@Cmaj7 Maybe you should zip up your whole patch (without sound files) and upload it to this thread.
I think you have a problem we will not find without the whole picture.
But try this patch first....... Microsound-granular.pd ..... as it is also reading from arrays (onscreen) and softening crossovers with a short fade down and back up again (duck)....... and maybe your problem is not the speed of the duck but when the duck occurs........ it needs to start before the crossover and be at zero already at the crossover point.

So the crossover needs to be delayed slightly...... the duck needs to know it is about to happen a few milliseconds earlier...... 5ms in the above patch....... see the "granular" part in the middle.
The delay will not affect your musical timing noticeably...... 5ms is like moving 5 feet away from your speakers..... or being 5 feet from another player in the band.
David.

@Cmaj7 Do you still get clicks if you hide the array in a sub-patch?
Are you writing to the array while you are reading from it?
Do you get clicks if you pause/resume playback from ram?....... cart_wav_simple.pd
David.

@cfry That's a bit tricky as it depends on the source. Some large / huge spaces are beautiful for a slow solo voice and impossible for a faster piece with multiple instruments and a beat.
The same is true for a reverb effect. Sometimes you have use unnatural settings to achieve what you want and some natural reverb settings will feel completely wrong in a space which looks nothing like what you are hearing.
And huge real spaces (cathedrals) tend to be dull........ long RT60 with loads of low frequencies and no treble..... and I often have to add a more "bright" reverb to make the music intelligible.
As @bocanegra says you are searching for a pleasing sound.
I wouldn't use more than 24ms of pre-delay as it can sound like an echo bouncing around and confuses the ears........ better to increase the reverb time (decay) for a large space and stay with about 14-17ms of pre-delay.
David.

@Coalman In 64-bit Pd........ list_test.pd

I remember Arrays were limited to 4 million elements by [soundfiler] in 32-bit Pd but that could be increased, although it would cause problems with single precision floats...... https://forum.pdpatchrepo.info/topic/3260/long-soundfiles-in-pd/10
For 64-bit Pd an array can be bigger.... about 2 billion elements........ see your Pd/doc/1.manual/x5.htm and scroll down to Pd 0.49.
David.

@cfry This one's not too heavy....... vfreeverb~.pd
Reverb just being delays and filters you can use it twice for quad and use only very slightly different settings for each stereo pair so as to give some different focus.
As your source is central the delays should probably be the same though and maybe you should bounce each pair very subtly back into the others input.
David.

@ingox said:

My opinion is that the whole selector system is dubious. [abc( should just be a symbol.

Yes. "all" is not special...... it is just that any single atom text is treated as text unless tagged as a symbol. A great source of confusion and changing that would be good.......
But it is not special....... it will trigger a bang through [route all] just like any other single atom text...... but will not trigger [sel all].
At the moment a single atom text will [route] and produce a bang........ but for it to trigger [select] it must first be tagged as a symbol........ crazy....

The other sources of problems can be that an automatically tagged list........ a list starting with a float..... can require a [list trim] before being passed on...... although for the user there is no indication that the message has had the list tag added.
And that a list of two symbols although routed correctly loses the symbol tag for the second atom when passing through [route] because it has become a singleton and reverted to text.... ...... That issue would also be addressed by your suggestion.

So yes....... a singleton text should be a symbol....... and can still be a selector that will trigger a bang.

Consistency between objects needs addressing and the selector system would be fixed I think.
And single atom text being a symbol might be all that is required, if any single atom is treated as a selector in the same way as any first atom is currently and [list] is made consistent between atom types.

We have had this discussion previously...... https://forum.pdpatchrepo.info/topic/12794/route-list-vs-array-problem/15

But nothing can actually be done without a serious plan....... we are wasting our time except when we document these anomalies.
The elephant in the room is the millions of patches already in existence.
If we can work out how to make the changes we would like....... without any changes to those existing patches being necessary..... then we will have a proposal for the list....

That might be possible........ but I think that the current inconsistency of the triggering of a bang will put a spanner in the works.
So maybe only alternative objects..... [routeNEW] [selectNEW] etc. would be possible while continuing with the old message system..
David.

@ingox Yes...... the same rule applies as for [route] ...... a tagged list of one symbol is converted to symbol.

Time to do a thorough analysis of [select] then......... but there is a lot more to unwrap because symbols and floats can...... sort of..... be mixed as arguments if tagged.
And worse (although sometimes useful I suppose) a second atom of a list goes to the right inlet of [select] when there is only one argument........ so [list all no-go(
...... but [list all 25( works as the creation argument is a symbol so cannot be changed to a float.
David.

@ingox Yes...... top right in the screenshot above.
David.

@ingox To be auto-tagged as a list the list has to start with a float.
See [pd special_cases_of_lists] and then the part at the bottom left......... route-help.pd
You will see that as a list passes through [route] the list header is removed but if the list starts with a float it is always a list and goes through as such to the second [route].
Somewhere it is documented that a list that starts with a float is treated as a list even without the list header.

I think I added that help-file to the forum in May last year..... frustrated by [route] like @ddw_music
It's a bit messy though.
David.

@oid That's interesting. I didn't know the [key] window focus had changed.
So not universal...... it won't work in extended...... but extended is dead apparently.
David.