Try Gemini 1.5 Pro, our most advanced multimodal model in Vertex AI, and see what you can build with a 1M token context window.Try Gemini 1.5 Pro, our most advanced multimodal model in Vertex AI, and see what you can build with a 1M token context window.
Stay organized with collections
Save and categorize content based on your preferences.
This section demonstrates how to transcribe streaming audio, like the
input from a microphone, to text.
Streaming speech recognition allows you to stream audio to
Speech-to-Text and receive a stream speech recognition results
in real time as the audio is processed. See also the
audio limits for streaming speech recognition requests.
Streaming speech recognition is available via gRPC only.
Perform streaming speech recognition on a local file
Below is an example of performing streaming speech recognition on a local audio
file. There is a 10 MB limit on all streaming requests sent to the API. This
limit applies to to both the initial StreamingRecognize request
and the size of each individual message in the stream. Exceeding this limit will
throw an error.
/**
* Performs streaming speech recognition on raw PCM audio data.
*
* @param fileName the path to a PCM audio file to transcribe.
*/
public static void streamingRecognizeFile(String fileName) throws Exception, IOException {
Path path = Paths.get(fileName);
byte[] data = Files.readAllBytes(path);
// Instantiates a client with GOOGLE_APPLICATION_CREDENTIALS
try (SpeechClient speech = SpeechClient.create()) {
// Configure request with local raw PCM audio
RecognitionConfig recConfig =
RecognitionConfig.newBuilder()
.setEncoding(AudioEncoding.LINEAR16)
.setLanguageCode("en-US")
.setSampleRateHertz(16000)
.setModel("default")
.build();
StreamingRecognitionConfig config =
StreamingRecognitionConfig.newBuilder().setConfig(recConfig).build();
class ResponseApiStreamingObserver<T> implements ApiStreamObserver<T> {
private final SettableFuture<List<T>> future = SettableFuture.create();
private final List<T> messages = new java.util.ArrayList<T>();
@Override
public void onNext(T message) {
messages.add(message);
}
@Override
public void onError(Throwable t) {
future.setException(t);
}
@Override
public void onCompleted() {
future.set(messages);
}
// Returns the SettableFuture object to get received messages / exceptions.
public SettableFuture<List<T>> future() {
return future;
}
}
ResponseApiStreamingObserver<StreamingRecognizeResponse> responseObserver =
new ResponseApiStreamingObserver<>();
BidiStreamingCallable<StreamingRecognizeRequest, StreamingRecognizeResponse> callable =
speech.streamingRecognizeCallable();
ApiStreamObserver<StreamingRecognizeRequest> requestObserver =
callable.bidiStreamingCall(responseObserver);
// The first request must **only** contain the audio configuration:
requestObserver.onNext(
StreamingRecognizeRequest.newBuilder().setStreamingConfig(config).build());
// Subsequent requests must **only** contain the audio data.
requestObserver.onNext(
StreamingRecognizeRequest.newBuilder()
.setAudioContent(ByteString.copyFrom(data))
.build());
// Mark transmission as completed after sending the data.
requestObserver.onCompleted();
List<StreamingRecognizeResponse> responses = responseObserver.future().get();
for (StreamingRecognizeResponse response : responses) {
// For streaming recognize, the results list has one is_final result (if available) followed
// by a number of in-progress results (if iterim_results is true) for subsequent utterances.
// Just print the first result here.
StreamingRecognitionResult result = response.getResultsList().get(0);
// There can be several alternative transcripts for a given chunk of speech. Just use the
// first (most likely) one here.
SpeechRecognitionAlternative alternative = result.getAlternativesList().get(0);
System.out.printf("Transcript : %s\n", alternative.getTranscript());
}
}
}
const fs = require('fs');
// Imports the Google Cloud client library
const speech = require('@google-cloud/speech');
// Creates a client
const client = new speech.SpeechClient();
/**
* TODO(developer): Uncomment the following lines before running the sample.
*/
// const filename = 'Local path to audio file, e.g. /path/to/audio.raw';
// const encoding = 'Encoding of the audio file, e.g. LINEAR16';
// const sampleRateHertz = 16000;
// const languageCode = 'BCP-47 language code, e.g. en-US';
const request = {
config: {
encoding: encoding,
sampleRateHertz: sampleRateHertz,
languageCode: languageCode,
},
interimResults: false, // If you want interim results, set this to true
};
// Stream the audio to the Google Cloud Speech API
const recognizeStream = client
.streamingRecognize(request)
.on('error', console.error)
.on('data', data => {
console.log(
`Transcription: ${data.results[0].alternatives[0].transcript}`
);
});
// Stream an audio file from disk to the Speech API, e.g. "./resources/audio.raw"
fs.createReadStream(filename).pipe(recognizeStream);
def transcribe_streaming(stream_file: str) -> speech.RecognitionConfig:
"""Streams transcription of the given audio file."""
