Pith. sign in

REVIEW 4 major objections 5 minor 31 references

Phir Hera Fairy: An English Fairytaler is a Strong Faker of Fluent Speech in Low-Resource Indian Languages

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Fine-tuning the English F5-TTS model on a small Indian-language dataset—about 1.4% of the English training data—produces a synthesizer that approaches human-level naturalness across 11 languages, gains polyglot, voice-cloning, code-mixed…

desk verdict Fine-tuning English F5 beats from-scratch and mixed training, but the 'human parity' claim is built on noisy references. read the letter →

arxiv 2505.20693 v1 pith:6MQSOVE4 submitted 2025-05-27 cs.CL cs.LGcs.SDeess.AS

Praveen Srinivasa Varadhan , Srija Anand , Soma Siddhartha , Mitesh M.Khapra This is my paper
classification cs.CLcs.LGcs.SDeess.AS
keywords speechsynthesislow-resourceTTSIndianlanguagesmultilingualzero-resourcetransferlearningvoicecloningcode-mixing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper asks whether a large English text-to-speech model can serve as a ready-made starting point for languages that lack data. It fine-tunes the English F5-TTS checkpoint on 1,417 hours of speech across 11 Indian languages—about 1.4% of the English pretraining data—and reports that the resulting IN-F5 model approaches or matches human recordings in naturalness, exceeds them on voice-cloning similarity, and gains polyglot, code-mixed, and expressive speech abilities. The paper also claims that this setup transfers to unseen languages: an hour of synthetically generated, native-verified speech in Bhojpuri or Tulu is enough to synthesize those languages at high quality. If correct, the result matters because low-resource TTS would no longer require hundreds of hours of studio data; a single strong English checkpoint plus a small amount of curated audio could bootstrap many languages at once.

What carries the argument

The object that carries the argument is IN-F5: the English F5-TTS checkpoint, a flow-matching text-to-speech model trained on roughly 100K hours of English, after direct fine-tuning on the IN11 corpus of 11 Indian languages. Adaptation works by expanding the token vocabulary to 685 characters spanning Indian scripts and initializing the new embeddings by random draws from the English embedding space, avoiding phoneme models. The data-scaling experiments show the recipe degrades gracefully; 10 hours per language keeps almost all emergent behaviour, while 1 hour does not. For unseen languages the mechanism is cross-lingual transfer through shared scripts—Bhojpuri through Devanagari and Tulu through the Kannada script—combined with synthetic speech generation from a related-language voice, native-speaker filtering, and self-training on the validated hour.

What would settle it

Re-run the same MUSHRA protocol with studio-quality, noise-free human reference recordings and a larger panel of native raters; if IN-F5's naturalness and speaker-similarity scores no longer reach or exceed the human references, the human-parity and zero-resource claims lose their support.

Watch

Extended reading notes

Core claim

The central claim is that "simply initializing from a large pretrained English model and fine-tuning on a small set of IN11 data (1.4% of EN data) is sufficient to reach human-level synthesis in Indian languages." Direct fine-tuning on Indian data alone (EN→IN) is the best of three strategies, scoring an overall MUSHRA of 73.4 against 43.2 for training from scratch and 66.2 for mixed English-Indian fine-tuning; on seen-speaker naturalness it scores 78.0 versus 75.9 for human recordings. The same model handles unseen speakers, polyglot speech across language families, code-mixing between pairs such as Hindi-Bengali and Kannada-Telugu, and six expressive styles. For zero-resource languages, IN-F5 synthesizes Tulu (93.6 MUSHRA) and Bhojpuri (82.0) without training data in those languages, using a donor voice from a script-sharing language, native-speaker validation, and self-training; in the simulated Bhojpuri setting it surpasses the human reference of 67.1. Benchmarking on the Rasa test set places IN-F5 at 80.5 MUSHRA, 8 points above VoiceCraft, making it the first Indian TTS system in the "Excellent" range.

