RealTimeSTT 0.1.16__tar.gz → 0.2.0__tar.gz

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
-Name: RealTimeSTT
-Version: 0.1.16
+Name: RealtimeSTT
+Version: 0.2.0
 Summary: A fast Voice Activity Detection and Transcription System
 Home-page: https://github.com/KoljaB/RealTimeSTT
 Author: Kolja Beigel
@@ -17,16 +17,21 @@ Requires-Dist: faster-whisper==1.0.2
 Requires-Dist: pvporcupine==1.9.5
 Requires-Dist: webrtcvad==2.0.10
 Requires-Dist: halo==0.0.31
-Requires-Dist: torch==2.3.0
-Requires-Dist: torchaudio==2.3.0
+Requires-Dist: torch==2.3.1
+Requires-Dist: torchaudio==2.3.1
 Requires-Dist: scipy==1.12.0
 Requires-Dist: websockets==v12.0
+Requires-Dist: openwakeword==0.6.0
 
 
 # RealtimeSTT
 
 *Easy-to-use, low-latency speech-to-text library for realtime applications*
 
+## New
+
+Custom wake words with [OpenWakeWord](#openwakeword). Thanks to the [developers](https://github.com/dscripka/openWakeWord) of this!
+
 ## About the Project
 
 RealtimeSTT listens to the microphone and transcribes voice into text.  
@@ -42,7 +47,7 @@ https://github.com/KoljaB/RealtimeSTT/assets/7604638/207cb9a2-4482-48e7-9d2b-072
 
 ### Updates
 
-Latest Version: v0.1.15
+Latest Version: v0.2.0
 
 See [release history](https://github.com/KoljaB/RealtimeSTT/releases).
 
@@ -66,10 +71,10 @@ This library uses:
 - **Speech-To-Text**
   - [Faster_Whisper](https://github.com/guillaumekln/faster-whisper) for instant (GPU-accelerated) transcription.
 - **Wake Word Detection**
-  - [Porcupine](https://github.com/Picovoice/porcupine) for wake word detection.
+  - [Porcupine](https://github.com/Picovoice/porcupine) or [OpenWakeWord](https://github.com/dscripka/openWakeWord) for wake word detection.
 
-*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
+*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
 ## Installation
 
@@ -83,22 +88,64 @@ Although it is possible to run RealtimeSTT with a CPU installation only (use a s
 
 ### GPU Support with CUDA (recommended)
 
-Additional steps are needed for a **GPU-optimized** installation. These steps are recommended for those who require **better performance** and have a compatible NVIDIA GPU.
+
+**Install RealtimeSTT with CUDA support**  
+
+*For Cuda 11.8:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu118
+```  
+
+
+*For Cuda 12.X:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu121
+```
+
+This command reinstalls PyTorch and Torchaudio with explicit support for CUDA.
+
+### Updating PyTorch Only for CUDA Support
+
+If you already have RealtimeSTT installed but want to upgrade your PyTorch installation to leverage GPU support with CUDA, follow these instructions based on your specific CUDA version. This process is useful if you initially installed RealtimeSTT with only CPU support and now wish to enhance its performance with CUDA capabilities.
+
+#### For CUDA 11.8:
+To update PyTorch and Torchaudio to support CUDA 11.8, use the following commands:
+
+```bash
+pip install torch==2.3.1+cu118 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu118
+```
+
+#### For CUDA 12.X:
+To update PyTorch and Torchaudio to support CUDA 12.X, execute the following:
+
+```bash
+pip install torch==2.3.1+cu121 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121
+```
+
+Replace `2.3.1` with the version of CUDA that matches your system and requirements.
+
+### Steps That Might Be Necessary Before
 
 > **Note**: *To check if your NVIDIA GPU supports CUDA, visit the [official CUDA GPUs list](https://developer.nvidia.com/cuda-gpus).*
 
-To use RealtimeSTT with GPU support via CUDA please follow these steps:
+If you didn't use CUDA models before, some additional steps might be needed one time before installation. These steps prepare the system for CUDA support and installation of the **GPU-optimized** installation. This is recommended for those who require **better performance** and have a compatible NVIDIA GPU. To use RealtimeSTT with GPU support via CUDA please also follow these steps:
 
-1. **Install NVIDIA CUDA Toolkit 11.8**:
-    - Visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-11-8-0-download-archive).
+1. **Install NVIDIA CUDA Toolkit**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-toolkit-archive) and select latest version.
+        - for 11.8 visit [NVIDIA CUDA Toolkit 11.8](https://developer.nvidia.com/cuda-11-8-0-download-archive).
     - Select operating system and version.
     - Download and install the software.
 
-2. **Install NVIDIA cuDNN 8.7.0 for CUDA 11.x**:
-    - Visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
-    - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
-    - Download and install the software.
-
+2. **Install NVIDIA cuDNN**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [cuDNN Downloads](https://developer.nvidia.com/cudnn-downloads).
+            - Select operating system and version.
+            - Download and install the software.
+        - for 11.8 visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
+            - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
+            - Download and install the software.
+    
 3. **Install ffmpeg**:
 
     > **Note**: *Installation of ffmpeg might not actually be needed to operate RealtimeSTT* <sup> *thanks to jgilbert2017 for pointing this out</sup>
@@ -137,12 +184,6 @@ To use RealtimeSTT with GPU support via CUDA please follow these steps:
         scoop install ffmpeg
         ```    
 
-4. **Install PyTorch with CUDA support**:
-    ```bash
-    pip uninstall torch
-    pip install torch==2.2.2+cu118 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cu118
-    ```
-
 ## Quick Start
 
 Basic usage:
@@ -279,6 +320,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **gpu_device_index** (int, default=0): GPU Device Index to use. The model can also be loaded on multiple GPUs by passing a list of IDs (e.g. [0, 1, 2, 3]).
 
