SlideShare a Scribd company logo
Abhyudai Yadav
Arnika Kumar
Dvizma Sinha
Jhalak R. Sharma
Mentors: Dr. Mandeep Goyal & Dr. Kumar Vishal
Performance Improvisation of Automatic
Speaker Recognition by Spectral Reverberation
Mitigation
Final Review
7th May, 2015
● Rationale & Objectives
● Methodology
● Literature Review
● Proposed Solution
● MatLab Simulations : Graphs
● Observations & Results
● Difficulties
● Future Work
● References
Contents
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Rationale
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Speaker Recognition
Process of automatically recognizing who is speaking on the basis of an individual’s
discrete information included in speech waves.
Problem
Presence of Reverberation in input speech signal
Reduces recognition rate for speech recognition engines.
TTo counter this problem Reverberation Mitigation is required.
Methodology
Problem Analysis
Feature Extraction
Training
Testing
Performance Analysis
Mitigation
Performance
Improvisation
Mid Term I Review Mid Term II Review Final Review
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Objectives
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
1.
• To mitigate reverberation from a speech signal.
2.
• To improve the performance of automatic speaker
recognition system in reverberant environments.
• Early reflections don’t distort the speech too much.
• Late reflections main cause of distortion.
Proposed Solution
Idea : To identify & remove the regions of the signal with most
reverberant energy.
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Figure: Reverberation
[Source: hyperphysics.phy]
• Envelope of late reflections -> described by an exponential decay curve.
• T60 found to get slope of decay curve.
• Regions identified in reverberant speech signal having same slope.
• Reverberant regions removed using a binary mask.
Proposed Solution
[Matlab Simulation]
𝑒 𝑟 𝑛 ≈ exp −1
𝜏
𝑛
= 𝑎 𝑛
● Reverberation time characterizes quality of auditory space
T60 Parameter
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Figure: A) Speech Signal
B) Reverberation tail
C) Decay Rate
D) Linear fit
[Source: Rama Ratnam; Blind estimation of
Reverberation Time; University Of Illinois]
● T30 and T20 are other such
parameters
● Higher value of T60 signifies
more reverberation and vice
versa
The T60 parameter measures the time taken for the sound level to drop 60 dB below the level at
sound cessation
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Algorithm
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Algorithm run on each band
individually
By passing through
Butterworth filter
Of each point on
the envelope
0 for exponential
1 otherwise
Using Hilbert
Transform
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Time Domain Plots
[A] [B]
Time Domain Representations of:
[A] Clean Speech
[B] Reverberated Signal
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
[Matlab Simulation]
Butterworth Filter Bank (Log Scale)
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Butterworth Filter Bank (Log Scale, 3 bins/octave)
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Channels obtained from filtering
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Envelope of the channels
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Estimating T60
[Matlab Simulation: Convergence of T60]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Mask for the channels
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Mask for one channel
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Reconstruct
signal
Smooth mask
Convert to log-
scale (dB)
Envelope
Follower
Filter bank
Calculate
binary mask
Estimate true
slope
Find local
slopes
yr(n)
yc(n)
Process of Dereverberation (in Time Domain)
Time Domain Plots
[A] [B]
Time Domain Representations of:
[A] Clean Speech
[B] Reverberated Signal
[C] De-reverberated Signal
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
[Matlab Simulation]
[C]
Process of Dereverberation
Spectrograms
[Matlab Simulation]
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Observations
• Higher T60 value -> Better accuracy.
• Percussive sounds easier to dereverberate
• Speech signals difficult to dereverberate.
 As the envelopes situated far apart there is
sufficient reverberant decay between
them.
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
[Matlab Simulation]
Results
Speech Recognition has two phases :
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Difficulties
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
• Sampling rate of input: > 10000 Hz
• Minimize effect of aliasing in the analog to digital conversion.
• Also known as signal processing front end.
Future Work
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
Thank You
Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation

More Related Content

What's hot

Hable John Uncharted2 Hdr Lighting
Hable John Uncharted2 Hdr LightingHable John Uncharted2 Hdr Lighting
Hable John Uncharted2 Hdr Lighting
ozlael ozlael
 
