speech recognition

- Added Whisper CLI
- Split Nerd Dictation into single file

wiki/linux/nerd-dictation.md

@@ -0,0 +1,48 @@
+# Nerd Dictation
+
+[Nerd Dictation](https://github.com/ideasman42/nerd-dictation/) is a real-time offline speech
+recognition software for [Linux](/wiki/linux.md)-based operating systems which uses the
+[VOSK API](/wiki/speech_recognition_and_synthesis.md#vosk-api).
+
+## Setup
+
+The Nerd Dictation program can be installed from source as described
+[on GitHub](https://github.com/ideasman42/nerd-dictation/).
+Some [Linux](/wiki/linux.md) [package managers](/wiki/linux/package_manager.md) package Nerd
+Dictation in the `nerd-dictation` package.
+
+### Configuration
+
+Nerd Dictation needs a model to recognize and transcribe speech.
+The default path for this is `~/.config/nerd-dictation/model`.
+In this directory a VOSK model can be placed to use as default model.
+
+## Usage
+
+This section addresses the usage of Nerd Dictation.
+
+### Basic Usage
+
+Nerd Dictation can be started using the following command.
+
+```sh 
+nerd-dictation begin
+```
+
+The model can also be specified using the `--vosk-model-dir` flag.
+Assuming the path to the model is `<model-path>` a command will look like the following.
+If no model path is specified it uses the default model as described in
+[the setup section](#configuration).
+
+```sh 
+nerd-dictation begin --vosk-model-dir=<model-path>
+```
+
+
+```sh 
+nerd-dictation end
+```
+
+## Troubleshooting 
+
+This section will focus on errors and the fixing of errors of /name/.

wiki/linux/whisper-cli.md

@@ -0,0 +1,44 @@
+# Whisper CLI
+
+[Whisper CLI](https://github.com/vatsalaggarwal/whisper-cli) is a non-real-time offline speech
+transcription software for [Linux](/wiki/linux.md)-based operating systems which uses the
+[Whisper API](/wiki/speech_recognition_and_synthesis.md#whisper-api).
+
+## Setup
+
+The Whisper CLI program can be installed from source as described
+[on GitHub](https://github.com/vatsalaggarwal/whisper-cli).
+Some [Linux](/wiki/linux.md) [package managers](/wiki/linux/package_manager.md) package Whisper CLI
+in the `whisper.cpp` package.
+
+### Configuration
+
+By default, Whisper CLI does not feature a global default model path.
+If not model path is given as shown in the [basic usage section](#basic-usage) the path
+`./models/ggml-base.en.bin` will be used.
+The behavior can not be changed using configuration of the program, but a simple alias can set a
+default path as shown below.
+
+```sh
+alias whisper="whisper-cli -m ~/.config/whisper-cli/ggml-large-v3-turbo-german-q5_0.bin"
+```
+
+## Usage
+
+This section addresses the usage of Whisper CLI.
+
+### Basic Usage
+
+Whisper CLI can be used to transcribe an [audio](/wiki/audio.md) file as shown in the following
+command.
+In this example `<model-path>` is the path to the Whisper model.
+
+```sh
+whisper-cli -m <model-path>
+```
+
+Additionally, a `.txt` file can be generated as shown below.
+
+```sh
+whisper-cli -m <model-path> -otxt
+```