In the next scenario, PS traffic was mostly allocated to
the BCCH layer, and CS WIC allocated to the nonBCCH
layer. From Figure 4, similar behavior can be
seen as in Figure 3 - the quality of CS traffic gets
clearly worse when increasing the PS traflic. In the
worst case, CS traffic of 14.1 Er1 and PS traffic of 9.1
ErI, the amount of bad quality samples of speech FER
>= 1% is about 7%. However, the result is much better
compared to the case where the BCCH TRX was not
preferred for PS traffc (Figure 3). When the PS
traffic load is not too high (less than 3 Erls), most of
the PS traftic is allocated to the BCCH layer, while
most of the CS traffic is allocated to the non-BCCH
layer. This interference separation is the reason behind
the small quality degmdation of speech. When the PS
trflic load is high, some PS Wic is allocated also to
the non-BCCH layer. More interference is introduced to CS service clearly decreasing the quality of speech