Questions tagged [speech-recognition]

Speech recognition is the process of converting the spoken word to text, usually without regard to a particular speaker (which is more commonly referred to as "voice recognition").

Filter by
Sorted by
Tagged with
3
votes
0answers
60 views

Frame lookup using audio

I'm an animator trying to get a program that will use machine learning to automate the lip syncing process. The program I need made will have two files given to it: a seed, and an audio file. The seed ...
5
votes
2answers
733 views

Dynamic Time Warping - Comparing Values

Ok, so I'm trying to compare two different speech signals and I have come into a problem. Here goes: I have split the signal into blocks, and I have computed the MFCC coefficients of each block. I ...
2
votes
1answer
532 views

MFCC Classification

I'm working on gender estimation from speech signal and I completed MFCC feature extraction. So now I'm trying to estimate gender from these features. But I have frames for an audio file and I ...
0
votes
1answer
141 views

How to build a simple model for voice-controlled system?

I want to build a speaker independent voice-controlled system. If I ask 10 people to say one unique word(for example"hello"), and record their voice and do LPC(linear predicting coding) upon them, I ...
1
vote
1answer
90 views

Audio from multiple microphones mixed to mono and ASR

I am trying to do speech recognition on an audio file which has 1 channel (mono). My problem is that the channel is comprised of the signal of 3 microphones which have captured the audio at different ...
1
vote
0answers
28 views

Understanding MFCC

I am not from electrical or electronic background so my knowledge will be lacking. MFCC is represented by 39 values for each window frame. 12 values are the mel filter-bank and we get 13th value by ...
0
votes
1answer
113 views

Speech recognition algorithm in GNU Octave to convert speech to text (and vice versa)

Recently, I am building an application in Octave and I saw some algorithms about speech recognition spread around the net, but most of them were implemented in Python, MATLAB, C++ and no one in Octave....
0
votes
0answers
18 views

Compare two spoken words with MFCC and DTW

Processing of the speech signal is done using Matlab. I want to use MFCC (Mel-frequency cepstral coefficients) is implemented using HTK and DTW (dynamic time warping v2.1) algorithms to make a speech ...
0
votes
2answers
263 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
2answers
409 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 ...
0
votes
2answers
795 views

Using k Nearest Neighbour & Dynamic Time Warping for keyword spotting

As stated in the title, I am currently researching a method using k-NN, specifically, as either an alternative or as a supplement to the DTW algorithm in keyword spotting based on MFCC. I have read ...
0
votes
1answer
786 views

Comparing band-pass filters and convolutional kernels?

I am currently working on extracting features for speech recognition purposes. I wanted to try an approach similar to MFCC features in which the center frequency of each band pass filters are placed ...
0
votes
1answer
58 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
23 views

MFCCs: magnitude spectrum vs power spectrum

As I know, the magnitude spectrum and power spectrum have the same information in this step before applying Mel-Filterbank, and the power spectrum may work slightly better in noise. However, I don't ...
0
votes
0answers
29 views

Audio in Image,Graph or WaveForm Representation

Our task is to feed the audio data into the deep learning tensor flow model in the form of image representation(or graphical,waveform). Question is what is the best way of representing audio in image ...
0
votes
0answers
193 views

Understanding liftering as the final step in MFCCs features extraction

In the book here, they apply liftering, as a final step of MFCCs features extraction, to isolate the system component by multiplying the whole cepstrum by a rectangular window centred on lower ...
0
votes
2answers
94 views

filterbank: understand the different responses at the center frequency of each filter

I'm computing filterbank by applying 26 triangular filters on a Mel-scale to the power spectrum of an audio frame to extract frequency bands, and I found that some references use filterbank with equal ...
1
vote
1answer
238 views

MFCCs and mean normalization

I'm reading a blog about extracting MFCCs features for Machine Learning applications, but I didn't understand the following points about the mean normalization: To balance the spectrum and improve ...
0
votes
0answers
18 views

How to generate d-vector or MFCC from wav for speaker diarization and then voice print idenfication?

This task of voice print identification is way beyond my current knowledge. I googled heavily and have basic algorithm: given: conversation btw customer and tech support operator(debt collection in ...
0
votes
0answers
34 views

Difference between Cepstrum and Spectral Envelope

I have no background in Signal processing and i don't understand the difference between cepstrum and spectral envelope. IDFT of the log magnitude of the DFT of a signal is cepstrum then what is ...
0
votes
0answers
64 views

Mixing signal with noise in frequency space (STFT)

I am working on a preprocessing function which is supposed to mix two audio signals clean and noise in the context of neural ...
0
votes
1answer
96 views

python3: speech recognition - calculating features of a WAV file

I am supposed to do this as a school project, but I am kind of lost when it comes to signals and would appreciate your help. Also, English is not my first language, and I will be translating the ...
-1
votes
1answer
53 views

audio classification based only on first few seconds

I record audio notes on my recorder, and each belongs to a certain category ("work", "parenting", "miscellaneous", etc.). At the beginning of each recording, I say the name of the category, then stay ...
1
vote
1answer
22 views

Problem getting DTW to work for speech recognition

I used my ipad to record the words "This. Is. A. Test. This. Is. A. Test.", then downloaded it to my computer as a .MOV file and used ffmpeg to extract audio as WAV. I then used the tuneR package to ...
1
vote
1answer
112 views

Calculate the Standard Deviation of Fundamental Frequency (MFCC)

