Or maybe, instead of constant 1.5 for all interlaced content, something like this: (assuming 1.5 is our desired value for horizontal resolution of 640 px and we don’t want to go over 2.0)
float icoeff = min(global.OriginalSize.x / (640 / 1.5), 2.0);
float fdivider = max(min(mix(prescalex.x, prescalex.y,TATE), 2.0), icoeff*float(interb));
It should function indeed. Meanwhile you can use such a solution if it works for you. But something similar will be included in the next release.
Not removed, it’s now automated, based on the pre-scale ratio. Works really well, no need for manual tuning tbh.
Feedback:
When using slotwidth 6.0, mask size 2.0 with shadowmask 2 and 6, instead having black bars with that pattern RRGGBB you get BRRGGB.
Mask looks like number 2 or 5 on background with only one color/shade. 
That’s really weird and is fixed in newest versions:
Perhaps, if someone else confirms the discrepancy also, i’ll have to take a closer look.
Settings:
shadowMask = "6.000000"
maskstr = "1.000000"
masksize = "2.000000"
slotmask = "1.000000"
slotmask1 = "1.000000"
slotwidth = "6.000000"
double_slot = "2.000000"Can you test this version? Only crt-guest-advanced version is altered…
https://mega.nz/file/wlIUBTTY#qp5heEoG4Z1Q8t0u6PQLMUbkjgfkGmLBq7pemn7T22A
Thanks for testing. Guess i needed to rework the slotmask function. Discrepancies happened before also…
Here is a new test version if you guys can test it…
https://mega.nz/file/5hhSCSYC#j5_hn_IPTFI5aXqYkQXU0e7Xw-ccq5bieRGr44t_TlU
Thanks for testing it again. These rounding errors and discrepancies are a real pita sometimes… Here is a version which works for me with different testings…
https://mega.nz/file/l9B0mQJJ#3OTmdxC8yEo-Vin0bi9S5Nq-LA-Njxnji_m-Bh3Q2jU
Yes this time it works, looks correct on my side. 
Also i suggest shifting the black bars 1 pixel down with double_slot 3.0
Everything will be centered and look better (black bar in center of scanline).
Hey, that’s good news. I’ll try to make a regular release tomorrow. Regarding that vertical shift, it’s really situational, but i’ll give it further thoughts. Some testings were already made to mitigate moire, but it’s not this trivial.
I was curious if that method kokoko3k was using to somewhat align the scanlines to the mask would be applicable here, he got some good results in his usecase
Probably not with the pass structure and code sequence, as i understand masks should be calculated before scanlines for this.
I have nevertheless implemented some slotmask mitigation code for better appearance, but it’s a trade-off regarding slotmask strength in scanlines…
New Release Version (2023-02-25-r1):
Notable changes:
Download link:
https://mega.nz/file/QoZlDJDB#zau7j7i5qExJ8h9bIvvUoNSbS1a074ZmK-ixuAWt77U
This is the most interesting part of the update for me. From what I understand it sounds like you can reduce moire by lowering slotmask strength, is that how I should interpret the “trade off”? As you and everyone that uses my presets know I use slot mask strengths at max 100 percent no exceptions. Please explain a little more about that part of the update, I’m not home at the moment to test this out myself.
Also a bit curious about this as well.
Slotmask presence is mitigated in scanlines for a more even appearance, also with darker pixels. But it’s purely optional and strength can be regulated. The trade-off is you can’t have full slotmask strenght on already dark sections with this.
Here is a comparison with a slotmask preset:
normal:
slotmask mitigation:
With just regulating slotmask strength you can’t mitigate the slotmask strength in scanlines, that’s new here.
No worries here, you can read the discussion a couple of posts above.
