It does not matter how many people would enjoy content that uses this feature if there isn't much content that uses it. The question being asked is if people would use it in their content.
I see what you mean, and of course it comes down to whatever factors are most important to you as the creator of NeoLemmix. However, let's say for example that only one user creates one pack using this feature. It could be that
every NeoLemmix player downloads and enjoys that one pack. The number of people that would enjoy the feature in other people's work is, therefore, a worthwhile factor as well as the number of people that would use it.
The converse of this is that
everybody could create packs using this feature, and then have 0 downloads of their content because nobody is interested in
playing packs which use it.
Granted, both of the scenarios that I've outlined here are extremes that are both very unlikely, but the point is to illustrate that end-user opinion can be just as important as creator opinion, particularly when it's a feature that not many people are likely to use in their own content.
Having said this, it seems that there are in fact a healthy enough number of people who would like to use this feature in their content anyway, which I have to say comes as a pleasant surprise!
Similarly, I'm worried that if custom sounds are tied to the theme.nxmi, this will create the situation where level authors who don't want the custom sounds (or don't want those particular custom sounds) will have to use the theme of another style -- again, messing up one-way arrow colours etc.
I can certainly see how that would be an issue.
To be fair,
if the end result of this suggestion is that the feature is implemented as a "per-pack" feature only, that would be more than good enough.
I just think that sounds, similar to background colours/sprites/mask colours/etc, are more of a 'theme' thing than a 'level pack' thing. I can see your point that tying it to the theme means that other content creators who
don't want those particular sounds have an extra step to take if they want to use that theme in their content. But, it could easily be argued that
if a creator wants to use a particular theme, it's because they
like all of the particular features of that theme.
As far as end-users are concerned - again, I can understand that if someone doesn't like a theme that a creator has used, they too would have to modify it. But, as a counterpoint to what you've said:
Surely it's better for an end-user to only have to modify a theme.nxmi just the once, and then whenever they download content using that theme,
it's already tailored to their preferences. If the feature was implemented 'per-pack', the user would have to modify each and every pack individually.
So, it's swings and roundabouts really. Neither option is necessarily better or worse than the other, but
it still seems to me that tying it to theme makes more sense and is potentially easier for everyone.