问题
My code is inspired by this php version of full duplex google speech API for speech-to-text : http://mikepultz.com/2013/07/google-speech-api-full-duplex-php-version/
I have few flac files that do work and give the array output as explained on Mike's post. But for few flac files it just doesn't return anything as output. For example : http://gavyadhar.com/video/upload/Pantry_Survey.flac, no output is returned.
But the same code works for this flac file: http://gavyadhar.com/video/upload/pantry_snack_video.flac and returns the array output as following:
---------response 1 : your face a pantry and one of the main stacks I usually Wesley Rd rasa multigrain crackers and I like them because they're pretty healthy for you and [...] ...and so on 5 responses.
Can anyone tell me, why is it that some flac files don't work? Or is there any way to catch the exceptions which are created by google API and display them to troubleshoot?
Thanks. Attaching my code here of two php files:
PHP file which passes the API key and the FLAC file with its sample rate.
<?php
include 'GoogleSpeechToText.php';
ini_set('display_errors',1);
ini_set('display_startup_errors',1);
error_reporting(-1);
// Your API Key goes here.
$apiKey = 'XXXXXXXXXXXXXXXXXXXXXXXXXXXX'; // my API server key
$speech = new GoogleSpeechToText($apiKey);
#var_dump($speech);
$file = realpath('upload/Pantry_Survey.flac'); // Full path to the file.
echo "<BR><BR>file : ", $file ;
#print_r("<br>-------------------");
#print_r($file1);
#print_r("<br>-------------------");
$bitRate = 44100; // The bit rate of the file.
$result = $speech->process($file, $bitRate, 'en-US');
print_r($result);
echo "<BR><BR>-------------";
$responses = $result[0]['alternative'];
$i = 0;
foreach ($responses as $response) {
$i = $i + 1;
#echo $response['transcript'];
echo "<br><br> ---------response $i : ", $response['transcript'];
}
?>GoogleSpeechToText.php
<?php
/**
* Convert FLAC files to Text using the Google Speech API
*
* Credit due to Mike (mike@mikepultz.com) for his first version of this.
*
* @version 0.1
* @author Roger Thomas
* @see
*
*/
class GoogleSpeechToText
{
/**
* URL of the Speech API
* @var string
*/
const SPEECH_BASE_URL = 'https://www.google.com/speech-api/full-duplex/v1/';
/**
* A 'unique' string to use for the requests
*
* @var string
*/
private $requestPair;
/**
* The Google Auth API Key
*
* @var string
*/
private $apiKey;
/**
* CURL Upload Handle
*
* @var resource
*/
private $uploadHandle;
/**
* CURL Download Handle
*
* @var resource
*/
private $downloadHandle;
/**
* Construct giving the Google Auth API Key.
*
* @param string $apiKey
* @throws Exception
*/
public function __construct($apiKey)
{
if (empty($apiKey)) {
throw new Exception('$apiKey should not be empty.');
}
$this->apiKey = $apiKey;
$this->requestPair = $this->getPair();
$this->setupCurl();
}
/**
* Setup CURL requests, both up and down.
*/
private function setupCurl()
{
$this->uploadHandle = curl_init();
$this->downloadHandle = curl_init();
curl_setopt($this->downloadHandle, CURLOPT_URL, self::SPEECH_BASE_URL . 'down?pair=' . $this->requestPair );
#echo "<br>downloadHandle :: ", self::SPEECH_BASE_URL . 'down?pair=' . $this->requestPair;
curl_setopt($this->downloadHandle, CURLOPT_RETURNTRANSFER, true);
curl_setopt($this->uploadHandle,CURLOPT_RETURNTRANSFER,true);
curl_setopt($this->uploadHandle,CURLOPT_POST,true);
//----added by asmi shah - 7 jan 2015
curl_setopt($this->uploadHandle, CURLOPT_MAX_SEND_SPEED_LARGE, 30000);
curl_setopt($this->uploadHandle, CURLOPT_LOW_SPEED_TIME, 9999);
curl_setopt($this->downloadHandle, CURLOPT_LOW_SPEED_TIME, 9999);
//----
}
/**
* Generate a Pair for the request. This identifies the requests later.
