## evs_encoder.png

The image is a detailed block diagram labeled "Encoder Block Diagram." It illustrates the process of encoding audio signals into a bitstream format for transmission over networks. The diagram is divided into several sections that represent different stages and components involved in this process.

### Left Side: Pre-Processing Section

1. **Input Audio**: This section starts with an input audio signal, which is the raw sound wave that needs to be processed.
2. **HP Filter (20 Hz)**: The audio signal passes through a high-pass filter set at 20 Hertz, which removes any low-frequency components below this threshold.
3. **Noise Update/Estimation**: This step involves estimating and updating noise levels in the audio signal.
4. **Filter-Bank & Resampling**: The filtered signal is then processed by a filter bank to separate it into different frequency bands, followed by resampling to adjust its sampling rate.
5. **Bandwidth Detector**: This component detects the bandwidth of the audio signal, which helps in determining how much data needs to be transmitted for efficient compression.
6. **Time-Domain Transient Detector**: Detects any sudden changes or transients in the audio signal that might require special handling during encoding.
7. **LP Analysis, Spectral Analysis**: Long-time and short-time spectral analysis is performed on the signal to understand its frequency characteristics over time.
8. **Channel Aware (CA) Configuration**: This step configures settings based on the channel conditions for optimal transmission.
9. **Signal Classifier**: The classifier identifies whether the input audio contains speech or music, which affects how it will be encoded differently.
10. **MDCT Selector**: Selects between different types of Discrete Cosine Transform (DCT) methods to process the signal.

### Middle Section: Encoder Core

The core encoder processes the pre-processed audio data and outputs a bitstream that can be transmitted over networks.

1. **Core and DTX Switching**: This section includes the main encoding logic, which switches between different modes based on the characteristics of the input audio.
2. **LP-based Encoder**: Encodes the signal using Linear Prediction (LP) techniques.
3. **MDCT-based Encoder**: Uses Modified Discrete Cosine Transform (MDCT) for encoding.
4. **DTX, CNG Encoder**: Implements Discontinuous Transmission (DTX) and Comfort Noise Generation (CNG) to manage silence periods in speech.

### Right Side: Encoding Modes

The diagram shows different modes of operation that the encoder can use based on specific requirements or standards:

1. **EVS PRIMARY MODES**:
   - **LP-based Encoder**: Uses Linear Prediction for encoding.
   - **MDCT-based Encoder**: Uses Modified Discrete Cosine Transform for encoding.
   - **DTX, CNG Encoder**: Implements DTX and CNG techniques.

2. **AMR-WB BACKWARD COMPATIBLE MODE**: This mode is designed to be compatible with the Adaptive Multi-Rate Wideband (AMR-WB) standard, ensuring backward compatibility while providing enhanced features.

### Output: Bitstream

The final output of this process is a bitstream that can be transmitted over networks. The bitstream contains encoded audio data in a format suitable for transmission and decoding at the receiver end.

### Logos and Credits:

At the bottom of the diagram, there are logos from various companies involved in the development or endorsement of this codec:
- **Fraunhofer IIS**
- **Huawei**
- **VoiceAge**

The source of the image is cited as "EVS Codec," with a link provided for further information: <https://www.aes.org/technical/documentDownloads.cfm?docID=548>.

This diagram provides a comprehensive overview of how audio signals are encoded into bitstreams, highlighting various stages and techniques used in modern audio compression standards.

This description was generated automatically from image files by a local LLM, and thus, may not be fully accurate. Please feel free to ask questions if you have further questions about the nature of the image or its meaning within the presentation.