Hi, I would like to have real stereo (or multi-channel) UGens in ChucK which can compute multichannel output and not just setting a pan or like that (like it is now??). The tick() function of an UGen should operate on array of samples (one sample on each channel) and produce also one sample for each channel as output. The audio engine should convert the channels if two ugens of different output/input channels are chucked together. Each UGen should either have a fixed number of in/out channels, or the user could specify the number of in/out channels for an UGen at construction time (to avoid extra multichannel processing when unnecessary). It would be not needed to access the individual channels of an UGen as UGens (this can be emulated e.g. by mono-to-stereo "split" UGens and stereo-to-mono "channel-select" UGens). I would like to implement something like this for myself because I think it is unconvenient that we don't have stereo ugens and some effects in which the channels are dependent of each other are not even possible to realize otherwise. Have you any other suggestions, advices how to implement multichannel ugens? It would be the best to have a solution which is good for everyone and which can be part of the official ChucK code. Or do I still not understand how multi-channel UGens work? Imre Kéri