About a year or 2 ago, there was a posting by someone by the name of Yan..
It was a very interesting issue.. which is to switch the slope of the spreading function so that the lower slope is less steeper than the upper slope..

Initially, I thought I found something interesting when I tried calculating the r_normalization factors and compared to the R_norm of the MP3 ISO specs..

I now think that Yan is wrong.. and MP3 ISO specs is wrong about the r_normal..
Both slopes should be identical in steepness.. Yan's proposal would result in audible artifacts..

Not only that.. refering to the orginal SNR equation :

SNR = (14.5 + i) * tb +(1-tb)* 5.5

this is when the spreading function slope is steeper in the upper slope than in the lower slope.. just as in Layer II and Layer I psychoacoustic model II..


when it was changed to SNR = 29 *tb + (1-tb)*6.0 in MP3, both slope of the spreading function has to be modified to make them identical in steepness to compensate for the constant TMN..

Try working out the mathematical equations..

wkwai