+- **device** (str, default="cuda"): Device for model to use. Can either be "cuda" or "cpu". 
+
 - **on_recording_start**: A callable function triggered when recording starts.
 
 - **on_recording_stop**: A callable function triggered when recording ends.
@@ -346,7 +389,13 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 #### Wake Word Parameters
 
-- **wake_words** (str, default=""): Wake words for initiating the recording. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator
+- **wakeword_backend** (str, default="pvporcupine"): Specifies the backend library to use for wake word detection. Supported options include 'pvporcupine' for using the Porcupine wake word engine or 'oww' for using the OpenWakeWord engine.
+
+- **openwakeword_model_paths** (str, default=None): Comma-separated paths to model files for the openwakeword library. These paths point to custom models that can be used for wake word detection when the openwakeword library is selected as the wakeword_backend.
+
+- **openwakeword_inference_framework** (str, default="onnx"): Specifies the inference framework to use with the openwakeword library. Can be either 'onnx' for Open Neural Network Exchange format or 'tflite' for TensorFlow Lite.
+
+- **wake_words** (str, default=""): Initiate recording when using the 'pvporcupine' wakeword backend. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator. For the 'openwakeword' backend, wake words are automatically extracted from the provided model files, so specifying them here is not necessary.
 
 - **wake_words_sensitivity** (float, default=0.6): Sensitivity level for wake word detection (0 for least sensitive, 1 for most sensitive).
 
@@ -354,6 +403,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **wake_word_timeout** (float, default=5): Duration in seconds after a wake word is recognized. If no subsequent voice activity is detected within this window, the system transitions back to an inactive state, awaiting the next wake word or voice activation.
 
+- **wake_word_buffer_duration** (float, default=0.1): Duration in seconds to buffer audio data during wake word detection. This helps in cutting out the wake word from the recording buffer so it does not falsely get detected along with the following spoken text, ensuring cleaner and more accurate transcription start triggers. Increase this if parts of the wake word get detected as text.
+
 - **on_wakeword_detected**: A callable function triggered when a wake word is detected.
 
 - **on_wakeword_timeout**: A callable function triggered when the system goes back to an inactive state after when no speech was detected after wake word activation.
@@ -362,6 +413,33 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **on_wakeword_detection_end**: A callable function triggered when stopping to listen for wake words (e.g. because of timeout or wake word detected)
 
+## OpenWakeWord  
+
+### Training models
+
+Look [here](https://github.com/dscripka/openWakeWord?tab=readme-ov-file#training-new-models) for information about how to train your own OpenWakeWord models. You can use a [simple Google Colab notebook](https://colab.research.google.com/drive/1q1oe2zOyZp7UsB3jJiQ1IFn8z5YfjwEb?usp=sharing) for a start or use a [more detailed notebook](https://github.com/dscripka/openWakeWord/blob/main/notebooks/automatic_model_training.ipynb) that enables more customization (can produce high quality models, but requires more development experience).
+
+### Convert model to ONNX format
+
+You might need to use tf2onnx to convert tensorflow tflite models to onnx format:
+
+```bash
+pip install -U tf2onnx
+python -m tf2onnx.convert --tflite my_model_filename.tflite --output my_model_filename.onnx
+```
+
+### Configure RealtimeSTT
+
+Suggested starting parameters for OpenWakeWord usage:
+```python
+    with AudioToTextRecorder(
+        wakeword_backend="oww",
+        wake_words_sensitivity=0.35,
+        openwakeword_model_paths="word1.onnx,word2.onnx",
+        wake_word_buffer_duration=1,
+        ) as recorder:
+```
+
 ## Contribution
 
 Contributions are always welcome! 

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/README.md

@@ -3,6 +3,10 @@
 
 *Easy-to-use, low-latency speech-to-text library for realtime applications*
 
+## New
+
+Custom wake words with [OpenWakeWord](#openwakeword). Thanks to the [developers](https://github.com/dscripka/openWakeWord) of this!
+
 ## About the Project
 
 RealtimeSTT listens to the microphone and transcribes voice into text.  
@@ -18,7 +22,7 @@ https://github.com/KoljaB/RealtimeSTT/assets/7604638/207cb9a2-4482-48e7-9d2b-072
 
 ### Updates
 
-Latest Version: v0.1.15
+Latest Version: v0.2.0
 
 See [release history](https://github.com/KoljaB/RealtimeSTT/releases).
 
@@ -42,10 +46,10 @@ This library uses:
 - **Speech-To-Text**
   - [Faster_Whisper](https://github.com/guillaumekln/faster-whisper) for instant (GPU-accelerated) transcription.
 - **Wake Word Detection**
-  - [Porcupine](https://github.com/Picovoice/porcupine) for wake word detection.
+  - [Porcupine](https://github.com/Picovoice/porcupine) or [OpenWakeWord](https://github.com/dscripka/openWakeWord) for wake word detection.
 
-*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
+*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
 ## Installation
 
@@ -59,22 +63,64 @@ Although it is possible to run RealtimeSTT with a CPU installation only (use a s
 
 ### GPU Support with CUDA (recommended)
 
-Additional steps are needed for a **GPU-optimized** installation. These steps are recommended for those who require **better performance** and have a compatible NVIDIA GPU.
+
+**Install RealtimeSTT with CUDA support**  
+
+*For Cuda 11.8:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu118
+```  
+
+
+*For Cuda 12.X:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu121
+```
+
+This command reinstalls PyTorch and Torchaudio with explicit support for CUDA.
+
+### Updating PyTorch Only for CUDA Support
+
+If you already have RealtimeSTT installed but want to upgrade your PyTorch installation to leverage GPU support with CUDA, follow these instructions based on your specific CUDA version. This process is useful if you initially installed RealtimeSTT with only CPU support and now wish to enhance its performance with CUDA capabilities.
+
+#### For CUDA 11.8:
+To update PyTorch and Torchaudio to support CUDA 11.8, use the following commands:
+
+```bash
+pip install torch==2.3.1+cu118 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu118
+```
+
+#### For CUDA 12.X:
+To update PyTorch and Torchaudio to support CUDA 12.X, execute the following:
+
+```bash
+pip install torch==2.3.1+cu121 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121
+```
+
+Replace `2.3.1` with the version of CUDA that matches your system and requirements.
+
+### Steps That Might Be Necessary Before
 
