Arun Prakash Singh

• Rosslund, Audun; Mayor, Julien; Mundry, Roger; Singh, Arun Prakash; Cristia, Alejandrina & Kartushina, Natalia (2024). A longitudinal investigation of the acoustic properties of infant-directed speech from 6 to 18 months. Royal Society Open Science. 11(11), s. 1–33. doi: 10.1098/rsos.240572. Fulltekst i vitenarkiv Vis sammendrag

  Caregivers often modulate their speech when interacting with infants, adapting a register that has been suggested to have attentional, affective and didactic purposes. The present preregistered study examined the longitudinal trajectories of a diverse range of acoustic features of infant-directed speech (IDS) and compared these with adult-directed speech (ADS), in Norwegian parents of 6- to 18-month-old infants. Sixty-nine families participated. Throughout five laboratory visits across one year, parents were recorded reading a picture-book to their infant (IDS) and an experimenter (ADS). The book was designed to tightly control for the linguistic content and context of speech between participants, timepoints and registers. Analyses of a total of 54 594 vowels and 22 958 phrases revealed, first, an overall effect of register: parents used higher pitch, wider pitch range, slower articulation rate, longer vowel duration and more variable and less distinct vowels in IDS than in ADS. Second, significant register-by-age interactions indicated that parents’ IDS, compared with their ADS, featured wider pitch range, larger vowel space and shorter vowel duration in older as compared with younger infants, while pitch, articulation rate and vowel variability and distinctiveness remained relatively stable with age. Results are discussed in the context of the proposed functions of IDS.

• Singh, Arun Prakash (2021). An improved method to localize simultaneously close and coherent sources based on symmetric-Toeplitz covariance matrix. Applied Acoustics. ISSN 0003-682X. 182. doi: 10.1016/j.apacoust.2021.108176. Fulltekst i vitenarkiv

Se alle arbeider i NVA

• Kartushina, Natalia; Mayor, Julien; White, Michelle Jennifer; Gundersen, Anni Sofie Myrtveit; Onnis, Luca & Hagelund, Silje [Vis alle 9 forfattere av denne artikkelen] (2025). Reliability of acoustic measures of infant-directed speech between lab and home environments and play and reading contexts. Fulltekst i vitenarkiv
• Rosslund, Audun; Mayor, Julien; Mundry, Roger; Singh, Arun Prakash; Cristia, Alejandrina & Kartushina, Natalia (2024). A longitudinal investigation of the acoustic properties of infant-directed speech from 6 to 18 months. Fulltekst i vitenarkiv
• Rosslund, Audun; Mayor, Julien; Mundry, Roger; Singh, Arun Prakash; Cristia, Alejandrina & Kartushina, Natalia (2024). A longitudinal investigation of the acoustic properties of infant-directed speech from 6 to 18 months. Fulltekst i vitenarkiv
• Singh, Arun Prakash & Kartushina, Natalia (2024). Classifying Infant Vocalizations in Audio Recordings through Transfer Learning and Image Processing Techniques. Fulltekst i vitenarkiv Vis sammendrag

  The identification and classification of infants' vocalizations (into canonical vs non-canonical babbling vs adult/other) in audio recordings can play crucial role in understanding infants’ language and speech development in various learning environments. Recent research indicates that deep learning models such as convolution neural network (CNN) trained from ‘scratch’ can be utilized to accomplish various audio classification tasks including classifying infants’ vocalizations. However, to achieve high classification accuracy, models trained from scratch require large number of vocalization instances. This study addresses this limitation and examines whether pre-trained vision deep learning models, using transfer learning with spectrogram images of vocalization audio clips as input, can achieve better accuracy in classifying infant vocalizations. An open-source labelled database of infant and adult vocalization recorded in a home environment for English-speaking families (infants wore Lena vests for 12-hour periods at 3, 6, 9, and 18 months of age) was used for this task. Despite few audio clips of infant and adult vocalizations being available, the classification accuracies of various vision deep learning models, used in the current study, ranged between 64.50% to 82.77%, which is higher, as compared to previously used models when trained with same number of audio clips, i.e., between 55% and 71.42% of accuracy.

• Kartushina, Natalia; Onnis, Luca; Rosslund, Audun; Singh, Arun Prakash; White, Michelle Jennifer & Theimann, Ane [Vis alle 14 forfattere av denne artikkelen] (2023). Babies and kids - how do they acquire and learn multiple lects? Fulltekst i vitenarkiv
• Singh, Arun Prakash & Kartushina, Natalia (2023). Use of transfer learning and image models to categorize infant vocalizations in audio recordings. Fulltekst i vitenarkiv
• Singh, Arun Prakash (2022). An improved two-stage dereverberation method based on bayesian estimation of a speech source. Fulltekst i vitenarkiv
• Singh, Arun Prakash (2017). A Robust Direction Of Arrival Estimation Algorithm For Planar Arrays To Detect Coherent Sources In Noisy Environment. Fulltekst i vitenarkiv
• Singh, Arun Prakash (2016). Source Localization Algorithm For Noisy Signal Using Sparse Optimization Technique. Fulltekst i vitenarkiv

Se alle arbeider i NVA