Soundwave: Less is More for Speech-Text Alignment in LLMs

Yuhao Zhang (张裕浩); Zhiheng Liu; Fan Bu; Ruiyu Zhang; Benyou Wang; Haizhou Li

doi:10.18653/v1/2025.acl-long.917

Soundwave: Less is More for Speech-Text Alignment in LLMs

Yuhao Zhang, Zhiheng Liu, Fan Bu, Ruiyu Zhang, Benyou Wang, Haizhou Li

Correct Metadata for

Important: The Anthology treat PDFs as authoritative. Please use this form only to correct data that is out of line with the PDF. See our corrections guidelines if you need to change the PDF.

Title Adjust the title. Retain tags such as <fixed-case>.

Authors Adjust author names and order to match the PDF.

Abstract Correct abstract if needed. Retain XML formatting tags such as <tex-math>.

Verification against PDF Ensure that the new title/authors match the snapshot below. (If there is no snapshot or it is too small, consult the PDF.)

Authors concatenated from the text boxes above:

ALL author names match the snapshot above—including middle initials, hyphens, and accents.

Abstract

Existing end-to-end speech large language models (LLMs) usually rely on large-scale annotated data for training, while data-efficient training has not been discussed in depth. We focus on two fundamental problems between speech and text: the representation space gap and sequence length inconsistency. We propose Soundwave, which utilizes an efficient training strategy and a novel architecture to address these issues. Results show that Soundwave outperforms other advanced speech LLMs in speech translation and AIR-Bench speech tasks with only a fraction of the training data. Further analysis shows that Soundwave still retains its intelligence during conversation.

Anthology ID:: 2025.acl-long.917
Volume:: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:: July
Year:: 2025
Address:: Vienna, Austria
Editors:: Wanxiang Che, Joyce Nabende, Ekaterina Shutova, Mohammad Taher Pilehvar
Venue:: ACL
SIG:
Publisher:: Association for Computational Linguistics
Note:
Pages:: 18718–18738
Language:
URL:: https://aclanthology.org/2025.acl-long.917/
DOI:: 10.18653/v1/2025.acl-long.917
Bibkey:
Cite (ACL):: Yuhao Zhang, Zhiheng Liu, Fan Bu, Ruiyu Zhang, Benyou Wang, and Haizhou Li. 2025. Soundwave: Less is More for Speech-Text Alignment in LLMs. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 18718–18738, Vienna, Austria. Association for Computational Linguistics.
Cite (Informal):: Soundwave: Less is More for Speech-Text Alignment in LLMs (Zhang et al., ACL 2025)
Copy Citation:
PDF:: https://aclanthology.org/2025.acl-long.917.pdf

PDF Cite Search Fix data

Export citation

BibTeX
MODS XML
Endnote
Preformatted

@inproceedings{zhang-etal-2025-soundwave,
    title = "Soundwave: Less is More for Speech-Text Alignment in {LLM}s",
    author = "Zhang, Yuhao  and
      Liu, Zhiheng  and
      Bu, Fan  and
      Zhang, Ruiyu  and
      Wang, Benyou  and
      Li, Haizhou",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2025.acl-long.917/",
    doi = "10.18653/v1/2025.acl-long.917",
    pages = "18718--18738",
    ISBN = "979-8-89176-251-0",
    abstract = "Existing end-to-end speech large language models (LLMs) usually rely on large-scale annotated data for training, while data-efficient training has not been discussed in depth. We focus on two fundamental problems between speech and text: the representation space gap and sequence length inconsistency. We propose Soundwave, which utilizes an efficient training strategy and a novel architecture to address these issues. Results show that Soundwave outperforms other advanced speech LLMs in speech translation and AIR-Bench speech tasks with only a fraction of the training data. Further analysis shows that Soundwave still retains its intelligence during conversation."
}