 > **Note**: *To check if your NVIDIA GPU supports CUDA, visit the [official CUDA GPUs list](https://developer.nvidia.com/cuda-gpus).*
 
-To use RealtimeSTT with GPU support via CUDA please follow these steps:
+If you didn't use CUDA models before, some additional steps might be needed one time before installation. These steps prepare the system for CUDA support and installation of the **GPU-optimized** installation. This is recommended for those who require **better performance** and have a compatible NVIDIA GPU. To use RealtimeSTT with GPU support via CUDA please also follow these steps:
 
-1. **Install NVIDIA CUDA Toolkit 11.8**:
-    - Visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-11-8-0-download-archive).
+1. **Install NVIDIA CUDA Toolkit**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-toolkit-archive) and select latest version.
+        - for 11.8 visit [NVIDIA CUDA Toolkit 11.8](https://developer.nvidia.com/cuda-11-8-0-download-archive).
     - Select operating system and version.
     - Download and install the software.
 
-2. **Install NVIDIA cuDNN 8.7.0 for CUDA 11.x**:
-    - Visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
-    - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
-    - Download and install the software.
-
+2. **Install NVIDIA cuDNN**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [cuDNN Downloads](https://developer.nvidia.com/cudnn-downloads).
+            - Select operating system and version.
+            - Download and install the software.
+        - for 11.8 visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
+            - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
+            - Download and install the software.
+    
 3. **Install ffmpeg**:
 
     > **Note**: *Installation of ffmpeg might not actually be needed to operate RealtimeSTT* <sup> *thanks to jgilbert2017 for pointing this out</sup>
@@ -113,12 +159,6 @@ To use RealtimeSTT with GPU support via CUDA please follow these steps:
         scoop install ffmpeg
         ```    
 
-4. **Install PyTorch with CUDA support**:
-    ```bash
-    pip uninstall torch
-    pip install torch==2.2.2+cu118 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cu118
-    ```
-
 ## Quick Start
 
 Basic usage:
@@ -255,6 +295,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **gpu_device_index** (int, default=0): GPU Device Index to use. The model can also be loaded on multiple GPUs by passing a list of IDs (e.g. [0, 1, 2, 3]).
 
+- **device** (str, default="cuda"): Device for model to use. Can either be "cuda" or "cpu". 
+
 - **on_recording_start**: A callable function triggered when recording starts.
 
 - **on_recording_stop**: A callable function triggered when recording ends.
@@ -322,7 +364,13 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 #### Wake Word Parameters
 
-- **wake_words** (str, default=""): Wake words for initiating the recording. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator
+- **wakeword_backend** (str, default="pvporcupine"): Specifies the backend library to use for wake word detection. Supported options include 'pvporcupine' for using the Porcupine wake word engine or 'oww' for using the OpenWakeWord engine.
+
+- **openwakeword_model_paths** (str, default=None): Comma-separated paths to model files for the openwakeword library. These paths point to custom models that can be used for wake word detection when the openwakeword library is selected as the wakeword_backend.
+
+- **openwakeword_inference_framework** (str, default="onnx"): Specifies the inference framework to use with the openwakeword library. Can be either 'onnx' for Open Neural Network Exchange format or 'tflite' for TensorFlow Lite.
+
+- **wake_words** (str, default=""): Initiate recording when using the 'pvporcupine' wakeword backend. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator. For the 'openwakeword' backend, wake words are automatically extracted from the provided model files, so specifying them here is not necessary.
 
 - **wake_words_sensitivity** (float, default=0.6): Sensitivity level for wake word detection (0 for least sensitive, 1 for most sensitive).
 
@@ -330,6 +378,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **wake_word_timeout** (float, default=5): Duration in seconds after a wake word is recognized. If no subsequent voice activity is detected within this window, the system transitions back to an inactive state, awaiting the next wake word or voice activation.
 
+- **wake_word_buffer_duration** (float, default=0.1): Duration in seconds to buffer audio data during wake word detection. This helps in cutting out the wake word from the recording buffer so it does not falsely get detected along with the following spoken text, ensuring cleaner and more accurate transcription start triggers. Increase this if parts of the wake word get detected as text.
+
 - **on_wakeword_detected**: A callable function triggered when a wake word is detected.
 
 - **on_wakeword_timeout**: A callable function triggered when the system goes back to an inactive state after when no speech was detected after wake word activation.
@@ -338,6 +388,33 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **on_wakeword_detection_end**: A callable function triggered when stopping to listen for wake words (e.g. because of timeout or wake word detected)
 
+## OpenWakeWord  
+
+### Training models
+
+Look [here](https://github.com/dscripka/openWakeWord?tab=readme-ov-file#training-new-models) for information about how to train your own OpenWakeWord models. You can use a [simple Google Colab notebook](https://colab.research.google.com/drive/1q1oe2zOyZp7UsB3jJiQ1IFn8z5YfjwEb?usp=sharing) for a start or use a [more detailed notebook](https://github.com/dscripka/openWakeWord/blob/main/notebooks/automatic_model_training.ipynb) that enables more customization (can produce high quality models, but requires more development experience).
+
+### Convert model to ONNX format
+
+You might need to use tf2onnx to convert tensorflow tflite models to onnx format:
+
+```bash
+pip install -U tf2onnx
+python -m tf2onnx.convert --tflite my_model_filename.tflite --output my_model_filename.onnx
+```
+
+### Configure RealtimeSTT
+
+Suggested starting parameters for OpenWakeWord usage:
+```python
+    with AudioToTextRecorder(
+        wakeword_backend="oww",
+        wake_words_sensitivity=0.35,
+        openwakeword_model_paths="word1.onnx,word2.onnx",
+        wake_word_buffer_duration=1,
+        ) as recorder:
+```
+
 ## Contribution
 
 Contributions are always welcome! 