I'm implementing a gunshot detector following the article "Algorithm for Gunshot Detection Using Mel-Frequency Cepstrum Coefficients (MFCC)" (paywall). In the article, the authors uses 22 features ...
0
votes
0answers
21 views

Targets for Deep Neural Network in acoustic model for speech recognition

I'm trying to come up to speed on the use of deep neural networks for speech recognition and I'm confused about what exactly are the targets for the DNN. I understand that the DNN tries to predict a ...
0
votes
0answers
25 views

Stack features in wakeword model

I'm not clear how is the input features are stacked in wakeword applications. Taking https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/43969.pdf as an example, the author ...
1
vote
1answer
293 views

Preprocess sound for voice recognition in Pocket Sphinx on Raspberry Pi 3

I want to improve my voice recognition system on a Raspberry Pi 3 running Pocket Spinx. It's pretty accurate (limited recognition vocabulary) but is terrible when other noise is around. Is there any ...
0
votes
0answers
56 views

How to use DTW for keyword spotting in human speech?

I am trying to use Dynamic Time Wrapping (DTW) + KNN to detect word occurrences in human speech audio. I am to implement the same approach suggested in this DSP StackExchange post, but I am seeing ...
2
votes
4answers
119 views

Cleaning telephone audio

I have telephone audio at 16kHz and I am trying to use it with a speech transcription engine that converts spoken letters to text. However, the engine has some difficulty transcribing certain letters ...
3
votes
1answer
609 views

Why is the Hamming and Hanning window the most widely used for speech spectograms?

When taking the Short-Time Fourier Transform of the speech signal, windowing is applied to taper away the signal frame at both ends and mitigate introduced artifact at expense of the maximum ...
7
votes
1answer
5k views

What's the correct graphical interpretation of a series of MFCC vectors?

I'm studying speech-recognition, in particular the use of MFCC for feature extraction. All examples I've found online tend to graph a series of MFCC extracted from a particular utterance as follows (...
3
votes
1answer
4k 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 ...
1
vote
1answer
33 views

Are delta and double delta features needed when classifying with LSTM

I understand that the combination of MFCCs, deltas and double deltas is a good feature to be used with HMMs for keyword detection problems. HMMs are limited by Markov Property and this limitation is ...
1
vote
2answers
566 views

How to edit audios so that they have the same length

I have a dataset of audio files, which differ on their lengths by a couple of milliseconds. Some are little shorter than 1s and some are little longer than 1s. They all have the same sampling rate, ...
0
votes
1answer
175 views

Mel separated filter banks

I am currently a bit confused. I am currently trying to understand the MFCC extraction process, and don't understand how the Mel filter banks only provide a scalar output from each filter bank. What ...
1
vote
1answer
117 views

Fast Vocal Tract Length Normalization(VTLN) implementation?

My team is building a speech recognition model. As far as I know, Vocal Tract Length Normalization (VTLN) is an effective warping method that may help improve the model performance by diminishing ...
2
votes
1answer
89 views

Is there any way that we can perform speech recognition without using Fourier transforms?

I am trying to research about speech recognition and why everyone uses Fourier transforms in going about the topic. I know that we get information related to the frequency of each sound uttered which ...
1
vote
1answer
198 views

Split audio file into audio phrases

I have couple of audio recordings of length 5 -7 min. I also have corresponding text to each audio file. I am not subject matter expert on speech to text but I am vaguely aware that much smaller ...
5
votes
3answers
2k views

VAD speech databases

Does anybody know any open voice/speech annotated database for Voice Activity Detection? Or any other corpus for speech evaluation purposes? The ones I know are private: AURORA, TIMIT, SpeechDat-Car, ...
1
vote
1answer
36 views

Services for phoneme recognition? [closed]

I wonder if anyone can point me to services or dev libraries that focus on or suit phoneme recognition? Target audience is children at learning to read stage. We have trialed IBM Watson (not good) ...
8
votes
3answers
1k views

How does Siri recognize me saying “Hey Siri”?

I am trying to understand how my iPhone can continually listening for me saying Hey Siri, Alexa, ...
-1
votes
2answers
171 views

why the MFCC is gender indenpend and used for isolated word recognition

I think female and male have total different MFCC feature for same word such as "one", but why the MFCC could be used for isolated word recognition which use the same HMM model for the same word ...
2
votes
1answer
957 views

Training a CNN-HMM model

I am currently trying to train a CNN-HMM acoustic model for speech recognition. The CNN model is able to detect a center monophone given a context window of x (limits has not been tested yet - but ...
2
votes
1answer
256 views

What Is the Point of Doing the Zero Padding? [duplicate]

What are the advantages and disadvantages of doing Zero-padding, in particular the case of speech signals?
2
votes
1answer
295 views

RNN as posterior probability estimation in speech recognition with HTK

I'm new to speech recognition and deep learning and in a learning phase. I'm trying to follow this paper to learn how to use RNN as posterior probability estimation in an HTK environment. The paper ...
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
0answers
20 views

Finding fundamental particles of sound In ASR systems? [closed]

Q1. Can we divide phonemes which are specific to a language into some more fundamental entities same across all languages ? Q2. Will this help us in achieving language independent acoustic modelling ...
0
votes
1answer
88 views

Signals - Removing Blocks

I'm after some advice on a problem that I am facing regarding signal processing. I am processing a voice clip (.wav) and this voice clip contains noise, and I need to remove the noise and just ...
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 ...