Merge pull request #119 from yulin-li/mrcp

add MRCP samples

Author: szhaomsft

MRCP/.gitignore

@@ -0,0 +1,27 @@
+MAINTAINERCLEANFILES   = Makefile.in
+
+SUBDIRS                =
+
+if DEMOSYNTH_PLUGIN
+SUBDIRS               += demo-synth
+endif
+
+if DEMORECOG_PLUGIN
+SUBDIRS               += demo-recog
+endif
+
+if DEMOVERIFIER_PLUGIN
+SUBDIRS               += demo-verifier
+endif
+
+if RECORDER_PLUGIN
+SUBDIRS               += mrcp-recorder
+endif
+
+if MSRECOG_PLUGIN
+SUBDIRS		      += ms-recog
+endif
+
+if MSSYNTH_PLUGIN
+SUBDIRS		      += ms-synth
+endif

MRCP/Makefile.am

@@ -0,0 +1,196 @@
+# Sample: Using Microsoft Cognitive Speech Service in UniMRCP
+
+This sample demonstrates how to use cognitive speech service (both SR and TTS) in [UniMRCP](http://www.unimrcp.org/) plugins.
+
+## Prerequisites
+
+* A subscription key for the Speech service. See [Try the speech service for free](https://docs.microsoft.com/azure/cognitive-services/speech-service/get-started).
+* An Ubuntu 18.04 or Windows PC with Visual Studio
+* On Ubuntu, install these packages to build and run this sample:
+
+  ```sh
+  sudo apt update
+  sudo apt install build-essential libssl1.0.0 libasound2 wget
+  sudo apt install pkg-config automake libtool libtool-bin
+  sudo apt install libpoco-dev rapidjson-dev
+  ```
+
+* On Windows, install and config [poco](https://pocoproject.org/) and `rapidjson`. The recommended method is to use [vcpkg](https://github.com/microsoft/vcpkg).
+
+  ```bash
+  vcpkg.exe install poco rapidjson
+  ```
+
+## Build the sample
+
+* Download and compile UniMRCP 1.6.0 on your PC, following the [installation instructions](http://www.unimrcp.org/index.php/project/get-started).
+
+* Clone or download the sample code to your development PC.
+
+* Copy the `ms-recog`, `ms-synth` and `ms-common` folders to `your_unimrcp_path/plugins`
+
+### On Linux
+
+* Download and extract the Speech SDK
+  * **By downloading the Microsoft Cognitive Services Speech SDK, you acknowledge its license, see [Speech SDK license agreement](https://aka.ms/csspeech/license201809).**
+  * Run the following commands after replacing the string `/your/path` with a directory (absolute path) of your choice:
+
+    ```sh
+    export SPEECHSDK_ROOT="/your/path"
+    mkdir -p "$SPEECHSDK_ROOT"
+    wget -O SpeechSDK-Linux.tar.gz https://aka.ms/csspeech/linuxbinary
+    tar --strip 1 -xzf SpeechSDK-Linux.tar.gz -C "$SPEECHSDK_ROOT"
+    ```
+
+* Navigate to the directory of UniMRCP
+* Open the file `configure.ac`, and add the following lines in correct position:
+
+    ```shell
+    dnl MS recognizer plugin.
+    UNI_PLUGIN_ENABLED(msrecog)
+
+    AM_CONDITIONAL([MSRECOG_PLUGIN],[test "${enable_msrecog_plugin}" = "yes"])
+
+    dnl MS synthesizer plugin.
+    UNI_PLUGIN_ENABLED(mssynth)
+
+    AM_CONDITIONAL([MSSYNTH_PLUGIN],[test "${enable_mssynth_plugin}" = "yes"])
+
+    ...
+
+    plugins/ms-recog/Makefile
+    plugins/ms-synth/Makefile
+
+    ...
+
+    echo MS recognizer plugin.......... : $enable_msrecog_plugin
+    echo MS synthesizer plugin......... : $enable_mssynth_plugin
+    ```
+
+* Edit the file `plugins/Makefile.am` and add this:
+
+    ```shell
+    if MSRECOG_PLUGIN
+    SUBDIRS		      += ms-recog
+    endif
+
+    if MSSYNTH_PLUGIN
+    SUBDIRS		      += ms-synth
+    endif
+    ```
+
+* Recompile and install the UniMRCP project.
+
+### On Windows
+
+* Add `ms-recog`, `ms-synth` and `ms-common` projects into the `unimrcp-2010.sln` solution.
+  
+* Restore the nuget packages.
+
+* Recompile the solution in `Release` and `x64` mode.
+
+## Run the sample
+
+* Edit the `unimrcpserver.xml` config file (in `/usr/local/unimrcp/conf` on Linux and `Porject_folder\x64\Release\conf`).
