For some reason the first screenshot looks better when viewed on my iPhone, probably due to whatever scaling/interpolation the iPhone is doing (everything looks better in screenshots).
Is there some additional interpolation or something we can throw on top, maybe?
NTSC shader passes are resolution sensitive and it’s a problem if the resolution get’s increased, because artifacting effect becomes negligible. Maybe you can throw in a g-sharp resampler horizontal pass.
https://mega.nz/file/otohlYYQ#EH4bNecPhCkhoy78S36Wd1YwRULpuxms2b0WLtddLj8
Meanwhile, as promised, i added NTSC blending modes to the NTSC composition and some of them can blend dithering very well. Be sure to check it out.
New Release Version (2021-06-24-r2):
Notable changes:
Download link:
https://mega.nz/file/4kBCFZDD#EC2TqnnTbD7svAMOL2F_YndPlLZEuraqLqYWqVgOxvU
Screenie: new blending mode, composite:
Blending modes doesn’t seem to do anything on my end
Is it because of the core, or because the game is not 2D?
Maybe try a situation where dithering is present? 
Here is a good example:
Vanilla:
Blending 0.0
Blending 2.0
I have a felling I’m making a fool out of myself now, but dithering is present in my example. I was under the impression that blending modes will produce the image that looks closer to the older version of the shader:
Maybe it’s a PSX thing and downsampling has to be used, I was just checking if it’s working as intended.
Setting the Prescale-X Factor to 2.0 does this trick, i like to mention it because it’s not this obvious. Another yaay goes for bolder uses of masks and ‘lighting’ tricks, makes the ntsc shader much more interesting imo.
I already had this on, but I’ve found out what was the issue: g-sharp_resampler_horizontal breaks the effect, replaced it with stock pass and blending modes now work as intended 
Same pic you can use with the image viewer:
@guest.r I noticed the interlaced modes being more dim than I remember from previous releases. Haven’t really done some testing versus previous releases, but within a game the interlaced content is now quite a bit dimmer than the non-interlaced content. I don’t remember it being that way some time ago when I last tested. Maybe I missed some changes in between, but just wanted to verify with you whether the interlaced mode brightness is still working as intended.
It’s balanced to a degree, but you can decrease the Interlacing Scanline Effect to get a brighter appearance with mixed contents in interlaced mode. Since many combinations of masks and scanline settings can be used the brightness appearance may vary when switching. In general it’s gameplay oriented since an user usually spends most of the time either on normal mode or on interlaced mode. Mixed situations are usually a niche where the focus should be on the non-interlaced mode.
All these are from guest advanced which is really cool. (The rainbowing is way less prominent during gameplay.) The composite is interesting because I can keep a phospur/mask sharpness instead of just blurring the entire image. It feels more true to life, but I don’t have a real life example to compare it to. Also I’m using GTU pass1 and pass2 with SVideo and I’m not sure if that’s duplicating color smearing the advanced shader already does.
I found the issue is only with the hires version of the shader, with the normal guest-advanced the brightness of the interlace mode is normal (good). The same picture in guest-advanced (using default settings) produces a dim picture.
Will post a screenshot comparison later.
Edit below a comparison, which shows the dim interlace issue with the -hires- version of the shader.
Edit 2 added comparison of just a plain white 640x480 image. For some reason guest advanced -hires- darkens / dims the image much more in interlaced mode
source:
guest-advanced:
guest-advanced-hires:
Raw image 640x480:
default guest-advanced preset:
default guest-advanced-hires preset:
Gee, thanks, that really helped to pin the issue. The update of the interlacing code with improved downsampling functionalities went fine, but the hires version missed an update on a portion of the code. Is fixed now and available for download at ‘usual location’. 
If Alex Kid can go to Shiniboi World, then Shiniboi can go to the disco. I imagine the Genesis soundtrack for that would be awesome.
Can anyone share all the shader pack in a zip file for example? I’m having problems with the last guest.r preset versions, and I don’t know where is the problem… Thx
Use the error log fam, it’ll make everyone’s life easier.
Thanks for the interlace fix in the hires shader. Have to say the shader looks absolutely marvelous all around 
For personal preference I would want to tweak gamma in the interlace mode only a bit. Do you use a separate gamma or adjustment for interlace currently? If so could that be made configurable in the interlace section? Or otherwise maybe point me to the relevant code part such that I can tinker a bit myself.
This is just nitpicking, but in a future update would there be a chance to have a switch to turn off the mask when interlace is enabled? It’s a bit nitpicking I know but it would be nice if it would be possible as currently the trinitron mask (setting 0.00) results visually in small vertical stripes when the screen is interlace. (no biggie that’s for sure)
I’m permanently using the 0.00 mask in my setup, just like your preset and think it’s lovely even for gameplaying sessions. Out of interest, what is currently your next best favourite to this mask ? Is it the one you used in this shot https://imgbb.com/hs0q49q ?
Anyway big thumbs up for the current state of the shader, feels really polished now, Can play for hours and not get tired of it 
Isn’t mask (-1) still a thing? Which was no mask iirc
Thanks, it’s nitpickiness which made it better i guess.
I try to avoid too many special circumstances, which, by auto, are not the same as an user would expect or too much sub-settings in general. I believe for interlaced content it should be enough to lower the input/output gamma like 20% and it could look better, closer to ‘linear interpolation’ look.
If you want it to tinker a bit, then it’s starts by adding a parameter in ‘linearize pass’ and a good coding compromise would be for output gamma to be the same as input gamma.
Good point, in fact the new interlacing approach allows it to fix the issue, will be added soon i guess.
Being an 1440p user, mask 0 alone is a bit too fine so i like to add something stronger. Lottes masks or ordinary masks with slotmask is what i like the most.
