Masking I Flashcards Preview

HESP606 > Masking I > Flashcards

Flashcards in Masking I Deck (15):

What is the definition of masking? How many types of masking are there? What is monoaural, binaural, and central masking? Temporal masking?

  • Definition of masking: the process by which the detectability of one sound, the signal, is impaired by the presence of another sound, the masker
    • Doesn't have to be at the same time
  • 18 different types of masking
  • Monaural (ipsilateral): signal and masker presented to same ear
  • Binaural: both signals and maskers presented to both ears
    • Diotic: S oNo: signals in phase, noise in phase. Same signal, same noise in both ears
    • Dichotic: SπNo: signals out-of-phase (180 degrees), noise in phase, or SoNπ
      • threshold is better by 10-12 db for low frequency
      • More improvement in SπNo than SoNπ
  • Central: signal to one ear, masker to contralateral ear
    • Causes poorer θ in test ear due to central interaction
    • Small threshold shift (5 dB) from quiet 
  • Temporal masking – a type of monaural masking
    • forward masking: θ shift by 20 dB at 10 ms
      • Masker before signal
      • More temporal masking efferct
    • backward masking: θ shift by 10 dB at 10 ms
      • SIgnal before masker 
    • Important for speech becasue vowels can mask consonants


What is laboratory masking? What is clinical masking?

  • Laboratory masking: usually signal and masker are presented to same ear
    • Ipsilateral masking
  • Clinical masking: signal is presented to one ear but crosses to non-test ear
    • Masker is presented to non-test ear
    • Purpose is to mask detectability of signal that has crossed over to non-test ear
      • Mask tone in non-test ear
    • Thus, masker to non-test ear creates threshold shift from Q to N conditions in that ear (a form of ipsilateral masking)
    • Not central masking


Describe a white noise masker. What does its spectrum look like? What does effective spectrum mean? What is good about a white noise masker? What is a problem with a white noise masker?

  • Spectrum:
    • all frequency components: 1 Hz–100k Hz
    • each component = in amplitude
  • Effective spectrum: determined by frequency response of transducers
    • Transducer has limited bandwidth, frequency response - white noise bandwidth becomes the same when passed through tansducer
  • Looks dense and flat across spectrum
  • Most predictable of all maskers
  • Problem: perceptually loud - loudness increases with bandwidth
    • But can pass only a band of noise - band pass filter


What is the critical band theory?

 Critical Band Theory: developed by Fletcher & Munson (1937); assumptions

  1. For a given tone, only the energy in the white noise (WN) masker around that pure tone’s frequency is effective in masking the pure tone
    • Some size band is equally effective
  2. At the masked θ, the power in the tone = the power in the critical band (CB)
  3. The bandwidth of the CB is determined by the pure tone you are masking
    • Lower frequencies - narrower than high


Describe a band of noise. What does the noise generator produce? What do earphones do? How is the bandwidth determined? What would the ideal noise band be?

  • Noise generator produces white noise (100,000 Hz wide)
  • earphones filter white noise (BW can be 5500 Hz, with TDH-39 earphones)
  • band of noise is even narrower
  • Bandwidth determined by halfpower points (3-dB down points)
    • At what point the power drops by 3 dB
  • Ideal noise band would have infinite drop edges


What is a band of noise described by? How is the relative intensity of the band calculated?

  • Described by:  
    • the width of the band of frequencies (f2 – f1, in Hz)
    • the relative intensity of the bandwidth re: one component in the band (in dB)
      • Relative intensity compared to one component in the bandwidth
        • Use power formula: 10 log (P1/P2)
          • P1 - Whole BW in Hz
        • 10 log (bandwidth/1)
      • Example:
        • BW = 500 Hz
        • What is relative intensity of BW?
          • dB = 10 log 500
          • log 500 =2.7 dB = 10(2.7) dB = 27


What are compnents of noise? How are they related (equation)?

