Course:PHYS341/2018/Calendar/Lecture 12
Phys341 Lecture 12: Summary and web references
2018.01.28
Textbook: 10.5-10.6
Slide List
- Project ideas
- Simple acoustic measurements on an instrument, the more unusual the better
- In this course at present, I have no examples of First Nations instruments, African instruments, and South Asian instruments. One of these instruments would be especially interesting.
- Singing styles
- All my examples are operatic; I have nothing from jazz, blues, or any popular style
- There are many musical instruments in MOA, and they are generally poorly described
- (e.g. a Chinese banhu is labelled simply as “violin”). Here’s an opportunity for an organological project.
- A great project is one that teaches me things I did not know. Musical acoustics is so large a field, that this is not as hard as you might think.
- See pages on the erhu, tambourica and room acoustics at http://acoustics.phas.ubc.ca/
- Send me your ideas and we can hone them into suitable projects (in the coming week).
- If you have a partner you wish to work with, let me know.
- If you don’t have a partner, let me know, and I can pair you off with another class member.
- Masking
- When one tone is played fairly loud, it becomes impossible to hear a range of tones of a higher frequency, unless they are played really loud.
- This is called Masking:
- For example, if a tone of 800 Hz reaches a listener with a SPL of 80 dB, a tone of 1000 Hz will have to be played at 50 dB above the normal threshold of hearing to be heard at all.
- This physiological effect is exploited in sound recording, where sound files are compressed to 10% of their original size by excluding sounds that we cannot hear due to masking (e.g. the mp3 format).
- Needs two separate function generators for effective demonstration
- Compression
- In an era of FFTs and fast computers, masking allows sound to be analyzed before being recorded :
- Parts of the sound spectrum that are fully perceived are recorded with high precision.
- Parts of the sound spectrum that are masked by other parts of the spectrum can be recorded with low precision.
- Frequency bands recorded separately in a recognized format (e.g. mp3).
- On playback the frequency spectra are reconstructed and inverse transformed back into a time-sequence of binary numbers.
- This time sequence is passed through a digital-analog-converter (DAC) and the voltage output is passed to a speaker.
- Hearing loss
- Audiologists measure hearing loss in dB, the difference between the patient’s threshold of hearing compared to that of a standard issue ISO 226 human being.
- Measurements are made at seven octave bands:
- Central frequencies 125, 250, 500, 1000, 2000, 4000 and 8000 Hz.
- The low tones especially have to be pure, otherwise the more easily heard harmonics will spoil the measurement.
- The threshold difference for considering hearing loss is ~ 25 dB
- Hearing loss is caused by age, noise exposure, disease, poverty, congenital conditions...
- Cochlear implants
- How can 22 frequency bands replace 15,000 hair cells connected to 30,000 nerve fibres?
- “Like playing a piano with boxing gloves on”.
- Low frequencies (deepest in the cochlea) are a problem.
- Need to use pitch cues.
- Note: we perceive the full range of colours with just three colour receptors in the eye.
- https://www.youtube.com/watch?v=SpKKYBkJ9Hw
- http://www.soundonsound.com/sound-advice/implanting-awareness