Load-bearing premise

The human-parity result rests on MUSHRA ratings in which the human reference recordings may contain noise, breath sounds, or microphone artifacts; if raters systematically prefer cleaner audio, IN-F5's scores above human recordings reflect reference quality rather than true human-level synthesis.

Editorial extensions

If this is right

  • Direct fine-tuning on Indian data alone beats mixed English-Indian fine-tuning and training from scratch, so the recommended recipe is to ignore English during fine-tuning and keep the original checkpoint for English use.
  • Ten hours of clean speech per language is a practical data floor: it retains voice cloning, polyglot fluency, and code-mixing with only a 0.8% average score drop relative to 100 hours, whereas 1 hour collapses intelligibility.
  • Unseen languages that share a script with a trained language can be synthesized with one hour of validated synthetic data, which means zero-resource TTS is reachable for script-sharing languages.
  • IN-F5 is the first Indian TTS system to enter the "Excellent" MUSHRA range on the Rasa test set, giving a new topline of 80.5 against 73.0 for VoiceCraft.
  • Code-mixed speech at near-human intelligibility, including unusual pairs like Punjabi-Telugu, is within reach, which directly addresses multilingual everyday use in India.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension is to test the same 1.4%-data recipe outside India, on other script-sharing language families (for example, languages sharing Cyrillic or Arabic scripts); the only new cost would be the native-speaker validation step.
  • Because EN→IN beats EN→EN+IN, the implied product architecture is a single strong English foundation checkpoint plus small per-language adapters, rather than one ever-growing multilingual model.
  • The margin over human recordings in voice-cloning scores may depend on the recording quality of the human references; a controlled comparison against studio-clean references would separate genuine synthesis gains from a preference for noise-free audio.
  • The zero-resource recipe's success likely hinges on the choice of donor voice—the authors deliberately select an expressive Maithili speaker for Bhojpuri—so matching donor expressiveness to the target language is a testable design variable.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper adapts the English F5-TTS model to 11 Indian languages, comparing three training strategies: training from scratch, direct fine-tuning on Indian data, and fine-tuning on Indian plus English data. It reports that direct fine-tuning performs best, and uses this model (IN-F5) to study voice cloning, polyglot synthesis, code-mixing, expressive speech, data-scaling effects, and a zero-resource recipe for Bhojpuri and Tulu. The central claims are that English pretraining provides a strong prior enabling human-level synthesis in low-resource Indian languages, and that a human-in-the-loop synthetic-data recipe can extend TTS to unseen languages.

Significance. If the central comparison holds, the paper makes a practical contribution: it demonstrates that a large English TTS checkpoint can be adapted with relatively small Indian-language data, it analyzes how performance scales with data, and it releases models and validated datasets for underserved languages. The relative ranking of training strategies (EN-to-IN better than EN-to-EN+IN and from-scratch) is defensible from Table 1, and the documentation of emergent behaviors is useful. However, the headline 'human parity' claim is not currently supported by the evaluation design, because the human reference recordings in the key comparison are not studio quality and the authors themselves attribute the model's higher scores to the cleanliness of synthetic audio. The zero-resource results also rest on a single expert rater per language. These are load-bearing issues that require re-analysis or reframing rather than mere editing.