client = speech.SpeechClient()
with open(stream_file, "rb") as audio_file:
content = audio_file.read()
# In practice, stream should be a generator yielding chunks of audio data.
stream = [content]
requests = (
speech.StreamingRecognizeRequest(audio_content=chunk) for chunk in stream
)
config = speech.RecognitionConfig(
encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
sample_rate_hertz=16000,
language_code="en-US",
)
streaming_config = speech.StreamingRecognitionConfig(config=config)
# streaming_recognize returns a generator.
responses = client.streaming_recognize(
config=streaming_config,
requests=requests,
)
for response in responses:
# Once the transcription has settled, the first result will contain the
# is_final result. The other results will be for subsequent portions of
# the audio.
for result in response.results:
print(f"Finished: {result.is_final}")
print(f"Stability: {result.stability}")
alternatives = result.alternatives
# The alternatives are ordered from most likely to least.
for alternative in alternatives:
print(f"Confidence: {alternative.confidence}")
print(f"Transcript: {alternative.transcript}")
While you can stream a local audio file to the Speech-to-Text API,
it is recommended that you perform synchronous or
asynchronous audio recognition for batch mode results.
Perform streaming speech recognition on an audio stream
Speech-to-Text can also perform recognition on streaming, real-time
audio.
Here is an example of performing streaming speech recognition on an audio stream
received from a microphone:
import queue
import re
import sys
from google.cloud import speech
import pyaudio
# Audio recording parameters
RATE = 16000
CHUNK = int(RATE / 10) # 100ms
class MicrophoneStream:
"""Opens a recording stream as a generator yielding the audio chunks."""
def __init__(self: object, rate: int = RATE, chunk: int = CHUNK) -> None:
"""The audio -- and generator -- is guaranteed to be on the main thread."""
self._rate = rate
self._chunk = chunk
# Create a thread-safe buffer of audio data
self._buff = queue.Queue()
self.closed = True
def __enter__(self: object) -> object:
self._audio_interface = pyaudio.PyAudio()
self._audio_stream = self._audio_interface.open(
format=pyaudio.paInt16,
# The API currently only supports 1-channel (mono) audio
# https://goo.gl/z757pE
channels=1,
rate=self._rate,
input=True,
frames_per_buffer=self._chunk,
# Run the audio stream asynchronously to fill the buffer object.
# This is necessary so that the input device's buffer doesn't
# overflow while the calling thread makes network requests, etc.
stream_callback=self._fill_buffer,
)
self.closed = False
return self
def __exit__(
self: object,
type: object,
value: object,
traceback: object,
) -> None:
"""Closes the stream, regardless of whether the connection was lost or not."""
self._audio_stream.stop_stream()
self._audio_stream.close()
self.closed = True
# Signal the generator to terminate so that the client's
# streaming_recognize method will not block the process termination.
self._buff.put(None)
self._audio_interface.terminate()
def _fill_buffer(
self: object,
in_data: object,
frame_count: int,
time_info: object,
status_flags: object,
) -> object:
"""Continuously collect data from the audio stream, into the buffer.
Args:
in_data: The audio data as a bytes object
frame_count: The number of frames captured
time_info: The time information
status_flags: The status flags
Returns:
The audio data as a bytes object
"""
self._buff.put(in_data)
return None, pyaudio.paContinue
def generator(self: object) -> object:
"""Generates audio chunks from the stream of audio data in chunks.
Args:
self: The MicrophoneStream object
Returns:
A generator that outputs audio chunks.
"""
while not self.closed:
# Use a blocking get() to ensure there's at least one chunk of
# data, and stop iteration if the chunk is None, indicating the
# end of the audio stream.
chunk = self._buff.get()
if chunk is None:
return
data = [chunk]
# Now consume whatever other data's still buffered.
while True:
try:
chunk = self._buff.get(block=False)
if chunk is None:
return
data.append(chunk)
except queue.Empty:
break
yield b"".join(data)
def listen_print_loop(responses: object) -> str:
"""Iterates through server responses and prints them.
The responses passed is a generator that will block until a response
is provided by the server.
Each response may contain multiple results, and each result may contain
multiple alternatives; for details, see https://goo.gl/tjCPAU. Here we
print only the transcription for the top alternative of the top result.
In this case, responses are provided for interim results as well. If the
response is an interim one, print a line feed at the end of it, to allow
the next result to overwrite it, until the response is a final one. For the
final one, print a newline to preserve the finalized transcription.
Args:
responses: List of server responses
Returns:
The transcribed text.
