Lecture speech
Download as PPT, PDF2 likes115 views
Sound is a vibrational energy that travels through a medium such as air, water or solid material. It is transmitted in the form of longitudinal waves of varying frequency and amplitude. Speech is a complex sound produced by humans for communication that contains different frequencies important for speech between 100-8000 Hz. Forensic voice identification analyzes the unique characteristics of an individual's voice such as fundamental frequency, intensity, wavelength, intonation and timbre to identify unknown speakers, which has applications for criminal cases where only a recorded voice is available as evidence.
Lecture speech
- 2. What is sound Sound is transmitted through a solid, liquid, or gas; particularly, sound means those vibrations composed of frequencies capable of being detected by ears (20Hz to 20KHz). Sound is a vibrational energy that travels in the form of signal (carrying information) in medium.
- 3. Sound Waves Longitudinal waves Travels through a medium in the form of compressions and rarefactions. Simple and complex Complex sound: Periodic, quasi periodic and non periodic
- 8. Sound v/s Speech All Speech is a sound but all Sound is not a speech Speech is a sound which is the form of communication among human being. Speech is a sound wave of air originated from the complex actions of human body. Speech is a complex waveform. It contains different frequencies.
- 9. Characteristicsof Sound/Speech/Voice 1. Frequency/Pitch: No of cycles per second. 100 Hz 200 Hz 1 second
- 10. Audible frequency range: 20 Hz to 20 KHz Frequencies important to speech signals are in the range of 100- 8000 Hz. Frequency related to the pitch directly. When frequency of vibration increased pitch increased and vice versa. Frequency is measurable but Pitch is a psychological phenomenon. Pitch is basically Fundamental Frequency of a person.
- 11. Male fundamental frequency/Pitch range : 80- 150 Hz Female fundamental frequency/Pitch range:150-250 Hz Child fundamental frequency/Pitch range: greater than 250 Hz
- 12. 2. Intensity/ Loudness: Amplitude/Energy of signal or Sound Pressure level. Intensity is measurable, Loudness is psychological phenomenon. A
- 13. More intensity more loudness and vice versa. Unit of intensity is dB (decibel). SPL = 20 log 10 (P0/Pr)
- 14. Wavelength: Length of complete one wave λ = c/fλ
- 15. Female’s voice has less wavelength than male’s voice because female’s fundamental frequency is more than male’s fundamental frequency.
- 16. Intonation: Variation in pitch pattern. Voice Quality/ Timbre: like Breathy Voice Diction: Clarity of speech Resonance: Human organ pipe is one end closed pipe First Harmonic/Formant f1 = λ/4 Second Harmonic/formant f2 = 3λ/4 Third Harmonic/formant f3 =5λ/4 ……… Only odd Harmonics
- 19. Voice Identification is one of the important emerging areas in forensic science The recent development in the area of information and communication technology has made infinite advantage of connecting people over exceptionally large distance at the fraction of second - The brighter side of the coin.
- 20. But the darker side of the coin flashes the use of communication technology in several criminal offences such making hoax calls to the police, ambulance or fire brigade, making threatening or harassing telephone calls, blackmail or extortion demands or match fixing etc.
- 21. Developed in the early years of 1940’s by Green, Potter and Kopp at the Bell Laboratories for the purpose of helping military during world war II
- 22. After criminal offence, no evidence was left except the recorded voice/voices of the suspect/suspects then we definitely need for Voice/Speaker Identification.
- 23. Voice Identification The task of determining an unknown speaker’s identity. (Voice identification determines which registered speaker provides a given utterance from amongst a set of known speakers.) Person’s speech/voice characteristics
- 24. MethodsofVoice/speakeridentification Aural examination of voice Spectrographic - visual examination of voice spectrogram Automatic method of voice identification