KKS - Posibility of making a 180 screenshot?

[jasonbarthes]

Hello! I've been using 360 screenshot to generate vids. However the file size and VRAM requirement makes the generation risky, which means that my PC always crashed during the process.
Recently I saw someone shared a modified dll for KK's screencap, which is able to capture 180 screenshots:(https://www.iwara.tv/post/20hnPDq0emf58K)
I tried that in KK and the alising was way better than 360, under the same resolution setting. Because my data are compatible with KKS only, I want to make a kks version plugin. Do you have any idea about how to turn the KKS plugin into a 180 version? I can code and contribute if you don't mind.

[ManlyMarco]

If someone makes a PR that adds this feature then by all means I will merge it.

I made an attempt at merging the changes made in that modified dll here https://github.com/IllusionMods/BepisPlugins/tree/3d-180 but got stuck on something, I don't remember what exactly was the issue. At the moment I don't plan on working on this any further.

[mellona0821]

@jasonbarthes
I’m not sure if it’s okay to write this here, but since I received a request, I’ll share it here
I was considering preparing a PR myself, but it seems there’s more to take care of than I expected.
Instead, I’ll just share the parts I modified here.
It’s been quite a while, so I don’t remember everything exactly, but I’ll do my best to write down as much as I can.

First, to change the 360-degree view to 180 degrees and apply tilt, I modified the shader at the following link:
https://github.com/mellona0821/BepisPlugins/blob/master/src/Core_Screencap/shader_source/EquirectangularConverter.shader

Shader "Hidden/I360CubemapToEquirectangular"
{
    Properties
    {
        _MainTex ("Cubemap (RGB)", CUBE) = "" {}
        _CameraRotationMatrix("Camera Rotation Matrix", Matrix) = (1,0,0,0,0,1,0,0,0,0,1,0,0,0,0,1)
    }

    Subshader
    {
        Pass
        {
            ZTest Always Cull Off ZWrite Off

            CGPROGRAM
            #pragma vertex vert
            #pragma fragment frag
            #pragma fragmentoption ARB_precision_hint_fastest
            #include "UnityCG.cginc"

            #define PI 3.141592653589793
            #define TWOPI 6.283185307179587

            struct v2f
            {
                float4 pos : POSITION;
                float2 uv : TEXCOORD0;
            };

            samplerCUBE _MainTex;
            float4x4 _CameraRotationMatrix;

            v2f vert(appdata_img v)
            {
                v2f o;
                o.pos = UnityObjectToClipPos(v.vertex);
                o.uv = v.texcoord.xy * float2(TWOPI, PI);
                return o;
            }

            fixed4 frag(v2f i) : COLOR 
            {
                float theta = i.uv.y;
                float phi = i.uv.x;

                float3 unit = mul(_CameraRotationMatrix, float4(
                    sin(phi) * sin(theta) * -1,
                    cos(theta) * -1,
                    cos(phi) * sin(theta) * -1, 
                    0
                )).xyz;

                return texCUBE(_MainTex, unit);
            }
            ENDCG
        }
    }
    Fallback Off
}

To match the modified shader, I also updated the code at the following link accordingly:
https://github.com/mellona0821/BepisPlugins/blob/master/src/Core_Screencap/Renderers/I360Render.cs

public static Texture2D CaptureTex(int width = 1024, Camera renderCam = null, bool faceCameraDirection = true)
{
	if (renderCam == null)
	{
		renderCam = Camera.main;
		if (renderCam == null)
		{
			throw new Exception("No camera detected");
		}
	}
	RenderTexture targetTexture = renderCam.targetTexture;
	int num = Mathf.Min(Mathf.NextPowerOfTwo(width), 20000);
	RenderTexture renderTexture = null;
	RenderTexture renderTexture2 = null;
	Texture2D result;
	try
	{
		renderTexture = RenderTexture.GetTemporary(num, num, 0, 0);
		renderTexture.dimension = 4;
		renderTexture2 = RenderTexture.GetTemporary(num, num, 0, 0);
		renderTexture2.dimension = 2;
		if (!renderCam.RenderToCubemap(renderTexture))
		{
			throw new Exception("Rendering to cubemap is not supported on device/platform");
		}
		I360Render.equirectangularConverter.SetFloat(I360Render.paddingX, faceCameraDirection ? (renderCam.transform.eulerAngles.y / 360f) : 0f);
		Matrix4x4 localToWorldMatrix = renderCam.transform.localToWorldMatrix;
		I360Render.equirectangularConverter.SetMatrix(I360Render.cameraRotationMatrixID, localToWorldMatrix);
		Graphics.Blit(renderTexture, renderTexture2, I360Render.equirectangularConverter);
		result = I360Render.RtToT2D(renderTexture2);
	}
	finally
	{
		renderCam.targetTexture = targetTexture;
		if (renderTexture != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture);
		}
		if (renderTexture2 != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture2);
		}
	}
	return result;
}

