Impact Factor (SJIF): 3.632 International Journal Advance Research in Engineering, Science & Technology e-issn: 2393-9877, p-issn: 2394-2444 (Special Issue for ITECE 2016) Review on High Frequency loss reduction methods in Digital Aid. Prashant G. Patil * Dr. A.K.Mittra * Dr.Vijay.S.Chourasia * Research Scholar Pressor & Head Assistant Pressor * Department Of Electronics Engineering * Manoharbhai Patel Institute Technology, Gondia (M.S) India Abstract -A high loss will affect a person s ability to underst speech. People with high loss can hear vowels just fine, but what they can t hear are the consonant sounds F, S, T, Z. Additionally, they are unable to hear higher octaves, like a woman s or a child s voice, or a bird chirping. Losing in those frequencies means that those sounds are harder to discern. For individuals with severe or pround loss at high frequencies, it is simply not possible to restore audibility to normal levels using only amplification. As conventional aids (HAs) have largely failed to provide such individuals with high cues, alternative methods signal processing have been considered that present information from high regions speech to lower regions. The method to be discussed in this article is commonly referred to as lowering (also known as shifting,, or transposition). Keywords- Hearing Aid, Transposition, High, Hearing Loss. I. Introduction The need to communicate has yielded a vast array inventions. The need to compute has similarly stimulated our inventiveness. According to World Health Organisation (WHO) 1 to 1.5 million Deaf peoples are found in India. The estimates vary according to how "deaf" is defined. Traditional view impaired market (Worldwide) says that 6 million peoples are wearing aid & 24 millions with no Aid.The reason behind not to wear Hearing aid is patient are not satisfying with aid, they also dems for in performance. Today 83% all aids sold which are based on digital signal processing. Many researchers have suggested we ve reached the limit what Digital Signal Processing can do; Digital Signal processors are as powerful as they can be No more digital features left to design. Redesign, Modification Processor will affect performance aid to overcome this drawback we need to make Hearing Aid Smart & Intelligent in certain environment. Now a day s many aids available in market are working on concept Multib, Feedback cancellation, Noise reduction & Directionality. Still many aid users need s in following factors. Poor benefit from aids in Background noise/noisy situations. Sound quality is poor. Volume control adjustment. Whistling feedback. High- loss not helped. Pround loss not helped. Too loud & Does not work on phone. II. Types Hearing loss. In general terms, there are two types loss, conductive sensor neural. A combination both is also seen as a mixed loss. Each is discussed below A. Conductive Hearing Loss Conductive loss is caused by any condition or disease that impedes the conveyance sound in its mechanical form through the middle ear cavity to the inner ear. A conductive loss can be the result a blockage in the external ear canal or can be caused by any disorder that unfavourably effects the middle ear's ability to transmit the mechanical energy to the stapes footplate. This result in reduction one the physical attributes sound called intensity (loudness), so the energy reaching the inner ear is lower or less intense than that in the original stimulus. Therefore, more energy is needed for the individual with a conductive loss to hear sound, but once it's loud 1
enough the mechanical impediment is overcome, that ear works in a normal way. Generally, the cause conductive loss can be identified treated resulting in a complete or partial in. Following the completion medical treatment for cause the conductive loss, aids are effective in correcting the remaining loss. The audiometric prile that indicates a conductive loss is the presence air-bone gaps (better by bone conduction than by air conduction), excellent word recognition at a comfortable listening level, evidence a middle ear dysfunction on. For situations where a blockage is noted in the external ear canal, testing is deferred until the canal is cleared. B. Sensorineural Hearing Loss The second type loss is called sensor neural loss. This word can be divided into its two components - sensory neural - to allow us more clarity in specifying the type loss. The comprehensive audiometric assessment supplemental tests can yield the information needed to differentiate between a sensory a neural loss, although they can co-exist in the same ear. Neural loss is another name for retro cochlear loss. Sensor neural loss results from inner ear or auditory nerve dysfunction. The sensory component may be from damage to the organ Corti or an inability the hair cells to stimulate the nerves or a metabolic problem in the fluids the inner ear. The neural or retro cochlear component can be the