KU Information Technology 4
Steps to analyze audio data involve multiple formats and methods.
Focus on identifying formants within the frequency range of 0 to 5,100 Hz.
Important to create a new object called the "formed object" to store analysis results.
LPC (Linear Predictive Coding) used for spectrum analysis.
Knowledge of the sampling frequency required:
Default frequency for CD-quality audio is 44,100 Hz.
Can find sampling frequency by checking audio file info if unfamiliar.
Make sure to input the correct sampling frequency for accurate measurements.
Adjust cursor position to derive a spectrum from audio clips.
Cursor should be positioned in the middle of the vowel sound.
Spectral average is calculated based on this cursor location.
The LPC spectrum displays frequency on the horizontal axis and amplitude on the vertical axis.
Cursor click gives precise measures for formants (f1, f2, f3).
Formant frequencies may vary based on recordings, showcasing the need for multiple approaches.
Recommended to practice recording yourself first before measuring others.
Discussion about the format used for recordings (WAV files recommended).
Importance of confirming ability to open and analyze audio files prior to project due dates.
Required submissions include:
A word document containing descriptions of speakers, word lists, glosses, and tables of vowel measurements.
Several organized WAV files for analysis, ideally not too lengthy and sorted logically.
Emphasize linguistic insights during file organization for clearer data analysis.
Recommended to round frequency measurements to the nearest Hertz and milliseconds for clarity.
Mention of MWC values related to average frequencies for male speakers as a reference point.
Check your measurements against these established averages to assess anomalies.
Students encouraged to communicate issues or ask for guidance on measurement discrepancies.
Final thoughts on the excitement of visualizing speech and the process of speech analysis.
Upcoming discussions on advanced topics like supersymmetry within phonetic study.
Steps to analyze audio data involve multiple formats and methods.
Focus on identifying formants within the frequency range of 0 to 5,100 Hz.
Important to create a new object called the "formed object" to store analysis results.
LPC (Linear Predictive Coding) used for spectrum analysis.
Knowledge of the sampling frequency required:
Default frequency for CD-quality audio is 44,100 Hz.
Can find sampling frequency by checking audio file info if unfamiliar.
Make sure to input the correct sampling frequency for accurate measurements.
Adjust cursor position to derive a spectrum from audio clips.
Cursor should be positioned in the middle of the vowel sound.
Spectral average is calculated based on this cursor location.
The LPC spectrum displays frequency on the horizontal axis and amplitude on the vertical axis.
Cursor click gives precise measures for formants (f1, f2, f3).
Formant frequencies may vary based on recordings, showcasing the need for multiple approaches.
Recommended to practice recording yourself first before measuring others.
Discussion about the format used for recordings (WAV files recommended).
Importance of confirming ability to open and analyze audio files prior to project due dates.
Required submissions include:
A word document containing descriptions of speakers, word lists, glosses, and tables of vowel measurements.
Several organized WAV files for analysis, ideally not too lengthy and sorted logically.
Emphasize linguistic insights during file organization for clearer data analysis.
Recommended to round frequency measurements to the nearest Hertz and milliseconds for clarity.
Mention of MWC values related to average frequencies for male speakers as a reference point.
Check your measurements against these established averages to assess anomalies.
Students encouraged to communicate issues or ask for guidance on measurement discrepancies.
Final thoughts on the excitement of visualizing speech and the process of speech analysis.
Upcoming discussions on advanced topics like supersymmetry within phonetic study.