Download as File

<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhang-etal-2025-soundwave">
    <titleInfo>
        <title>Soundwave: Less is More for Speech-Text Alignment in LLMs</title>
    </titleInfo>
    <name type="personal">
        <namePart type="given">Yuhao</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Zhiheng</namePart>
        <namePart type="family">Liu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Fan</namePart>
        <namePart type="family">Bu</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Ruiyu</namePart>
        <namePart type="family">Zhang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Benyou</namePart>
        <namePart type="family">Wang</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <name type="personal">
        <namePart type="given">Haizhou</namePart>
        <namePart type="family">Li</namePart>
        <role>
            <roleTerm authority="marcrelator" type="text">author</roleTerm>
        </role>
    </name>
    <originInfo>
        <dateIssued>2025-07</dateIssued>
    </originInfo>
    <typeOfResource>text</typeOfResource>
    <relatedItem type="host">
        <titleInfo>
            <title>Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
        </titleInfo>
        <name type="personal">
            <namePart type="given">Wanxiang</namePart>
            <namePart type="family">Che</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Joyce</namePart>
            <namePart type="family">Nabende</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Ekaterina</namePart>
            <namePart type="family">Shutova</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <name type="personal">
            <namePart type="given">Mohammad</namePart>
            <namePart type="given">Taher</namePart>
            <namePart type="family">Pilehvar</namePart>
            <role>
                <roleTerm authority="marcrelator" type="text">editor</roleTerm>
            </role>
        </name>
        <originInfo>
            <publisher>Association for Computational Linguistics</publisher>
            <place>
                <placeTerm type="text">Vienna, Austria</placeTerm>
            </place>
        </originInfo>
        <genre authority="marcgt">conference publication</genre>
        <identifier type="isbn">979-8-89176-251-0</identifier>
    </relatedItem>
    <abstract>Existing end-to-end speech large language models (LLMs) usually rely on large-scale annotated data for training, while data-efficient training has not been discussed in depth. We focus on two fundamental problems between speech and text: the representation space gap and sequence length inconsistency. We propose Soundwave, which utilizes an efficient training strategy and a novel architecture to address these issues. Results show that Soundwave outperforms other advanced speech LLMs in speech translation and AIR-Bench speech tasks with only a fraction of the training data. Further analysis shows that Soundwave still retains its intelligence during conversation.</abstract>
    <identifier type="citekey">zhang-etal-2025-soundwave</identifier>
    <identifier type="doi">10.18653/v1/2025.acl-long.917</identifier>
    <location>
        <url>https://aclanthology.org/2025.acl-long.917/</url>
    </location>
    <part>
        <date>2025-07</date>
        <extent unit="page">
            <start>18718</start>
            <end>18738</end>
        </extent>
    </part>
</mods>
</modsCollection>

Download as File

%0 Conference Proceedings
%T Soundwave: Less is More for Speech-Text Alignment in LLMs
%A Zhang, Yuhao
%A Liu, Zhiheng
%A Bu, Fan
%A Zhang, Ruiyu
%A Wang, Benyou
%A Li, Haizhou
%Y Che, Wanxiang
%Y Nabende, Joyce
%Y Shutova, Ekaterina
%Y Pilehvar, Mohammad Taher
%S Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2025
%8 July
%I Association for Computational Linguistics
%C Vienna, Austria
%@ 979-8-89176-251-0
%F zhang-etal-2025-soundwave
%X Existing end-to-end speech large language models (LLMs) usually rely on large-scale annotated data for training, while data-efficient training has not been discussed in depth. We focus on two fundamental problems between speech and text: the representation space gap and sequence length inconsistency. We propose Soundwave, which utilizes an efficient training strategy and a novel architecture to address these issues. Results show that Soundwave outperforms other advanced speech LLMs in speech translation and AIR-Bench speech tasks with only a fraction of the training data. Further analysis shows that Soundwave still retains its intelligence during conversation.
%R 10.18653/v1/2025.acl-long.917
%U https://aclanthology.org/2025.acl-long.917/
%U https://doi.org/10.18653/v1/2025.acl-long.917
%P 18718-18738

Download as File

Markdown (Informal)

[Soundwave: Less is More for Speech-Text Alignment in LLMs](https://aclanthology.org/2025.acl-long.917/) (Zhang et al., ACL 2025)

Soundwave: Less is More for Speech-Text Alignment in LLMs (Zhang et al., ACL 2025)

ACL

Yuhao Zhang, Zhiheng Liu, Fan Bu, Ruiyu Zhang, Benyou Wang, and Haizhou Li. 2025. Soundwave: Less is More for Speech-Text Alignment in LLMs. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 18718–18738, Vienna, Austria. Association for Computational Linguistics.