AI-Made Speech Teaches Machines to Hear Estonia and Slovenia

Automatic speech recognition has become astonishingly good for English, Spanish, and Mandarin, largely because those languages command hundreds of thousands of hours of transcribed audio. For smaller languages, the picture is far bleaker. Estonian and Slovenian, two morphologically intricate European languages, have only a few crowd-sourced hours of labeled speech available in public corpora, and their recognition accuracy lags well behind that of well-resourced languages. A new study published in Machine Learning with Applications tackles this gap with a strikingly modern recipe: generate novel training sentences with a large language model, validate them with a second language-model pass, synthesize them into roughly twenty-one hours of artificial speech, and fine-tune two competing speech recognition architectures on the resulting mix. The results show that synthetic data helps every model tested, but also that the two architectures fail in characteristically different ways, a distinction that a single word error rate figure conceals.

The research team, led by Yuriy Perezhohin with Mauro Castelli and Aleš Popovič, fine-tuned two multilingual foundation models on identical data for each language. The first was Parakeet-TDT-0.6B-v3, a compact 0.6-billion-parameter transducer built on a FastConformer encoder that decodes speech frame by frame, predicting each token and the number of audio frames to skip before the next one. The second was Whisper-large-v3, OpenAI’s 1.55-billion-parameter encoder-decoder Transformer, whose autoregressive decoder attends over the full encoded audio and conditions every token on the complete history of what it has already written, effectively acting as an audio-conditioned language model. Both models had seen Estonian and Slovenian during pretraining, so neither entered the comparison blind, which distinguishes this study from earlier architecture comparisons in English and Polish where systems were trained from scratch or paired asymmetrically.

The synthetic data pipeline began with GPT-5-mini generating roughly six thousand candidate sentences per language across three diversity categories: paraphrases of existing CommonVoice transcripts, domain-expanded sentences introducing vocabulary absent from the training distribution, and morphology-targeted sentences engineered to stress rare Estonian case forms and Slovenian dual-number constructions. A separate LLM-as-judge validation call then screened each candidate for grammaticality, naturalness, appropriate length, and language purity, passing 68.3 percent of Estonian and 74.5 percent of Slovenian candidates on the first attempt. Failed items were regenerated with corrective feedback, recovering more than ninety percent of them, and near-duplicates were removed using word-level Jaccard similarity. The final validated corpora comprised 5,852 Estonian and 5,848 Slovenian sentences, synthesized with eleven distinct voices into approximately 10.4 and 10.3 hours of audio respectively.

The gains were unambiguous in-domain. Fine-tuning on CommonVoice alone left Parakeet at a raw word error rate of 22.35 percent on Estonian and 14.08 percent on Slovenian; adding the full synthetic corpus pushed these down to 21.03 and 11.56 percent respectively. Every one of the sixteen comparisons between real-only and real-plus-synthetic training was statistically significant under paired bootstrap resampling with one hundred thousand iterations, at p less than ten to the minus five. Notably, Parakeet, with 2.5 times fewer parameters than Whisper, achieved the lower word error rate in every fine-tuned condition and on every test set under each model’s standard recipe. Before fine-tuning, the two systems had divided the spoils: Whisper dominated Slovenian zero-shot at 21.20 percent versus Parakeet’s 50.23 percent, while Parakeet led on Estonian at 27.19 versus 34.40 percent. Fine-tuning erased that division entirely.

Out-of-domain behavior told a sharper story. Parakeet improved on the FLEURS benchmark in every fine-tuned condition for both languages. Whisper did not. Fully fine-tuning Whisper on the 1,388 Slovenian CommonVoice sentences raised its FLEURS word error rate from 37.02 to 46.79 percent, a 9.8-point degradation relative to leaving the model untouched. The synthetic mix recovered part of this loss, with the best augmented configuration reaching 40.46 percent, but no configuration returned the model to its zero-shot level. The authors attribute this partly to a register mismatch: the Slovenian training sentences average 5.9 words in a conversational style, while the Slovenian FLEURS references average 19.7 words of encyclopedic prose. Estonian, whose training and evaluation lengths are closer, showed no comparable effect.

Beyond aggregate scores, the study deployed a two-pronged error analysis: automated signature detectors applied to all 95,070 re-aligned sentence pairs, and an exhaustive manual reading of a stratified 1,735-utterance sample. The detectors revealed that Whisper produces substantially more complete misrecognitions than Parakeet, at 1.6 to 2.0 times the rate in every fine-tuned regime, and also more Estonian case-suffix confusions and diacritic-only errors, patterns that were present before any fine-tuning and retained their direction throughout. Parakeet, by contrast, exhibited a stable language-conditioned boundary bias: it split Estonian compound words at morpheme boundaries, turning

Subject of Research: Synthetic speech data augmentation for low-resource Estonian and Slovenian automatic speech recognition

Article Title: Synthetic speech augmentation for low-resource Estonian and Slovenian ASR: Comparing Parakeet-TDT and Whisper

Article References: Perezhohin, Y., Castelli, M., & Popovič, A. (2026). Synthetic speech augmentation for low-resource Estonian and Slovenian ASR: Comparing Parakeet-TDT and Whisper. Machine Learning with Applications, 26, Article 101043. https://doi.org/10.1016/j.mlwa.2026.101043

Image Credits: AI Generated

DOI: 10.1016/j.mlwa.2026.101043

Keywords: automatic speech recognition, synthetic speech, data augmentation, Estonian, Slovenian, Whisper, Parakeet-TDT, text-to-speech, low-resource languages, large language models, morphology, error analysis