Deferred rendering in Dying Light
Deferred rendering in Dying LightDeferred rendering in Dying Light
Deferred rendering in Dying Light
Maciej Jamrozik
 
Lighting Shading by John Hable
Lighting Shading by John HableLighting Shading by John Hable
Lighting Shading by John Hable
Naughty Dog
 
Multrate dsp
Multrate dspMultrate dsp
Epic_GDC2011_Samaritan
Epic_GDC2011_SamaritanEpic_GDC2011_Samaritan
Epic_GDC2011_Samaritan
MinGeun Park
 
Dss
Dss Dss
Dss
nil65
 
Stochastic Screen-Space Reflections
Stochastic Screen-Space ReflectionsStochastic Screen-Space Reflections
Stochastic Screen-Space Reflections
Electronic Arts / DICE
 
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
Philip Hammer
 
Digital Filters Part 1
Digital Filters Part 1Digital Filters Part 1
Digital Filters Part 1
Premier Farnell
 
Implementation of Digital Filters
Implementation of Digital FiltersImplementation of Digital Filters
Implementation of Digital Filters
op205
 
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Barco
 
IIR filter design, Digital signal processing
IIR filter design, Digital signal processingIIR filter design, Digital signal processing
IIR filter design, Digital signal processing
Abhishek Thakkar
 
Ultrasound Modular Architecture
Ultrasound Modular ArchitectureUltrasound Modular Architecture
Ultrasound Modular Architecture
Jose Miguel Moreno
 
Filter design techniques ch7 iir
Filter design techniques ch7 iirFilter design techniques ch7 iir
Filter design techniques ch7 iir
Falah Mohammed
 
Warping Concept (iir filters-bilinear transformation method)
Warping Concept  (iir filters-bilinear transformation method)Warping Concept  (iir filters-bilinear transformation method)
Warping Concept (iir filters-bilinear transformation method)
ZunAib Ali
 
Fir and iir filter_design
Fir and iir filter_designFir and iir filter_design
Fir and iir filter_design
shrinivasgnaik
 
Techfest jan17
Techfest jan17Techfest jan17
Techfest jan17
Ramin Anushiravani
 
FILTER DESIGN
FILTER DESIGNFILTER DESIGN
FILTER DESIGN
naimish12
 
Dsp ppt madhuri.anudeep
Dsp ppt madhuri.anudeepDsp ppt madhuri.anudeep
Dsp ppt madhuri.anudeep
Pawan_Kumar_Kumar
 
1 digital filters (fir)
1 digital filters (fir)1 digital filters (fir)
1 digital filters (fir)
Ahmed A. Thabit
 

What's hot (20)

Hable John Uncharted2 Hdr Lighting
Hable John Uncharted2 Hdr LightingHable John Uncharted2 Hdr Lighting
Hable John Uncharted2 Hdr Lighting
 
Deferred rendering in Dying Light
Deferred rendering in Dying LightDeferred rendering in Dying Light
Deferred rendering in Dying Light
 
Lighting Shading by John Hable
Lighting Shading by John HableLighting Shading by John Hable
Lighting Shading by John Hable
 
Multrate dsp
Multrate dspMultrate dsp
Multrate dsp
 
Epic_GDC2011_Samaritan
Epic_GDC2011_SamaritanEpic_GDC2011_Samaritan
Epic_GDC2011_Samaritan
 
Dss
Dss Dss
Dss
 
Stochastic Screen-Space Reflections
Stochastic Screen-Space ReflectionsStochastic Screen-Space Reflections
Stochastic Screen-Space Reflections
 
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
The Rendering Technology of 'Lords of the Fallen' (Game Connection Europe 2014)
 
Digital Filters Part 1
Digital Filters Part 1Digital Filters Part 1
Digital Filters Part 1
 
Implementation of Digital Filters
Implementation of Digital FiltersImplementation of Digital Filters
Implementation of Digital Filters
 
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
Demystifying laser projection for cinema: 5 frequently asked questions, 125+ ...
 