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/RealTimeSTT.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
-Name: RealTimeSTT
-Version: 0.1.16
+Name: RealtimeSTT
+Version: 0.2.0
 Summary: A fast Voice Activity Detection and Transcription System
 Home-page: https://github.com/KoljaB/RealTimeSTT
 Author: Kolja Beigel
@@ -17,16 +17,21 @@ Requires-Dist: faster-whisper==1.0.2
 Requires-Dist: pvporcupine==1.9.5
 Requires-Dist: webrtcvad==2.0.10
 Requires-Dist: halo==0.0.31
-Requires-Dist: torch==2.3.0
-Requires-Dist: torchaudio==2.3.0
+Requires-Dist: torch==2.3.1
+Requires-Dist: torchaudio==2.3.1
 Requires-Dist: scipy==1.12.0
 Requires-Dist: websockets==v12.0
+Requires-Dist: openwakeword==0.6.0
 
 
 # RealtimeSTT
 
 *Easy-to-use, low-latency speech-to-text library for realtime applications*
 
+## New
+
+Custom wake words with [OpenWakeWord](#openwakeword). Thanks to the [developers](https://github.com/dscripka/openWakeWord) of this!
+
 ## About the Project
 
 RealtimeSTT listens to the microphone and transcribes voice into text.  
@@ -42,7 +47,7 @@ https://github.com/KoljaB/RealtimeSTT/assets/7604638/207cb9a2-4482-48e7-9d2b-072
 
 ### Updates
 
-Latest Version: v0.1.15
+Latest Version: v0.2.0
 
 See [release history](https://github.com/KoljaB/RealtimeSTT/releases).
 
@@ -66,10 +71,10 @@ This library uses:
 - **Speech-To-Text**
   - [Faster_Whisper](https://github.com/guillaumekln/faster-whisper) for instant (GPU-accelerated) transcription.
 - **Wake Word Detection**
-  - [Porcupine](https://github.com/Picovoice/porcupine) for wake word detection.
+  - [Porcupine](https://github.com/Picovoice/porcupine) or [OpenWakeWord](https://github.com/dscripka/openWakeWord) for wake word detection.
 
-*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
+*These components represent the "industry standard" for cutting-edge applications, providing the most modern and effective foundation for building high-end solutions.*
 
 ## Installation
 
@@ -83,22 +88,64 @@ Although it is possible to run RealtimeSTT with a CPU installation only (use a s
 
 ### GPU Support with CUDA (recommended)
 
-Additional steps are needed for a **GPU-optimized** installation. These steps are recommended for those who require **better performance** and have a compatible NVIDIA GPU.
+
+**Install RealtimeSTT with CUDA support**  
+
+*For Cuda 11.8:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu118
+```  
+
+
+*For Cuda 12.X:*  
+```bash
+pip install RealtimeSTT --index-url https://download.pytorch.org/whl/cu121
+```
+
+This command reinstalls PyTorch and Torchaudio with explicit support for CUDA.
+
+### Updating PyTorch Only for CUDA Support
+
+If you already have RealtimeSTT installed but want to upgrade your PyTorch installation to leverage GPU support with CUDA, follow these instructions based on your specific CUDA version. This process is useful if you initially installed RealtimeSTT with only CPU support and now wish to enhance its performance with CUDA capabilities.
+
+#### For CUDA 11.8:
+To update PyTorch and Torchaudio to support CUDA 11.8, use the following commands:
+
+```bash
+pip install torch==2.3.1+cu118 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu118
+```
+
+#### For CUDA 12.X:
+To update PyTorch and Torchaudio to support CUDA 12.X, execute the following:
+
+```bash
+pip install torch==2.3.1+cu121 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu121
+```
+
+Replace `2.3.1` with the version of CUDA that matches your system and requirements.
+
+### Steps That Might Be Necessary Before
 
 > **Note**: *To check if your NVIDIA GPU supports CUDA, visit the [official CUDA GPUs list](https://developer.nvidia.com/cuda-gpus).*
 
-To use RealtimeSTT with GPU support via CUDA please follow these steps:
+If you didn't use CUDA models before, some additional steps might be needed one time before installation. These steps prepare the system for CUDA support and installation of the **GPU-optimized** installation. This is recommended for those who require **better performance** and have a compatible NVIDIA GPU. To use RealtimeSTT with GPU support via CUDA please also follow these steps:
 
-1. **Install NVIDIA CUDA Toolkit 11.8**:
-    - Visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-11-8-0-download-archive).
+1. **Install NVIDIA CUDA Toolkit**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [NVIDIA CUDA Toolkit Archive](https://developer.nvidia.com/cuda-toolkit-archive) and select latest version.
+        - for 11.8 visit [NVIDIA CUDA Toolkit 11.8](https://developer.nvidia.com/cuda-11-8-0-download-archive).
     - Select operating system and version.
     - Download and install the software.
 
