Update voicemail_transcription.rst

fusionpbx · Nov 4, 2021 · db6d63b · db6d63b
1 parent e563e7f
commit db6d63b
Showing 1 changed file with 59 additions and 21 deletions.
diff --git a/source/applications/voicemail_transcription.rst b/source/applications/voicemail_transcription.rst
@@ -6,11 +6,12 @@ Voicemail Transcription
 
 Uses API services to transcribe voicemails into text to be used in the app-sms and the voicemail to email options. Bing's Speech API or other generic APIs can be used.
 
-Bing API
+
+IBM Watson API
 ====================
 
 
-Sign up and language information is located on `Microsoft Site <https://www.microsoft.com/cognitive-services/en-us/Speech-api/documentation/API-Reference-REST/BingVoiceRecognition>`_ Note: The Bing Speech API is deprecated as of October 2018, this works for now but needs to be ported to `the new API <https://github.com/MicrosoftDocs/azure-docs/blob/master/articles/cognitive-services/Speech-Service/how-to-migrate-from-bing-speech.md>`_
+Sign up and language information is located on `IBM Watson's Site <https://cloud.ibm.com/catalog/services/speech-to-text>`_ 
 
 .. warning:: We cannot use mod_shout to record Voicemails because the transcription service needs an uncompressed version of the audio. Therefore we will record in WAV and then use LAME to re-encode in MP3. This could cause added resource utilization to your system.
 
@@ -20,44 +21,45 @@ Add the following entries
   +-------------+-----------------------+-----------+---------------------------+-----------+
   |  Category   |  Subcategory          |  Type     |  Value                    |  Enabled  |
   +=============+=======================+===========+===========================+===========+
-  |  voicemail  |  transcribe_provider  |  text     |  microsoft                |  True     |
+  |  voicemail  |  transcribe_provider  |  text     |  watson                   |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
-  |  voicemail  |  microsoft_key1       |  text     |  {your microsoft key #1}  |  True     |
+  |  voicemail  |  watson_key           |  text     |  { your watson key }      |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
-  |  voicemail  |  microsoft_key2       |  text     |  {your microsoft key #2}  |  True     |
+  |  voicemail  |  watson_url           |  text     |  { *watson url }          |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
   |  voicemail  |  transcribe_language  |  text     |  en-US                    |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
   |  voicemail  |  transcribe_enabled   |  boolean  |  true                     |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  voicemail  |  json_enabled         |  boolean  |  true                     |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+
+*NOTE: Watson URL used for testing was the following:
+https://stream.watsonplatform.net/speech-to-text/api/v1/recognize?model=en-US_NarrowbandModel*
 
  Click "Reload" at the top of the page.
 
 **Goto Status > Sip Status.**
 
-Click "Flush Memcache", "Reload XML" and "Rescan".
+Click "Flush Cache", "Reload XML" and "Rescan".
 
 If you entered your key's correctly, you should now start getting transcriptions delivered in your voicemail to email and you will also see them on the Messages page.
 
-IBM Watson API
-====================
-
-
-Sign up and language information is located on `IBM Watson's Site <https://cloud.ibm.com/catalog/services/speech-to-text>`_ 
 
-.. warning:: We cannot use mod_shout to record Voicemails because the transcription service needs an uncompressed version of the audio. Therefore we will record in WAV and then use LAME to re-encode in MP3. This could cause added resource utilization to your system.
+Azure API
+====================
 
 **Goto Advanced > Default Settings.**
 Add the following entries
 
   +-------------+-----------------------+-----------+---------------------------+-----------+
   |  Category   |  Subcategory          |  Type     |  Value                    |  Enabled  |
   +=============+=======================+===========+===========================+===========+
-  |  voicemail  |  transcribe_provider  |  text     |  watson                   |  True     |
+  |  voicemail  |  transcribe_provider  |  text     |  azure                    |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
-  |  voicemail  |  watson_key           |  text     |  { your watson key }      |  True     |
+  |  voicemail  |  azure_key            |  text     |  { your zaure key }       |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
-  |  voicemail  |  watson_url           |  text     |  { *watson url }          |  True     |
+  |  voicemail  |  azure_server_region  |  text     |  { your server region }   |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
   |  voicemail  |  transcribe_language  |  text     |  en-US                    |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
@@ -66,17 +68,15 @@ Add the following entries
   |  voicemail  |  json_enabled         |  boolean  |  true                     |  True     |
   +-------------+-----------------------+-----------+---------------------------+-----------+
 
