Partial loudness - if the background is a complex signal

[Christian Uhle <Christian.Uhle@xxxxxxxxxxxxxxxxx>]

Dear readers,

this is my first post here: Hello everyone.
The question which puzzles me for quite some time is related to partial 
loudness.
The loudness of a sound is reduced if it is presented in noise.
Models exist which are capable of predicting the partial loudness, e.g. [1].

My understanding is that in [1] the interferer (the masker, or 
background) is stationary noise.
Are there also models for partial loudness if the interferer is a 
complex signal, e.g. music or speech?

I assume that such a model needs to determine a timbral similarity 
between the target and the interferer.
(For example: In an extreme case where both signals sound (are) the 
same, the interferer would not reduce but increase the loudness of the 
signal.
If we start to modulate the frequency of one of the signals slightly, 
this will affect the partial loudness, but how (quantitatively)?)
The similarity of both signals depends on comodulation of sub-band 
amplitude envelopes of signal and interferer, and on other cues as well, 
as in stream segregation mechanisms.

Pointers to related work and papers as well as your thougths about this 
are highly appreciated.
Another question would be if experimental data about partial loudness in 
complex signals exist, and where to find them.

All the best,

Christian

[1] Moore, Glasberg, Baer, "A Model for the Prediction of Thresholds, 
Loudness, and Partial Loudness", JAES, 1997.

--
Dr.-Ing. Christian Uhle

IIS Labs
Dept. Audio
Fraunhofer Institute for Integrated Circuits IIS