[Edited to reply the the above while I was typing the message below.. ^ Yep, sort of. The stuff picked up by the single 8 AUD portrayed in a pseudo stereo "wider and enveloping" sense rather than a discrete left/right imaging sense. It won't cancel out the audience though unless you sum the resulting stereo output to mono.]
A Mid/Side pair works fine as an AUD source mixed with the SBD. Think of it as similar to running an X/Y supercard AUD. It will definitely work, just as any other coincident AUD setup will. If you prefer it or not is a different question.
Technically, you cannot substitute the SBD feed for the Mid signal in combination with an AUD positioned Side fig-8 and expect virtual X/Y-like stereo patterns from that. The Mid and Side "sampling points" need to be fully coincident both with respect to time-of-arrival and position-in-space for that. But we aren't taping to please theoreticians nor derive textbook virtual pickup patterns. Yet in practice it can work fine, as long as it sounds good.. And In actuality, that is partly what you are doing by Mixing the SBD in with the AUD anyway. Consider that by matrixing the (primarily monophonic) SBD with the AUD you are effectively modifying the Mid component of the summed Left/Right stereo output more than its Side component. How much depends on the mix ratio of SBD to AUD. The difference is a matter of degree, and how you think about it.
Really, Mid/Side decoding is little more than mixing, with the twist of using only one channel for Side, and inverting polarity of that to one speaker versus the other. When mixing for 2 channel stereo, I sometimes use my single rear-facing ambience microphone as Side channel, "mixed" in via Mid/Side with my single forward facing directional center microphone. I typically do that to help spread the auditory center a bit so that it blends more smoothly into the stereo panorama with my other Left/Right mic channels instead of seeming to stand apart as a focused center-forward point-source. Thought of another way, it helps spread out the rear-facing microphone information out so as make the ambient contribution from it more wide and diffuse in the playback image, rather than being reproduced as a narrow monophonic central point-source of ambience competing with what the forward-facing center microphone is providing.
You can use any microphone pattern as the Side channel- an omni, a cardioid or whatever. When approaching Mid/Side in this way, "mixing to add width, diffuseness and depth" rather than as a way of achieving clear and predictable Left/Right locational imaging, it becomes advantageous to limit as much as possible how much Mid information gets into the Side channel. A sideways-facing figure-8 does that somewhat, yet still picks up a lot of information from the front as well to either side of the central null. Here is one of the basic things I've come to realize- Side channel used this way, and Ambience channels in general, become far more useful when they exclude the direct sound from the stage and PA to the greatest degree possible, as long as you already have sufficient direct sound. But that's somewhat different than AUD + SBD when both are focused on picking up the band on stage to the exclusion of all else. If the SBD is not a complete mix, you'll need the AUD to provide what isn't in the SBD. If the SBD is a complete and good sounding (if dry) mix, all you really want from the AUD is good diffuse ambience and audience reaction. An omni Mid worked well for ambience in Healy's Ultramix because it was placed back at the SBD, far enough from the PA that it was primarily in the diffuse field. Mid/Side with an omni Mid decodes to sideways-facing opposing cardioids patterns, remaining omnidirectional in an overall combined sense.
There was a similar technique to Healy's from the late 80's / early 90's which I've heard about, which was used by some folks recording material for matrixed Dolby surround. Analog matrixed Dolby surround was a 4 channel format- LRCS = Left,Center,Right,Surround.. with the surround channel being monophonic (essentially Side channel in the sense I describe using it above). This technique for recording the S channel was to place a pair of sideways-facing coincident cardioids well back in the room, and to flip polarity of one of them before summing them and feeding that into the S input of the matrix encoder. If you think about that, it is essentially the same as using a single sideways-facing figure-8.
All good, but..