The ABX test is a method of comparing two sensory stimuli to determine if there is a detectable difference between them. First play two sections of samples A and B respectively, followed by an unknown sample X. X is randomly selected from A or B. Participants need to give a clear judgment on whether X is A or B. If If X cannot be judged reliably by a low p-value, it cannot prove that there is a noticeable difference between A and B.
The ABX test can be easily performed in the form of a double-blind listening test, eliminating any potential influence from the researcher or test organizer. Since sample A and sample B are provided before sample X, there is no need to distinguish the difference between them based on assumptions of long-term memory or past experience. Therefore, the ABX test answers whether there is a difference in perception under ideal circumstances.
The ABX test is usually used for the evaluation of digital audio data compression methods; sample A is usually an uncompressed sample, and sample B is a compressed version of A. Audio compression artifacts that show that the compression algorithm is defective can be identified through subsequent tests. The ABX test can also be used to compare the fidelity loss of two different audio formats at a given bit rate.
ABX testing can be used for audio input, processing, output components and wiring: almost any audio product or prototype design.