From 628ad2aaa2e207aa2836e9b68f7bfbd890dff766 Mon Sep 17 00:00:00 2001 From: makefu Date: Tue, 5 Nov 2013 18:14:21 +0100 Subject: speech to text: initial commit --- util/lib/stt/google.sh | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+) create mode 100644 util/lib/stt/google.sh (limited to 'util/lib/stt/google.sh') diff --git a/util/lib/stt/google.sh b/util/lib/stt/google.sh new file mode 100644 index 00000000..a78579d5 --- /dev/null +++ b/util/lib/stt/google.sh @@ -0,0 +1,35 @@ +#!/bin/sh + +_get_content_type(){ + file -b --mime-type "$1" +} +_get_audio_rate(){ + file "$1" | sed -n -e 's/.* \([.0-9]\+\) kHz.*/\1/p' \ + | awk '{print int($1 *1000)}' +} + +record_audio(){ + # usage : _record_audio num_seconds + # echoes the output file + tmpfile=$(mktemp) + : ${1?please provide number of seconds to record} + arecord -d "$1" -r 16000 -t wav -q -f cd | flac -s -f - -o "$tmpfile" && echo "$tmpfile" +} +stt(){ + # usage: (lang=de-de stty recorded_file) + : ${1? please provide recorded file} + infile="$1" + lang=${lang:-en-us} + _get_content_type "$1" | (! grep -q "x-flac" ) \ + && echo "infile needs to be in flac format" \ + && return 1 + # only flac seems to be working... + wget -q -O - \ + -U 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_6_8) AppleWebKit/535.7 (KHTML, like Gecko) Chrome/16.0.912.77 Safari/535.7' \ + --post-file "$infile" \ + --header "Content-Type: `_get_content_type $infile`; rate=`_get_audio_rate $infile`;" \ + "http://www.google.com/speech-api/v1/recognize?lang=${lang}&client=chromium&maxresults=1" \ + | sed -n 's/.*utterance":"\([^"]*\)".*/\1/p' + + # returns {"status":0,"id":"d9269e6f741997161e41a4d441b34ba1-1","hypotheses":[{"utterance":"hallo Welt","confidence":0.7008959}]} +} -- cgit v1.2.3