Questions tagged [speech]
The speech tag has no usage guidance.
120
questions
0
votes
0answers
11 views
Help in LBG Vector Quantization - Splitting Factor
I'm currently making a program for speech recognition. In the step of codebook generation using the LBG (Linde-Buzo-Gray) algorithm, I've read that the splitting factor $\varepsilon = 0.01$ (generally)...
0
votes
2answers
34 views
Audio keeps echoing continuously when using 2 separate devices for audio and recording on video calls
I had used 2 separate devices for a video call. Desktop for video and audio and phone for recording, since the desktop microphone is not working properly. Unfortunately when I started speaking, my ...
2
votes
1answer
73 views
Convolution : Anechoic Speech signal and Impulse response
I am trying to perform convolution in MATLAB of an anechoic speech signal (2.1 s) and a impulse response (20 s) using the following code:
...
0
votes
1answer
88 views
Removing background noise algorithm
What is the algorithm for removing background noise from an audio clip of human speech? Say I have a 10 second audio clip of somebody talking, and there is background noise of them tapping on their ...
0
votes
0answers
49 views
Simulating packet loss in audio
I have loaded a wav file into a Python Numpy array.
I would like to "perturb the signal" such that the audio sounds akin to audio streamed over a poor internet connection.
How can I mutate the ...
0
votes
0answers
19 views
Is audio normalisation required for a speaker recognition/speech classification task? How to do it?
Not entirely sure if this Stack or a ML-related stack would be a fit for this question, but I focus on the signal processing here on this question, so I decided this one
I'm developing a speech ...
2
votes
1answer
44 views
How many bps would be required to reproduce the words in speech recognizably?
In Robert Gallager's lecture notes for 6.450 Principles of Digital Communications I, Exercise 2.1 asks an interesting conceptual question:
Voice waveforms could be converted to binary data by ...
3
votes
1answer
6k views
What are i-vectors and x-vectors in the context of Speech Recognition?
I have read that i-vectors and x-vectors are widely used in speaker recognition tasks but I don't get the difference between them and how exactly they work. Can someone explain it starting from the ...
0
votes
1answer
36 views
MATLAB & Power Spectral Density - Error - Getting the same range of values regardless of sound type?
I wrote code [attached at end of question] to take .wav files of child vocalizations and tried to see if there are valid ranges for the formants F1:F2:F3 for the 5 ...
4
votes
2answers
2k views
Mel Cepstral Distortion
I am working on a speech synthesis model and I am looking to evaluate my synthesized speech. I found that most people use the Mel Cepstral Distortion (MCD) which can be calculated by the formula:
$$\...
0
votes
1answer
776 views
Audio/speech detection, verification
I would like to write a program to compare two '.wav' files, one of which is the original played file with human speech and the other is a recorded '.wav' file. The program I wish to write in python ...
1
vote
2answers
252 views
Converting speech audio to telephone audio
I am trying to train a machine learning algorithm on telephony speech audio. However, there isn't really enough data for this anywhere that I can find. My solution is to just use speech audio from ...
0
votes
1answer
27 views
Characterization of the amount of modulation in a signal
We have used the ICRA babble noise to generate an 8-talker babble. As a background - babble noise is used as a substitute for competing talkers as babble noise lacks lexical content, but contains the ...
0
votes
1answer
189 views
Frequency detection in audio and what parameter or set of parameters help/limit frequency detection
I'm trying to learn audio and voice/speech processing techniques out of interest. I intend to learn, experiment with algorithms and design, implementation of such systems, frameworks using programming ...
1
vote
1answer
1k views
MFCC window size at different sampling rates
The general recommendation for window size when calculating MFCC seems to be 20-40 msec. This is most often recommended in a context of 16000 samples per second, so leading to a window containing 320-...
1
vote
1answer
92 views
About speech preprocessing in mobile phones
I understand that cell phones transmit speech signals in digital form through 2G/3G or other standards. But I could not find anything about the signal processing method lying in between the analog ...
