The paper presents an orthogonal frequency division multiplexing (OFDM)-based digital semantic communication (SemCom) system that integrates with existing digital communication infrastructures. The system extracts and transmits task-oriented semantics, reducing data transmission size while maintaining task performance. Key contributions include: 1. **System Model**: The system uses scalar quantizers to quantize extracted semantics, transforms them into OFDM signals, and transmits over frequency-selective channels. 2. **Semantic Importance Measurement**: A method is proposed to measure the importance of different semantics based on their correlation with tasks and inter-signal correlations. 3. **Optimization Problem**: A sub-carrier and bit allocation problem is formulated to maximize communication performance while minimizing semantic distortion. 4. **Deep Reinforcement Learning (DRL)**: A DRL-based bit allocation algorithm with a dynamic action space is introduced to optimize the allocation strategy. 5. **Simulation Results**: The proposed system achieves 9.7% and 28.7% performance gains compared to analog SemCom and conventional bit-based communication systems, respectively, across different signal-to-noise ratios (SNRs) and bit constraints. The paper addresses challenges such as multipath fading, semantic importance evaluation, and efficient resource allocation, demonstrating the effectiveness of the proposed system in enhancing communication efficiency and robustness.The paper presents an orthogonal frequency division multiplexing (OFDM)-based digital semantic communication (SemCom) system that integrates with existing digital communication infrastructures. The system extracts and transmits task-oriented semantics, reducing data transmission size while maintaining task performance. Key contributions include: 1. **System Model**: The system uses scalar quantizers to quantize extracted semantics, transforms them into OFDM signals, and transmits over frequency-selective channels. 2. **Semantic Importance Measurement**: A method is proposed to measure the importance of different semantics based on their correlation with tasks and inter-signal correlations. 3. **Optimization Problem**: A sub-carrier and bit allocation problem is formulated to maximize communication performance while minimizing semantic distortion. 4. **Deep Reinforcement Learning (DRL)**: A DRL-based bit allocation algorithm with a dynamic action space is introduced to optimize the allocation strategy. 5. **Simulation Results**: The proposed system achieves 9.7% and 28.7% performance gains compared to analog SemCom and conventional bit-based communication systems, respectively, across different signal-to-noise ratios (SNRs) and bit constraints. The paper addresses challenges such as multipath fading, semantic importance evaluation, and efficient resource allocation, demonstrating the effectiveness of the proposed system in enhancing communication efficiency and robustness.