geoffkait 22,183 posts Who likes running their mouth? 22,183 posts. Sorry, though, my fault, when I write number like 0.1db, I am communicating numbers that are well known by those "skilled in the art". geoffkait 22,183 posts. This is a very well known number for volume matching when doing human perception testing in audio as it is known to eliminate volume biasing in testing. geoffkait 22,183 posts. Some researchers would prefer to use 0.05db to ensure more tolerance between identified just noticeable differences and what is used in the test. geoffkait 22,183 posts
https://secure.aes.org/forum/pubs/journal/?elib=5163
That 0.1db is with noise spectra where we have much higher sensitivity to differences. With actual music, it is much harder to detect volume differences. Those numbers are for rapid switching too. With longer differences in switching time, small differences in volume are harder to detect.
https://secure.aes.org/forum/pubs/journal/?elib=5163
That 0.1db is with noise spectra where we have much higher sensitivity to differences. With actual music, it is much harder to detect volume differences. Those numbers are for rapid switching too. With longer differences in switching time, small differences in volume are harder to detect.