If you google "jriver bitstreaming" you'll see that questions like that have been asked countless times. I'll try to give a short, simplified explanation, but do more research there if you want to know more.
Simply put, Bitstreaming lets JRiver output encoded audio without decoding it first. This would make no sense at all if you're talking about 2-channel PCM, which is what is on a CD. It only makes sense if you're talking about an encoded format; Dolby Digital or DTS, or Dolby Atmos, would be examples.
Usually it it is best to let MC do the decoding. It then outputs the signal as multi-channel PCM. This requires your connection support that. A TOSLINK or S/PDIF optical connection supports 2-channel audio, period. It cannot transmit multi-channel (more than 2 channel audio). So formats like DTS exist that take 5.1 channel audio, and mathematically encode it into a 2-channel PCM stream that S/PDIF can physically carry. Then your receiver decodes it back into 5.1 channel audio.
So Bitstreaming sends the encoded data, without decoding it. Why would you want to do this? Well, over HDMI, you usually wouldn't, because HDMI can support multichannel PCM audio. But there are still a couple of cases where you might:
- Dolby Atmos: Only hardware receivers licensed by Dolby can fully decode Atmos with all the object coordinate data; so to get the full Atmos experience, you have to bitstream it.
- Metadata preservation: Other formats, like DTS and DD, contain metadata about the audio stream that some processors can use to better manipulate the signal. (For example, some processors, like Yamaha, use volume metadata to provide very sophisticated volume normalization and dynamic range compression, far superior to what MC does.) Decoding in MC and outputting as PCM would strip and destroy this metadata. Bitstreaming preserves it. So now your receiver can act on it, but MC cannot. It's a trade-off.
And if you're NOT using HDMI, you would use bitstreaming to send encoded 5.1 audio over an S/PDIF connection.
Atmos and multi-channel over S/PDIF would be the most common use cases by far.
Except for these very specific use cases, it is better to let MC decode the audio. Then you have access to the full MC feature set, like all the audio processing it can do: equalization, room correction, , mixing, effects, etc. MC can do none of those things to a bitstreamed signal.
That's not as simple an answer as I'd hoped, but there is a certain amount of irreducible complexity in the topic. I hope this helped.