Even though it worked fine with just these changes, I analyzed the code to improve performance and found that the part where the two images (for the left and right eye) are combined was quite inefficient.
So, I combined the capturing and image-stitching code into a single method.

public static Texture2D CaptureStereoTex(int width = 1024, Camera renderCam = null, bool faceCameraDirection = true, float eyeSeparation = 0.6f, bool flipEyes = false)
{
	if (renderCam == null)
	{
		renderCam = Camera.main;
		if (renderCam == null)
		{
			throw new Exception("No camera detected");
		}
	}
	RenderTexture targetTexture = renderCam.targetTexture;
	int num = Mathf.Min(Mathf.NextPowerOfTwo(width), 20000);
	RenderTexture renderTexture = null;
	RenderTexture renderTexture2 = null;
	RenderTexture renderTexture3 = null;
	RenderTexture renderTexture4 = null;
	Texture2D result;
	try
	{
		renderTexture = RenderTexture.GetTemporary(num, num, 0, 0);
		renderTexture.dimension = 4;
		renderTexture2 = RenderTexture.GetTemporary(num, num, 0, 0);
		renderTexture2.dimension = 2;
		renderTexture3 = RenderTexture.GetTemporary(num, num, 0, 0);
		renderTexture4 = RenderTexture.GetTemporary(num, num, 0, 0);
		I360Render.equirectangularConverter.SetFloat(I360Render.paddingX, faceCameraDirection ? (renderCam.transform.eulerAngles.y / 360f) : 0f);
		I360Render.equirectangularConverter.SetMatrix(I360Render.cameraRotationMatrixID, renderCam.transform.localToWorldMatrix);
		renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
		if (!renderCam.RenderToCubemap(renderTexture))
		{
			throw new Exception("Rendering to cubemap is not supported on device/platform");
		}
		Graphics.Blit(renderTexture, renderTexture4, I360Render.equirectangularConverter);
		renderCam.transform.position -= renderCam.transform.right * eyeSeparation;
		if (!renderCam.RenderToCubemap(renderTexture))
		{
			throw new Exception("Rendering to cubemap is not supported on device/platform");
		}
		Graphics.Blit(renderTexture, renderTexture3, I360Render.equirectangularConverter);
		renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
		Texture2D texture2D = new Texture2D(renderTexture2.width * 2, renderTexture2.height, 5, false);
		RenderTexture.active = (flipEyes ? renderTexture4 : renderTexture3);
		texture2D.ReadPixels(new Rect(0f, 0f, (float)renderTexture2.width, (float)renderTexture2.height), 0, 0);
		RenderTexture.active = (flipEyes ? renderTexture3 : renderTexture4);
		texture2D.ReadPixels(new Rect(0f, 0f, (float)renderTexture2.width, (float)renderTexture2.height), renderTexture2.width, 0);
		texture2D.Apply();
		result = texture2D;
	}
	finally
	{
		renderCam.targetTexture = targetTexture;
		if (renderTexture != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture);
		}
		if (renderTexture2 != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture2);
		}
		if (renderTexture3 != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture3);
		}
		if (renderTexture4 != null)
		{
			RenderTexture.ReleaseTemporary(renderTexture4);
		}
	}
	return result;
}

I originally modified this for personal use, so the code isn’t exactly high quality.
Also, since I rewrote this based on memory, it might not be entirely accurate.
Still, I hope it can be helpful in some way!

[jasonbarthes]

Thank you for the response. But I'm stuck where I try to compile the plugin

[jasonbarthes]