*
* @return string
*/
private function getPair()
{
$c = '0123456789';
$s = '';
for ($i=0; $i<16; $i++) {
$s .= $c[rand(0, strlen($c) - 1)];
}
echo "pair : ",$s;
return $s;
}
/**
* Make the request, returning either an array, or boolean false on
* failure.
*
* @param string $file the file name to process
* @param integer $rate the bitrate of the flac content (example: 44100)
* @param string $language the ISO language code
* (en-US has been confirmed as working)
* @throws Exception
* @return array|boolean false for failure.
*/
public function process($file, $rate, $language = 'en-US')
{
if (!$file || !file_exists($file) || !is_readable($file)) {
throw new Exception(
'$file must be specified and be a valid location.'
);
}
else { echo "<br>file exists"; }
$data = file_get_contents($file);
#var_dump($rate);
if (!$data) {
throw new Exception('Unable to read ' . $file);
}
else { echo "<br>file is readable"; }
$upload_data = array(
"Content_Type" => "audio/x-flac; rate=". $rate,
"Content" => $data,
);
if (empty($rate) || !is_integer($rate)) {
throw new Exception('$rate must be specified and be an integer');
}
else { echo "<br>sample rate is received :" . $rate ; }
curl_setopt($this->uploadHandle,CURLOPT_URL,self::SPEECH_BASE_URL . 'up?lang=' .$language . '&lm=dictation&timeout=20&client=chromium&pair=' .$this->requestPair . '&key=' . $this->apiKey);
curl_setopt($this->uploadHandle,CURLOPT_HTTPHEADER,array('Transfer-Encoding: chunked','Content-Type: audio/x-flac; rate=' . $rate));
curl_setopt($this->uploadHandle,CURLOPT_POSTFIELDS,$upload_data);
#echo "<br><br>------ data : ", $data;
#echo "<br><br>------";
#echo "<BR><BR> URL made up : ", self::SPEECH_BASE_URL . 'up?lang=' .$language . '&lm=dictation&client=chromium&pair=' .$this->requestPair . '&key=' . $this->apiKey;
$curlMulti = curl_multi_init();
curl_multi_add_handle($curlMulti, $this->downloadHandle);
curl_multi_add_handle($curlMulti, $this->uploadHandle);
$active = null;
do {
curl_multi_exec($curlMulti, $active);
} while ($active > 0);
$res = curl_multi_getcontent($this->downloadHandle);
#var_dump($this->downloadHandle);
#echo "resource type ::".get_resource_type($this->downloadHandle);
$output = array();
$results = explode("\n", $res);
#var_dump($results);
#$i = 0;
foreach ($results as $result) {
#$i = $i + 1;
#echo "<BR><BR><BR>--------------- string ||$i|| : ", $result;
$object = json_decode($result, true);
if (
(isset($object['result']) == true) &&
(count($object['result']) > 0)
) {
foreach ($object['result'] as $obj) {
$output[] = $obj;
}
}
}
curl_multi_remove_handle($curlMulti, $this->downloadHandle);
curl_multi_remove_handle($curlMulti, $this->uploadHandle);
curl_multi_close($curlMulti);
if (empty($output)) {
echo "<BR><br>output is empty<BR>";
return false;
}
echo "<BR><BR>";
return $output;
}
/**
* Close any outstanding connections in the destruct
*/
public function __destruct()
{
curl_close($this->uploadHandle);
curl_close($this->downloadHandle);
}
}
?>回答1:
The audio file you provided uses 2 channels. As speech currently only supports single channel audio, you will need to convert it down to a single channel. All encodings support only 1 channel (mono) audio. Audio encoding detail of Google speech API
After converting the audio file down to 1 channel I was able to successfully get a response with the provided audio file.
回答2:
For your information the SPEECH_BASE_URL = 'https://www.google.com/speech-api/full-duplex/v1/'; is now deprecated, the new one is: https://speech.googleapis.com/v1/speech:recognize?key=YOUR_API_KEY
as per https://cloud.google.com/speech/docs/auth#using_an_api_key
来源：https://stackoverflow.com/questions/27982625/google-speech-api-php-does-not-return-anything