Speech and audio AI

Publishing authors

Base scenario: 3.3k publishing authors. Paper count × 3.90 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 3,346 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 1,348 papers (74.2%). 12.1% of authorship records lack an Author ID; 11.7% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 3.90; 95% bootstrap interval 3.66–4.17. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $876M.

Internet of things: $14.6B × 5.00%
Ed tech: $1.44B × 10.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

ACL 2026 — research topics

Scope and related research

Voice generation and non-speech audio analysis are separate areas.

Organizations

Selected examples

OpenAI Models
Speech models transcribe audio into text.
Google Cloud Tools & infrastructure
Chirp and Speech-to-Text APIs provide multilingual speech recognition.
Deepgram Product
Voice AI infrastructure.

More organizations (6)

Abridge Product
Clinical documentation.
AssemblyAI Product
Speech recognition.
ELSA Product
AI English speaking tutor.
Inworld AI Product
Speech models and inference APIs.
Nabla Product
Clinical assistant.
Speechmatics Product
Speech recognition.

Speech generation and voice AI

Text-to-speech, voice conversion, expressive speech, and speech-to-speech models.

TTS Text-to-speech Speech-to-speech

2,193 papers

1.5× vs. 2023

Authors · scenario: 3.3k–7.9k
Investment · scenario: $282M–$1.13B

Sources and methodology

Publication activity

2019: 831
2020: 1,042
2021: 1,133
2022: 1,240
2023: 1,417
2024: 1,644
2025: 2,193

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("text to speech" OR "speech synthesis" OR "voice conversion" OR "speech to speech")

Retrieved 2026-09-26.

Publishing authors

Base scenario: 3.3k publishing authors. Paper count × 3.58 known authors per paper ÷ 3 papers per author per year. The range assumes 1–5 papers per author; the base and both ends of the range are at least the 3,304 distinct Author IDs already observed in the sample.

The publication rate is an assumption, not a measured rate for this field. The range is a scenario, not a confidence interval.

Random sample: 1,000 of 2,193 papers (45.6%). 12.6% of authorship records lack an Author ID; 9.8% of retrieved works lack an abstract. Retrieved 2026-09-26.

Mean known authors per paper: 3.58; 95% bootstrap interval 3.37–3.83. This describes sampling variability, not search accuracy or uncertainty in the number of researchers.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $563M.

Creative, music, video content: $4.19B × 10.00%
Ed tech: $1.44B × 10.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

ACL 2026 — research topics

Scope and related research

Focuses on speech; music generation is listed separately.

Organizations

Selected examples

ElevenLabs Product
Speech and voice agents.
OpenAI Models
Text-to-speech models generate expressive spoken audio.
Cartesia Product
Real-time audio models.

More organizations (3)

ByteDance Seed Models
Multimodal model research.
Deepgram Product
Voice AI infrastructure.
Inworld AI Product
Speech models and inference APIs.

Music AI

Music generation, transcription, structural analysis, and music retrieval.

Music generation Music retrieval

525 papers

1.7× vs. 2023

Authors · observed: 1,311
Investment · scenario: $392M–$1.57B

Sources and methodology

Publication activity

2019: 144
2020: 208
2021: 225
2022: 256
2023: 317
2024: 409
2025: 525

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("music generation" OR "automatic music transcription" OR "music information retrieval" OR "music understanding" OR "music source separation")

Retrieved 2026-09-26.

Publishing authors

1,311 distinct OpenAlex Author IDs across all 525 matching papers. This is an observed count within the query result; no publication-rate assumption is used.

Full query result: 525 of 525 papers (100.0%). 17.1% of authorship records lack an Author ID; 8.0% of retrieved works lack an abstract. Retrieved 2026-09-26.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $785M.

Creative, music, video content: $4.19B × 10.00%
Entertainment: $3.66B × 10.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.

SIGIR-AP 2026 — research topics

Scope and related research

Groups tasks that work with musical signals, across both analysis and generation.

Organizations

Selected examples

Google DeepMind Models
Lyria models generate music and audio.
Suno Product
Music generation.
Udio Product
Music generation.

More organizations (3)

AudioShake Product
Audio source separation.
Boomy Product
AI music.
SOUNDRAW Product
AI music.

Audio perception and processing

Sound events, speaker recognition, diarization, source separation, enhancement, and spatial audio.

Diarization Audio separation Spatial audio

956 papers

0.9× vs. 2023

Authors · observed: 2,958
Investment · scenario: $732M–$2.93B

Sources and methodology

Publication activity

2019: 628
2020: 786
2021: 821
2022: 901
2023: 1,010
2024: 1,044
2025: 956

Matches include papers applying these methods. Search precision and recall have not been systematically measured.

Title and abstract matches for articles, preprints and reviews; retracted work is excluded. Versions and overlapping areas may be counted more than once.

("sound event detection" OR "audio classification" OR "speaker recognition" OR "speaker diarization" OR "speech enhancement" OR "audio source separation" OR "acoustic scene classification")

Retrieved 2026-09-26.

Publishing authors

2,958 distinct OpenAlex Author IDs across all 956 matching papers. This is an observed count within the query result; no publication-rate assumption is used.

Full query result: 956 of 956 papers (100.0%). 10.2% of authorship records lack an Author ID; 19.7% of retrieved works lack an abstract. Retrieved 2026-09-26.

Applied coauthors are included. Missing or incorrectly linked author records affect these figures. They do not measure research jobs or everyone working in a field.

Private investment

An assumed allocation of broader investment segments. The range is 0.5–2× the base case. Shares are editorial assumptions, not measured deals or research spending.

Base case: $1.46B.

Internet of things: $14.6B × 10.00%

Company investment, not revenue or research spending. Ranges are scenarios, not confidence intervals.