Speech Compression Using Gsm
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View Speech Compression Using Gsm as PDF for free.
More details
- Words: 964
- Pages: 23
Speech Compression using GSM RPE-LTP Faiza Nawaz Bisma Hashmi Mehrin Kiani
Introduction to GSM
The Global System for Mobile Communications is the most popular standard for mobile phones in the world.
GSM service is used by over 2 billion people across more than 212 countries and territories.
The ubiquity of the GSM standard makes international roaming very common between mobile phone operators.
GSM differs significantly from its predecessors in that both signaling and speech channels are Digital call quality. (so it is considered a second generation (2G) mobile phone system.) 2
Architecture Of GSM
3
What is Speech?
Speech Generation:
4
GSM 6.10 Vocoder
Key principle: mathematical modeling of the human vocal tract, leading to an efficient compression method for transmitting speech.
A vocoder (combination of voice and coder) is used to describe GSM systems tailored for the compression of speech.
The sampling rate is 8000 sample/s leading to an average bit rate for the encoded bit stream of 13 K bit/s
5
GSM 6.10 Vocoder
Coding scheme used by GSM 6.10 Vocoder is the Regular Pulse Excitation - Long Term prediction - Linear Predictive Coder (RPE-LTP)
Vocoder sends three kinds of information to the receiver:
Voiced or unvoiced signal (If it is voiced) The period of the excitation signal The parameters of the prediction filter.
6
Linear Predictive Coder (LPC)
LPC algorithm assumes that each speech sample is a linear combination of previous samples.
Speech is sampled, stored and analyzed.
Coefficients calculated from the sample are transmitted and processed in the receiver.
Receiver accurately processes and categorizes voiced and unvoiced sounds.
7
Residual Pulse Excited (RPE) Coder
Determines if the signal is voiced or unvoiced
Determines the period for voiced sounds, encodes periodicity and transmits the coefficient
When the signal changes from voiced to unvoiced, RPE transmits a code that stops the receiver from generating periodic pulses
Starts generating random pulses to correspond to the noise like nature of unvoiced
8
GSM Compression Technologies
Four compression technologies are:
Full Rate Enhanced Full Rate (EFR) Adaptive Multi-Rate (AMR) Half Rate
9
GSM Full Rate Vocoder Using RPE-LTP
Described as an RPE-LTP linear predictive coder.
Models the human vocal tract as a series of cylinders of different widths.
By forcing air through these cylinders, speech sounds can be generated— the LPC coder models this with a set of simultaneous equations.
10
GSM Full Rate Vocoder Using RPE-LTP (…contd)
The input data to the RPE-LTP coder is 20ms of speech composed of 160 samples, each with 13bit resolution.
The data is first passed through a pre-emphasis filter:
Enhances high-frequency components of the signal. (better transmission efficiency.) Also removes any offset on the signal. (Simplifies computation.)
11
LPC Speech Generation
The model of speech generation can be thought of as air passing through a set of different size cylinders.
12
Short Term Analysis Stage
Uses autocorrelation to calculate a set of eight reflection coefficients.
Schur recursion is used to efficiently solve the set of equations resulting from it.
The parameters are then converted into log-area ratios (LARs) -- that allow better quantizing in a smaller number of bits — the first eight parameters of the transmission stream.
13
Short Term Analysis Stage (… contd)
The coded LARs is then decoded back to coefficients and used to filter the input samples.
The reason for decoding the LARs is to ensure that the encoder uses the same information available at the decoder to perform the filtering.
An array of weights lpc[P] is computed such that s[n] ~ lpc[0]*s[n--1]+lpc[1]*s[n--2]+_+lpc[P--1]*s[n--P] (P is usually between 8 and 14, GSM uses 8.)
14
Long Term Prediction Stage
The 160 samples are split into 4 sub-windows of 40 samples each.
15
Long Term Prediction Stage (…contd)
The long-term predictor produces two parameters for each sub window: the lag and the gain.
The LTP lag describes the source of the copy in time.
The LTP gain describes the scaling factor.
16
Calculating Lag and Gain
LAG: Compute resemblance by correlation. correlation of x[n] and y[n] = Sum of products x[n]*y[n-lag]
GAIN: Maximum correlation divided by the energy of the reconstructed short-term residual signal.
17
Residual Pulse Encoding
To remove the long-term predictable signal from its input, the algorithm then subtracts the scaled 40 samples.
The residual signal is either weak or random and consequently cheaper to encode and transmit.
18
Residual Signal(…contd)
The algorithm down-samples by a factor of three, discarding two out of three sample values.
Results in four evenly spaced 13-value subsequences to choose from, starting with samples 1, 2, 3, and 4.
The algorithm picks the sequence with the most energy.
That leaves us with 13 3-bit sample values and a 6-bit scaling factor that turns the PCM encoding into an APCM 19
Speech Decoder
Decoder consists of three parts
RPE Decoding
LTP synthesis filter
LPC short term synthesis filter
20
Speech Decoder(…contd)
21
Speech Decoder (…contd)
Algorithm multiplies the 13 3-bit samples by the scaling factor and expands them back into 40 samples, zero-padding the gaps
Resulting residual pulse is fed to the long-term synthesis filter
40-sample segment is cut from the old estimated short-term residual signal, scaled by the LTP gain and added to the incoming pulse
Estimated short-term residual signal passes through the short-term synthesis filter whose reflection coefficients are calculated by the LPC module
Noise from the excited long-term synthesis filter passes through the tubes of the simulated vocal tract--and emerges as speech 22
QUESTIONS ???