result severe damage to the organ Corti that causes the nerves to degenerate or it can be an inability the nerves themselves to convey neurochemical information through the central auditory pathways. The reason for sensorineural loss sometimes cannot be determined, it does not typically respond favourably to medical treatment, it is typically described as an irreversible, permanent condition. Like conductive loss, sensorineural loss reduces the intensity sound, but it might also introduce an element distortion into what is heard resulting in sounds being unclear even when they are loud enough. Once any medically treatable conditions have been ruled out, the treatment for sensorineural loss is amplification through aids. C. Mixed Hearing Loss A mixed loss can be thought as a sensorineural loss with a conductive component overlaying all or part the audiometric range tested. So, in addition to some irreversible loss caused by an inner ear or auditory nerve disorder, there is also a dysfunction the middle ear mechanism that makes the worse than the sensorineural loss alone. Hearing aids can be beneficial for persons with a mixed loss, but caution must be exercised by the care pressional patient if the conductive component is due to an active ear infection. III. Causes high loss: People with high loss can hear vowels just fine, but what they can't hear are the consonant sounds F, S, T, Z. Additionally, they are unable to hear higher octaves, like a woman's or a child's voice, or a bird chirping Figure 1- Loudness & Pitch map with Hearing Loss in db Extended exposure to loud sounds Diabetes due to neuropathy Genetics family history Age hair cells in the cochlea die f as we get older Acoustic neuroma. Infections colds, flu, meningitis Otitis Media (inflammation the middle ear) 2
A. Symptoms: Unable to hear consonants Inability to hear higher pitched sounds Tinnitus (ringing or buzzing in your ears) Difficulty talking in groups & Unable to hear when background noise is present B. Prevention & Treatment High Frequency Hearing Loss There are many ways to prevent high- loss. People concerned about risks can consider various methods protection, such as ear plugs. High- loss can be managed by using aids. Open-fit aids, which leave your ear canal at least partially open, have become popular for high- loss. Open-fit aids allow low- mid- sounds into the ear normally, so that only high- sounds are amplified. People with loss may experience some or all the following problems; Difficulty conversations, especially where there is background noise Hissing, roaring, or ringing in the ears (tinnitus) Difficulty the television or radio at a normal volume Fatigue irritation caused by the effort to hear Dizziness or problems with balance IV. Hearing Loss Range : When we speak loss, we mean threshold points that are higher than normal. If, for example, a person has a 45 db loss in the 4000 Hz range, it means that for him to be able to hear a sound at that or pitch, the sound must be at least 45 db in loudness. He/she cannot hear sounds below that volume at that. A. Hearing Loss Speech Intelligibility In the audiogram below, we can see where our basic speech sounds lies when engaging in normal conversation. Two things are important to recognize. For the most part: 1. Consonants are higher pitched than vowels (they lie more to the right on the chart). 2. Consonants are spoken more stly than vowels (they lie higher on the chart, in the lower decibel ranges). Most vowels consonants lies in the following regions: So a person with loss will have trouble the consonants in the first place. He may be hanging on by a thread. 3. Consonants convey most the word information; they are much more important to putting it together: Consonants are more important than vowels in understing speech. Consonants are spoken more stly than vowels, they tend to get drowned out in noisy environments. Consonants are higher-pitched than vowels most loss occurs in the higher frequencies. Table 1- Types Hearing Loss & Symptoms. 3
Category Normal Loss International Journal Advance Research in Engineering, Science & Technology (IJAREST) Hearing Loss Symptoms Range in db 0-19 db Normal People also having this kind loss Mild loss 20-39 db. Unable to hear st sounds. Cannot hear a whispered conversation in a quiet room. Can hear a normal conversation in a quiet room but has difficulty in a noisy environment Moderate loss 40-59 db. Has considerable difficulty a normal conversation in a quiet room.if there is background noise, he/she will not be able to underst many the words, unless he lip reads. Severe loss 60-89dB. Cannot hear a conversation unless the speaker speaks loudly Pround loss. 90+ db. Cannot underst speech even if the person speaks very loudly. Can only hear very loud sounds such as a chainsaw. Figure 2- Alphabets mapping with pitch & HTL in db V. Signal processing for removing high loss : For individuals with severe or pround