Just as before, I've tried to keep the code as lean as I reasonably could, and performance seems to be quite good. That being said though, I did encounter a noticeable issue with the audio implementation in the form of playback latency. Since I was already using SFML for the video, I chose to use it for the audio as well, and I'm not entirely sure, but it seems to me at least like it may be what's causing the latency issue. To put it short, even when interfacing with the SFML SoundStream with the least amount of overhead and in the most direct way possible (even going so far as to temporarily remove any considerations for variable length sample size per read), I still found that the audio was slowly falling out of sync with the video no matter what I did. Mind you, the video plays back quite fast, almost having no perceivable latency to me at least, which is what leads me to believe it might be the SFML SoundStream implementation itself that's causing the majority of the audio latency.
My solution to get around this without interfering with the consistency of the audio too much was to just introduce a "tolerance" level for how much the audio could get out of sync and drop a read's worth of samples when it's exceeded. This does mean that there is still some amount of audio latency and that there's also some audio loss now, but I seem to have manged to balance it to where the audio isn't too out of sync but also the frequency of dropped samples isn't too often either. Based on what I've observed so far, the audio performance is still very reasonable and the software is more than usable like this. I really would like to have complete and consistent audio though, so I will be looking into ways that I might be able to improve this (or hopefully just fix it outright) and will continue to post about my progress.
Anyway, thanks again to anyone whose taken an interest in this. I may have made this software out of my own need, but I'm glad if even just one other person's getting something out of it!