"""
num_chars_printed = 0
for response in responses:
if not response.results:
continue
# The `results` list is consecutive. For streaming, we only care about
# the first result being considered, since once it's `is_final`, it
# moves on to considering the next utterance.
result = response.results[0]
if not result.alternatives:
continue
# Display the transcription of the top alternative.
transcript = result.alternatives[0].transcript
# Display interim results, but with a carriage return at the end of the
# line, so subsequent lines will overwrite them.
#
# If the previous result was longer than this one, we need to print
# some extra spaces to overwrite the previous result
overwrite_chars = " " * (num_chars_printed - len(transcript))
if not result.is_final:
sys.stdout.write(transcript + overwrite_chars + "\r")
sys.stdout.flush()
num_chars_printed = len(transcript)
else:
print(transcript + overwrite_chars)
# Exit recognition if any of the transcribed phrases could be
# one of our keywords.
if re.search(r"\b(exit|quit)\b", transcript, re.I):
print("Exiting..")
break
num_chars_printed = 0
return transcript
def main() -> None:
"""Transcribe speech from audio file."""
# See http://g.co/cloud/speech/docs/languages
# for a list of supported languages.
language_code = "en-US" # a BCP-47 language tag
client = speech.SpeechClient()
config = speech.RecognitionConfig(
encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
sample_rate_hertz=RATE,
language_code=language_code,
)
streaming_config = speech.StreamingRecognitionConfig(
config=config, interim_results=True
)
with MicrophoneStream(RATE, CHUNK) as stream:
audio_generator = stream.generator()
requests = (
speech.StreamingRecognizeRequest(audio_content=content)
for content in audio_generator
)
responses = client.streaming_recognize(streaming_config, requests)
# Now, put the transcription responses to use.
listen_print_loop(responses)
if __name__ == "__main__":
main()
import queue
import re
import sys
import time
from google.cloud import speech
import pyaudio
# Audio recording parameters
STREAMING_LIMIT = 240000 # 4 minutes
SAMPLE_RATE = 16000
CHUNK_SIZE = int(SAMPLE_RATE / 10) # 100ms
RED = "\033[0;31m"
GREEN = "\033[0;32m"
YELLOW = "\033[0;33m"
def get_current_time() -> int:
"""Return Current Time in MS.
Returns:
int: Current Time in MS.
"""
return int(round(time.time() * 1000))
class ResumableMicrophoneStream:
"""Opens a recording stream as a generator yielding the audio chunks."""
def __init__(
self: object,
rate: int,
chunk_size: int,
) -> None:
"""Creates a resumable microphone stream.
Args:
self: The class instance.
rate: The audio file's sampling rate.
chunk_size: The audio file's chunk size.
returns: None
"""
self._rate = rate
self.chunk_size = chunk_size
self._num_channels = 1
self._buff = queue.Queue()
self.closed = True
self.start_time = get_current_time()
self.restart_counter = 0
self.audio_input = []
self.last_audio_input = []
self.result_end_time = 0
self.is_final_end_time = 0
self.final_request_end_time = 0
self.bridging_offset = 0
self.last_transcript_was_final = False
self.new_stream = True
self._audio_interface = pyaudio.PyAudio()
self._audio_stream = self._audio_interface.open(
format=pyaudio.paInt16,
channels=self._num_channels,
rate=self._rate,
input=True,
frames_per_buffer=self.chunk_size,
# Run the audio stream asynchronously to fill the buffer object.
# This is necessary so that the input device's buffer doesn't
# overflow while the calling thread makes network requests, etc.
stream_callback=self._fill_buffer,
)
def __enter__(self: object) -> object:
"""Opens the stream.
Args:
self: The class instance.
returns: None
"""
self.closed = False
return self
def __exit__(
self: object,
type: object,
value: object,
traceback: object,
) -> object:
"""Closes the stream and releases resources.
Args:
self: The class instance.
type: The exception type.
value: The exception value.
traceback: The exception traceback.
returns: None
"""
self._audio_stream.stop_stream()
self._audio_stream.close()
self.closed = True
# Signal the generator to terminate so that the client's
# streaming_recognize method will not block the process termination.
self._buff.put(None)
self._audio_interface.terminate()
def _fill_buffer(
self: object,
in_data: object,
*args: object,
**kwargs: object,
) -> object:
"""Continuously collect data from the audio stream, into the buffer.
Args:
self: The class instance.
in_data: The audio data as a bytes object.
args: Additional arguments.
kwargs: Additional arguments.
returns: None
"""
self._buff.put(in_data)
return None, pyaudio.paContinue
def generator(self: object) -> object:
"""Stream Audio from microphone to API and to local buffer
Args:
self: The class instance.
returns:
The data from the audio stream.
