Spectral Selectivity & Masking Flashcards
What is masking?
- The interaction of sounds, keeping you from hearing a target
- Masking occurs when the presence of a second sound (masker) in addition to a target signal results in increased threshold
What is energetic masking?
-Peripheral for “certain” or “predictable” noise
What is informational masking?
-Central for “uncertain” noise
What is a critical band?
- Narrowband of frequencies surrounding the CF of a target tone
- As per the power spectrum model, only those frequencies within the critical band contribute to the masking of the target tone
- Each auditory filter along the BM has its own CF and, therefore, its own CB
What is an excitation pattern.
- Derived from auditory filter outputs as a function of their CFs
- Broadens with increased frequency and level
- Asymmetrical due to the asymmetric nature of the traveling wave
Compare PTCs vs. NTCs.
- In PTCs, bot a target and masker are presented
- Can perform PTCs in living humans (don’t need to impale single ANF)
- Response is from multiple ANFs
- More behavioral
- Because of off-frequency listening, PTCs have sharper tips than NTCs
Describe BM motion.
- Vibrates in response to incoming acoustic stimuli
- Traveling wave grows in amplitude until hitting the CF of the CB, at which point it dissipates quickly
What is the upward spread of masking?
- Lower frequency maskers are more effective that higher frequency
- Because masking grows nonlinearly on the high-frequency side
Compare notched-noise experiments vs. band-widening experiments.
- Band widening experiments: expand the width of the noise for masking to see if there is any effect on threshold
- Notched noise experiments: bring noise in from off-frequency regions (so it gets closer to CB of the CF) to see if there is any effect on threshold
What is auditory filter shape effected by level?
- Width of CB around filter CF increases with level due to AN recruitment
- More ANFs are responding/firing
What is tonal masking?
- Presenting 2 tones (1 signal, 1 masker) simultaneously
- Two tones interact to create beats, if close in frequency
- Strongest at 3-5 Hz and equal amplitudes
- Fixed by using shorter durations
What is co-modulation masking release?
- In the presence of random noise, thresholds worsen as noise is increased
- Expected from a band-widening experiment, since total noise power is increased
- In multiplied noise, thresholds worsen until a certain point where they actually being to slightly improve
- Therefore, power spectrum model is wrong because energy is not all that matters in a single envelope (temporal information matters too)
- Only true for co-modulated noise because, in random noise, each filter is different
What is co-modulated noise?
- Nonrandom, filtered noise in different bands obscuring the perception of a signal
- EX: filtered noise with AM noise
What is profile analysis?
- Spectral shape discrimination
- More tones requires more filters for comparison
- Too many tones, however, worsens thresholds due to masking
- Level roving done to ensure that the listener is paying attention to spectral shape and not just loudness using energy for detection
- Evidence for across-frequency filter comparisons
What are mechanisms for forward masking?
- BM ringing
- Short-term adaptation in auditory nerve
- Persistent activity in CNS
- Central inhibition
- Efferent system
Compare NH and HI PTCs. Describe their importance for masking.
- At certain masker frequencies, NH listeners could detect target tones in the presence of masking even at 100 dB SPL
- With HI listeners, PTCs were flat, erratic, and/or inverted compared to NH PTCs, especially in hearing loss regions
- Therefore, level does matter because NH listeners’ PTCs were broader at higher levels than lower levels
Describe the effect of informational masking (i.e. speech on speech masking).
- Confusability
- In noise, performance decreased with lower target-to-masker ratio (aka closer in level, more confusability)
How can you measure compression perceptually?
-