IIR filter design, Digital signal processing
IIR filter design, Digital signal processingIIR filter design, Digital signal processing
IIR filter design, Digital signal processing
 
Ultrasound Modular Architecture
Ultrasound Modular ArchitectureUltrasound Modular Architecture
Ultrasound Modular Architecture
 
Filter design techniques ch7 iir
Filter design techniques ch7 iirFilter design techniques ch7 iir
Filter design techniques ch7 iir
 
Warping Concept (iir filters-bilinear transformation method)
Warping Concept  (iir filters-bilinear transformation method)Warping Concept  (iir filters-bilinear transformation method)
Warping Concept (iir filters-bilinear transformation method)
 
Fir and iir filter_design
Fir and iir filter_designFir and iir filter_design
Fir and iir filter_design
 
Techfest jan17
Techfest jan17Techfest jan17
Techfest jan17
 
FILTER DESIGN
FILTER DESIGNFILTER DESIGN
FILTER DESIGN
 
Dsp ppt madhuri.anudeep
Dsp ppt madhuri.anudeepDsp ppt madhuri.anudeep
Dsp ppt madhuri.anudeep
 
1 digital filters (fir)
1 digital filters (fir)1 digital filters (fir)
1 digital filters (fir)
 

Viewers also liked

Speaker Recognition
Speaker RecognitionSpeaker Recognition
Speaker Recognition
niranjan kumar
 
Automatic Speaker Recognition system using MFCC and VQ approach
Automatic Speaker Recognition system using MFCC and VQ approachAutomatic Speaker Recognition system using MFCC and VQ approach
Automatic Speaker Recognition system using MFCC and VQ approach
Abdullah al Mamun
 
Speaker recognition using MFCC
Speaker recognition using MFCCSpeaker recognition using MFCC
Speaker recognition using MFCC
Hira Shaukat
 
practica 1
practica 1practica 1
HCAD 600 Case Study Paper
HCAD 600 Case Study PaperHCAD 600 Case Study Paper
HCAD 600 Case Study Paper
Lindsay Allen
 
Anthony Davis
Anthony DavisAnthony Davis
Anthony Davis
Austin Chan
 
Contos Infantis Para a Família - Apresentação Instituições
Contos Infantis Para a Família - Apresentação InstituiçõesContos Infantis Para a Família - Apresentação Instituições
Contos Infantis Para a Família - Apresentação Instituições
Daniele Esprega
 
Presentation_Team_Backdated
Presentation_Team_BackdatedPresentation_Team_Backdated
Presentation_Team_Backdated
Tareque Hasan
 
A+P Design Studio
A+P Design StudioA+P Design Studio
A+P Design Studio
Puran Mistry
 
Portfolio Book mini
Portfolio Book miniPortfolio Book mini
Portfolio Book mini
Colin Swenson-Healey
 
Coursework Research: Headlines
Coursework Research: HeadlinesCoursework Research: Headlines
Coursework Research: Headlines
williamellishughes
 
Iman 2016
Iman 2016Iman 2016
Iman 2016
Iman Aburegeila
 

Viewers also liked (13)

Speaker Recognition
Speaker RecognitionSpeaker Recognition
Speaker Recognition
 
Automatic Speaker Recognition system using MFCC and VQ approach
Automatic Speaker Recognition system using MFCC and VQ approachAutomatic Speaker Recognition system using MFCC and VQ approach
Automatic Speaker Recognition system using MFCC and VQ approach
 
Speaker recognition using MFCC
Speaker recognition using MFCCSpeaker recognition using MFCC
Speaker recognition using MFCC
 
practica 1
practica 1practica 1
practica 1
 
HCAD 600 Case Study Paper
HCAD 600 Case Study PaperHCAD 600 Case Study Paper
HCAD 600 Case Study Paper
 
Anthony Davis
Anthony DavisAnthony Davis
Anthony Davis
 
Contos Infantis Para a Família - Apresentação Instituições
Contos Infantis Para a Família - Apresentação InstituiçõesContos Infantis Para a Família - Apresentação Instituições
Contos Infantis Para a Família - Apresentação Instituições
 
