CML-TTS is a recursive acronym for CML-Multi-Lingual-TTS, a Text-to-Speech (TTS) dataset developed at the Center of Excellence in Artificial Intelligence (CEIA) of the Federal University of Goias (UFG). CML-TTS is a dataset comprising audiobooks sourced from the public domain books of Project Gutenberg, read by volunteers from the LibriVox project. The dataset includes recordings in Dutch, German, French, Italian, Polish, Portuguese, and Spanish, all at a sampling rate of 24kHz. The data archives were restructured from the original ones from OpenSLR to make it easier to stream. - text-to-speech, text-to-audio: The dataset can also be used to train a model for Text-To-Speech (TTS). The…
Publicly accessible
cc-by-4.0
1M<n<10M
Chaashini — Hindi/Urdu for sugar syrup — is a continuously growing corpus of clean, single-speaker, studio-grade Indian-language speech built for training speech models (text-to-speech, speech recognition, speech language models). Every clip in the corpus has passed a strict multi-stage quality gate; the aim is purity over volume. The corpus grows automatically: new shards are appended every ~2 hours of newly accepted audio. Audio is sourced from publicly available spoken-word recordings (talks, interviews, narration, lectures, podcasts and similar long-form speech). Each recording then passes through: 1. Source-level screening – recordings dominated by music, singing, or non-speech content…
Access requested at publisher
apache-2.0
1M<n<10M
Incrementally published generated audio and WER, CER, DNSMOS, WavLM-large ECAPA speaker SIM and UTMOS22 measurements. The full campaign is still running. Each generation method is evaluated separately using its publisher's native API. Full evaluations contain all 1,088 English Seed-TTS-Eval targets; eight-target diagnostic pilots are stored separately and must not be treated as full scores. experiments/ / / / contains result.json, records.json, contract.json, verification.json, and audio.tar. records.json contains a rows array. Each successful row specifies its WAV's audioarchive, audiooffset, audiobytes, and audiosha256. Request exactly that byte range from the archive at an immutable…
Publicly accessible
AI-generated multilingual speech using 1,124 synthetic English reference voices. Audio is mono 64kbps CBR Opus in approximately 1GB WebDataset tar shards. Source text: dbdp961202/emilia-translategemma-4b-55-locales (ttsaccepted). Reference voices: dbdp961202/voice-generator-reference-en. Model: openbmb/VoxCPM2. Original reference WAVs are not distributed here. Generation settings and speakermap.json are preserved in metadata/generations/. Audio and manifests are under data/ /part-xxxxx/; generation IDs remain in each sample metadata. Legacy task records are archived under metadata/legacytasks/. Verified legacy runs/ paths have been removed; production data uses the simplified data/ layout.…
Publicly accessible
Synthetic, expressive, multilingual speech for cross-lingual dubbing research. Each example pairs a style-annotated text with generated audio that clones an English - ~1.9M examples in the onespeaker config - 17 languages - ~900 distinct reference speakers Each sample shows the generated audio followed by the English reference voice that conditioned it. All nine subsets below are registered as loadable configurations with a train split. File counts reflect the repository snapshot inspected on 2026-09-26; they are not row counts. Some generation runs were stopped before exhausting their source data, so Step 3 and ambientv4 should not be assumed to cover all rows in earlier versions.…
Access requested at publisher
1M<n<10M
RVCBench Accepted to NeurIPS 2026. RVCBench is a benchmark dataset for studying robustness in voice cloning, text-to-speech, speaker privacy, audio protection, adversarial audio perturbations, and related audio generation pipelines. RVCBench is designed for evaluating how modern voice cloning (VC), TTS, and audio
Publicly accessible
cc0-1.0
W
Dataset · Text to speech
Wynn
Millions of Burmese sentences synthesized to audio using MyanmarTTS — a from-scratch 31M-parameter flow-matching TTS model (CC0 1.0, pip install myanmartts). Derived from freococo/myanmarspokencorpus, which merges Burmese text from FineWeb2, DCAI, DCAD200, and other web corpora. All files are Parquet, each holding ~5,000 rows: CC0 1.0 — public domain.