-2. **Install NVIDIA cuDNN 8.7.0 for CUDA 11.x**:
-    - Visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
-    - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
-    - Download and install the software.
-
+2. **Install NVIDIA cuDNN**:
+    - select between CUDA 11.8 or CUDA 12.X Toolkit
+        - for 12.X visit [cuDNN Downloads](https://developer.nvidia.com/cudnn-downloads).
+            - Select operating system and version.
+            - Download and install the software.
+        - for 11.8 visit [NVIDIA cuDNN Archive](https://developer.nvidia.com/rdp/cudnn-archive).
+            - Click on "Download cuDNN v8.7.0 (November 28th, 2022), for CUDA 11.x".
+            - Download and install the software.
+    
 3. **Install ffmpeg**:
 
     > **Note**: *Installation of ffmpeg might not actually be needed to operate RealtimeSTT* <sup> *thanks to jgilbert2017 for pointing this out</sup>
@@ -137,12 +184,6 @@ To use RealtimeSTT with GPU support via CUDA please follow these steps:
         scoop install ffmpeg
         ```    
 
-4. **Install PyTorch with CUDA support**:
-    ```bash
-    pip uninstall torch
-    pip install torch==2.2.2+cu118 torchaudio==2.2.2 --index-url https://download.pytorch.org/whl/cu118
-    ```
-
 ## Quick Start
 
 Basic usage:
@@ -279,6 +320,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **gpu_device_index** (int, default=0): GPU Device Index to use. The model can also be loaded on multiple GPUs by passing a list of IDs (e.g. [0, 1, 2, 3]).
 
+- **device** (str, default="cuda"): Device for model to use. Can either be "cuda" or "cpu". 
+
 - **on_recording_start**: A callable function triggered when recording starts.
 
 - **on_recording_stop**: A callable function triggered when recording ends.
@@ -346,7 +389,13 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 #### Wake Word Parameters
 
-- **wake_words** (str, default=""): Wake words for initiating the recording. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator
+- **wakeword_backend** (str, default="pvporcupine"): Specifies the backend library to use for wake word detection. Supported options include 'pvporcupine' for using the Porcupine wake word engine or 'oww' for using the OpenWakeWord engine.
+
+- **openwakeword_model_paths** (str, default=None): Comma-separated paths to model files for the openwakeword library. These paths point to custom models that can be used for wake word detection when the openwakeword library is selected as the wakeword_backend.
+
+- **openwakeword_inference_framework** (str, default="onnx"): Specifies the inference framework to use with the openwakeword library. Can be either 'onnx' for Open Neural Network Exchange format or 'tflite' for TensorFlow Lite.
+
+- **wake_words** (str, default=""): Initiate recording when using the 'pvporcupine' wakeword backend. Multiple wake words can be provided as a comma-separated string. Supported wake words are: alexa, americano, blueberry, bumblebee, computer, grapefruits, grasshopper, hey google, hey siri, jarvis, ok google, picovoice, porcupine, terminator. For the 'openwakeword' backend, wake words are automatically extracted from the provided model files, so specifying them here is not necessary.
 
 - **wake_words_sensitivity** (float, default=0.6): Sensitivity level for wake word detection (0 for least sensitive, 1 for most sensitive).
 
@@ -354,6 +403,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **wake_word_timeout** (float, default=5): Duration in seconds after a wake word is recognized. If no subsequent voice activity is detected within this window, the system transitions back to an inactive state, awaiting the next wake word or voice activation.
 
+- **wake_word_buffer_duration** (float, default=0.1): Duration in seconds to buffer audio data during wake word detection. This helps in cutting out the wake word from the recording buffer so it does not falsely get detected along with the following spoken text, ensuring cleaner and more accurate transcription start triggers. Increase this if parts of the wake word get detected as text.
+
 - **on_wakeword_detected**: A callable function triggered when a wake word is detected.
 
 - **on_wakeword_timeout**: A callable function triggered when the system goes back to an inactive state after when no speech was detected after wake word activation.
@@ -362,6 +413,33 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **on_wakeword_detection_end**: A callable function triggered when stopping to listen for wake words (e.g. because of timeout or wake word detected)
 
+## OpenWakeWord  
+
+### Training models
+
+Look [here](https://github.com/dscripka/openWakeWord?tab=readme-ov-file#training-new-models) for information about how to train your own OpenWakeWord models. You can use a [simple Google Colab notebook](https://colab.research.google.com/drive/1q1oe2zOyZp7UsB3jJiQ1IFn8z5YfjwEb?usp=sharing) for a start or use a [more detailed notebook](https://github.com/dscripka/openWakeWord/blob/main/notebooks/automatic_model_training.ipynb) that enables more customization (can produce high quality models, but requires more development experience).
+
+### Convert model to ONNX format
+
+You might need to use tf2onnx to convert tensorflow tflite models to onnx format:
+
+```bash
+pip install -U tf2onnx
+python -m tf2onnx.convert --tflite my_model_filename.tflite --output my_model_filename.onnx
+```
+
+### Configure RealtimeSTT
+
+Suggested starting parameters for OpenWakeWord usage:
+```python
+    with AudioToTextRecorder(
+        wakeword_backend="oww",
+        wake_words_sensitivity=0.35,
+        openwakeword_model_paths="word1.onnx,word2.onnx",
+        wake_word_buffer_duration=1,
+        ) as recorder:
+```
+
 ## Contribution
 
 Contributions are always welcome! 