Presentation_Team_Backdated
Presentation_Team_BackdatedPresentation_Team_Backdated
Presentation_Team_Backdated
 
A+P Design Studio
A+P Design StudioA+P Design Studio
A+P Design Studio
 
Abstract_EN_FR
Abstract_EN_FRAbstract_EN_FR
Abstract_EN_FR
 
Portfolio Book mini
Portfolio Book miniPortfolio Book mini
Portfolio Book mini
 
Coursework Research: Headlines
Coursework Research: HeadlinesCoursework Research: Headlines
Coursework Research: Headlines
 
Iman 2016
Iman 2016Iman 2016
Iman 2016
 

Similar to Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation

ASR_final
ASR_finalASR_final
ASR_final
Bidhan Barai
 
Design of digital filters
Design of digital filtersDesign of digital filters
Design of digital filters
Naila Bibi
 
Cryosat2013 scagliola
Cryosat2013 scagliolaCryosat2013 scagliola
Cryosat2013 scagliola
michelescagliola
 
Analysis of Speech Enhancement Incorporating Speech Recognition
Analysis of Speech Enhancement Incorporating Speech RecognitionAnalysis of Speech Enhancement Incorporating Speech Recognition
Analysis of Speech Enhancement Incorporating Speech Recognition
IRJET Journal
 
Signal & Image Processing And Analysis For Scientists And Engineers Technical...
Signal & Image Processing And Analysis For Scientists And Engineers Technical...Signal & Image Processing And Analysis For Scientists And Engineers Technical...
Signal & Image Processing And Analysis For Scientists And Engineers Technical...
Jim Jenkins
 
Speech Compression using LPC
Speech Compression using LPCSpeech Compression using LPC
Speech Compression using LPC
Disha Modi
 
IJCER (www.ijceronline.com) International Journal of computational Engineerin...
IJCER (www.ijceronline.com) International Journal of computational Engineerin...IJCER (www.ijceronline.com) International Journal of computational Engineerin...
IJCER (www.ijceronline.com) International Journal of computational Engineerin...
ijceronline
 
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLABA GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
sipij
 
Speech Compression using LPC
Speech Compression using LPCSpeech Compression using LPC
Speech Compression using LPC
Disha Modi
 
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, BregenzFreedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Wojtek Cieplik
 
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital ConverterDesign of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
Karthik Rathinavel
 
H0814247
H0814247H0814247
H0814247
IOSR Journals
 
Linear Predictive Coding
Linear Predictive CodingLinear Predictive Coding
Linear Predictive Coding
Srishti Kakade
 
Design of iir digital highpass butterworth filter using analog to digital map...
Design of iir digital highpass butterworth filter using analog to digital map...Design of iir digital highpass butterworth filter using analog to digital map...
Design of iir digital highpass butterworth filter using analog to digital map...
Subhadeep Chakraborty
 
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
rijksmuseum
 
Introduction to deep learning based voice activity detection
Introduction to deep learning based voice activity detectionIntroduction to deep learning based voice activity detection
Introduction to deep learning based voice activity detection
NAVER Engineering
 
Integrating Sound level meter
Integrating Sound level meterIntegrating Sound level meter
Integrating Sound level meter
ECRD IN
 
Slm_1
Slm_1Slm_1
Slm_1
ECRD2015
 
Lecture 2- Practical AD and DA Conveters (Online Learning).pptx
Lecture 2- Practical AD and DA Conveters (Online Learning).pptxLecture 2- Practical AD and DA Conveters (Online Learning).pptx
Lecture 2- Practical AD and DA Conveters (Online Learning).pptx
HamzaJaved306957
 
Linear Predictive Coding
Linear Predictive CodingLinear Predictive Coding
Linear Predictive Coding
Shruti Bhatnagar Dasgupta
 

Similar to Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation (20)

ASR_final
ASR_finalASR_final
ASR_final
 
Design of digital filters
Design of digital filtersDesign of digital filters
Design of digital filters
 