0
votes
0answers
97 views
Passband filter before Hilbert transform
I have a .wav file of human speech sampled at 48kHz. Before making the analytic signal with:
hilbert(s_real)
lets say in range 200 - 3,500 Hz, I probably need to do pass-band filtering, say 300 - 3,...
2
votes
3answers
274 views
Phase change rate of human speech
I have a .wav file of human speech, sampled at 48kHz, with useful signal up to 5kHz, and almost no noise. If I do Hilbert transform in MatLab, I get the Analytic signal:
$$ s_\mathrm{a}[n] = s_\...
1
vote
1answer
2k views
How to convert NIST to RIFF wav file
I have a dataset of audios(namely TIDIGITS, collected at Texas Instruments, Inc.) and they are in .wav format. However, when I try to read them from python, I get the following error:
...
-2
votes
1answer
176 views
How do i remove the DC component from an audio signal using Matlab2016?
I have been doing speech processing in matlab2016, and am stack on how to remove DC component. I tried using the the function y=x-x(mean) but it seems not to ...
0
votes
1answer
194 views
Normalize speech features per dimensions or across dimensions
I am working on pipelines for audio processing (audio features that will go into a neural net to create some embeddings).
We are using two features : Filter banks, or MFCC.
I was wondering what was ...
0
votes
1answer
2k views
2
votes
1answer
72 views
How exactly does sound get encoded by a cochlear implant processor?
I've been wanting to know more about how sound is processed by a cochlear implant processor, how they are programmed, and how they stimulate the magnets and electrodes inside.
How do you take pure ...
2
votes
0answers
125 views
Speaker diarization tool for acoustic analysis
I'm researching certain acoustic prosodic features during live clinical interactions. Because of the setting, the recordings always have cross channel bleed, and in some cases everything is recorded ...
2
votes
1answer
242 views
Speech signal formant spectrogram. How can we determine F1?
The formant spectrogram in this page points F1 and F2 as below;
In bottom-right F1 movement, I thought F1 could be below blue one. But how can we determine F1 is upper one not below one?
Is it ...
-1
votes
1answer
195 views
symmetrical wav audio in time domain and Phase spectrum significance
I have plotted many wav files in MALTAB and tried to understand their time domain plot , Magnitude spectrum and Phase spectrum.
But I don't get why time domain plots of speech wav files are ...
0
votes
2answers
277 views
MFCC in Speech Recognition and ANN
I have a dataset with over 2000 examples of digits (0 to 9) recorded on wav files. I'm trying to process these files with MFCC so that later I could train an Artificial Neural Network (ANN).
My ...
4
votes
1answer
1k views
Speech recognition: words vs. phonemes
One could train a speech recognizer with words or phonemes. I guess that they both have their pros and cons:
with words there would be a lower error rate as phonemes can be put together building a ...
1
vote
1answer
445 views
autoregressive moving average code implementation
I am new to DSP and i am trying to take a wav (human speech) file and apply ARMA filter and plot its PSD graph in python.
I see that there are a lot of AR implementations but almost none ARMA. I ...
0
votes
1answer
138 views
Trying to compare two speech samples that have been recorded by smartphones in pockets
I'm trying to compare two audio samples of speech from two different Android smartphones to determine if the people are co-located. I'm doing this using the frequency bins of an FFT between 300 Hz - ...
1
vote
1answer
76 views
Is there a method to combine speech signals recorded by different microphones?
I have 3 speech files of the same speech recorded by 3 different microphones at the same time.
I'm only concerned to improve the speech quality (in terms of SNR and intelligibility) by combining the ...
1
vote
1answer
1k views
The minimal signal to noise ratio (SNR) for people to understand a speech in the noisy background
This question seems easy to find an answer but I got lost in the end. I am trying to find the threshold for people to understand speech in a noisy environment. For example, assume the background noise ...
2
votes
0answers
1k views
Speex echo cancellation takes a long time (~20s) to converge
I have been experimenting with Speex's echo canceller during the past few days.
I got the echo canceller to work, however, it takes a long time until it actually begins filtering out the echo. It ...
