I hooked up the audio feed from a radio that picks up the local police, etc. But to do it, I had to come up with a way to create a feed that BI recognizes as a legitimate data stream. It turns out that while you can buy devices that are made to take an audio stream, digitize it, and then encode it into a type of data stream that BI wants to see. But they're ridiculously expensive.
So I took one of the lovely Reolink cameras I have which have microphones and thus audio, disconnected its internal microphone and connected to the input made for the mic. That input has power to bias the electret element of their mic, so I isolated it with a capacitor to block that DC. Then I played with the setup with a signal generator to find out what range of signal levels are acceptable. Then I installed a small audio coupling transformer, a resistive divider, and then the DC-blocking capacitor to create an input that will accept normal "line level" audio. The problem is that the camera has its own audio compression built in, and I already have a professional audio compressor as part of the "police radio" interface that we use. So the signal gets "double compressed".
However, it works reasonably well. The beauty of it is that because this audio gets recorded the same way as any other video/audio stream would. So you can use the same settings, and it will trigger, record, etc., with the same settings available.
So I set it to record a few seconds before the start of any detected audio, and keep recording for ten seconds after the audio ends. And that means that if some event happens, I can go back through the recordings just as I would when looking for video of some event. And then, play it back and find out what really happened.
Further, because it doesn't record except when triggered, as I play back, unless there's a lot going on, the playback zips past the periods of "dead air", so I can review an hour of time in a lot less time. So it's nice.
But what I'd really like to find is a better audio digitizer gadget that creates the appropriately encoded data stream, but does NOT compress the audio. But that's just me because I already have far better compression on this audio than what I'd likely get from some camera or dedicated audio "capture" device. For most folks, if you just have a scanner, you may well want the compression that's built into a typical inexpensive camera.
Police scanner audio really benefits from audio compression because a lot of the transceivers used have no compression on their mic inputs, and different officers either sit on the mic and mumble, or alternately hold the mic right up to their lips and scream into it. So to listen to pretty much any police, fire, etc., channel, you need to turn it up loud enough to hear the soft-talking mumblers who hold the mic between their legs, and thus when someone speaks normally or yells, you have your ears burned off. Before you dare connect the audio from a "scanner" to, perhaps, a stereo system, you MUST compress it or risk having your speakers burned to a crisp.
But the compression in my particular camera, when added to the already compressed audio that I provide to it, is a bit too much. But I digress.
Now I know that this does not answer the OP's question, because they somehow already have a digital stream of scanner audio that they wanted to bring into
Blue Iris. But keep this in mind if you want to have BI record an audio stream. A cheap camera with microphone will likely cost you a LOT less than any dedicated audio-only gadget. At least the ones I found. They're probably sold to be very high-fi, for professional audio applications, and thus they really charge a lot for them. But for what we likely want, the audio encoding from an inexpensive camera is good enough.
So that's one way to get it done, and then Blue Iris just thinks you're connecting another camera - because you are.
I have the camera mounted so that it doesn't see anything interesting. But you could either aim it somewhere that you want to view only when people are talking on the scanner (LOL), or just block it's view entirely. The video from it is unimportant. It's the audio stream that I'm interested in using.
Crude, but effective.