Cryosat2013 scagliola
Cryosat2013 scagliolaCryosat2013 scagliola
Cryosat2013 scagliola
 
Analysis of Speech Enhancement Incorporating Speech Recognition
Analysis of Speech Enhancement Incorporating Speech RecognitionAnalysis of Speech Enhancement Incorporating Speech Recognition
Analysis of Speech Enhancement Incorporating Speech Recognition
 
Signal & Image Processing And Analysis For Scientists And Engineers Technical...
Signal & Image Processing And Analysis For Scientists And Engineers Technical...Signal & Image Processing And Analysis For Scientists And Engineers Technical...
Signal & Image Processing And Analysis For Scientists And Engineers Technical...
 
Speech Compression using LPC
Speech Compression using LPCSpeech Compression using LPC
Speech Compression using LPC
 
IJCER (www.ijceronline.com) International Journal of computational Engineerin...
IJCER (www.ijceronline.com) International Journal of computational Engineerin...IJCER (www.ijceronline.com) International Journal of computational Engineerin...
IJCER (www.ijceronline.com) International Journal of computational Engineerin...
 
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLABA GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
A GAUSSIAN MIXTURE MODEL BASED SPEECH RECOGNITION SYSTEM USING MATLAB
 
Speech Compression using LPC
Speech Compression using LPCSpeech Compression using LPC
Speech Compression using LPC
 
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, BregenzFreedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
 
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital ConverterDesign of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
Design of 17-Bit Audio Band Delta-Sigma Analog to Digital Converter
 
H0814247
H0814247H0814247
H0814247
 
Linear Predictive Coding
Linear Predictive CodingLinear Predictive Coding
Linear Predictive Coding
 
Design of iir digital highpass butterworth filter using analog to digital map...
Design of iir digital highpass butterworth filter using analog to digital map...Design of iir digital highpass butterworth filter using analog to digital map...
Design of iir digital highpass butterworth filter using analog to digital map...
 
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
2+3D Photography 2017 – INV 7 The 3D Image Capture Moonshot: Managing the Ene...
 
Introduction to deep learning based voice activity detection
Introduction to deep learning based voice activity detectionIntroduction to deep learning based voice activity detection
Introduction to deep learning based voice activity detection
 
Integrating Sound level meter
Integrating Sound level meterIntegrating Sound level meter
Integrating Sound level meter
 
Slm_1
Slm_1Slm_1
Slm_1
 
Lecture 2- Practical AD and DA Conveters (Online Learning).pptx
Lecture 2- Practical AD and DA Conveters (Online Learning).pptxLecture 2- Practical AD and DA Conveters (Online Learning).pptx
Lecture 2- Practical AD and DA Conveters (Online Learning).pptx
 
Linear Predictive Coding
Linear Predictive CodingLinear Predictive Coding
Linear Predictive Coding
 

Recently uploaded

Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
IJECEIAES
 
Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...
bijceesjournal
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
171ticu
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Sinan KOZAK
 
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
MadhavJungKarki
 
Mechanical Engineering on AAI Summer Training Report-003.pdf
Mechanical Engineering on AAI Summer Training Report-003.pdfMechanical Engineering on AAI Summer Training Report-003.pdf
Mechanical Engineering on AAI Summer Training Report-003.pdf
21UME003TUSHARDEB
 
CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1
PKavitha10
 
An Introduction to the Compiler Designss
An Introduction to the Compiler DesignssAn Introduction to the Compiler Designss
An Introduction to the Compiler Designss
ElakkiaU
 
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
RamonNovais6
 
Digital Twins Computer Networking Paper Presentation.pptx
Digital Twins Computer Networking Paper Presentation.pptxDigital Twins Computer Networking Paper Presentation.pptx
Digital Twins Computer Networking Paper Presentation.pptx
aryanpankaj78
 
TIME TABLE MANAGEMENT SYSTEM testing.pptx
TIME TABLE MANAGEMENT SYSTEM testing.pptxTIME TABLE MANAGEMENT SYSTEM testing.pptx
TIME TABLE MANAGEMENT SYSTEM testing.pptx
CVCSOfficial
 
