ABX Blind Test
Double-blind ABX audio comparison. Load two files, guess which is X, and get statistical significance results.
What this tool tells you
What it does
Use this when you want to test, under double-blind conditions, whether you can actually distinguish between two audio files.
How to use it
Load files A and B, listen to the randomly-assigned X sample, and guess whether X is A or B. Repeat multiple trials; the tool tracks your results and computes statistical significance.
How to interpret it
The p-value is a one-tailed binomial test against random guessing (p=0.5), not a measure of how different the files sound. At least 10–16 trials are recommended before drawing conclusions — a single lucky run can appear significant by chance.
Frequently Asked Questions
- What is an ABX test?
- An ABX test is a double-blind methodology where you compare two audio samples (A and B) with an unknown sample (X). X is randomly assigned as A or B, and you must identify which one it is.
- How many trials do I need?
- At least 10-16 trials are recommended for statistical significance. With 12 correct out of 16, you achieve p < 0.05, indicating a statistically significant ability to distinguish.
Selected files are processed locally in the browser and are not uploaded to TrueRes for processing. Open this page in a modern browser to use the interactive tool.