[LAD] Best way to insert DSP module in audio chain in Linux

Hi, I want to make is a DSP module that improves the sound quality of the sound coming from the laptop builtin speakers by applying DSP (FIR/IIR filters). I want to route all audio played back (from for instance an offline WAV or MP3 file, from a movie, or streamed from the internet) to pass through the DSP plugin whenever headphones are NOT connected. When headphones are connected no processing should be applied. I want to place the DSP module as close to the hardware as possible to make sure all audio is really routed through the DSP plugin so I always can enjoy the DSP improvements made to the sound. Where should I place this module/DSP code? - As a plugin to PulseAudio? - As a plugin to ALSA? - As its own virtual sound card? - In the audio driver for my built-in sound card? - As a kernel module? I have tried loading the LADSPA-module using the module-ladspa-sink in PulseAudio but I am not sure this is the best solution. I need to remove some limitations in module-ladspa-sink to get it working properly ( multichannel audio, only DSP processing for internal speakers). My requirements: 1) Must be able to detect if headphones are connected 2) Must be able to process stereo and multichannel (5.1/7.1) formats. I need the multichannel formats to perform binaural downmixing to stereo. So it is important that I can receive multichannel audio and mix it down to stereo and output it to a stereo soundcard. /Kim