A wireless audio device may realize power savings when processing multi-channel data, e.g. 2-channel stereo audio data, by splitting the multi-channel block data into separate data blocks for each channel, e.g. Left-channel data block and Right-channel data block, and processing each separate data block independently in the source device to generate respective data packets targeting corresponding sink devices, e.g. a Left earphone and a Right earphone. The source device may then transmit each data packet to a different corresponding sink device, e.g. a Left-channel packet to a Left earphone and a Right-channel packet to a Right earphone. The data packets may also include header information indicative of whether a next packet is intended for a given sink device, enabling sink devices to enter sleep mode when the header information indicates that a next packet is not intended for the given sink device, to save power.