Speech Recognition
Library for performing speech recognition with support for Google Speech Recognition, Wit.ai, IBM Speech to Text, and AT&T Speech to Text.
Links:
Quickstart: pip install SpeechRecognition
. See the "Installing" section for more details.
To quickly try it out, run python -m speech_recognition
after installing.
How to cite this library (APA style):
Zhang, A. (2015). Speech Recognition (Version 3.1) [Software]. Available from https://github.com/Uberi/speech_recognition#readme.
How to cite this library (Chicago style):
Zhang, Anthony. 2015. Speech Recognition (version 3.1).
Also check out the Python Baidu Yuyin API, which is based on an older version of this project, and adds support for Baidu Yuyin.
Examples
See the examples/
directory for usage examples:
- Recognize speech input from the microphone
- Transcribe a WAV audio file
- Show extended recognition results
- Calibrate the recognizer energy threshold for ambient noise levels (see
recognizer_instance.energy_threshold
for details) - Listening to a microphone in the background
Installing
First, make sure you have all the requirements listed in the "Requirements" section.
The easiest way to install this is using pip install SpeechRecognition
.
Otherwise, download the source distribution from PyPI, and extract the archive.
In the folder, run python setup.py install
.
Requirements
Python
The first software requirement is Python 2.6, 2.7, or Python 3.3+. This is required to use the library.
PyAudio (for microphone users)
If you want to use audio input from microphones, PyAudio is also necessary. If not installed, the library will still work, but Microphone
will not be defined.
The official PyAudio builds seem to be broken on Windows. As a result, in the installers
folder you will find unofficial PyAudio builds for Windows that actually work. Run the installer corresponding to your Python version to install PyAudio.
On Debain-based distributions such as Ubuntu, you can generally install PyAudio by running sudo apt-get install python-pyaudio python3-pyaudio
, which will install it for both Python 2 and Python 3.
On other POSIX-based systems, simply use the packages provided on the downloads page linked above, or compile and install it from source.
FLAC (for some systems)
A FLAC encoder is required to encode the audio data to send to the API. If using Windows, OS X, or Linux on an i385-compatible architecture, the encoder is already bundled with this library - you do not need to install anything else.
Otherwise, ensure that you have the flac
command line tool, which is often available through the system package manager.
In summary, this library requires:
- Python 2.6, 2.7, or 3.3+
- PyAudio (required only if you need to use microphone input)
- FLAC encoder (required only if the system is not x86-based Windows/Linux/OS X)
Troubleshooting
Microphone
class is missing/not defined!
The This class is not defined when PyAudio is not available.
Make sure you have PyAudio installed, and make sure you can import it correctly. Test this out by opening a Python console (make sure to use the same version you're running your program with!) and typing in import pyaudio
. If you get an error, PyAudio is not installed or not configured correctly.
See the "Requirements" section for more information about installing PyAudio.
The recognizer tries to recognize speech even when I'm not speaking.
Try increasing the recognizer_instance.energy_threshold
property. This is basically how sensitive the recognizer is to when recognition should start. Higher values mean that it will be less sensitive, which is useful if you are in a loud room.
This value depends entirely on your microphone or audio data. There is no one-size-fits-all value, but good values typically range from 50 to 4000.
The recognizer can't recognize speech right after it starts listening for the first time.
The recognizer_instance.energy_threshold
property is probably set to a value that is too high to start off with, and then being adjusted lower automatically by dynamic energy threshold adjustment. Before it is at a good level, the energy threshold is so high that speech is just considered ambient noise.
The solution is to decrease this threshold, or call recognizer_instance.adjust_for_ambient_noise
beforehand, which will set the threshold to a good value automatically.
The recognizer doesn't understand my particular language/dialect.
Try setting the recognition language to your language/dialect. To do this, see the documentation for recognizer_instance.recognize_google
, recognizer_instance.recognize_wit
, recognizer_instance.recognize_ibm
, and recognizer_instance.recognize_att
.
For example, if your language/dialect is British English, it is better to use "en-GB"
as the language rather than "en-US"
.
UnicodeEncodeError: 'ascii' codec can't encode character
when run.
