Adi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dinesh</namePart>
<namePart type="family">Manocha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Leibny</namePart>
<namePart type="given">Paola</namePart>
<namePart type="family">Garcia Perera</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-496-5</identifier>
</relatedItem>
<abstract>Instruction-guided text-to-speech (TTS) systems are among the most intuitive for controllable speech synthesis, but writing descriptive instructions in the form of style prompts is too cumbersome for users to maintain natural conversation rhythm. We introduce TurnStyle, a prompt generation system, that produces style prompts for instruction-guided TTS, leveraging a bimodal dialogue history (speech and text transcripts) and a user script query. To address the unintuitive interface problem, TurnStyle creates a style prompt by conditioning the Llama-3.2-3B-Instruct model on a prefix vector generated from an embedded bimodal representation of a dialogue. Computational experiments with performance measured across standard metrics like BERTscore and ROUGE-L show that TurnStyle’s architecture improved on a comparative baseline model for the understudied task of style prompt generation. The introduced model also consumed over 15 times less energy than the baseline and required less memory, suggesting that the TurnStyle framework may feasibly run on edge devices.</abstract>
<identifier type="citekey">mclaughlin-alm-2026-turnstyle</identifier>
<location>
<url>https://aclanthology.org/2026.salma-1.1/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>1</start>
<end>10</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T TurnStyle: Suggesting Style Prompts for Instruction-Guided Text-to-Speech Using Bimodal Dialogue Contexts
%A McLaughlin, Jack
%A Alm, Cecilia O.
%Y T.Y.S.S., Santosh
%Y Ghosh, Sreyan
%Y Duraiswami, Ramani
%Y Watanabe, Shinji
%Y Arora, Siddhant
%Y Kumar, Sonal
%Y Adi, Yossi
%Y Manocha, Dinesh
%Y Garcia Perera, Leibny Paola
%S The Proceedings of Speech and Audio Language Models Workshop (2nd Edition)
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-496-5
%F mclaughlin-alm-2026-turnstyle
%X Instruction-guided text-to-speech (TTS) systems are among the most intuitive for controllable speech synthesis, but writing descriptive instructions in the form of style prompts is too cumbersome for users to maintain natural conversation rhythm. We introduce TurnStyle, a prompt generation system, that produces style prompts for instruction-guided TTS, leveraging a bimodal dialogue history (speech and text transcripts) and a user script query. To address the unintuitive interface problem, TurnStyle creates a style prompt by conditioning the Llama-3.2-3B-Instruct model on a prefix vector generated from an embedded bimodal representation of a dialogue. Computational experiments with performance measured across standard metrics like BERTscore and ROUGE-L show that TurnStyle’s architecture improved on a comparative baseline model for the understudied task of style prompt generation. The introduced model also consumed over 15 times less energy than the baseline and required less memory, suggesting that the TurnStyle framework may feasibly run on edge devices.
%U https://aclanthology.org/2026.salma-1.1/
%P 1-10
Markdown (Informal)
[TurnStyle: Suggesting Style Prompts for Instruction-Guided Text-to-Speech Using Bimodal Dialogue Contexts](https://aclanthology.org/2026.salma-1.1/) (McLaughlin & Alm, SALMA 2026)
ACL