0
votes
1answer
357 views
Raw speech data into spectrogram
I am using Delphi trying to make a very simple speech recognition system. The system collects raw data well, but I have a hard time turning the data into a spectrogram:
The system collects an array ...
-1
votes
1answer
3k views
Compare two audio tracks
I need to compare two audio tracks and compute the difference. The difference is how close they are in terms of pronunciation, rhythm etc. ignoring speaker voice characteristics.
The first track is ...
2
votes
3answers
1k views
Understanding LPC for Formant Estimation
I went through the Matlab tutorial on Formant Estimation using LPC Coefficients. Though I vaguely understand the details, it's not entirely clear why we need to do this. From http://person2.sol.lu.se/...
1
vote
1answer
834 views
Is Cross correlation between stationary random signal and noise zero for all lags?
If $x(n)$ is a stationary random process (let say small 20ms window of speech which can be assumed stationary) and $v(n)$ is said to be uncorrelated noise, can we say that cross correlation of $x(n)$ ...
3
votes
0answers
4k views
What's spectral “tilt”?
I am looking at article Speech-in-noise intelligibility improvement based on spectral shaping and dynamic range compression. In paragraph 2.2 the article mentions "tilt" of the spectral envelope. The ...
0
votes
1answer
68 views
Fine-grained evaluation of speech recognition results
I know that usually ASR systems are evaluated using WER (word error rate), and I was wondering if it makes sense to use a more fine grained approach. E.g. transforming both the ASR hypothesis and the ...
1
vote
1answer
403 views
audio quality evaluation
I am creating a dataset, where in many speakers are recording their speech signals using different microphones and in different environment. I would like to evaluate the quality of audio and discard ...
1
vote
2answers
703 views
Why discrete cosine transform may not maintain locality
In paper Convolutional neural network for speech recognition, they say
I didn't understand the highlighted sentence.
0
votes
2answers
3k views
FFT spectrogram in log frequency space. How?
I was reading a paper and the authors say that they extract FFT spectrograms and log-scale in frequency domain. I understand how the general spectrogram is computed, but how to make it log-scale in ...
3
votes
0answers
433 views
Implementing blind deconvolution in MATLAB
I want to implement blind deconvolution for the signal $ r(n) = h(n) \star s(n) + a(n)$ in MATLAB where
$r(n)$ is the recorded speech
$h(n)$ is impulse response of room acoustics
$s(n)$ is desired ...
4
votes
2answers
422 views
Matching LPC frequency response to raw sample data frequency response
Background:
The intention is feature/formant extraction from speech input.
I have used an FFT on raw samples in the past and have done simple peak picking from a weighted average of frequencies over ...
2
votes
3answers
2k views
Speech corpora for VAD
I'm doing a student research project connected with VAD.
And I want to find speech corpora with added noise.
I found many references to the NIST Speaker Recognition Corpora. But unfortunately, all ...
4
votes
1answer
480 views
What does MATLAB display when plotting a sound signal?
When you load the speech file 'mtlb' in MATLAB (which is by the way the voice of a woman saying the word 'MATLAB'), and plot the data, you get the following plot:
My question is this: what does the ...
0
votes
1answer
1k views
Calculating Delta MFCC, correct formula?
I have seen different formulations of how to calculate delta MFCCs. In general, I want to be consistent in such a way that if I have other features to look at other than MFCCs, the way I would ...
2
votes
1answer
201 views
Using 8 bit samples for speech comparison
I am making a speech comparison app in javascript.
I found this dsp.SE post:
Speech comparison algorithm for rating on similarities,
where I read it is best to use mfcc and dtw algorithms.
I ...
0
votes
1answer
113 views
Mask a voice tone
What are the best ways to mask a voice tone, like what they do during news shows on television sometimes? Is there a way that is not reversibile? I think that a ...
0
votes
1answer
376 views
Obtaining signal energy modulation for voice activity detection
Im currently working on VAD (voice activity detection) topic. Since I'm beginner in digital signal processing, I've started to wonder if calculating single frame energy (after aplying FFT to it) and ...
