During a specified time interval, the bit error rate (BER) is the number of times that a transmission of bits received has been modified by interference, noise, bit synchronization errors, phase jitter, or distortion. The number of errors in that time interval is then compared to the total number of bits transmitted to come up with the percentage of errors. As such, the BER is a network performance ratio for digital transmissions via radio data links, Ethernet, or fiber optic data networks. For example, if a transmission packet sent contains 10 bits of binary code and two of those bits are altered in transmission, the BER would be 20 percent. In telecommunications over fiber optics, this BER is calculated differently as user-visible error-rates calculations are needed; measurements are of errored seconds, found by measuring the intervals of a second during which any bit errors occur.
As BER measurements can be taken on transmitters, receivers, and the communication networks linking them, BER is a total system assessment tool for detection of system integrity in effective performance. Analysis of bit error rate on systems is usually done utilizing simulation models. The results of simulations determines what forward-error-correction codes a system administrator needs to apply to improve raw channel transmission performance.
Sometimes the bit error rate can be improved by using a stronger signal; however, this may cause more cross-talk errors as well as bit errors. If bit errors have already been addressed with forward-error-correction coding and the BER is still too high, it is best to address the factors that cause the bit errors. The main culprit is usually noise and radio propagation path changes. In fiber optics networks, the problems are usually in the components of the network itself, requiring the network to be tested minutely. Noise can come from optical receivers themselves when photodiodes or amplifiers are not responding to very small changes and produce high noise levels.
To test for bit error rate causes, one of the simulators used is a pseudo-random binary sequence of numbers sent in pattern sequences to check for phase jitter in the system. A similar test is when a quasi-random signal source generates and sends every possible combination of a 20-bit word and repeats these every 1,048,575 bits. At the same time the source generator would be suppressing consecutive zeros to less than 14 and shifting between high and low density changes to measure for phase jitter. Another test, called All-Ones, sends packets of ones only and repeats to consume the maximum power to see if direct current to the repeater is regulated correctly and to test span power. Many simulations can test all the components of any transmission system.