major comments (4)
  1. [§4.1, Table 1] The claim that EN→IN 'reaches human-level synthesis' is not supported by the data as presented. In the voice-cloning condition, IN-F5 scores 78.0/76.6 versus human recordings at 75.9/74.6, but these human recordings are not studio quality; the same table reports studio-quality human naturalness of 89.7, against which IN-F5 scores 77.6, a 12.1-point gap. Section 4.2 concedes that the human references 'may contain subtle background noise, breath sounds, or microphone artifacts' and that IN-F5 generates 'clean, noise-free speech,' and Section 3.3 describes a no-mentioned-reference MUSHRA variant in which raters are not anchored to a labeled reference. The observed margins therefore measure a preference for cleaner audio over degraded references, not parity with high-quality human speech. Please either reframe the claim as 'comparable to non-studio human recordings in the voice-cloning condition' or add a studio-quality human control to the same MUSHRA condition.
  2. [§4.4, Table 5] The zero-resource results for Bhojpuri and Tulu are each based on MUSHRA scores from a single native-speaker expert, as the authors acknowledge. A single rater cannot support quantitative comparisons such as 82.0 vs. 67.1 or the 82.0-to-83.9 self-training improvement, because no inter-rater agreement or score distribution is available. Please present these as pilot or case-study evidence, report utterance-level scores and rater reliability, or collect additional raters; the current framing overstates the strength of the zero-resource conclusion.
  3. [§2.4, §4.4] The zero-resource recipe fine-tunes IN-F5 on its own synthetic outputs and then evaluates the resulting model on the same languages, which creates a circularity risk for the post-self-training scores. The manuscript does not state whether the MUSHRA test utterances overlap with the generated training samples, and it provides no external intelligibility measure (e.g., ASR WER or human transcription) for Bhojpuri or Tulu. Please specify the train/eval split and add a non-circular intelligibility check to support the claim that self-training improves quality.
  4. [§4.5, Table 6] The SOTA comparison on the Rasa test set reports Human MUSHRA 91.8 and IN-F5 80.5, which is inconsistent with the abstract's 'human parity' statement. If the human reference in Table 6 is studio quality, then the proper summary of the paper is that IN-F5 is a strong system that approaches but does not reach studio-quality human speech. The headline should be reconciled with this table, either by qualifying the notion of 'human-level' or by explaining why the two human references differ so substantially.
minor comments (5)
  1. [§3.3] 'atleast' should be 'at least'.
  2. [§4.3] The phrase 'average reduction of 0.8% in performance' is not defined; Table 4 shows a MUSHRA drop from 64.3 to 61.5 (about 4.4% relative) and an improvement in WER-M, so please specify which metrics are averaged and whether the percentage is absolute or relative.
  3. [§3.1] The name 'IN11-Test-Set' is inconsistently formatted, and the description of the 1100 held-out utterances would benefit from a per-language breakdown.
  4. [References] Reference [13] is a self-citation to the authors' own MUSHRA variant; consider also citing the ITU-R BS.1534 standard for the original MUSHRA protocol.
  5. [Title] The title pun 'Phir Hera Fairy' is not transparent to non-Hindi readers; a descriptive subtitle would improve accessibility.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the results are empirical comparisons with independent external baselines; the main risk is an evaluation confound rather than a circular derivation.

full rationale

The paper does not present a derivation chain; it reports fine-tuning experiments and MUSHRA evaluations. The central comparisons are against human recordings and against external systems (VoiceCraft, FastSpeech2-HS), so the SOTA claim has independent content. The zero-resource recipe uses the model's own synthetic outputs for self-training, but the initial zero-shot scores (82.0/93.6) are measured before self-training and the final scores are human-rated; this is a self-training loop, not a by-construction equivalence. The paper's own limitation statements (single-expert evaluation, degraded human references, no-mentioned-reference MUSHRA from [13]) point to evaluation risks that could inflate the 'human parity' claim, but they do not make the claim equivalent to its inputs by definition. Self-citations to [13], [16], and [19] provide the evaluation protocol and datasets but are not used to prove the core result via a self-referential theorem. Hence no significant circularity.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

This is an empirical adaptation study, not a derivation. The central claims rest on measured performance, so the ledger records the key modeling and measurement assumptions the experiments rely on, plus hand-chosen budgets. No new theoretical entities or fitted parameters are introduced, but several evaluation choices (MUSHRA protocol, single-expert zero-resource ratings, data budgets) carry the conclusions.

