I'm trying to set up windowing in the sampler from my earlier post: https://forum.pdpatchrepo.info/topic/12699/problems-with-ableton-link-sampler

I've tried to incorporate the example of windowing from the load bang book, section 3.4.2.6: http://www.pd-tutorial.com/english/ch03s04.html

I've tried various changes but nothing seems to work. In fact introducing the windowing table has made it so that there is a click when the sample isn't playing, and the dip in volume caused by the windowing seems more conspicuous than the click from the sampler without windowing.

I'm also confused by the crown table in the example from pd-tutorial. I don't understand how anything is written to the crown table, is it that I'm supposed to by drawing it by hand? That doesn't seem likely to me but it's as close as I've come to figuring out where it got its data.

@paste Drawing the crown by hand will not be easy..... but that can be automated.
[tabread4~] does 4-point interpolation..... so the arrays should have 3 more samples so that they wrap correctly........ https://forum.pdpatchrepo.info/topic/11208/why-does-floss-manual-give-sinesum-a-value-of-2051
You can get a better understanding working through the Pd/doc/3.audio.examples from B01.wavetables.pd onwards.

To get you going again....... @weightless posted [sample-smooth] here....... https://forum.pdpatchrepo.info/topic/10713/smoothly-change-start-and-end-of-looped-sample
which automates the drawing of the crown and could provide a complete solution.
I think you can ignore the "pitch variation" problem in that post as you are not varying the frequency of the [phasor~]
Inside [pd makeCrown] you will see that the right inlet allows you to make steeper ramps.
Don't use the [line~] into [*~] parts post [phasor~] which will cause pitch variations as discussed in that thread. You could use that method post [tabread4~] to do the same as the crown though.
Remember that [table] should always be hidden in a subpatch.
David.

@whale-av I'm looking through the examples starting at B01 and working onward through B, as well as the sample-smooth patch you linked and I think I'm making progress but I'm still a little confused.

So with my sampler, I'm adjusting the loop length by sending different beat lengths to [s beats]. From what I've been looking at it seems that I should be changing the table size to fit the size of the loop I'm making, +3 for tabread4, is that correct? From what I understand I had just made the table long enough to accommodate all lengths I would be sending through [s beats] and only reading the part of the table I was using rather than going through the whole table, which seems like it might have been one of my problems.

Another thing that is confusing me is the concept of making the last 3 points of a table copies of the first 3 points. I don't understand how this is accomplished, is this something that is done in the sample-smooth patch?

As for the example from pd-tutorial.com, one of the things that was throwing me off is that they don't show a mechanism for writing to the crown table, which is what led me towards the idea of drawing the table (I did try and it was difficult lol). I feel pretty confident now that there wasn't anything in that patch that gave the crown table its parameters but I wanted to double check with you just to make sure. It just seems awkward to have a tutorial for windowing that doesn't actually show how the windowing table was made so I'm second guessing my assessment and wondering if I'm missing something.

@paste So it was @Maelstorm that patched a crown generator (originally?) ........ https://forum.pdpatchrepo.info/topic/5790/windowing
Yes, there is nothing in floss about how to do it......... which is a major omission.
You can probably ignore the 3 extra table indexes for interpolation in this case as the crown will have taken the output down close to zero at the wrap point and so any glitch should be inaudible (a longer crown table will suffer less from interpolation effects too...... see the link above).
I have not thought about how to copy the first 3 values to the last 3 in a live record situation.
It is easy when drawing a wave table.... as in the examples.
David.

@paste I think I may have used @Maelstorm's windowing technique that @whale-av linked to above — it's been a long time so I can't remember for sure, but it does sound like the same approach (amplified/clipped cosine as crown window), which I find works pretty well!

This was the looper patch I used it for:
https://forum.pdpatchrepo.info/topic/11353/tabread4-example-or-alternative/3
The full patch probably won't work for you as-is because the iem_dp library can be hard to get working, but you could certainly check out the [pd window] subpatch & modify it for your own sampler.