Latest update Musescore 2.1 on Ubuntu

You reports a bug, you don't want to fix it, so you need to help those that are willing to fix it, simple as that. You could have testes 2.1 more than a year ago, there was an RC prior to release. And even before that numerous users helped in test it.
If you want to report a bug, by all means do so, but please explain exactly what that bug is, what you do, what you expect and what you get instead, and how to reproduce.
If you don't want to help, fine, but then please stop complaining.

The default soundfont did not change, not significantly, nothing that explains why you believe to sense a decrease in quality, it did change some drumset assignment and did fix a few pitches for some instruments that had bad samples or where out of tune.
There was indeed a regression reg. Tenor drums, and about some bass synthesizers, the latter not related to the soundfont itself at all, the former has a valid workaround of using an adjusted drm setting.
ANother sound related problem had been fixed the other day, again not related to the soundfont itself, but to fluid, the synthesizer, yet another is still in the works (but very unlikely yours)
I'm not aware of any changes to cymbals or snares.

As mentioned, the soundfont barely changed at all, and the few changes that were made were mostly improvements to specific notes on specific instruments that were out of tune or whatever. The changes you are perceiving are almost certainly more due to the code than the soundfont, and those changes did fix problems, so that for very many scores, the sound is improved. But indeed, there are some small subset of scores that sound worse because they use particular instruments in particular ways that these code changes did not handle well. So for that small subset of scores, indeed, the perception might be that the sound has gotten worse. It's unfortunate that at no time during the extensive test phase for 2.1 did someone who had a score that would trigger these problems try it out for themselves, because if they had and had reported the problem, it could have been fixed. But as it is, the very many people who volunteer their time to work on these releases had only their own scores to test with, not yours, and so no one was able to detect the problem.

The lesson here is, open source software development really is a community effort, and it depends on the community of users to help test as much as it depends on the community of programmers to do the development. That's the way open source software works. We would welcome your participation in the 2.2 development process, I strongly encourage you to try a build of 2.2 to see if the fixes already made solve all your problems or not. If they don't, again, please help us help you by sharing a score that demonstrates a problem.

I can only assure youwant to see this problem fixed. I assume you do as well. So all I can do is repeat myself - if you want to see a problem fixed, we need to be able to reproduce it. Otherwise, we probably can't fix it. A verbal description of the problem can be enough to give us an idea of where to look, and since others have posted scores, we've been able to fix several problems. Maybe those are the same as the ones you are seeing, but the only way to be sure we've fixed each of the specific problems you are seeing is to have a score or scores that reproduces the specific problem.

To borrow your analogy: if a CD player plays 1,000,000 CD's with no difficulty, but six users say they have a problem, that does suggest one or more faults in the CD player indeed. But the CD manufacturer probably won't be able to identify or fix any of those faults without a copy of the CD that triggers the problem. If five of those six users choose to share CD's that trigger the problems they are seeing, and the CD manufacturer is able to identify and fix three problems that turn out to the ones responsible for all of the issues reported by those five users, great. But there still be no way of knowing if the problems experienced by that sixth user - the one who refused to share his CD - were all covered by those three fixes, or if one of the problems he is experiencing is perhaps unique to his score. Either the fixes that the manufacturer makes address his problems or they don't - no way to tell unless he shares his CD.

Up to you on whether you want to take the chance that the fixes already made fix your problems or not. If they do, you will be happy when 2.2 comes out and your problems are over. But if not, you'll have only yourself to blame.

I was going to say this is like that thing people will tell you where if you don't vote in an election, you have no right to complain about the result. But actually, the situation here is more direct. With an election, if you vote, the election might still go against your wish - there is no guarantee that participating in the process will actually lead to the result you want. It's really barely better than lottery odds of it. In the case at hand, though, there is excellent chance that helping us out here by sharing a score will allow us to verify if the problems you personally are experiencing are fixed already or not - and if not, your sharing a score will be exactly what makes the difference in our being able to fix it. it's a no-lose, huge-win scenario.

Your choice.

If you believe we're asking for a sample score without reason and you don't want to provide one nor provide exact steps on how to reproduce the perceived defect, this is your problem and yours only, and I really can't fathom how this simple fact is eluding you.