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
Paris Salesforce Developer Group
 
Applications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdfApplications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdf
Atif Razi
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
IJECEIAES
 
Properties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptxProperties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptx
MDSABBIROJJAMANPAYEL
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
Gino153088
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
Yasser Mahgoub
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
VICTOR MAESTRE RAMIREZ
 
Software Engineering and Project Management - Software Testing + Agile Method...
Software Engineering and Project Management - Software Testing + Agile Method...Software Engineering and Project Management - Software Testing + Agile Method...
Software Engineering and Project Management - Software Testing + Agile Method...
Prakhyath Rai
 
Design and optimization of ion propulsion drone
Design and optimization of ion propulsion droneDesign and optimization of ion propulsion drone
Design and optimization of ion propulsion drone
bjmsejournal
 

Recently uploaded (20)

Embedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoringEmbedded machine learning-based road conditions and driving behavior monitoring
Embedded machine learning-based road conditions and driving behavior monitoring
 
Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...Rainfall intensity duration frequency curve statistical analysis and modeling...
Rainfall intensity duration frequency curve statistical analysis and modeling...
 
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样学校原版美国波士顿大学毕业证学历学位证书原版一模一样
学校原版美国波士顿大学毕业证学历学位证书原版一模一样
 
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
Optimizing Gradle Builds - Gradle DPE Tour Berlin 2024
 
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
1FIDIC-CONSTRUCTION-CONTRACT-2ND-ED-2017-RED-BOOK.pdf
 
Mechanical Engineering on AAI Summer Training Report-003.pdf
Mechanical Engineering on AAI Summer Training Report-003.pdfMechanical Engineering on AAI Summer Training Report-003.pdf
Mechanical Engineering on AAI Summer Training Report-003.pdf
 
CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1CEC 352 - SATELLITE COMMUNICATION UNIT 1
CEC 352 - SATELLITE COMMUNICATION UNIT 1
 
An Introduction to the Compiler Designss
An Introduction to the Compiler DesignssAn Introduction to the Compiler Designss
An Introduction to the Compiler Designss
 
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURSCompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
CompEx~Manual~1210 (2).pdf COMPEX GAS AND VAPOURS
 
Digital Twins Computer Networking Paper Presentation.pptx
Digital Twins Computer Networking Paper Presentation.pptxDigital Twins Computer Networking Paper Presentation.pptx
Digital Twins Computer Networking Paper Presentation.pptx
 
TIME TABLE MANAGEMENT SYSTEM testing.pptx
TIME TABLE MANAGEMENT SYSTEM testing.pptxTIME TABLE MANAGEMENT SYSTEM testing.pptx
TIME TABLE MANAGEMENT SYSTEM testing.pptx
 
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
AI + Data Community Tour - Build the Next Generation of Apps with the Einstei...
 
Applications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdfApplications of artificial Intelligence in Mechanical Engineering.pdf
Applications of artificial Intelligence in Mechanical Engineering.pdf
 
Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...Advanced control scheme of doubly fed induction generator for wind turbine us...
Advanced control scheme of doubly fed induction generator for wind turbine us...
 
Properties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptxProperties Railway Sleepers and Test.pptx
Properties Railway Sleepers and Test.pptx
 
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
4. Mosca vol I -Fisica-Tipler-5ta-Edicion-Vol-1.pdf
 
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
2008 BUILDING CONSTRUCTION Illustrated - Ching Chapter 02 The Building.pdf
 
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student MemberIEEE Aerospace and Electronic Systems Society as a Graduate Student Member
IEEE Aerospace and Electronic Systems Society as a Graduate Student Member
 
Software Engineering and Project Management - Software Testing + Agile Method...
Software Engineering and Project Management - Software Testing + Agile Method...Software Engineering and Project Management - Software Testing + Agile Method...
Software Engineering and Project Management - Software Testing + Agile Method...
 
