[laughs]

But so will TH- I think there's, there's a, there's an, um, there's an element of audio testing where THD is perfectly appropriate, where pure tones are perfectly appropriate, and those types of applications typically revolve around systems that are completely linear in behavior.

The thing that is, you know, a little bit tricky is when devices are highly nonlinear or involve AI-based, neural network-based type processing on the back end where it's very hard for us to even send the stimulus signal through the system.

So one thing is a loudspeaker test, a driver, that's a pretty straightforward test.

We have good knowledge and good understanding of what happens when we send a one-kilohertz tone into it and the behaviors that we're seeing as a result of that one-kilohertz tone.

Another thing is if you have a headset, a Bluetooth loudspeaker, a microphone array with some noise processing on the back end, what happens to the signal we're sending into that type of device? That's a little bit of a challenge.

And so as I'm, you know, as, uh, as, as I come into this, the two primary examples that we are experiencing are codec issues, so specifically telecommunications codecs when we're transcoding signals and, uh, sending, sending them over the network.

So the NLM, uh, what do you get out of it? Is it the traditional measurement data, like a frequency response, uh, some distortion metric?