  1. Overall intensity:  SPL generated by all components in the noise
    • Measured by SLM
  2. LPC (level-per-cycle, or spectrum level): SPL generated by one component of the noise
    • Need to calculate
  3. Bandwidth (in dB or Hz): how wide is the band of noise (in Hz) or the relative level of the band of noise compared to 1 cycle (in dB) 

These are related by: OL = LPC + BW (in dB)


How do you convert between BW in Hz and BW in dB? How do you calculate LPC?

  • BW (in dB) = 10 log BW (in Hz)
    • example: BW (in Hz) = 5500
    • how wide is BW in dB?
  • Need to know overall level & BW (in dB or in Hz) and calculate LPC
    • example: OL = 100 dB SPL
    • BW (in Hz) = 5500 Hz
    • what is LPC?


Describe the Fletcher and Munson experiment. Who were their subjects? What was their masker? What was their test tone? What was the frequency response of the earphones? What was measured? What was the masked threshold?

  • Normal subjects
  • white noise masker (100 dB SPL)
  • 1 kHz test tone
  • Earphones used were 7000 Hz wide
  • measured threshold of tone in the white noise masker
  • masked threshold was 80 dB SPL
  • what is the critical band?


How is the critical band calculated? What is this calculation now called?

  • Step 1: calculate the LPC of white noise masker
    • LPC = OL - 10 log BW (in Hz)
    • LPC = ?
  • Step 2: calculate width of CB, in dB
    • BW of CB = OL of CB - LPC of noise
    • BW of CB = ?  (in Hz?  in dB?)
  • This calculation method is now called the Critical Ratio

Calculate the critical band with the values from the Fletcher Munson experiment

  • OL = 100 dB SPL
  • BW of earphones = 7000 Hz
  • OL of CB = 80 dB SPL because rule #2 says power in critical band s threshold for tone


How is the critical band measured directly? What ar the findings? What is the width of the NBN masker in audiometers based on?

  • Method: measure masked threshold in noise with decreasing bandwidth
    • narrow bandwidth until threshold gets better
  • Findings: width of CB > CR by 2.5 times (Hz), or by + 4 dB (dB), above 500 Hz
    • wider than critical ratio
    • gets wider with frequency
  • width of NBN masker in audiometer: used to be based on critical ratios, then critical bands

A image thumb

Describe narrowband noise in the audiometer. What does ANSI specify? What needs to be checked and compared?

  • ANSI (S3.6, 2010)
  • requires bandwidths to be wider than CBs
  • Bandwidths are now between a 1/3octave band and a ½-octave band (see Table 4, ANSI, 2010)
  • Need to check audiometer specs and compare NBN BW to audiometer 


What is the definition of effective masking level? What is an alternate defintion?

  • Definition: SPL of a band of noise whose geometric center frequency coincides with that of a specific pure tone that masks the pure tone to 50% probability of detection. (dB EM)
    • EML is analogous to hearing level
    • For normal-hearing listeners, EML = # dB that a given band of noise shifts a pure-tone Θ
      • EML - SPL of noise that masks pure tone for YNH
    • Alternate definition: how much a masker shifts threshold, relative to threshold in quiet (0 dB HL), for normal-hearing listeners
      • 20 dB EML -> 20 dB threshold shift


What is the relationship between EML and masked threshold? What are the advantages of NBN?

  • Relationship between EM level (EML) and masked threshold (M)
    • Above 10 dB, EML = M
    • For normal hearing EML = masked threshold 1000 Hz in same ear
    • If they can hear masking, masked threshold=level of masker
  • examples of applying EM
  • Summary: advantages of narrow band noise
    • has equivalent masking effectiveness as white noise
    • Loudness of NBN is < loudness of WN


How do you calibrate the masking unit?

  • Know the RETSPLS for pure tones at each frequency, for your earphones
  • Know the bandwidth of the narrowband noise for your audiometer - spec sheet
  • Add correction factor for noise bandwidth to pure tone RETSPL
    • ANSI (S3.6, 2010), Table 4
    • 1000 Hz
      • 7 dB RETSPL + 6 dB correction factor = 13 dB SPL for 0 dB EML
  • Provides RETSPL for each noise band