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/RealTimeSTT.egg-info/SOURCES.txt

@@ -7,4 +7,9 @@ RealTimeSTT.egg-info/dependency_links.txt
 RealTimeSTT.egg-info/requires.txt
 RealTimeSTT.egg-info/top_level.txt
 RealtimeSTT/__init__.py
-RealtimeSTT/audio_recorder.py
+RealtimeSTT/audio_recorder.py
+RealtimeSTT.egg-info/PKG-INFO
+RealtimeSTT.egg-info/SOURCES.txt
+RealtimeSTT.egg-info/dependency_links.txt
+RealtimeSTT.egg-info/requires.txt
+RealtimeSTT.egg-info/top_level.txt

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/RealTimeSTT.egg-info/requires.txt

@@ -3,7 +3,8 @@ faster-whisper==1.0.2
 pvporcupine==1.9.5
 webrtcvad==2.0.10
 halo==0.0.31
-torch==2.3.0
-torchaudio==2.3.0
+torch==2.3.1
+torchaudio==2.3.1
 scipy==1.12.0
 websockets==v12.0
+openwakeword==0.6.0

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/RealtimeSTT/audio_recorder.py

@@ -31,9 +31,11 @@ import torch.multiprocessing as mp
 import torch
 from typing import List, Union
 from ctypes import c_bool
+from openwakeword.model import Model
 from scipy.signal import resample
 from scipy import signal
 import faster_whisper
+import openwakeword
 import collections
 import numpy as np
 import pvporcupine
@@ -67,6 +69,7 @@ INIT_WAKE_WORDS_SENSITIVITY = 0.6
 INIT_PRE_RECORDING_BUFFER_DURATION = 1.0
 INIT_WAKE_WORD_ACTIVATION_DELAY = 0.0
 INIT_WAKE_WORD_TIMEOUT = 5.0
+INIT_WAKE_WORD_BUFFER_DURATION = 0.1
 ALLOWED_LATENCY_LIMIT = 10
 
 TIME_SLEEP = 0.02
@@ -90,8 +93,9 @@ class AudioToTextRecorder:
                  model: str = INIT_MODEL_TRANSCRIPTION,
                  language: str = "",
                  compute_type: str = "default",
-                 input_device_index: int = 0,
+                 input_device_index: int = None,
                  gpu_device_index: Union[int, List[int]] = 0,
+                 device: str = "cuda",
                  on_recording_start=None,
                  on_recording_stop=None,
                  on_transcription_start=None,
@@ -128,12 +132,16 @@ class AudioToTextRecorder:
                  on_vad_detect_stop=None,
 
                  # Wake word parameters
+                 wakeword_backend: str = "pvporcupine",
+                 openwakeword_model_paths: str = None,
+                 openwakeword_inference_framework: str = "onnx",
                  wake_words: str = "",
                  wake_words_sensitivity: float = INIT_WAKE_WORDS_SENSITIVITY,
                  wake_word_activation_delay: float = (
                     INIT_WAKE_WORD_ACTIVATION_DELAY
                  ),
                  wake_word_timeout: float = INIT_WAKE_WORD_TIMEOUT,
+                 wake_word_buffer_duration: float = INIT_WAKE_WORD_BUFFER_DURATION,
                  on_wakeword_detected=None,
                  on_wakeword_timeout=None,
                  on_wakeword_detection_start=None,
@@ -173,6 +181,8 @@ class AudioToTextRecorder:
             IDs (e.g. [0, 1, 2, 3]). In that case, multiple transcriptions can
             run in parallel when transcribe() is called from multiple Python
             threads
+        - device (str, default="cuda"): Device for model to use. Can either be 
+            "cuda" or "cpu".
         - on_recording_start (callable, default=None): Callback function to be
             called when recording of audio to be transcripted starts.
         - on_recording_stop (callable, default=None): Callback function to be
@@ -241,11 +251,26 @@ class AudioToTextRecorder:
             be called when the system listens for voice activity.
         - on_vad_detect_stop (callable, default=None): Callback function to be
             called when the system stops listening for voice activity.
+        - wakeword_backend (str, default="pvporcupine"): Specifies the backend
+            library to use for wake word detection. Supported options include
+            'pvporcupine' for using the Porcupine wake word engine or 'oww' for
+            using the OpenWakeWord engine.
+        - openwakeword_model_paths (str, default=None): Comma-separated paths
+            to model files for the openwakeword library. These paths point to
+            custom models that can be used for wake word detection when the
+            openwakeword library is selected as the wakeword_backend.
+        - openwakeword_inference_framework (str, default="onnx"): Specifies
+            the inference framework to use with the openwakeword library.
+            Can be either 'onnx' for Open Neural Network Exchange format 
+            or 'tflite' for TensorFlow Lite.
         - wake_words (str, default=""): Comma-separated string of wake words to
-            initiate recording. Supported wake words include:
-            'alexa', 'americano', 'blueberry', 'bumblebee', 'computer',
-            'grapefruits', 'grasshopper', 'hey google', 'hey siri', 'jarvis',
-            'ok google', 'picovoice', 'porcupine', 'terminator'.
+            initiate recording when using the 'pvporcupine' wakeword backend.
+            Supported wake words include: 'alexa', 'americano', 'blueberry',
+            'bumblebee', 'computer', 'grapefruits', 'grasshopper', 'hey google',
+            'hey siri', 'jarvis', 'ok google', 'picovoice', 'porcupine',
+            'terminator'. For the 'openwakeword' backend, wake words are
+            automatically extracted from the provided model files, so specifying
+            them here is not necessary.
         - wake_words_sensitivity (float, default=0.5): Sensitivity for wake
             word detection, ranging from 0 (least sensitive) to 1 (most
             sensitive). Default is 0.5.
@@ -257,6 +282,12 @@ class AudioToTextRecorder:
             wake word is recognized. If no subsequent voice activity is
             detected within this window, the system transitions back to an
             inactive state, awaiting the next wake word or voice activation.
+        - wake_word_buffer_duration (float, default=0.1): Duration in seconds
+            to buffer audio data during wake word detection. This helps in
+            cutting out the wake word from the recording buffer so it does not
+            falsely get detected along with the following spoken text, ensuring
+            cleaner and more accurate transcription start triggers.
+            Increase this if parts of the wake word get detected as text.
         - on_wakeword_detected (callable, default=None): Callback function to
             be called when a wake word is detected.
         - on_wakeword_timeout (callable, default=None): Callback function to
@@ -298,9 +329,11 @@ class AudioToTextRecorder:
         self.compute_type = compute_type
         self.input_device_index = input_device_index
         self.gpu_device_index = gpu_device_index
+        self.device = device
         self.wake_words = wake_words
         self.wake_word_activation_delay = wake_word_activation_delay
         self.wake_word_timeout = wake_word_timeout
+        self.wake_word_buffer_duration = wake_word_buffer_duration
         self.ensure_sentence_starting_uppercase = (
             ensure_sentence_starting_uppercase
         )
@@ -368,6 +401,7 @@ class AudioToTextRecorder:
         self.last_transcription_bytes = None
         self.initial_prompt = initial_prompt
         self.suppress_tokens = suppress_tokens
+        self.use_wake_words = wake_words or wakeword_backend in {'oww', 'openwakeword', 'openwakewords'}
 
