Using a forced aligner for prosody research
Bibliographic record
Abstract
Abstract Forced alignment is a speech technique that can automatically align audio files with transcripts. With the help of forced alignment tools, annotating audio files and creating annotated speech databases have become much more accessible and efficient. Researchers have recently started to evaluate the benefits and accuracy of forced aligners in speech research and have provided insightful suggestions for improvement. However, current work has so far paid little attention to evaluating forced aligners in prosody research, which focuses on suprasegmental features. In this paper, we take ambiguous sentence-level audio input in Mandarin Chinese, which can be disambiguated prosodically, to evaluate the alignment accuracy of the Montreal Forced Aligner (MFA). With a satisfactory result for syllable-by-syllable alignment, we further explore the possibility and benefits of using the forced alignment tool to generate phrase-by-phrase alignment. This topic has barely been studied in previous research on forced alignment. Our paper demonstrates that the forced alignment tool can effectively generate accurate alignment at both syllable and phrase levels for tonal languages, such as Mandarin. We found that the average differences between human annotators and MFA were smaller than the gold standard, indicating a satisfactory level of performance by the tool. Moreover, the MFA-assisted annotation rate by human transcribers was at least 20 times faster than previously reported manual annotation efficiency, providing significant time and resource savings for prosody researchers. Our results also suggest that phrase-level alignment accuracy of MFA can be affected by the quality of the recording, calling prosody researchers’ attention to controlling the audio quality in the recording. The finding that de-stressed words/phrases pose challenges for MFA also provides a reference for improving forced aligners.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.013 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.015 | 0.005 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".