How to set up additive blending on a surface shader?

But this doesn't answer the question. This would require NOT using a surface shader, unless I misunderstand something?

Actually, it exactly answers the question. x4000 wanted to know how to set up a surface shader to use additive blending, and the way to do it is to use the "Blend SrcAlpha One" tag as with a non-CGPROGRA$$anonymous$$ shader, and leave out the "alpha", "decal:add", and "decal:blend" tags from the CGPROGRA$$anonymous$$. In the post, x4000 mentions that the "Blend" tag does nothing when writing a surface shader, and my explanation makes it do something - thereby allowing x4000 to do exactly what they want.

Are you wondering about something else that my answer doesn't fix?

Hm. I'm new to this, sorry. $$anonymous$$y understanding is this (and correct me where I'm wrong): if you're using a surface shader, then you don't include a Pass { } block. But if you want to use the old-style Blend settings you have to put them in a Pass.

To test, I have this shader that works already - it's an emissive shader that uses both texture alpha and alpha based on normal falloff.

 Shader "Falloff/GhostFlat" {
     Properties {
         _Color ("$$anonymous$$ain Color", Color) = (1,1,1,1)
         _$$anonymous$$ainTex ("Base (RGB) Trans (A)", 2D) = "white" {}
         _RimPower ("Rim Exponent", Range(0.5,25.0)) = 3.0
         _BaseAlpha ("Base Alpha", Range(0.0,1.0)) = 1.0
     }
 
 
 
     SubShader {
         Tags {"Queue"="Transparent" "IgnoreProjector"="True" "RenderType"="Transparent"}
 
         CGPROGRA$$anonymous$$
         #pragma surface surf Lambert alpha
         
 
         sampler2D _$$anonymous$$ainTex;
         fixed4 _Color;
         float _RimPower;
         float _BaseAlpha;
 
         struct Input {
             float2 uv_$$anonymous$$ainTex;
             float3 viewDir;
         };
 
         void surf (Input IN, inout SurfaceOutput o) {
             fixed4 c = tex2D(_$$anonymous$$ainTex, IN.uv_$$anonymous$$ainTex) * _Color;
             half rim = saturate(dot (normalize(IN.viewDir), o.Normal));
             //o.Albedo = c.rgb;        //No albedo, only emission
             o.Emission = c.rgb;
             o.Alpha =  pow (rim, _RimPower) * _BaseAlpha * c.a;
         }
         ENDCG
     }
     FallBack "Self-Illu$$anonymous$$/Diffuse"
 }

But if I wanted to make it additive as well, how would I do that? I tried taking the "alpha" out of the #pragma line, and adding this under the Tags {} block: Pass { Blend One One } but that caused a bunch of errors. So how can you combine these?

To the best of my knowledge, there is no such thing as "additive as well". There is only additive or not.

The Blend mode does not need to be in a Pass of its own, you can just put it under the SubShader block.

A surface shader basically creates its own Pass block, so when you put the Blend mode in a Pass block, it's a completely separate pass from the CGPROGRA$$anonymous$$ section, and therefore does not affect it. When you do not specify a blend pragma (using the alpha or decal options), it would appear that the shader defaults to whatever was specified in the subshader. As I understand it, this is how it works with fixed function Pass blocks as well - they default to the SubShader Blend mode, which is opaque (Zero One, I believe) when unspecified.