free parameters (3)
  • Fine-tuning step budget = 150K steps
    All fine-tuning runs use up to 150K steps (Sections 2.3 and 3.2). The value is a hand-chosen budget, not derived from the data or optimized against the metrics.
  • Learning rate = 5e-05
    AdamW learning rate of 5e-5 used throughout (Section 3.2); a standard fine-tuning value chosen by the authors.
  • Zero-resource synthetic data budget = 1 hour per language
    The self-training recipe generates and validates one hour of synthetic speech per zero-resource language (Section 2.4). The amount is fixed by convention; no ablation studies this budget.
assumptions (5)
  • domain assumption English pretraining on roughly 100K hours provides a transferable prior for Indian-language speech synthesis.
    This is the premise of the EN-to-IN strategy (Sections 2.2, 4.1). It is tested comparatively against training from scratch but is not independently established.
  • domain assumption Indian languages are sufficiently phonetic that character-based modeling without grapheme-to-phoneme conversion is adequate.
    Stated in Section 2.1, citing prior work [14]; drives the vocabulary design of 685 character tokens.
  • domain assumption The MUSHRA variant with no-mentioned-reference and the new MUSHRA-S and MUSHRA-I scales yield valid measures of naturalness, speaker similarity, and intelligibility.
    All human-parity and 'surpasses human' claims are read from these scales (Section 3.3); the anchors are defined by the authors.
  • domain assumption Shared scripts and related languages enable cross-lingual transfer for zero-resource synthesis.
    The zero-resource recipe maps Tulu to Kannada script and Bhojpuri to Devanagari and assumes transfer through these shared representations (Section 2.4).
  • domain assumption IndicConformer WER and WavLM embeddings validly measure intelligibility and speaker similarity on synthesized, accented, and code-mixed speech.
    Objective metrics in Section 3.3; ASR WER can be brittle on non-native and code-mixed TTS output.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Phir Hera Fairy: An English Fairytaler is a Strong Faker of Fluent Speech in Low-Resource Indian Languages." pith.science (2026). https://pith.science/paper/6MQSOVE4

@misc{pith2026250520693,
  author       = {Pith},
  title        = {Pith review of: Phir Hera Fairy: An English Fairytaler is a Strong Faker of Fluent Speech in Low-Resource Indian Languages},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6MQSOVE4}},
  note         = {Machine review of arXiv:2505.20693}
}
read the original abstract

What happens when an English Fairytaler is fine-tuned on Indian languages? We evaluate how the English F5-TTS model adapts to 11 Indian languages, measuring polyglot fluency, voice-cloning, style-cloning, and code-mixing. We compare: (i) training from scratch, (ii) fine-tuning English F5 on Indian data, and (iii) fine-tuning on both Indian and English data to prevent forgetting. Fine-tuning with only Indian data proves most effective and the resultant IN-F5 is a near-human polyglot; that enables speakers of one language (e.g., Odia) to fluently speak in another (e.g., Hindi). Our results show English pretraining aids low-resource TTS in reaching human parity. To aid progress in other low-resource languages, we study data-constrained setups and arrive at a compute optimal strategy. Finally, we show IN-F5 can synthesize unseen languages like Bhojpuri and Tulu using a human-in-the-loop approach for zero-resource TTS via synthetic data generation.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

31 extracted references · 13 canonical work pages

  1. [1]

    These breakthroughs are largely driven by the scaling of both data and model size

    Introduction Text-to-Speech (TTS) synthesis has seen remarkable advance- ments, particularly in English, where state-of-the-art (SOTA) models now approach human parity in naturalness [1, 2, 3], speaker adaptation [4, 5, 6], and expressiveness [7, 8, 9]. These breakthroughs are largely driven by the scaling of both data and model size. For instance, the F5...

    work page

  2. [2]

    To un- derstand effects of scaling data, we ablate fine-tuning on 1, 10, and 100-hour splits per language

    Methodology In this section, we describe how we adapt F5-TTS for Indian languages by expanding its vocabulary and evaluating different training strategies for achieving human-like synthesis. To un- derstand effects of scaling data, we ablate fine-tuning on 1, 10, and 100-hour splits per language. Finally, we propose a zero- resource TTS recipe, leveraging...

    arXiv 2025

  3. [3]