The code examples throw When you're using Python 2, and your language uses non-ASCII characters, and the terminal or file-like object you're printing to only supports ASCII, an error is thrown when trying to write non-ASCII characters.
This is because in Python 2, recognizer_instance.recognize_google
, recognizer_instance.recognize_wit
, recognizer_instance.recognize_ibm
, and recognizer_instance.recognize_att
return unicode strings (u"something"
) rather than byte strings ("something"
). In Python 3, all strings are unicode strings.
To make printing of unicode strings work in Python 2 as well, replace all print statements in your code of the following form:
print SOME_UNICODE_STRING
With the following:
print SOME_UNICODE_STRING.encode("utf8")
This change, however, will prevent the code from working in Python 3.
PyInstaller.
The program doesn't run when compiled withPyInstaller doesn't know that the FLAC converters need to be bundled with the application. To resolve this, we need to make a PyInstaller hook to include those files and tell PyInstaller where that hook is:
Create a folder in your project directory to store PyInstaller hooks, if the project doesn't already have one. For example, a folder
pyinstaller-hooks
in the project root directory.Create a file called
hook-speech_recognition.py
in that folder, with the following contents:from PyInstaller.hooks.hookutils import collect_data_files datas = collect_data_files("speech_recognition")
When building the project using something like
pyinstaller SOME_SCRIPT.py
, simply supply the--additional-hooks-dir
option set to the PyInstaller hooks folder. For example,pyinstaller --additional-hooks-dir pyinstaller-hooks/ SOME_SCRIPT.py
.
Note that the development versions of PyInstaller already include this hook, and will therefore work out of the box.
On Ubuntu/Debian, I get errors like "jack server is not running or cannot be started" or "Cannot lock down [...] byte memory area (Cannot allocate memory)".
The Linux audio stack is pretty fickle. There are a few things that can cause these issues.
First, make sure JACK is installed - to install it, run sudo apt-get install multimedia-jack
You will then want to configure the JACK daemon correctly to avoid that "Cannot allocate memory" error. Run sudo dpkg-reconfigure -p high jackd2
and select "Yes" to do so.
Now, you will want to make sure your current user is in the audio
group. You can add your current user to this group by running sudo adduser $(whoami) audio
.
Unfortunately, these changes will require you to reboot before they take effect.
After rebooting, run pulseaudio --kill
, followed by jack_control start
, to fix the "jack server is not running or cannot be started" error.
On Ubuntu/Debian, I get annoying output in the terminal saying things like "bt_audio_service_open: [...] Connection refused" and various others.
The "bt_audio_service_open" error means that you have a Bluetooth audio device, but as a physical device is not currently connected, we can't actually use it - if you're not using a Bluetooth microphone, then this can be safely ignored. If you are, and audio isn't working, then double check to make sure your microphone is actually connected. There does not seem to be a simple way to disable these messages.
For errors of the form "ALSA lib [...] Unknown PCM", see this StackOverflow answer. Basically, to get rid of an error of the form "Unknown PCM cards.pcm.rear", simply comment out pcm.rear cards.pcm.rear
in /usr/share/alsa/alsa.conf
, ~/.asoundrc
, and /etc/asound.conf
.
Reference
Microphone(device_index = None, sample_rate = 16000, chunk_size = 1024)
This is available if PyAudio is available, and is undefined otherwise.
Creates a new Microphone
instance, which represents a physical microphone on the computer. Subclass of AudioSource
.
If device_index
is unspecified or None
, the default microphone is used as the audio source. Otherwise, device_index
should be the index of the device to use for audio input.
A device index is an integer between 0 and pyaudio.get_device_count() - 1
(assume we have used import pyaudio
beforehand) inclusive. It represents an audio device such as a microphone or speaker. See the PyAudio documentation for more details.
The microphone audio is recorded in chunks of chunk_size
samples, at a rate of sample_rate
samples per second (Hertz).
Higher sample_rate
values result in better audio quality, but also more bandwidth (and therefore, slower recognition). Additionally, some machines, such as some Raspberry Pi models, can't keep up if this value is too high.
