It's a sound idea, as long as you can justify the development costs while recognizing that it still represents a risk. The main danger there is in the potential for industry changes to hit you in the signal path (such as HDMI v1.2/v1.3/v1.4). If the latest features that owners want include things like CEC, ARC, 3D video, and 4k video, no amount of over-engineering of the DSP chip and memory resources will help when you also need to re-design around a new HDMI transceiver chip or a video signal path with much more bandwidth.