    Experimental Setup 3.1. Datasets IN11.We compile a diverse and representative speech-text dataset for 11 Indian Languages, totaling 1417 hours across Assamese, Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Odia, Punjabi, Tamil, and Telugu. To ensure high- quality synthesis, we incorporate studio-quality speech from In- dicTTS [20], LIMMITS [17], ...

    work page 2000

  4. [4]

    Results We present our findings on the adaptation of F5 to Indian lan- guages, examining fine-tuning strategy, emergent behaviors, data scaling effects, and performance against prior baselines. 4.1. Are English speech foundation models good priors for multilingual adaptation? We present the overall MUSHRA scores for the three fine- tuning strategies in Ta...

    work page

  5. [5]

    Excellent

    While performance drops for more challenging combina- tions, such as Assamese-Telugu (52.5), this remains a signif- icant milestone, as even human speakers may struggle with such pairings. Moreover, IN-F5 performs exceptionally well in highly spoken languages such as Hindi (hi-X: 85.5), and Tamil (ta-X: 76.7), showingintelligibility scoresthat are compara...

    work page

  6. [6]

    Conclusion We establish IN-F5 as a state-of-the-art TTS system for Indian languages, demonstrating that fine-tuning a large-scale English model enables high-quality speech synthesis while unlocking polyglot fluency, voice cloning, and code-mixing. Contrary to expectations, fine-tuning without English yields the most natu- ral adaptation, We also realize z...

    work page

  7. [7]

    Maskgct: Zero-shot text-to-speech with masked generative codec transformer,

    Y . Wang, H. Zhan, L. Liu, R. Zeng, H. Guo, J. Zheng, Q. Zhang, S. Zhang, and Z. Wu, “Maskgct: Zero-shot text-to-speech with masked generative codec transformer,” CoRR, vol. abs/2409.00750, 2024. [Online]. Available: https: //doi.org/10.48550/arXiv.2409.00750

    arXiv doi:10.48550/arxiv.2409.00750 2024

  8. [8]

    Naturalspeech: End-to-end text-to-speech synthesis with human-level quality,

    X. Tan, J. Chen, H. Liu, J. Cong, C. Zhang, Y . Liu, X. Wang, Y . Leng, Y . Yi, L. He, S. Zhao, T. Qin, F. K. Soong, and T. Liu, “Naturalspeech: End-to-end text-to-speech synthesis with human-level quality,”IEEE Trans. Pattern Anal. Mach. Intell., vol. 46, no. 6, pp. 4234–4245, 2024. [Online]. Available: https://doi.org/10.1109/TPAMI.2024.3356232

    arXiv 2024

Show all 31 references
  1. [9]

    Seed-tts: A family of high-quality versatile speech generation models,

    P. Anastassiou, J. Chen, J. Chen, Y . Chen, Z. Chen, Z. Chen, J. Cong, L. Deng, C. Ding, L. Gaoet al., “Seed-tts: A family of high-quality versatile speech generation models,”arXiv preprint arXiv:2406.02430, 2024

    2024 arXiv

  2. [10]

    Naturalspeech 2: Latent diffusion models are natural and zero-shot speech and singing synthesizers,

    K. Shen, Z. Ju, X. Tan, E. Liu, Y . Leng, L. He, T. Qin, sheng zhao, and J. Bian, “Naturalspeech 2: Latent diffusion models are natural and zero-shot speech and singing synthesizers,” inThe Twelfth International Conference on Learning Representations,

  3. [11]

    Yourtts: Towards zero-shot multi-speaker tts and zero-shot voice conversion for everyone,

    E. Casanova, J. Weber, C. D. Shulby, A. C. Junior, E. G ¨olge, and M. A. Ponti, “Yourtts: Towards zero-shot multi-speaker tts and zero-shot voice conversion for everyone,” inInternational Con- ference on Machine Learning. PMLR, 2022, pp. 2709–2720

    2022

  4. [12]

    V oicecraft: Zero-shot speech editing and text-to-speech in the wild,

    P. Peng, P. Huang, S. Li, A. Mohamed, and D. Harwath, “V oicecraft: Zero-shot speech editing and text-to-speech in the wild,” inProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok, Thailand, August ...