"""
while not self.closed:
data = []
if self.new_stream and self.last_audio_input:
chunk_time = STREAMING_LIMIT / len(self.last_audio_input)
if chunk_time != 0:
if self.bridging_offset < 0:
self.bridging_offset = 0
if self.bridging_offset > self.final_request_end_time:
self.bridging_offset = self.final_request_end_time
chunks_from_ms = round(
(self.final_request_end_time - self.bridging_offset)
/ chunk_time
)
self.bridging_offset = round(
(len(self.last_audio_input) - chunks_from_ms) * chunk_time
)
for i in range(chunks_from_ms, len(self.last_audio_input)):
data.append(self.last_audio_input[i])
self.new_stream = False
# Use a blocking get() to ensure there's at least one chunk of
# data, and stop iteration if the chunk is None, indicating the
# end of the audio stream.
chunk = self._buff.get()
self.audio_input.append(chunk)
if chunk is None:
return
data.append(chunk)
# Now consume whatever other data's still buffered.
while True:
try:
chunk = self._buff.get(block=False)
if chunk is None:
return
data.append(chunk)
self.audio_input.append(chunk)
except queue.Empty:
break
yield b"".join(data)
def listen_print_loop(responses: object, stream: object) -> object:
"""Iterates through server responses and prints them.
The responses passed is a generator that will block until a response
is provided by the server.
Each response may contain multiple results, and each result may contain
multiple alternatives; for details, see https://goo.gl/tjCPAU. Here we
print only the transcription for the top alternative of the top result.
In this case, responses are provided for interim results as well. If the
response is an interim one, print a line feed at the end of it, to allow
the next result to overwrite it, until the response is a final one. For the
final one, print a newline to preserve the finalized transcription.
Arg:
responses: The responses returned from the API.
stream: The audio stream to be processed.
Returns:
The transcript of the result
"""
for response in responses:
if get_current_time() - stream.start_time > STREAMING_LIMIT:
stream.start_time = get_current_time()
break
if not response.results:
continue
result = response.results[0]
if not result.alternatives:
continue
transcript = result.alternatives[0].transcript
result_seconds = 0
result_micros = 0
if result.result_end_time.seconds:
result_seconds = result.result_end_time.seconds
if result.result_end_time.microseconds:
result_micros = result.result_end_time.microseconds
stream.result_end_time = int((result_seconds * 1000) + (result_micros / 1000))
corrected_time = (
stream.result_end_time
- stream.bridging_offset
+ (STREAMING_LIMIT * stream.restart_counter)
)
# Display interim results, but with a carriage return at the end of the
# line, so subsequent lines will overwrite them.
if result.is_final:
sys.stdout.write(GREEN)
sys.stdout.write("\033[K")
sys.stdout.write(str(corrected_time) + ": " + transcript + "\n")
stream.is_final_end_time = stream.result_end_time
stream.last_transcript_was_final = True
# Exit recognition if any of the transcribed phrases could be
# one of our keywords.
if re.search(r"\b(exit|quit)\b", transcript, re.I):
sys.stdout.write(YELLOW)
sys.stdout.write("Exiting...\n")
stream.closed = True
break
else:
sys.stdout.write(RED)
sys.stdout.write("\033[K")
sys.stdout.write(str(corrected_time) + ": " + transcript + "\r")
stream.last_transcript_was_final = False
return transcript
def main() -> None:
"""start bidirectional streaming from microphone input to speech API"""
client = speech.SpeechClient()
config = speech.RecognitionConfig(
encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16,
sample_rate_hertz=SAMPLE_RATE,
language_code="en-US",
max_alternatives=1,
)
streaming_config = speech.StreamingRecognitionConfig(
config=config, interim_results=True
)
mic_manager = ResumableMicrophoneStream(SAMPLE_RATE, CHUNK_SIZE)
print(mic_manager.chunk_size)
sys.stdout.write(YELLOW)
sys.stdout.write('\nListening, say "Quit" or "Exit" to stop.\n\n')
sys.stdout.write("End (ms) Transcript Results/Status\n")
sys.stdout.write("=====================================================\n")
with mic_manager as stream:
while not stream.closed:
sys.stdout.write(YELLOW)
sys.stdout.write(
"\n" + str(STREAMING_LIMIT * stream.restart_counter) + ": NEW REQUEST\n"
)
stream.audio_input = []
audio_generator = stream.generator()
requests = (
speech.StreamingRecognizeRequest(audio_content=content)
for content in audio_generator
)
responses = client.streaming_recognize(streaming_config, requests)
# Now, put the transcription responses to use.
listen_print_loop(responses, stream)
if stream.result_end_time > 0:
stream.final_request_end_time = stream.is_final_end_time
stream.result_end_time = 0
stream.last_audio_input = []
stream.last_audio_input = stream.audio_input
stream.audio_input = []
stream.restart_counter = stream.restart_counter + 1
if not stream.last_transcript_was_final:
sys.stdout.write("\n")
stream.new_stream = True
if __name__ == "__main__":
main()
If you're new to Google Cloud, create an account to evaluate how
Speech-to-Text performs in real-world
scenarios. New customers also get $300 in free credits to run, test, and
deploy workloads.