Design and optimization of ion propulsion drone
Design and optimization of ion propulsion droneDesign and optimization of ion propulsion drone
Design and optimization of ion propulsion drone
 

Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation

  • 1. Abhyudai Yadav Arnika Kumar Dvizma Sinha Jhalak R. Sharma Mentors: Dr. Mandeep Goyal & Dr. Kumar Vishal Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Final Review 7th May, 2015
  • 2. ● Rationale & Objectives ● Methodology ● Literature Review ● Proposed Solution ● MatLab Simulations : Graphs ● Observations & Results ● Difficulties ● Future Work ● References Contents Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 3. Rationale Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Speaker Recognition Process of automatically recognizing who is speaking on the basis of an individual’s discrete information included in speech waves. Problem Presence of Reverberation in input speech signal Reduces recognition rate for speech recognition engines. TTo counter this problem Reverberation Mitigation is required.
  • 4. Methodology Problem Analysis Feature Extraction Training Testing Performance Analysis Mitigation Performance Improvisation Mid Term I Review Mid Term II Review Final Review Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 5. Objectives Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation 1. • To mitigate reverberation from a speech signal. 2. • To improve the performance of automatic speaker recognition system in reverberant environments.
  • 6. • Early reflections don’t distort the speech too much. • Late reflections main cause of distortion. Proposed Solution Idea : To identify & remove the regions of the signal with most reverberant energy. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Figure: Reverberation [Source: hyperphysics.phy]
  • 7. • Envelope of late reflections -> described by an exponential decay curve. • T60 found to get slope of decay curve. • Regions identified in reverberant speech signal having same slope. • Reverberant regions removed using a binary mask. Proposed Solution [Matlab Simulation] 𝑒 𝑟 𝑛 ≈ exp −1 𝜏 𝑛 = 𝑎 𝑛
  • 8. ● Reverberation time characterizes quality of auditory space T60 Parameter Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Figure: A) Speech Signal B) Reverberation tail C) Decay Rate D) Linear fit [Source: Rama Ratnam; Blind estimation of Reverberation Time; University Of Illinois] ● T30 and T20 are other such parameters ● Higher value of T60 signifies more reverberation and vice versa The T60 parameter measures the time taken for the sound level to drop 60 dB below the level at sound cessation
  • 9. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Algorithm Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n) Algorithm run on each band individually By passing through Butterworth filter Of each point on the envelope 0 for exponential 1 otherwise Using Hilbert Transform
  • 10. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 11. Time Domain Plots [A] [B] Time Domain Representations of: [A] Clean Speech [B] Reverberated Signal Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation [Matlab Simulation]
  • 12. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 13. [Matlab Simulation] Butterworth Filter Bank (Log Scale) Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 14. Butterworth Filter Bank (Log Scale, 3 bins/octave) [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 15. Channels obtained from filtering [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 16. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 17. Envelope of the channels [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 18. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 19. Estimating T60 [Matlab Simulation: Convergence of T60] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 20. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 21. Mask for the channels [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 22. Mask for one channel [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 23. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation Reconstruct signal Smooth mask Convert to log- scale (dB) Envelope Follower Filter bank Calculate binary mask Estimate true slope Find local slopes yr(n) yc(n)
  • 24. Process of Dereverberation (in Time Domain)
  • 25. Time Domain Plots [A] [B] Time Domain Representations of: [A] Clean Speech [B] Reverberated Signal [C] De-reverberated Signal Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation [Matlab Simulation] [C]
  • 27. Spectrograms [Matlab Simulation] Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 28. Observations • Higher T60 value -> Better accuracy. • Percussive sounds easier to dereverberate • Speech signals difficult to dereverberate.  As the envelopes situated far apart there is sufficient reverberant decay between them. Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation [Matlab Simulation]
  • 29. Results Speech Recognition has two phases : Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 30. Difficulties Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation • Sampling rate of input: > 10000 Hz • Minimize effect of aliasing in the analog to digital conversion.
  • 31. • Also known as signal processing front end. Future Work Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation
  • 32. Thank You Performance Improvisation of Automatic Speaker Recognition by Spectral Reverberation Mitigation