         # Initialize the logging configuration with the specified level
         log_format = 'RealTimeSTT: %(name)s - %(levelname)s - %(message)s'
@@ -397,7 +431,7 @@ class AudioToTextRecorder:
             logging.debug("Explicitly setting the multiprocessing start method to 'spawn'")
             mp.set_start_method('spawn')
         except RuntimeError as e:
-            logging.debug("Start method has already been set. Details:", e)
+            logging.debug(f"Start method has already been set. Details: {e}")
 
         logging.info("Starting RealTimeSTT")
 
@@ -406,6 +440,9 @@ class AudioToTextRecorder:
         self.main_transcription_ready_event = mp.Event()
         self.parent_transcription_pipe, child_transcription_pipe = mp.Pipe()
 
+        # Set device for model
+        self.device = "cuda" if self.device == "cuda" and torch.cuda.is_available() else "cpu"
+
         self.transcript_process = mp.Process(
             target=AudioToTextRecorder._transcription_worker,
             args=(
@@ -413,6 +450,7 @@ class AudioToTextRecorder:
                 model,
                 self.compute_type,
                 self.gpu_device_index,
+                self.device,
                 self.main_transcription_ready_event,
                 self.shutdown_event,
                 self.interrupt_stop_event,
@@ -452,7 +490,7 @@ class AudioToTextRecorder:
                              )
                 self.realtime_model_type = faster_whisper.WhisperModel(
                     model_size_or_path=self.realtime_model_type,
-                    device='cuda' if torch.cuda.is_available() else 'cpu',
+                    device=self.device,
                     compute_type=self.compute_type,
                     device_index=self.gpu_device_index
                 )
@@ -467,33 +505,82 @@ class AudioToTextRecorder:
                           "transcription model initialized successfully")
 
         # Setup wake word detection
-        if wake_words:
+        if wake_words or wakeword_backend in {'oww', 'openwakeword', 'openwakewords'}:
+            self.wakeword_backend = wakeword_backend
 
             self.wake_words_list = [
                 word.strip() for word in wake_words.lower().split(',')
             ]
-            sensitivity_list = [
+            self.wake_words_sensitivity = wake_words_sensitivity
+            self.wake_words_sensitivities = [
                 float(wake_words_sensitivity)
                 for _ in range(len(self.wake_words_list))
             ]
 
-            try:
-                self.porcupine = pvporcupine.create(
-                    keywords=self.wake_words_list,
-                    sensitivities=sensitivity_list
-                )
-                self.buffer_size = self.porcupine.frame_length
-                self.sample_rate = self.porcupine.sample_rate
+            match self.wakeword_backend:
+                case 'pvp' | 'pvporcupine':
+                    try:
+                        self.porcupine = pvporcupine.create(
+                            keywords=self.wake_words_list,
+                            sensitivities=self.wake_words_sensitivities
+                        )
+                        self.buffer_size = self.porcupine.frame_length
+                        self.sample_rate = self.porcupine.sample_rate
 
-            except Exception as e:
-                logging.exception("Error initializing porcupine "
-                                  f"wake word detection engine: {e}"
-                                  )
-                raise
+                    except Exception as e:
+                        logging.exception(
+                            "Error initializing porcupine "
+                            f"wake word detection engine: {e}"
+                        )
+                        raise
+
+                    logging.debug(
+                        "Porcupine wake word detection engine initialized successfully"
+                    )
+
+                case 'oww' | 'openwakeword' | 'openwakewords':
+                    openwakeword.utils.download_models()
+
+                    try:
+                        if openwakeword_model_paths:
+                            model_paths = openwakeword_model_paths.split(',')
+                            self.owwModel = Model(
+                                wakeword_models=model_paths,
+                                inference_framework=openwakeword_inference_framework
+                            )
+                            logging.info(
+                                "Successfully loaded wakeword model(s): "
+                                f"{openwakeword_model_paths}"
+                            )
+                        else:
+                            self.owwModel = Model(
+                                inference_framework=openwakeword_inference_framework)
+                        
+                        self.oww_n_models = len(self.owwModel.models.keys())
+                        if not self.oww_n_models:
+                            logging.error(
+                                "No wake word models loaded."
+                            )
+   
+                        for model_key in self.owwModel.models.keys():
+                            logging.info(
+                                "Successfully loaded openwakeword model: "
+                                f"{model_key}"
+                            )
+
+                    except Exception as e:
+                        logging.exception(
+                            "Error initializing openwakeword "
+                            f"wake word detection engine: {e}"
+                        )
+                        raise
+
+                    logging.debug(
+                        "Open wake word detection engine initialized successfully"
+                    )
+                case _:
+                    logging.exception("Wakeword engine {} unknown/unsupported. Please specify one of: pvporcupine, openwakeword.")
 
