gpt4 book ai didi

c++ - LookAt(glm)返回错误的转换z值

转载 作者:行者123 更新时间:2023-12-03 07:26:12 24 4
gpt4 key购买 nike

您好,我对glm lookAt函数的返回值有疑问。当我以 Debug模式执行时,此时
glm函数的 ... Result[3][2] = dot(f, eye); ...在矩阵的平移z位置中的值错误。值是-2,向我表明前向和眼向 vector 处于相反的位置。我的眼睛,中心 vector 和向上 vector 分别是eye(0,0,2),center(0,0,-1)和up(0,1,0)。凸轮余弦 vector 为:f(0,0,-1),s(1,0,0)和u(0,1,0)。用户查看的优势点是(0,0,0)。因此,正确的 View 矩阵应为:

1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1

但我得到这个:

1 -0 -0
-0 1 -0 -0
0 0 0 1-2
0 0 0 0 1

我的代码是:

struct camera {
vec3 position = vec3(0.0f); // position of the camera
vec3 view_direction = vec3(0.0f); // forward vector (orientation)
vec3 side = vec3(0.0f); // right vector (side)
vec3 up = vec3(0.0f, 1.0f, 0.0f); // up vector
float speed = 0.1;
float yaw = 0.0f; // y-rotation
float cam_yaw_speed = 10.0f; // 10 degrees per second
float pitch = 0.0f; // x-rotation
float roll = 0.0f;

...
// calculate the orientation vector (forward)
vec3 getOrientation(vec3 vantage_point) {
// calc the difference and normalize the resulting vector
vec3 result = vantage_point - position;
result = normalize(result);
return result;
}

// calculate the right (side) vector of the camera, by given orientation(forward) and up vectors
mat4 look_at_point(vec3 vantage_point) {
view_direction = getOrientation(vantage_point);

// calculate the lookat matrix
return lookAt(position, position + view_direction, up);
}
};

我试图弄清楚如何解决这个问题,但我仍然不知道。有人能帮我吗?

我正在执行main_cam.look_at_point(vantage_point)函数的主要函数如下所示:
...
GLfloat points[] = {
0.0f, 0.5f, 0.0f,
0.5f, 0.0f, 0.0f,
-0.5f, 0.0f, 0.0f };

float speed = 1.0f; // move at 1 unit per second

float last_position = 0.0f;

// init camera
main_cam.position = vec3(0.0f, 0.0f, 2.0f); // don't start at zero, or will be too close
main_cam.speed = 1.0f; // 1 unit per second
main_cam.cam_yaw_speed = 10.0f; // 10 degrees per second
vec3 vantage_point = vec3(0.0f, 0.0f, 0.0f);

mat4 T = translate(mat4(1.0), main_cam.position);
//mat4 R = rotate(mat4(), -main_cam.yaw, vec3(0.0, 1.0, 0.0));
mat4 R = main_cam.look_at_point(vantage_point);
mat4 view_matrix = R * T;

// input variables
float near = 0.1f; // clipping plane
float far = 100.0f; // clipping plane
float fov = 67.0f * ONE_DEG_IN_RAD; // convert 67 degrees to radians
float aspect = (float)g_gl_width / (float)g_gl_height; // aspect ratio

mat4 proj_matrix = perspective(fov, aspect, near, far);

use_shader_program(shader_program);
set_uniform_matrix4fv(shader_program, "view", 1, GL_FALSE, &view_matrix[0][0]);
set_uniform_matrix4fv(shader_program, "proj", 1, GL_FALSE, &proj_matrix[0][0]);
...

用glm的旋转功能测试的三角形如右图所示。

Triangle shown with the rotate function of glm

最佳答案

我怀疑问题出在这里:

mat4 view_matrix = R * T; // <---
lookAt返回的矩阵已经完成翻译。
尝试在三角形内部的(0,0,0)点上手动应用变换。 T会将其转换为(0,0,2),但现在它与相机重合,因此 R会将其发送回(0,0,0)。现在,您将得到射影除以零事故的除法。
因此,通过 T删除乘法:
mat4 view_matrix = R;
现在(0,0,0)将被映射到(0,0,-2),即相机所看的方向。 (在相机空间中,投影中心在(0,0,0),并且相机朝负Z方向看)。

编辑:我想指出的是,从 view_direction计算 vantage_point,然后将 position + view_direction反馈回 lookAt是实现目标的一种相当人为的方式。您在 getOrientation函数中执行的操作是 lookAt在内部已执行的操作。相反,您可以从 view_direction的结果中获取 lookAt:
mat4 look_at_point(vec3 vantage_point) {
// calculate the lookat matrix
mat4 M = lookAt(position, vantage_point, up);
view_direction = -vec3(M[2][0], M[2][1], M[2][2]);
return M;
}
但是,考虑到最终您将尝试实现偏航/俯仰/侧倾摄像机控制,因此您就是 better off to not using lookAt at all

关于c++ - LookAt(glm)返回错误的转换z值,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/39943833/

24 4 0
Copyright 2021 - 2024 cfsdn All Rights Reserved 蜀ICP备2022000587号
广告合作:1813099741@qq.com 6ren.com