Speech: Use all results, instead of just the first.

Author: Jerjou Cheng

speech/recognize.js

@@ -60,8 +60,10 @@ function syncRecognize (filename, encoding, sampleRateHertz, languageCode) {
 
  // Detects speech in the audio file
  speech.recognize(request)
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
+ .then((data) => {
+ const response = data[0];
+ const transcription = response.results.map(result =>
+ result.alternatives[0].transcript).join('\n');
  console.log(`Transcription: `, transcription);
  })
  .catch((err) => {
@@ -106,8 +108,10 @@ function syncRecognizeGCS (gcsUri, encoding, sampleRateHertz, languageCode) {
 
  // Detects speech in the audio file
  speech.recognize(request)
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
+ .then((data) => {
+ const response = data[0];
+ const transcription = response.results.map(result =>
+ result.alternatives[0].transcript).join('\n');
  console.log(`Transcription: `, transcription);
  })
  .catch((err) => {
@@ -154,18 +158,20 @@ function syncRecognizeWords (filename, encoding, sampleRateHertz, languageCode)
 
  // Detects speech in the audio file
  speech.recognize(request)
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
- console.log(`Transcription: `, transcription);
- results[0].results[0].alternatives[0].words.forEach((wordInfo) => {
- // NOTE: If you have a time offset exceeding 2^32 seconds, use the
- // wordInfo.{x}Time.seconds.high to calculate seconds.
- const startSecs = `${wordInfo.startTime.seconds}` + `.` +
- (wordInfo.startTime.nanos / 100000000);
- const endSecs = `${wordInfo.endTime.seconds}` + `.` +
- (wordInfo.endTime.nanos / 100000000);
- console.log(`Word: ${wordInfo.word}`);
- console.log(`\t ${startSecs} secs - ${endSecs} secs`);
+ .then((data) => {
+ const response = data[0];
+ response.results.forEach((result) => {
+ console.log(`Transcription: `, result.alternatives[0].transcript);
+ result.alternatives[0].words.forEach((wordInfo) => {
+ // NOTE: If you have a time offset exceeding 2^32 seconds, use the
+ // wordInfo.{x}Time.seconds.high to calculate seconds.
+ const startSecs = `${wordInfo.startTime.seconds}` + `.` +
+ (wordInfo.startTime.nanos / 100000000);
+ const endSecs = `${wordInfo.endTime.seconds}` + `.` +
+ (wordInfo.endTime.nanos / 100000000);
+ console.log(`Word: ${wordInfo.word}`);
+ console.log(`\t ${startSecs} secs - ${endSecs} secs`);
+ });
  });
  })
  .catch((err) => {
@@ -212,13 +218,16 @@ function asyncRecognize (filename, encoding, sampleRateHertz, languageCode) {
  // Detects speech in the audio file. This creates a recognition job that you
  // can wait for now, or get its result later.
  speech.longRunningRecognize(request)
- .then((results) => {
- const operation = results[0];
+ .then((data) => {
+ const response = data[0];
+ const operation = response;
  // Get a Promise representation of the final result of the job
  return operation.promise();
  })
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
+ .then((data) => {
+ const response = data[0];
+ const transcription = response.results.map(result =>
+ result.alternatives[0].transcript).join('\n');
  console.log(`Transcription: ${transcription}`);
  })
  .catch((err) => {
@@ -265,13 +274,15 @@ function asyncRecognizeGCS (gcsUri, encoding, sampleRateHertz, languageCode) {
  // Detects speech in the audio file. This creates a recognition job that you
  // can wait for now, or get its result later.
  speech.longRunningRecognize(request)
- .then((results) => {
- const operation = results[0];
+ .then((data) => {
+ const operation = data[0];
  // Get a Promise representation of the final result of the job
  return operation.promise();
  })
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
+ .then((data) => {
+ const response = data[0];
+ const transcription = response.results.map(result =>
+ result.alternatives[0].transcript).join('\n');
  console.log(`Transcription: ${transcription}`);
  })
  .catch((err) => {
@@ -319,23 +330,25 @@ function asyncRecognizeGCSWords (gcsUri, encoding, sampleRateHertz, languageCode
  // Detects speech in the audio file. This creates a recognition job that you
  // can wait for now, or get its result later.
  speech.longRunningRecognize(request)
- .then((results) => {
- const operation = results[0];
+ .then((data) => {
+ const operation = data[0];
  // Get a Promise representation of the final result of the job
  return operation.promise();
  })
- .then((results) => {
- const transcription = results[0].results[0].alternatives[0].transcript;
- console.log(`Transcription: ${transcription}`);
- results[0].results[0].alternatives[0].words.forEach((wordInfo) => {
- // NOTE: If you have a time offset exceeding 2^32 seconds, use the
- // wordInfo.{x}Time.seconds.high to calculate seconds.
- const startSecs = `${wordInfo.startTime.seconds}` + `.` +
- (wordInfo.startTime.nanos / 100000000);
- const endSecs = `${wordInfo.endTime.seconds}` + `.` +
- (wordInfo.endTime.nanos / 100000000);
- console.log(`Word: ${wordInfo.word}`);
- console.log(`\t ${startSecs} secs - ${endSecs} secs`);
+ .then((data) => {
+ const response = data[0];
+ response.results.forEach((result) => {
+ console.log(`Transcription: ${result.alternatives[0].transcript}`);
+ result.alternatives[0].words.forEach((wordInfo) => {
+ // NOTE: If you have a time offset exceeding 2^32 seconds, use the
+ // wordInfo.{x}Time.seconds.high to calculate seconds.
+ const startSecs = `${wordInfo.startTime.seconds}` + `.` +
+ (wordInfo.startTime.nanos / 100000000);
+ const endSecs = `${wordInfo.endTime.seconds}` + `.` +
+ (wordInfo.endTime.nanos / 100000000);
+ console.log(`Word: ${wordInfo.word}`);
+ console.log(`\t ${startSecs} secs - ${endSecs} secs`);
+ });
  });
  })
  .catch((err) => {