This disclosure describes techniques to measure the quality of digital TV signals. Publicly available television broadcasts are received using high-quality reception equipment to obtain an error free training set. Transmission errors such as signal corruption and packet loss are simulated by injecting common faults into the error free training set to generate faulty signal training data sets. A machine learning model is trained using a combination of error-free and faulty digital TV signal training data sets. The trained machine learning model is used to measure signal quality of an MPEG transport stream of a digital TV signal and generate a metric indicative of signal quality.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.