Text-to-Speech (TTS) system for the Urdu language presents significant challenges, primarily due to the scarcity of high-quality datasets and an insufficient focus on modeling pronunciation. Urdu is spoken by 250 million people worldwide, but its research on computational linguistics remains underrepresented. In this paper, we introduce URDUTTS, a comprehensive and publicly available Urdu TTS dataset containing 89 hours of studio-quality speech, with accompanying transcriptions in three formats: Urdu Script, Phonemized Script, and Romanized Script. The dataset includes both mono-speaker and multi-speaker configurations. As Urdu relies heavily on phonetic features, accurate pronunciation is highly essential for the language. Therefore, we benchmark our dataset using VITS and GlowTTS models to compare the widely used Romanized script format with the Phonemized representation. To make the evaluation highly comprehensive, we combined both objective and subjective evaluation strategies. For objective evaluation, Mel-Cepstral Distortion (MCD with Plain, Dynamic Time-Warping, and Slope-Limitation variants), Signal-to-Noise Ratio (SNR), Word Error Rate (WER), and Character Error Rate (CER) were taken. Subjective evaluation was governed by Mean Opinion Score (MOS) ratings from 40 native speakers. Results show that using VITS and GlowTTS with Phonemized transcriptions performs significantly better than Romanized ones, with an improvement of 9.6% and 26.5% in MOS. The data and code are available at github.com/KAABSHAHID/URDUTTS.

India’s tribal (*Janjatiya/Adivasi*) languages remain absent from public ASR evaluation, despite their importance for oral history, ecological knowledge, and cultural continuity. We present **AdiDhwani**, a preservation-oriented ASR benchmark for **Bhili** (ISO 639-3: ‘bhb‘), India’s largest non-scheduled language, and **Mundari** (ISO 639-3: ‘unr‘), a Vulnerable North Munda language. The benchmark comprises 56.5 hours of community-validated Bhili speech and 11.7 hours of Mundari speech with fixed evaluation splits and reproducible baselines. We evaluate multilingual ASR under zero-shot transfer, target-language fine-tuning, donor-language initialization, synthetic supervision, and KenLM-assisted decoding. Target-language supervision substantially improves WER, yet lower edit distance does not always reflect better documentation fidelity: models replace attested Bhili forms with Hindi cognates or delete grammatically meaningful Mundari suffixes while still achieving competitive scores. We introduce **Hindi-Normalization Bias (HNB)** and its alignment-constrained variant **(HNBA)**, **Morphological Fragmentation Rate (MFR)**, and **Suffix Deletion Rate (SDR)** as preservation-oriented diagnostics beyond WER. Human transcriber baselines show that these gaps are not fixed limits. **AdiDhwani** provides the first controlled ASR evaluation framework for these two under-documented Indian languages and supports language-preserving speech recognition research in endangered-language settings.