    2024 doi

  5. [13]

    Soundstorm: Efficient parallel audio gen- eration,

    Z. Borsos, M. Sharifi, D. Vincent, E. Kharitonov, N. Zeghidour, and M. Tagliasacchi, “Soundstorm: Efficient parallel audio gen- eration,”arXiv preprint arXiv:2305.09636, 2023

    2023 arXiv

  6. [14]

    Naturalspeech 3: Zero-shot speech synthesis with factorized codec and diffusion models,

    Z. Ju, Y . Wang, K. Shen, X. Tan, D. Xin, D. Yang, Y . Liu, Y . Leng, K. Song, S. Tang, Z. Wu, T. Qin, X.-Y . Li, W. Ye, S. Zhang, J. Bian, L. He, J. Li, and S. Zhao, “Naturalspeech 3: Zero-shot speech synthesis with factorized codec and diffusion models,” 2024

    2024

  7. [15]

    V oicebox: Text-guided multilingual universal speech generation at scale,

    M. Le, A. Vyas, B. Shi, B. Karrer, L. Sari, R. Moritz, M. Williamson, V . Manohar, Y . Adi, J. Mahadeokaret al., “V oicebox: Text-guided multilingual universal speech generation at scale,”Advances in neural information processing systems, vol. 36, 2024

    2024

  8. [16]

    Cosyvoice 2: Scalable stream- ing speech synthesis with large language models,

    Z. Du, Y . Wang, Q. Chen, X. Shi, X. Lv, T. Zhao, Z. Gao, Y . Yang, C. Gao, H. Wanget al., “Cosyvoice 2: Scalable stream- ing speech synthesis with large language models,”arXiv preprint arXiv:2412.10117, 2024

    2024 arXiv

  9. [17]

    F5-tts: A fairytaler that fakes fluent and faithful speech with flow matching,

    Y . Chen, Z. Niu, Z. Ma, K. Deng, C. Wang, J. Zhao, K. Yu, and X. Chen, “F5-tts: A fairytaler that fakes fluent and faithful speech with flow matching,”arXiv preprint arXiv:2410.06885, 2024

    2024 arXiv

  10. [18]

    A Tulu resource for machine translation,

    M. Narayanan and N. Aepli, “A Tulu resource for machine translation,” inProceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), N. Calzolari, M.-Y . Kan, V . Hoste, A. Lenci, S. Sakti, and N. Xu...

    2024

  11. [19]

    Towards building text-to-speech systems for the next billion users,

    G. K. Kumar, V . PraveenS., P. Kumar, M. M. Khapra, and K. Nandakumar, “Towards building text-to-speech systems for the next billion users,”ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 1–5, 2022. [Online]. Available:...

    2023

  12. [20]

    Rethinking mushra: Addressing modern challenges in text- to-speech evaluation,

    P. S. Varadhan, A. Gulati, A. Sankar, S. Anand, A. Gupta, A. Mukherjee, S. K. Marepally, A. Bhatia, S. Jaju, S. Bhooshan et al., “Rethinking mushra: Addressing modern challenges in text- to-speech evaluation,”arXiv preprint arXiv:2411.12719, 2024

    2024 arXiv

  13. [21]

    Enhancing out-of-vocabulary performance of indian tts systems for practical applications through low-effort data strate- gies,

    S. Anand, P. Srinivasa Varadhan, A. Sankar, G. Raju, and M. M. Khapra, “Enhancing out-of-vocabulary performance of indian tts systems for practical applications through low-effort data strate- gies,” inInterspeech 2024, 2024, pp. 1200–1204

    2024

  14. [22]

    Emilia: An extensive, multilingual, and diverse speech dataset for large-scale speech generation,

