Loud is relative. You wouldn't hear it standing on deck of the support ship. And the hydrophones they were using for communications and pings were possibly (likely?) passed into an FFT, band-pass filtered to look for the expected frequencies of pings, and triggered on a signal spike in that range. I doubt they had somebody just listening to a straight up amplified signal straight from the hydrophones. Even if they did, someone unfamiliar with what they were hearing might not recognize it as an implosion event and attribute it to something else. And given the apparent attitude and methodology of the whole operation, it wouldn't surprise me if they didn't bother making recordings of the raw data. So it's entirely possible they wouldn't hear/notice an implosion event.
As for the banging sounds heard days later, there are a thousand possible sources of those, especially with a dozen or more other vessels on the scene. Anything from a buoy with a loose part clanking away, to some waves breaking and clapping on the surface, to a whale slapping its tail, to extraneous noise from one of the search vessels or its equipment, to other sources of sea life making a racket. It's a bit like trying to listen for the buzzing of a bee in a basketball stadium with a game going on. Any signal you get is far more likely to be something else, but when you have literally no better options, every signal, no matter how unlikely, is worth investigating. Then they report they investigated and found nothing, and in a bit of target fixation, people assume it must have been people in the sub. Understandable, but it's not necessarily the most likely/logical conclusion.
Exactly this, "imploded straightaway" doesn't explain no implosion registered AND probable banging noise heard a few days later, so it doesn't explain everything.
"Lost then imploded" explains both by adding only a small extra assumption so by occam's razor is strictly superior to imploded straightaway as far as I can see.
People are putting a lot of weight on the whole "one vessel heard a rhythmic sound while exploring". From what I've seen of these investigations the ocean is a noisy place and sometimes it gets mistaken for signal. We saw a lot of similar reports from the MH370 investigation.
My money is on simple catastrophic failure of the hull and it not being detected either because the private company is run by jokers who weren't listening for it or because they have been running around like chickens with their heads cut off because the CEO and paying customers just died and they don't want to have to report that to the family, government, media, insurance company, etc... Either explanation is plausible, but I'm slightly more inclined to go with the second simply because they were actively trying to communicate with the sub when it happened and it seems so improbable that they could miss it.
As for the banging sounds heard days later, there are a thousand possible sources of those, especially with a dozen or more other vessels on the scene. Anything from a buoy with a loose part clanking away, to some waves breaking and clapping on the surface, to a whale slapping its tail, to extraneous noise from one of the search vessels or its equipment, to other sources of sea life making a racket. It's a bit like trying to listen for the buzzing of a bee in a basketball stadium with a game going on. Any signal you get is far more likely to be something else, but when you have literally no better options, every signal, no matter how unlikely, is worth investigating. Then they report they investigated and found nothing, and in a bit of target fixation, people assume it must have been people in the sub. Understandable, but it's not necessarily the most likely/logical conclusion.