public static Texture2D CaptureStereoTex(int width = 1024, Camera renderCam = null, bool faceCameraDirection = true, float eyeSeparation = 0.6f, bool flipEyes = false)
{
if (renderCam == null)
{
renderCam = Camera.main;
if (renderCam == null)
{
throw new Exception("No camera detected");
}
}
RenderTexture targetTexture = renderCam.targetTexture;
int num = Mathf.Min(Mathf.NextPowerOfTwo(width), 20000);
RenderTexture renderTexture = null;
RenderTexture renderTexture2 = null;
RenderTexture renderTexture3 = null;
RenderTexture renderTexture4 = null;
Texture2D result;
try
{
renderTexture = RenderTexture.GetTemporary(num, num, 0, 0);
renderTexture.dimension = 4;
renderTexture2 = RenderTexture.GetTemporary(num, num, 0, 0);
renderTexture2.dimension = 2;
renderTexture3 = RenderTexture.GetTemporary(num, num, 0, 0);
renderTexture4 = RenderTexture.GetTemporary(num, num, 0, 0);
I360Render.equirectangularConverter.SetFloat(I360Render.paddingX, faceCameraDirection ? (renderCam.transform.eulerAngles.y / 360f) : 0f);
I360Render.equirectangularConverter.SetMatrix(I360Render.cameraRotationMatrixID, renderCam.transform.localToWorldMatrix);
renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
if (!renderCam.RenderToCubemap(renderTexture))
{
throw new Exception("Rendering to cubemap is not supported on device/platform");
}
Graphics.Blit(renderTexture, renderTexture4, I360Render.equirectangularConverter);
renderCam.transform.position -= renderCam.transform.right * eyeSeparation;
if (!renderCam.RenderToCubemap(renderTexture))
{
throw new Exception("Rendering to cubemap is not supported on device/platform");
}
Graphics.Blit(renderTexture, renderTexture3, I360Render.equirectangularConverter);
renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
Texture2D texture2D = new Texture2D(renderTexture2.width * 2, renderTexture2.height, 5, false);
RenderTexture.active = (flipEyes ? renderTexture4 : renderTexture3);
texture2D.ReadPixels(new Rect(0f, 0f, (float)renderTexture2.width, (float)renderTexture2.height), 0, 0);
RenderTexture.active = (flipEyes ? renderTexture3 : renderTexture4);
texture2D.ReadPixels(new Rect(0f, 0f, (float)renderTexture2.width, (float)renderTexture2.height), renderTexture2.width, 0);
texture2D.Apply();
result = texture2D;
}
finally
{
renderCam.targetTexture = targetTexture;
if (renderTexture != null)
{
RenderTexture.ReleaseTemporary(renderTexture);
}
if (renderTexture2 != null)
{
RenderTexture.ReleaseTemporary(renderTexture2);
}
if (renderTexture3 != null)
{
RenderTexture.ReleaseTemporary(renderTexture3);
}
if (renderTexture4 != null)
{
RenderTexture.ReleaseTemporary(renderTexture4);
}
}
return result;
}

I tried these steps but the result was the code wants to access properties that are not accessible for Unity 5.6.x.(eg.RenderTexture.dimension). I could not compile it until I replaced these lines. However, the dll I generated was still giving me 360 results

[mellona0821]

Are you still working on the issue?
If so, could you give this one a try?

public static Texture2D CaptureStereoTex(int width = 1024, Camera renderCam = null, bool faceCameraDirection = true, float eyeSeparation = 0.6f, bool flipEyes = false)
{
  if (renderCam == null)
  {
    renderCam = Camera.main;
    if (renderCam == null)
    {
      throw new Exception("No camera detected");
    }
  }
  RenderTexture targetTexture = renderCam.targetTexture;
  int num = Mathf.NextPowerOfTwo(width);
  RenderTexture renderTexture = null;
  RenderTexture renderTexture2 = null;
  RenderTexture renderTexture3 = null;
  Texture2D result;
  try
  {
    renderTexture = RenderTexture.GetTemporary(num, num, 0, RenderTextureFormat.ARGB32);
    renderTexture2 = RenderTexture.GetTemporary(num, num, 0, RenderTextureFormat.ARGB32);
    renderTexture3 = RenderTexture.GetTemporary(num, num, 0, RenderTextureFormat.ARGB32);
    renderTexture.dimension = TextureDimension.Cube;
    I360Render.equirectangularConverter.SetFloat(I360Render.paddingX, faceCameraDirection ? (renderCam.transform.eulerAngles.y / 360f) : 0f);
    I360Render.equirectangularConverter.SetMatrix(I360Render.cameraRotationMatrixID, renderCam.transform.localToWorldMatrix);
    renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
    if (!renderCam.RenderToCubemap(renderTexture))
    {
      throw new Exception("Rendering to cubemap is not supported on device/platform");
    }
    Graphics.Blit(renderTexture, renderTexture3, I360Render.equirectangularConverter);
    renderCam.transform.position -= renderCam.transform.right * eyeSeparation;
    if (!renderCam.RenderToCubemap(renderTexture))
    {
      throw new Exception("Rendering to cubemap is not supported on device/platform");
    }
    Graphics.Blit(renderTexture, renderTexture2, I360Render.equirectangularConverter);
    renderCam.transform.position += renderCam.transform.right * (eyeSeparation / 2f);
    Texture2D texture2D = new Texture2D(num * 2, num, TextureFormat.ARGB32, false);
    RenderTexture.active = (flipEyes ? renderTexture3 : renderTexture2);
    texture2D.ReadPixels(new Rect(0f, 0f, (float)num, (float)num), 0, 0);
    RenderTexture.active = (flipEyes ? renderTexture2 : renderTexture3);
    texture2D.ReadPixels(new Rect(0f, 0f, (float)num, (float)num), num, 0);
    texture2D.Apply();
    result = texture2D;
  }
  finally
  {
    renderCam.targetTexture = targetTexture;
    RenderTexture.active = null;
    if (renderTexture != null)
    {
      RenderTexture.ReleaseTemporary(renderTexture);
    }
    if (renderTexture2 != null)
    {
      RenderTexture.ReleaseTemporary(renderTexture2);
    }
    if (renderTexture3 != null)
    {
      RenderTexture.ReleaseTemporary(renderTexture3);
    }
  }
  return result;
}