loss at high frequencies, it is simply not possible to restore audibility to normal levels using only amplification. As conventional aids (HAs) have largely failed to provide such individuals with high cues, alternative methods signal processing have been considered that present information from high- regions speech to lower regions. The method to be discussed in this article is commonly referred to as lowering (also known as shifting,, or transposition. These terms are ten used interchangeably when describing lowering techniques). For convenience, the second section this review describes these processing techniques under the headings channel vocoder, slow playback, transposition,. Currently five techniques are implemented in commercially available aids: Linear by AVR Sonovation Israel (Introduced 1991) Linear transposition (LFT) by Widex Lynge, Denmark (Introduced 2006) (NFC) by Phonak Stafa, Switzerl (introduced 2008). NFC also used by Unitron Kitchener, Ontario, Canada (starting 2012) Spectral envelope warping by Starkey Hearing Technologies Minnesota,USA (introduced 2011) The need for The problem with linear aids the same gain is applied to all levels input signal. We need high gain for low input levels, low gain for high input levels. We need some way automatically turning down the gain the aid as the input intensity increases. Following figure & table will illustrate operating concept & principle all techniques. There are certain advantages & Disadvantages each technique. 4
Figure 3- A visual representation how different lowering methods affect the signal. Table 2- Comparison all techniques Method Way Processing Advantages Disadvantages Transposition Shifts high sounds to lower quality, preserves high low frequencies can More natural sound If continuously active, overlap Transposition frequencies adds harmonic relationship mask useful low- transposed signal to between information, can transpose unprocessed lower components unwanted high- signal background noise Frequency shifting Slow playback Lower frequencies unprocessed, higher frequencies compressed in greater amounts Lowers all components downward by a constant factor Records segments the speech signal then plays these segments back at a slower speed than that used in the original recording More natural sound quality, preservation vowel intelligibility, no overlap in information Preserves harmonic relationship between components Preserves harmonic relationship between components Does not preserve harmonic relationship between components Lowers pitch speaker, unnatural sound quality Segment deletion can cause distortion or discard useful speech information, signal stretched in time Sr No 1 Author/s (Year Published) Ling (1968) Table 3- Literature review on high loss reduction in Digital Hearing Aid Signal Processing method Aid Channel vocoder in No. HA User Outcome Measures Training Details Results 8 Vowel consonant discrimination, word recognition No training 2 Ling Channel 18 Nonsense syllables No training 5
3 4 Maretic (1971) Beasley, Mosher, Orchik (1976) Rosenhouse (1990) vocoder Slow playback 9 Picture plates No training 1 Aided thresholds, environmental Ling sounds No training 5 6 7 8 Rees Velmans (1993) Parent, Chmiel, Jerger (1997) Turner Hurtig (1999), Dorkos, Dean, & Ching (1999) Transposition 8 Monosyllabic words, nonsense syllables Linear shifting 4 Spondees, vowel consonant recognition, monosyllabic words, sentence 15 Closed-set testing consonant vowel nonsense syllables 5 Sentences consonants in quiet No training 4 to 6 weeks Up to 10 training sessions with materials similar to that used for testing 12 weeks 2 participants showed average 22% for sentences with device 50 % group showed s & 7% on average when listening to a female speaker 2 participants showed statistically 10% to 30% 9 Dean (2000) Linear shifting 6 Monosyllabic words A total 10 training sessions 1 hour duration using vowel stimuli No statistically s reported for the group for shiftedspeech 10 Sakamoto, Goto, Tateno, Kaga (2000) 5 Sentences recognition 1 to 2 weeks. No participant showed statistically with device 11 12 13 Knight (2001) MacArdle et al. (2001) Miller- Hansen, AVR ImpaCt 3 Monosyllabic words, consonants, sentences in noise 11 Aided thresholds, speech intelligibility ratings, closed-set speech testing 19 Aided thresholds, word 6 weeks Live Demonstration Live Demonstration No participant showed statistically with device 6