Publicly accessible
cc0-1.0
1M<n<10M
Dataset · Text to speech
Jerry
A public snapshot of conversational speech research code, synthetic speech data, model checkpoints, manuscripts, and evaluation records. Upload is in progress. Volumes preserve directory paths beneath a chosen restore root. Runtime environments, credentials and Git internals are excluded. Files held for content review are not yet included. Existing upstream licenses still apply; no new blanket license is asserted. Each volume is an independent tar.zst archive and includes a manifest with per-file SHA256 hashes. Verify archive hashes from the upload index, then extract into an empty restore directory. Do not extract directly over an existing installation. Absolute filesystem links require…
Publicly accessible
German in-the-wild speech (talk shows, interviews, podcasts, vlogs and other YouTube content) paired with its transcript, acoustic measurements, discrete German descriptor tags and, for female speakers, a free-text German description of the voice and recording conditions. The dataset is intended for training description-conditioned TTS models such as The data was built as part of research work. It is a random subset of the German part of Emilia-YODAS, processed until the female split reached about 350 hours. Both splits have the same columns; only the female split is described. The split a clip lands in is decided by a gender classifier (see Columns), not by human labels, so a small share…
Access requested at publisher
cc-by-4.0
100K<n<1M
German European Parliament plenary speech paired with a transcript, acoustic measurements, discrete German descriptor tags and a free-text German description of the voice and recording conditions. The dataset is intended for training description-conditioned TTS models such as The data was built as part of research work. It is a random subset of the German unlabelled part of VoxPopuli (release v1, from the VoxPopuli GitHub repository), not the transcribed subset on the Hub (facebook/voxpopuli), which it does not overlap (see below). Most of the audio is simultaneous interpretation. VoxPopuli's German unlabelled data is the Parliament's German audio channel: when a member speaks German it…
Access requested at publisher
cc-by-4.0
100K<n<1M
HV
Dataset · Text to speech
Ha Vu
Kho dữ liệu âm thanh đa ngôn ngữ (Tiếng Việt & Tiếng Anh) chuẩn phòng thu SOTA, được tiền xử lý và thẩm định toàn diện (SQUIM noise assessment, trích xuất vector âm sắc ECAPA-TDNN 192-d, căn chỉnh phụ đề JSON/timestamps, Mel-spectrogram & F0 pitch contours) phục vụ tối ưu cho việc huấn luyện và suy luận mô hình Matcha-TTS kết hợp Vocoder Vocos. Toàn bộ dữ liệu âm thanh lớn được đóng gói thành các Shard tiêu chuẩn 2.5 GB – 3.0 GB (.tar.gz) để tối ưu hoá băng thông, tránh lỗi timeout và tương thích hoàn hảo với giới hạn Git LFS của Hugging Face. Số lượng: 36 vector đặc trưng âm sắc cho tất cả các giọng nói. Định dạng:.npy (NumPy Array), Shape (192,), Dtype float32, chuẩn hoá $L2$-norm = 1.0.…
Publicly accessible
apache-2.0
100K<n<1M
This dataset contains text-and-audio pairs selected for TTS training. The initial audio/uzbek split has 2,184 utterances. Other languages and additional rows can be added as separate, append-only shards. Each row retains the complete structured text specification alongside audio and the OmniVoice fields utt, text, spk, and instruct. ttstaggedtext preserves the local delivery and vocal-event controls separately from the clean text field. For this release, the saved synthesis request matches the selected transcript and speaker, and each WAV passed structural checks. The audio has not been independently transcribed or human-rated; qastatus records that distinction. The matching source…
Access requested at publisher
other
Single-speaker synthetic Tajik speech, generated 2026-09-22..24. Sources: khazina (book readings), law (legal texts), books (book texts). CC-BY-NC-4.0. Research use only. Generated with an OmniVoice LoRA (omnivoice-lora-sl-panther2); OmniVoice is CC-BY-NC, so this corpus inherits the non-commercial restriction. Every manifest row carries licensenote: "CC-BY-NC-derived, research-only, never Sado product". This data must not be used in a commercial product. Manifest fields: clipid, text, wav, duration, refvoice, generator, synthetic, rmsdbfs, normalizationtargetdbfs, peakceilingdbfs, normalizationmethod, licensenote, sourcerow. All audio is loudness-normalised to -16 dBFS RMS with a -1 dBFS…
Access requested at publisher
cc-by-nc-4.0
Узбекская речь для обучения TTS: ~690 часов, 24 kHz, с транскриптами. Две версии аудио с одинаковыми id. Доступ — по заявке, одобряется вручную. Подробное описание — в DETAILS.md (доступно после одобрения).
Access requested at publisher
