Manually line up an early transient spike that is present is both recordings then using time stretch drag the audience recording so it lines up to the end of the soundboard recording. If the audience recording is slightly late by tiny amounts it should be fine since the audience recording is providing an ambience missing from the soundboard recording.
If you need more precision, you can places cuts between each song in the audience recording, then manually line up the starts of each song and then time stretch drag each audience recording song to line up with the end of each soundboard song. It’s a lot more work, but definitely more precise.