Higher chunk_size
values help avoid triggering on rapidly changing ambient noise, but also makes detection less sensitive. This value, generally, should be left at its default.
This class is a context manager, and is designed to be used with with
statements:
with Microphone() as source: # open the microphone and start recording
pass # do things here - ``source`` is the Microphone instance created above
# the microphone is automatically released at this point
WavFile(filename_or_fileobject)
Creates a new WavFile
instance, which represents a WAV audio file. Subclass of AudioSource
.
If filename_or_fileobject
is a string, then it is interpreted as a path to a WAV audio file (mono or stereo) on the filesystem. Otherwise, filename_or_fileobject
should be a file-like object such as io.BytesIO
or similar. In either case, the specified file is used as the audio source.
This class is a context manager, and is designed to be used with with
statements:
with WavFile("test.wav") as source: # open the WAV file for reading
pass # do things here - ``source`` is the WavFile instance created above
wavfile_instance.DURATION
Represents the length of the audio stored in the WAV file in seconds. This property is only available when inside a context - essentially, that means it should only be accessed inside a with wavfile_instance ...
statement. Outside of contexts, this property is None
.
This is useful when combined with the offset
parameter of recognizer_instance.record
, since when together it is possible to perform speech recognition in chunks.
However, note that recognizing speech in multiple chunks is not the same as recognizing the whole thing at once. If spoken words appear on the boundaries that we split the audio into chunks on, each chunk only gets part of the word, which may result in inaccurate results.
Recognizer()
Creates a new Recognizer
instance, which represents a collection of speech recognition settings and functionality.
recognizer_instance.energy_threshold = 300
Represents the energy level threshold for sounds. Values below this threshold are considered silence, and values above this threshold are considered speech. Can be changed.
This is adjusted automatically if dynamic thresholds are enabled (see recognizer_instance.dynamic_energy_threshold
). A good starting value will generally allow the automatic adjustment to reach a good value faster.
This threshold is associated with the perceived loudness of the sound, but it is a nonlinear relationship. The actual energy threshold you will need depends on your microphone sensitivity or audio data. Typical values for a silent room are 0 to 100, and typical values for speaking are between 150 and 3500. Ambient (non-speaking) noise has a significant impact on what values will work best.
If you're having trouble with the recognizer trying to recognize words even when you're not speaking, try tweaking this to a higher value. If you're having trouble with the recognizer not recognizing your words when you are speaking, try tweaking this to a lower value. For example, a sensitive microphone or microphones in louder rooms might have a ambient energy level of up to 4000:
import speech_recognition as sr
r = sr.Recognizer()
r.energy_threshold = 4000
# rest of your code goes here
The dynamic energy threshold setting can mitigate this by increasing or decreasing this automatically to account for ambient noise. However, this takes time to adjust, so it is still possible to get the false positive detections before the threshold settles into a good value.
To avoid this, use recognizer_instance.adjust_for_ambient_noise(source, duration = 1)
to calibrate the level to a good value. Alternatively, simply set this property to a high value initially (4000 works well), so the threshold is always above ambient noise levels: over time, it will be automatically decreased to account for ambient noise levels.
recognizer_instance.dynamic_energy_threshold = True
Represents whether the energy level threshold (see recognizer_instance.energy_threshold
) for sounds should be automatically adjusted based on the currently ambient noise level while listening. Can be changed.
Recommended for situations where the ambient noise level is unpredictable, which seems to be the majority of use cases. If the ambient noise level is strictly controlled, better results might be achieved by setting this to False
to turn it off.
recognizer_instance.dynamic_energy_adjustment_damping = 0.15
If the dynamic energy threshold setting is enabled (see recognizer_instance.dynamic_energy_threshold
), represents approximately the fraction of the current energy threshold that is retained after one second of dynamic threshold adjustment. Can be changed (not recommended).
Lower values allow for faster adjustment, but also make it more likely to miss certain phrases (especially those with slowly changing volume). This value should be between 0 and 1. As this value approaches 1, dynamic adjustment has less of an effect over time. When this value is 1, dynamic adjustment has no effect.
recognizer_instance.dynamic_energy_adjustment_ratio = 1.5
If the dynamic energy threshold setting is enabled (see recognizer_instance.dynamic_energy_threshold
), represents the minimum factor by which speech is louder than ambient noise. Can be changed (not recommended).