-*NOTE: Watson URL used for testing was the following:
-https://stream.watsonplatform.net/speech-to-text/api/v1/recognize?model=en-US_NarrowbandModel*
-
  Click "Reload" at the top of the page.
 
 **Goto Status > Sip Status.**
 
-Click "Flush Memcache", "Reload XML" and "Rescan".
+Click "Flush Cache", "Reload XML" and "Rescan".
 
 If you entered your key's correctly, you should now start getting transcriptions delivered in your voicemail to email and you will also see them on the Messages page.
 
+
 Google API
 ====================
 
@@ -103,13 +103,51 @@ Add the following entries
 
 **Goto Status > Sip Status.**
 
-Click "Flush Memcache", "Reload XML" and "Rescan".
+Click "Flush Cache", "Reload XML" and "Rescan".
 
 If you entered your key's correctly, you should now start getting transcriptions delivered in your voicemail to email and you will also see them on the Messages page.
 
+
+
+Bing API
+====================
+
+Recommend using Azure as an alternative to Bing.
+
+Sign up and language information is located on `Microsoft Site <https://www.microsoft.com/cognitive-services/en-us/Speech-api/documentation/API-Reference-REST/BingVoiceRecognition>`_ Note: The Bing Speech API is deprecated as of October 2018, this works for now but needs to be ported to `the new API <https://github.com/MicrosoftDocs/azure-docs/blob/master/articles/cognitive-services/Speech-Service/how-to-migrate-from-bing-speech.md>`_
+
+.. warning:: We cannot use mod_shout to record Voicemails because the transcription service needs an uncompressed version of the audio. Therefore we will record in WAV and then use LAME to re-encode in MP3. This could cause added resource utilization to your system.
+
+**Goto Advanced > Default Settings.**
+Add the following entries
+
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  Category   |  Subcategory          |  Type     |  Value                    |  Enabled  |
+  +=============+=======================+===========+===========================+===========+
+  |  voicemail  |  transcribe_provider  |  text     |  microsoft                |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  voicemail  |  microsoft_key1       |  text     |  {your microsoft key #1}  |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  voicemail  |  microsoft_key2       |  text     |  {your microsoft key #2}  |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  voicemail  |  transcribe_language  |  text     |  en-US                    |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+  |  voicemail  |  transcribe_enabled   |  boolean  |  true                     |  True     |
+  +-------------+-----------------------+-----------+---------------------------+-----------+
+
+ Click "Reload" at the top of the page.
+
+**Goto Status > Sip Status.**
+
+Click "Flush Cache", "Reload XML" and "Rescan".
+
+If you entered your key's correctly, you should now start getting transcriptions delivered in your voicemail to email and you will also see them on the Messages page.
+
+
 Custom API
 ====================
 
+Currently does not work with the FusionPBX email_queue.
 
 API info from the Speech to Text provider of your choice is needed, or you can self host a transcription engine like `Mozilla DeepSpeech <https://git.callpipe.com/fusionpbx/deepspeech_frontend>`_ or `Kaldi ASR <https://github.com/dialogflow/asr-server>`_
 
@@ -136,6 +174,6 @@ Add the following entries
 
 **Goto Status > Sip Status.**
 
-Click "Flush Memcache", "Reload XML" and "Rescan".
+Click "Flush Cache", "Reload XML" and "Rescan".
 
 If you entered your key's correctly, you should now start getting transcriptions delivered in your voicemail to email and you will also see them on the Messages page.