408 Request timed out Microsoft Speech to Text

Question

My .wav file length is just 4 seconds. Even after multiple retries and running it on cloud i am constantly getting following error

  * upload completely sent off: 12 out of 12 bytes
  < HTTP/1.1 408 Request timed out (> 14000 ms)
  < Transfer-Encoding: chunked
  < Content-Type: text/plain
  < Server: Microsoft-IIS/8.5
  < X-MSEdge-Ref: 

Has anybody faced this issue? This is my request

  `curl -v "https://speech.platform.bing.com/recognize?
  scenarios=catsearch&appid=D4D52672-91D7-4C74-8AD8-42B1D98141A5&locale=en-  
  US&device.os=wp7&version=3.0&format=json&requestid=1d4b6030-9099-12e0-91e4-
  0800200c9a67&instanceid=1d4b6030-9099-12e0-91e5-0800200c9a68" -H 
  "Authorization: Bearer $1" -H "Content-Type: audio/wav; samplerate=8000" -- 
  data-binary $2`

Answer 1

I also ran into a few problems getting it to work. The following BASH script "bingrec.sh" may help make it more clear; enter your SUBSCRIPTION_KEY & adjust SAMPLERATE etc. as required. As others have pointed out, locale & scenarios need to be set to supported values, and instance_id and request_id need to be in GUID format. The audio file should be less than 10 seconds long, and have a sampling rate of 8000 or 16000. Also the curl "--data-binary" parameter requires an "@" in front of the audio filename.

#!/bin/bash
#  Usage:  ./bingrec.sh  /path/to/file 
#  Send audio file $1 through Bing speech recognition API.
#
SUBSCRIPTION_KEY=<your-key-here>
LOCALE=en-US
SCENARIOS=ulm
SAMPLERATE=8000
CODEC=audio/pcm

TARGET_FILE=$1
if [ ! -f "$TARGET_FILE" ]; then
  echo Error:  file $TARGET_FILE does not exist!
  exit 1
fi

INSTANCE_ID=`uuidgen`    # random GUID for instance
REQUEST_ID=`uuidgen`     # random GUID for request
APPID=D4D52672-91D7-4C74-8AD8-42B1D98141A5   # APPID for Bing Speechrec API, don't change
DEVICE_OS=linux          # arbitraty
FORMAT=json

AUTH_TOKEN=`curl -v -X POST "https://api.cognitive.microsoft.com/sts/v1.0/issueToken" -H "Content-type: application/x-www-form-urlencoded" -H "Content-Length: 0" -H "Ocp-Apim-Subscription-Key: ${SUBSCRIPTION_KEY}"`

curl -v -X POST "https://speech.platform.bing.com/recognize?scenarios=${SCENARIOS}&appid=${APPID}&locale=${LOCALE}&device.os=${DEVICE_OS}&version=3.0&format=${FORMAT}&instanceid=${INSTANCE_ID}&requestid=${REQUEST_ID}" -H "Authorization: Bearer ${AUTH_TOKEN}" -H "Content-type: audio/wav; codec='${CODEC}'; samplerate=${SAMPLERATE}" --data-binary @${TARGET_FILE}

Answer 2

I got this working. There were couple of issues. One was with the locale, which i changed to en-IN. and then scenarios=ulm. This seems to have done the trick. I was able to detect speech very clearly.