For example, the default value of 1.5 means that speech is at least 1.5 times louder than ambient noise. Smaller values result in more false positives (but fewer false negatives) when ambient noise is loud compared to speech.
recognizer_instance.pause_threshold = 0.8
Represents the minimum length of silence (in seconds) that will register as the end of a phrase. Can be changed.
Smaller values result in the recognition completing more quickly, but might result in slower speakers being cut off.
recognizer_instance.record(source, duration = None, offset = None)
Records up to duration
seconds of audio from source
(an AudioSource
instance) starting at offset
(or at the beginning if not specified) into an AudioData
instance, which it returns.
If duration
is not specified, then it will record until there is no more audio input.
recognizer_instance.adjust_for_ambient_noise(source, duration = 1)
Adjusts the energy threshold dynamically using audio from source
(an AudioSource
instance) to account for ambient noise.
Intended to calibrate the energy threshold with the ambient energy level. Should be used on periods of audio without speech - will stop early if any speech is detected.
The duration
parameter is the maximum number of seconds that it will dynamically adjust the threshold for before returning. This value should be at least 0.5 in order to get a representative sample of the ambient noise.
recognizer_instance.listen(source, timeout = None)
Records a single phrase from source
(an AudioSource
instance) into an AudioData
instance, which it returns.
This is done by waiting until the audio has an energy above recognizer_instance.energy_threshold
(the user has started speaking), and then recording until it encounters recognizer_instance.pause_threshold
seconds of non-speaking or there is no more audio input. The ending silence is not included.
The timeout
parameter is the maximum number of seconds that it will wait for a phrase to start before giving up and throwing an speech_recognition.WaitTimeoutError
exception. If timeout
is None
, it will wait indefinitely.
recognizer_instance.listen_in_background(source, callback)
Spawns a thread to repeatedly record phrases from source
(an AudioSource
instance) into an AudioData
instance and call callback
with that AudioData
instance as soon as each phrase are detected.
Returns a function object that, when called, requests that the background listener thread stop, and waits until it does before returning. The background thread is a daemon and will not stop the program from exiting if there are no other non-daemon threads.
Phrase recognition uses the exact same mechanism as recognizer_instance.listen(source)
.
The callback
parameter is a function that should accept two parameters - the recognizer_instance
, and an AudioData
instance representing the captured audio. Note that callback
function will be called from a non-main thread.
recognizer_instance.recognize_google(audio_data, key = None, language = "en-US", show_all = False)
Performs speech recognition on audio_data
(an AudioData
instance), using the Google Speech Recognition API.
The Google Speech Recognition API key is specified by key
. If not specified, it uses a generic key that works out of the box. This should generally be used for personal or testing purposes only, as it may be revoked by Google at any time.
To obtain your own API key, simply following the steps on the API Keys page at the Chromium Developers site. In the Google Developers Console, Google Speech Recognition is listed as "Speech API".
The recognition language is determined by language
, an IETF language tag like "en-US"
or "en-GB"
, defaulting to US English. A list of supported language codes can be found here. Basically, language codes can be just the language (en
), or a language with a dialect (en-US
).
Returns the most likely transcription if show_all
is false (the default). Otherwise, returns the raw API response as a JSON dictionary.
Raises a speech_recognition.UnknownValueError
exception if the speech is unintelligible. Raises a speech_recognition.RequestError
exception if the key isn't valid, the quota for the key is maxed out, or there is no internet connection.
recognizer_instance.recognize_wit(audio_data, key, show_all = False)
Performs speech recognition on audio_data
(an AudioData
instance), using the Wit.ai API.
The Wit.ai API key is specified by key
. Unfortunately, these are not available without signing up for an account and creating an app.
To get the API key for a Wit.ai app, go to the app settings, go to the section titled "API Details", and look for "Server Access Token" or "Client Access Token". If the desired field is blank, click on the "Reset token" button on the right of the field. Wit.ai API keys are 32-character uppercase alphanumeric strings.