14 15 16 17 18 Nelson, Widen, (2003) Simpson, Hersbach, (2005) Simpson, Hersbach, (2006) recognition testing 17 Monosyllabic words 4 to 5 weeks 7 Monosyllabic words, sentences in noise 3 to 4 weeks Kuk (2007) Transposition 13 Nonsense syllables 2 weeks Robinson, Baer, Moore (2007) Gifford, Dorman, Spahr, McKarns (2007) Transposition 7 Nonsense syllables Three to four training sessions 2 hours duration with materials similar to that used for testing AVR Nano Xp 6 Monosyllabic words, sentences in quiet in noise 5 weeks Eight participants showed statistically with device with a mean group 6% for phoneme scores No participant showed statistically Average group 3-6% for vowels consonants Average statistically group 20% for affricates 2 participants showed with device, with average 17% 19 20 21 Glista et al. (2009) Glista, Scollie, Bagatto, Seewald, Johnson (2009) Xianbo xieo Hui zhuang Guangshu Hu Chanhong Hu Jia liu (2009) By setting parameters for Vowels & consonants uses overlapping,squeezed method 13 Ling sound test, consonant vowel recognition, plural recognition 11 consonant vowel recognition, & plural recognition 6 Vowels Consonants Syllables Mono syllables Words Short Sentences 4 weeks 2-3 Weeks (take Home) Comparing with users previous Aid & new algorithm Eight adults showed statistically with device Significant Improvement In all types 22 Hugh J. LFT, NLFC -- Sinusoids, flute Spectral analyses Significant 7
23 (2011) Mohammed Alnahwi, Zeinab A AlQudehy (2015) ( According to User Parameter Fitting) Comparison Between FT & FL method sounds, speech material. 376 Vowels Consonants Syllables Mono syllables Words Short Sentences on the output signals produced by the aids in each condition in both LFT & NLFC FC showed benefit (91%) compared with children who benefited from FT (70%). VI. Conclusion Providing high- information successfully is important for both male & female aid user who uses these cues to learn articulation sounds as well as many grammatical rules. Despite improved technology, today s conventional devices are still far from providing improved speech recognition for severe high loss. With this challenge in mind, an alternative to current amplification techniques is to shift high- components a signal into a lower region. These schemes are beginning to show successful outcomes for individuals with high- loss. Mean while there are lot required in terms consonants, vowels recognition. These algorithm needs to work & modification required for different languages. VII. References [1] FranciscoJ Fraga,Letica Pimenta C,S prates, Alan M Marotta, lowering Algorithms for Hearing Impaired, A Text book Speech technologies, Book Chapter 18, pp. 361-388. [2] Alexer J.M. (2012). : Balancing start ratio. 39th Annual meeting the American Auditory Society, Scottsdale, AZ. [3] Alexer J.M. (2013). Individual variability in recognition -lowered speech. Seminars in Hearing, 34, 86-109. [4] Danielle Glista, Susan Scollie, Marlene Bagatto, Richard Seewald, Vijay Parsa, Andrew Johnson, Evaluation nonlinear : Clinical outcomes. International Journal Audiology 2009; 48: pp 632644 [5] Harry Levitt, A Historical Perspective on Digital Hearing Aids: How Digital Technology Has Changed Modern Hearing Aids, Trends in Amplification Volume 11 Number 1,March 2007.pp 7-24 [6] Joshua M. Alexer, Individual Variability in Recognition Frequency-Lowered Speech, Seminars in Hearing, Volume 34, Number 2,2013 [7] Francis Kuk, Considerations in Verifying Frequency Lowering, Published on January 19, 2013. International Journal Audiology 2013(online). [8] Francis Kuk, Denise Keenan, Jane Auriemmo, Petri Korhonen, Heidi Peeters, Chi Lau, Bryan Crose, Interpreting the efficacy -lowering algorithms. The Hearing Journal, April 2010, Vol. 63, No. 4 [9] Xianbo Xiao, Hui Zhang, Guangshu Hu, Chunhong Liu, Jia Liu, Evaluation -lowering algorithms for intelligibility Chinese speech in -aid users, Progress in Natural Science 19 (2009), 741 749, Available online at www.sciencedirect.com. [10] Hugh J., A Technical Comparison Digital Frequency-Lowering Algorithms Available in Two Current Hearing Aids,. Published at PLoS ONE July 2011, Volume 6,Issue 7, e22358. [11] Hartmut Traunmüller Anders Eriksson, The range the voice fundamental in the speech male female adults. Acustica 50, pp 118 125. [12] J Gou, J Smith, J Valero, I Rubio, The Effect Frequency Transposition on Speech Perception in Adolescents Young Adults with Pround Hearing Loss. Deafness & education international, Vol. 13 No. 1, March, 2011, 17 33. [13] Andrea Simpson, Frequency-Lowering Devices for Managing High-Frequency Hearing Loss: A Review, Trends in Amplification, Volume 13, Number 2, June 2009, 87-106. [14] Vijay Parsa, Susan Scollie, Danielle Glista Andreas Seelisch, Frequency Compression: Effects on Sound Quality Ratings Speech Music, Trends in Amplification 17(1), 54 68. [15] Silvia Allegro, Olegs Timms, Hersbach, Method for Frequency Transposition use the method in a device & communication device. United States Patent, US007248711B2. 8
[16] Andreas Tiefenau,Zaam, Method for Transposition for Hearing Aid, United States Patent, US 20110142271Al. [17] Ying-Yee Kong, Ala Mullangi, On the development a -lowering system that enhances place--articulation perception, Speech Communication 54 (2012) 147 160, Available online at www.sciencedirect.com. [18] http://www.languagesgulper.com/eng/marathi.html [19] http://www.audiologyonline.com/articles/highs--lows- 9