    H. He, Z. Shang, C. Wang, X. Li, Y . Gu, H. Hua, L. Liu, C. Yang, J. Li, P. Shiet al., “Emilia: An extensive, multilingual, and diverse speech dataset for large-scale speech generation,” in2024 IEEE Spoken Language Technology Workshop (SLT). IEEE, 2024, pp. 885–890

    2024

  15. [23]

    Indicvoices-r: Unlocking a massive multilingual multi-speaker speech corpus for scaling indian TTS,

    A. Sankar, S. Anand, P. S. Varadhan, S. Thomas, M. Singal, S. Kumar, D. Mehendale, A. Krishana, G. Raju, and M. M. Khapra, “Indicvoices-r: Unlocking a massive multilingual multi-speaker speech corpus for scaling indian TTS,”NeurIPS 2024 Datasets and Benchmarks track, Vancouver...

    2024 doi arXiv

  16. [24]

    Limmits’24: Multi- speaker, multi-lingual indic tts with voice cloning,

    A. Singh, A. Nagireddi, D. G, J. Bandekar, R. R, S. Badiger, S. Udupa, P. K. Ghosh, H. A. Murthy, P. Kumar, K. Tokuda, M. Hasegawa-Johnson, and P. Olbrich, “Limmits’24: Multi- speaker, multi-lingual indic tts with voice cloning,” in2024 IEEE International Conference on Acousti...

    2024

  17. [26]

    Rasa: Building Expressive Speech Synthesis Systems for Indian Lan- guages in Low-resource Settings,

    P. S. Varadhan, A. Sankar, G. Raju, and M. M. Khapra, “Rasa: Building Expressive Speech Synthesis Systems for Indian Lan- guages in Low-resource Settings,” inProc. INTERSPEECH 2024, 2024

    2024

  18. [27]

    Re- sources for indian languages,

    A. Baby, A. L. Thomas, N. Nishanthi, T. Consortiumet al., “Re- sources for indian languages,” inProceedings of Text, Speech and Dialogue, 2016

    2016

  19. [28]

    Crowdsourcing speech data for low-resource languages from low-income workers,

    B. Abraham, D. Goel, D. Siddarth, K. Bali, M. Chopra, M. Choudhury, P. Joshi, P. Jyoti, S. Sitaram, and V . Seshadri, “Crowdsourcing speech data for low-resource languages from low-income workers,” inProceedings of the 12th Language Re- sources and Evaluation Conference, 2020,...

    2020

  20. [29]

    Indicvoices: Towards building an inclusive mul- tilingual speech dataset for indian languages,

    T. Javed, J. A. Nawale, E. I. George, S. Joshi, K. S. Bhogale, D. Mehendale, I. V . Sethi, A. Ananthanarayanan, H. Faquih, P. Palitet al., “Indicvoices: Towards building an inclusive mul- tilingual speech dataset for indian languages,”arXiv preprint arXiv:2403.01926, 2024

    2024 arXiv

  21. [30]

    Wavlm: Large-scale self-supervised pre-training for full stack speech processing,

    S. Chen, C. Wang, Z. Chen, Y . Wu, S. Liu, Z. Chen, J. Li, N. Kanda, T. Yoshioka, X. Xiao, J. Wu, L. Zhou, S. Ren, Y . Qian, Y . Qian, J. Wu, M. Zeng, X. Yu, and F. Wei, “Wavlm: Large-scale self-supervised pre-training for full stack speech processing,”IEEE J. Sel. Top. Signal...

    2022

  22. [31]

    Towards developing state-of-the-art TTS synthesisers for 13 indian languages with signal processing aided alignments,

    A. Prakash, S. Umesh, and H. A. Murthy, “Towards developing state-of-the-art TTS synthesisers for 13 indian languages with signal processing aided alignments,” inIEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2023, Taipei, Taiwan, December 16-20, 2023. IEEE...

    2023

  23. [2024]

    Available: https://openreview.net/forum?id=Rc 7dAwVL3v

    [Online]. Available: https://openreview.net/forum?id=Rc 7dAwVL3v

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.