The number of DVE channels that can be used in Viz Engine depends on the video hardware. The limiting factor is the amount of memory on the card, but also the available bandwidth for transfers between the FPGA and the on-board memory. For example, on a Matrox X.mio3 card, you can scale 8 1080i surfaces to 99.9% and compose those 8 surfaces at the same time. If the scaling factor is reduced, the overall DVE performance will increase, due to the lowered amount of bandwidth needed. At a certain point, it might even be possible to add an additional DVE input.
Looking at a single channel system fit with an X.mio3, 9 inputs can be used as DVE, as one is always consumed by the background layer. However, setting the system up as dual channel, reduces the amount of DVE layers to 3 in each channel, considering that 2 background layers are used.