+
+  * In `<plugin-factory>` section, add configure for the plugins, and disable demo plugins:
+
+  ```xml
+  <engine id="Demo-Synth-1" name="demosynth" enable="false"/>
+  <engine id="MS-Synth-1" name="mssynth" enable="true"/>
+  <engine id="Demo-Recog-1" name="demorecog" enable="false"/>
+  <engine id="MS-Recog-1" name="msrecog" enable="true"/>
+  ```
+
+  * Add 16kHz/16bit Codec support for higher accuracy and quality.
+
+  ```xml
+  <codecs own-preference="false">PCMU/97/16000 PCMA/98/16000 L16/99/16000 PCMU PCMA L16/96/8000 telephone-event/101/8000</codecs>
+  ```
+
+  * Other settings like IP address and port.
+
+* Add a conf file `config.json` in the same folder of `unimrcpserver.xml`, the sample file is in `sample-conf` path of this project. Remember to replace `YourSubscriptionKey` and `YourServiceRegion` with your own key and region of you subscription.
+
+### On Linux
+
+To run the sample, you'll need to configure the loader's library path to point to the Speech SDK library.
+
+* On an `x64` machine, run:
+
+  ```sh
+  export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$SPEECHSDK_ROOT/lib/x64"
+  ```
+
+* On an `x86` machine, run:
+
+  ```sh
+  export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:$SPEECHSDK_ROOT/lib/x86"
+  ```
+
+Run the application:
+
+  ```sh
+  /usr/local/unimrcp/bin/unimrcpserver
+  ```
+
+### On Windows
+
+Run the executable file `Porject_folder\x64\Release\bin\unimrcpserver.exe`.
+
+## Build and Run with docker
+
+You can use docker to build and deploy the UniMRCP with Microsoft SR/TTS plugins easily with docker.
+
+To build the docker, run:
+
+  ```shell
+  ./docker/build.sh
+  ```
+
+Then, run following script to start the UniMRCP server:
+
+*Note: it's better to use the host network to run the container, for easy configuration and high network efficiency.*
+
+  ```shell
+  docker run -dt -v ~/conf:/usr/local/unimrcp/conf --network=host unimrcp_ms:latest --name unimrcp_ms
+  ```
+
+The `~/conf` folder contains configure files of UniMRCP and the Microsoft plugins. You can find the sample configure files in `sample-conf` of this project.
+Remember to replace `YourSubscriptionKey` and `YourServiceRegion` with your own key and region of you subscription.
+
+## Test
+
+For easily test, you can use `umc` in the UniMRCP project.
+In `umc` console, type `run recog` and `run synth` for testing SR and TTS, respectively.
+
+*Tips: remember to correct configurations for codec. The configs in `sample-conf` are tested, you can use them after configuring the IP.*
+
+If you want to test the performance of the MRCP as well as the plugins, a easy way is to adopt `umc` to call SR/TTS continuously and concurrently.
+Change the `UmcConsole::RunCmdLine()` method in `umcconsole.cpp` with
+
+  ```c++
+  bool UmcConsole::RunCmdLine()
+  {
+      for(int i = 0; i < 100; i++)
+      {
+          char cmdline[1024] = "run recog\0";
+          std::this_thread::sleep_for(std::chrono::milliseconds(1000));
+          ProcessCmdLine(cmdline);
+      }
+      std::this_thread::sleep_for(std::chrono::seconds(20));
+      return true;
+  }
+  ```
+  
+Also append `#include <thread>` and `#include <chrono>` in the header.
+You can change the repeat time and interval to run your own test.
+Then, you can use the python scripts in `test_scripts` to analyze the log file to get the latency.
+
+
+Alternatively, you can use FreeSWITCH to make a real call to test the plugins. See [this]("./test_with_freeswitch.md) for details.
+
+## References
+
+* [UniMRCP](http://www.unimrcp.org/)
+* [Speech SDK API reference for C++](https://aka.ms/csspeech/cppref)