-            logging.debug("Porcupine wake word detection "
-                          "engine initialized successfully"
-                          )
 
         # Setup voice activity detection model WebRTC
         try:
@@ -566,6 +653,7 @@ class AudioToTextRecorder:
                               model_path,
                               compute_type,
                               gpu_device_index,
+                              device,
                               ready_event,
                               shutdown_event,
                               interrupt_stop_event,
@@ -593,6 +681,7 @@ class AudioToTextRecorder:
             compute_type (str): Specifies the type of computation to be used
                 for transcription.
             gpu_device_index (int): Device ID to use.
+            device (str): Device for model to use.
             ready_event (threading.Event): An event that is set when the
               transcription model is successfully initialized and ready.
             shutdown_event (threading.Event): An event that, when set,
@@ -616,7 +705,7 @@ class AudioToTextRecorder:
         try:
             model = faster_whisper.WhisperModel(
                 model_size_or_path=model_path,
-                device='cuda' if torch.cuda.is_available() else 'cpu',
+                device=device,
                 compute_type=compute_type,
                 device_index=gpu_device_index,
             )
@@ -698,6 +787,9 @@ class AudioToTextRecorder:
         """
         try:
             audio_interface = pyaudio.PyAudio()
+            if input_device_index is None:
+                default_device = audio_interface.get_default_input_device_info()
+                input_device_index = default_device['index']
             stream = audio_interface.open(
                 rate=sample_rate,
                 format=pyaudio.paInt16,
@@ -852,6 +944,46 @@ class AudioToTextRecorder:
             logging.error(result)
             raise Exception(result)
 
+    def _process_wakeword(self, data):
+        """
+        Processes audio data to detect wake words.
+        """
+        match self.wakeword_backend:
+            case 'pvp' | 'pvporcupine':
+                pcm = struct.unpack_from(
+                    "h" * self.buffer_size,
+                    data
+                )
+                porcupine_index = self.porcupine.process(pcm)
+                if self.debug_mode:
+                    print (f"wake words porcupine_index: {porcupine_index}")
+                return self.porcupine.process(pcm)
+
+            case 'oww' | 'openwakeword' | 'openwakewords':
+                pcm = np.frombuffer(data, dtype=np.int16)
+                prediction = self.owwModel.predict(pcm)
+                max_score = -1
+                max_index = -1
+                wake_words_in_prediction = len(self.owwModel.prediction_buffer.keys())
+                self.wake_words_sensitivities
+                if wake_words_in_prediction:
+                    for idx, mdl in enumerate(self.owwModel.prediction_buffer.keys()):
+                        scores = list(self.owwModel.prediction_buffer[mdl])
+                        if scores[-1] >= self.wake_words_sensitivity and scores[-1] > max_score:
+                            max_score = scores[-1]
+                            max_index = idx
+                    if self.debug_mode:
+                        print (f"wake words oww max_index, max_score: {max_index} {max_score}")
+                    return max_index  
+                else:
+                    if self.debug_mode:
+                        print (f"wake words oww_index: -1")
+                    return -1
+
+        if self.debug_mode:        
+            print("wake words no match")
+        return -1
+
     def text(self,
              on_transcription_finished=None,
              ):
@@ -1109,14 +1241,14 @@ class AudioToTextRecorder:
                     if wake_word_activation_delay_passed \
                             and not delay_was_passed:
 
-                        if self.wake_words and self.wake_word_activation_delay:
+                        if self.use_wake_words and self.wake_word_activation_delay:
                             if self.on_wakeword_timeout:
                                 self.on_wakeword_timeout()
                     delay_was_passed = wake_word_activation_delay_passed
 
                     # Set state and spinner text
                     if not self.recording_stop_time:
-                        if self.wake_words \
+                        if self.use_wake_words \
                                 and wake_word_activation_delay_passed \
                                 and not self.wakeword_detected:
                             self._set_state("wakeword")
@@ -1126,14 +1258,10 @@ class AudioToTextRecorder:
                             else:
                                 self._set_state("inactive")
 
-                    # Detect wake words if applicable
-                    if self.wake_words and wake_word_activation_delay_passed:
+                    #self.wake_word_detect_time = time.time()
+                    if self.use_wake_words and wake_word_activation_delay_passed:
                         try:
-                            pcm = struct.unpack_from(
-                                "h" * self.buffer_size,
-                                data
-                                )
-                            wakeword_index = self.porcupine.process(pcm)
+                            wakeword_index = self._process_wakeword(data)
 
                         except struct.error:
                             logging.error("Error unpacking audio data "
@@ -1144,14 +1272,14 @@ class AudioToTextRecorder:
                             logging.error(f"Wake word processing error: {e}")
                             continue
 
-                        # If a wake word is detected
+                        # If a wake word is detected                        
                         if wakeword_index >= 0:
 
                             # Removing the wake word from the recording
-                            samples_for_0_1_sec = int(self.sample_rate * 0.1)
+                            samples_time = int(self.sample_rate * self.wake_word_buffer_duration)
                             start_index = max(
                                 0,
-                                len(self.audio_buffer) - samples_for_0_1_sec
+                                len(self.audio_buffer) - samples_time
                                 )
                             temp_samples = collections.deque(
                                 itertools.islice(
@@ -1164,12 +1292,13 @@ class AudioToTextRecorder:
 
                             self.wake_word_detect_time = time.time()
                             self.wakeword_detected = True
+                            #self.wake_word_cooldown_time = time.time()
                             if self.on_wakeword_detected:
                                 self.on_wakeword_detected()
 
                     # Check for voice activity to
                     # trigger the start of recording
-                    if ((not self.wake_words
+                    if ((not self.use_wake_words
                          or not wake_word_activation_delay_passed)
                             and self.start_recording_on_voice_activity) \
                             or self.wakeword_detected:

{RealTimeSTT-0.1.16 → RealtimeSTT-0.2.0}/setup.py

@@ -8,8 +8,8 @@ with open('requirements.txt') as f:
     requirements = f.read().splitlines()
 
 setuptools.setup(
-    name="RealTimeSTT",
-    version="0.1.16",
+    name="RealtimeSTT",
+    version="0.2.0",
     author="Kolja Beigel",
     author_email="kolja.beigel@web.de",
     description="A fast Voice Activity Detection and Transcription System",