It’d be really great if we knew what the PPUs were doing. Is there a big difference in gamma between SNES and NES? There are no settings that look right for both systems. Seems like a 0.20 - 0.30 difference in output gamma is required. I think I recall reading somewhere that the SNES had a very saturated output, so it kinda makes sense that a higher output gamma would be needed, if true.
2 Likes
FYI one of the NES presets uses 1.80 gamma. I wonder if, other than using a power-pc Macintosh, there is any reason for this 
\presets\nes-color-decoder+colorimetry+pixellate
I feel like we haven’t paid enough attention to NES color palettes on this forum. Then again it is a can of worms!
3 Likes
Is this PAL or NTSC?
This is from a PAL Wii. Note though that I didn’t turn the sharpness control all the way down for these photos, that would make a difference for composite, though not the extent it would suddenly mimic Genesis/Megadrive output.
Wasn’t aware of this until I stumbled over it recently. At least some Sony WEGAs have adjustment settings for their comb and notch filters in service mode.
I also read that on some sets you can even disable interlacing, though probably not on the WEGAs.
4 Likes
This brings back memories of me stumbling my way through the service menu 
It’s the AEG CTX 480 preset on my shader (Sony Megatron) underneath the shaders/HDR folder. Simply get the latest version of RetroArch and then go to online update and update slang shaders (I think you need to use D3D11, d3d12 or vulkan). The shader does require a very bright screen - normally HDR capable but supports bright SDR only displays like laptops.
3 Likes
@MajorPainTheCactus Oh sure, I’m familiar with your shader. But I was talking about what settings you’re using to achieve that NTSC look that’s a bit sharper than the NTSC-Adpative.
So I decided to attempt to go down the rabbit hole of trying to replicate the NES composite output. Unfortunately I no longer have access to real hardware, but when I last did, it was a front loader NTSC NES hooked up to a Sony 27FS100. I recall despite it having access to a 3D comb filter IIRC, the grunginess of the composite output shone through regardless, and it is this I am now trying to achieve.
I found a pretty good close-up picture that I think encapsulates rather well how I remember the NES looking:
I got pretty close using CRT-Guest-Advanced-NTSC, but I’m still playing with it. For sure it requires a decent bit of fringing with no field merging, as I definitely recall the image being flickery, but without much artifacting. I’m torn on the blend mode, though. It almost looks correct with it on, but then things like the brown of the bricks bleeding into the black gap between the bricks, whereas they appear almost wholly black in the picture. I’ll have to mess with it some.
4 Likes
Did you achieve that with any kind of sharpening? Where is the fringing coming from?
2 Likes
That’s not my shot, I got it off some forum. It’s an actual CRT. I’ll see about posting an approximation using guest-advanced-ntsc in a little while.
1 Like
Sure, I know the photo is not yours.
Anxiously awaiting it. 
I want to see how you guys are making NTSC sharper without ruining dithering.
2 Likes
That’s exactly what I was trying to achieve with my presets, I think I’ve gotten close enough while still being able to get the Sonic waterfall effect with rainbow banding
1 Like
Here is the current iteration of my NTSC preset, the goal is to be as sharp as possible while blending dithering completely, minimizing sharpening artifacts and maintaining adequate color and brightness with 100% mask strength.
I see very bad sharpening artifacts in a lot of NTSC presets that are attempting to be sharp. This is just as bad as being too blurry, IMO. It’s very difficult to strike the right balance.
4 Likes
That’s looking really nice and sharp! Thanks for sharing. I’ll play around with it.
1 Like
Here’s my take:
Load Full Resolution, zoom in or view full screen.
2 Likes
That mask kinda makes it hard to tell what’s what, can you post settings with a higher TVL mask? I think that maybe lighter scanlines are needed for improved phosphor shape and maybe an adjustment to saturation is needed, but that’s not what we’re focused on right now.
1 Like