It's been more than a year that this topic shows as 'being on the roadmap', but for now the reality is that people bypass PulseAudio to hear multichannel sound. I thought solving this would be a nice summer skunkworks project while everyone is busy increasing their chances of skin cancer. Here's the deal: formatting an AC3 file into the required IEC format takes about 30 lines of code, this can be done in the application that connects to PulseAudio. Then the problem is to pass raw data internally without applying any volume controls and with mixing disabled. There's no issue with timing or variable bitrate since the formatting adds headers and zero-padding to reach the non-compressed PCM bitrate, i.e. all the bytes-to-ms conversions still apply. I started looking into this but the volume control code is spread out between sink-input, sink, alsa, and it's hard to figure out how to go about this. Likewise, preventing mixing isn't self-explanatory. If anyone has pointers on these points or is willing to contribute, let me know. Cheers - Pierre