The recognition language is configured in the Wit.ai app settings.
Returns the most likely transcription if show_all
is false (the default). Otherwise, returns the raw API response as a JSON dictionary.
Raises a speech_recognition.UnknownValueError
exception if the speech is unintelligible. Raises a speech_recognition.RequestError
exception if the key isn't valid, the quota for the key is maxed out, or there is no internet connection.
recognizer_instance.recognize_ibm(audio_data, username, password, language="en-US", show_all = False)
Performs speech recognition on audio_data
(an AudioData
instance), using the IBM Speech to Text API.
The IBM Speech to Text username and password are specified by username
and password
, respectively. Unfortunately, these are not available without an account. IBM has published instructions for obtaining these credentials in the IBM Watson Developer Cloud documentation.
The recognition language is determined by language
, an IETF language tag with a dialect like "en-US"
or "es-ES"
, defaulting to US English. At the moment, this supports the tags "en-US"
, "es-ES"
, and "ja-JP"
.
Returns the most likely transcription if show_all
is false (the default). Otherwise, returns the raw API response as a JSON dictionary.
Raises a speech_recognition.UnknownValueError
exception if the speech is unintelligible. Raises a speech_recognition.RequestError
exception if the key isn't valid, or there is no internet connection.
recognizer_instance.recognize_att(audio_data, app_key, app_secret, language="en-US", show_all = False)
Performs speech recognition on audio_data
(an AudioData
instance), using the AT&T Speech to Text API.
The AT&T Speech to Text app key and app secret are specified by app_key
and app_secret
, respectively. Unfortunately, these are not available without signing up for an account and creating an app.
To get the app key and app secret for an AT&T app, go to the My Apps page and look for "APP KEY" and "APP SECRET". AT&T app keys and app secrets are 32-character lowercase alphanumeric strings.
The recognition language is determined by language
, an IETF language tag with a dialect like "en-US"
or "es-ES"
, defaulting to US English. At the moment, this supports the tags "en-US"
, "es-ES"
, and "ja-JP"
.
Returns the most likely transcription if show_all
is false (the default). Otherwise, returns the raw API response as a JSON dictionary.
Raises a speech_recognition.UnknownValueError
exception if the speech is unintelligible. Raises a speech_recognition.RequestError
exception if the key isn't valid, or there is no internet connection.
AudioSource
Base class representing audio sources. Do not instantiate.
Instances of subclasses of this class, such as Microphone
and WavFile
, can be passed to things like recognizer_instance.record
and recognizer_instance.listen
.
AudioData
Storage class for audio data. Do not instantiate.
Instances of this class are returned from recognizer_instance.record
and recognizer_instance.listen
, and are passed to callbacks of recognizer_instance.listen_in_background
.
Developing
To hack on this library, first make sure you have all the requirements listed in the "Requirements" section.
- Most of the library code lives in
speech_recognition/__init__.py
. - Examples live under the
examples/
directory, and the demo script lives inspeech_recognition/__main__.py
. - The FLAC encoder binaries are in the
speech_recognition/
directory.
To install/reinstall the library locally, run python setup.py install
in the project root directory.
Releases are done by running either build.sh
or build.bat
. These are bash and batch scripts, respectively, that build Python source packages and Python Wheels, then upload them to PyPI.
Features and bugfixes should be tested, at minimum, on Python 2.7 and a recent version of Python 3. It is highly recommended to test features on Python 2.6, 2.7, 3.3, and the latest version of Python 3.
Authors
Uberi <azhang9@gmail.com> (Anthony Zhang) bobsayshilol arvindch <achembarpu@gmail.com> (Arvind Chembarpu) kevinismith <kevin_i_smith@yahoo.com> (Kevin Smith) haas85 DelightRun <changxu.mail@gmail.com> maverickagm
Please report bugs and suggestions at the issue tracker!
License
Copyright 2014-2015 Anthony Zhang (Uberi).
The source code is available online at GitHub.
This program is made available under the 3-clause BSD license. See LICENSE.txt
in the project's root directory for more information.