23
Introduction to GSM
The Global System for Mobile Communications is the most popular standard for mobile phones in the world.
GSM service is used by over 2 billion people across more than 212 countries and territories.
The ubiquity of the GSM standard makes international roaming very common between mobile phone operators.
GSM differs significantly from its predecessors in that both signaling and speech channels are Digital call quality. (so it is considered a second generation (2G) mobile phone system.) 2
Architecture Of GSM
3
What is Speech?
Speech Generation:
4
GSM 6.10 Vocoder
Key principle: mathematical modeling of the human vocal tract, leading to an efficient compression method for transmitting speech.
A vocoder (combination of voice and coder) is used to describe GSM systems tailored for the compression of speech.
The sampling rate is 8000 sample/s leading to an average bit rate for the encoded bit stream of 13 K bit/s
5
GSM 6.10 Vocoder
Coding scheme used by GSM 6.10 Vocoder is the Regular Pulse Excitation - Long Term prediction - Linear Predictive Coder (RPE-LTP)
Vocoder sends three kinds of information to the receiver:
Voiced or unvoiced signal (If it is voiced) The period of the excitation signal The parameters of the prediction filter.
6
Linear Predictive Coder (LPC)
LPC algorithm assumes that each speech sample is a linear combination of previous samples.
Speech is sampled, stored and analyzed.
Coefficients calculated from the sample are transmitted and processed in the receiver.
Receiver accurately processes and categorizes voiced and unvoiced sounds.
7
Residual Pulse Excited (RPE) Coder
Determines if the signal is voiced or unvoiced
Determines the period for voiced sounds, encodes periodicity and transmits the coefficient
When the signal changes from voiced to unvoiced, RPE transmits a code that stops the receiver from generating periodic pulses
Starts generating random pulses to correspond to the noise like nature of unvoiced
8
GSM Compression Technologies
Four compression technologies are:
Full Rate Enhanced Full Rate (EFR) Adaptive Multi-Rate (AMR) Half Rate
9
GSM Full Rate Vocoder Using RPE-LTP
Described as an RPE-LTP linear predictive coder.
Models the human vocal tract as a series of cylinders of different widths.
By forcing air through these cylinders, speech sounds can be generated— the LPC coder models this with a set of simultaneous equations.
10
GSM Full Rate Vocoder Using RPE-LTP (…contd)
The input data to the RPE-LTP coder is 20ms of speech composed of 160 samples, each with 13bit resolution.
The data is first passed through a pre-emphasis filter:
Enhances high-frequency components of the signal. (better transmission efficiency.) Also removes any offset on the signal. (Simplifies computation.)
11
LPC Speech Generation
The model of speech generation can be thought of as air passing through a set of different size cylinders.
12
Short Term Analysis Stage
Uses autocorrelation to calculate a set of eight reflection coefficients.
Schur recursion is used to efficiently solve the set of equations resulting from it.
The parameters are then converted into log-area ratios (LARs) -- that allow better quantizing in a smaller number of bits — the first eight parameters of the transmission stream.
13
Short Term Analysis Stage (… contd)
The coded LARs is then decoded back to coefficients and used to filter the input samples.
The reason for decoding the LARs is to ensure that the encoder uses the same information available at the decoder to perform the filtering.
An array of weights lpc[P] is computed such that s[n] ~ lpc[0]*s[n--1]+lpc[1]*s[n--2]+_+lpc[P--1]*s[n--P] (P is usually between 8 and 14, GSM uses 8.)
14
Long Term Prediction Stage
The 160 samples are split into 4 sub-windows of 40 samples each.
15
Long Term Prediction Stage (…contd)
The long-term predictor produces two parameters for each sub window: the lag and the gain.
The LTP lag describes the source of the copy in time.
The LTP gain describes the scaling factor.
16
Calculating Lag and Gain
LAG: Compute resemblance by correlation. correlation of x[n] and y[n] = Sum of products x[n]*y[n-lag]
GAIN: Maximum correlation divided by the energy of the reconstructed short-term residual signal.
17
Residual Pulse Encoding
To remove the long-term predictable signal from its input, the algorithm then subtracts the scaled 40 samples.
The residual signal is either weak or random and consequently cheaper to encode and transmit.
18
Residual Signal(…contd)
The algorithm down-samples by a factor of three, discarding two out of three sample values.
Results in four evenly spaced 13-value subsequences to choose from, starting with samples 1, 2, 3, and 4.
The algorithm picks the sequence with the most energy.
That leaves us with 13 3-bit sample values and a 6-bit scaling factor that turns the PCM encoding into an APCM 19
Speech Decoder
Decoder consists of three parts
RPE Decoding
LTP synthesis filter
LPC short term synthesis filter
20
Speech Decoder(…contd)
21
Speech Decoder (…contd)
Algorithm multiplies the 13 3-bit samples by the scaling factor and expands them back into 40 samples, zero-padding the gaps
Resulting residual pulse is fed to the long-term synthesis filter
40-sample segment is cut from the old estimated short-term residual signal, scaled by the LTP gain and added to the incoming pulse
Estimated short-term residual signal passes through the short-term synthesis filter whose reflection coefficients are calculated by the LPC module
Noise from the excited long-term synthesis filter passes through the tubes of the simulated vocal tract--and emerges as speech 22
QUESTIONS ???
23
Related Documents
Speech Compression Using Gsm
November 2019 4
Speech Compression Using Gsm Q&a
November 2019 1
9.gsm Speech & Channel Coding
June 2020 3
Speech Separation Using Ica
June 2020 3
Compression
April 2020 21