1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
import math
import torch
import torch.nn.functional as F
from onb import ONB


class GGX_BRDF(object):

    @classmethod
    def eval(
        cls,
        local_onb: ONB,
        wi: torch.Tensor,
        wo: torch.Tensor,
        ax: torch.Tensor,
        ay: torch.Tensor,
        pd: torch.Tensor,
        ps: torch.Tensor,
        is_diff: torch.Tensor = None,
        specular_component: str = "D_F_G_B",
    ) -> torch.Tensor:
        """
        Evaluate brdf in shading coordinate(ntb space).

        For each shading point, we will build a coordinate system to calculate
        the brdf. The coordinate is usually expressed as ntb. n is the normal
        of the shading point.

        Args:
            local_onb: a coordinate system to do shade
            wi: incident light in world space, of shape (batch, lightnum, 3)
            wo: outgoing light in world space, of shape (batch, 3)
            ax: shape (batch, 1)
            ay: shape (batch, 1)
            pd: shape (batch, channel), range [0, 1]
            ps: shape (batch, channel), range [0, 10]
            is_diff: shape (batch, lightnum), if the value is 0, eval "pd_only", else if
                the value is 1, eval "ps_only". If `is_diff` is None, means "both"
            specular_component: the ingredient of BRDF, usually "D_F_G_B", B means bottom

        Returns:
            brdf: (batch, lightnum, channel)
            meta: 
        """
        N = wi.size(1)

        # transform wi and wo to local frame
        wi_local = local_onb.transform(wi)    # (batch, lightnum, 3)
        wo_local = local_onb.transform(wo)    # (batch, 3)

        meta = {}

        a = torch.unsqueeze(pd / math.pi, dim=1)  # (batch, 1, 1)
        b = cls.ggx_brdf_aniso(wi_local, wo_local, ax, ay, specular_component)  # (batch, lightnum, 1)
        ps = torch.unsqueeze(ps, dim=1)  # (batch, 1, channel)

        if is_diff is None:
            brdf = a + b * ps
        else:
            is_diff_ = is_diff.unsqueeze(2)
            brdf = a.repeat(1, N, 1) * (1 - is_diff_) + b * ps * is_diff_

        meta['pd'] = a
        meta['ps'] = b * ps

        return brdf, meta

    @classmethod
    def ggx_brdf_aniso(
        cls,
        wi: torch.Tensor,
        wo: torch.Tensor,
        ax: torch.Tensor,
        ay: torch.Tensor,
        specular_component: str
    ) -> torch.Tensor:
        """
        Calculate anisotropy ggx brdf in shading coordinate.
        
        Args:
            wi: incident light in ntb space, of shape (batch, lightnum, 3)
            wo: emergent light in ntb space, of shape (batch, 3)
            ax: shape (batch, 1)
            ay: shape (batch, 1)
            specular_component: the ingredient of BRDF, usually "D_F_G_B"
        
        Returns:
            brdf: shape (batch, lightnum, 1)
        """

        lightnum = wi.size(1)
        wo = torch.unsqueeze(wo, dim=1).repeat(1, lightnum, 1)

        wi_z = wi[:, :, [2]]  # (batch, lightnum, 1)
        wo_z = wo[:, :, [2]]
        denom = 4 * wi_z * wo_z  # (batch, lightnum, 1)
        vhalf = F.normalize(wi + wo, dim=2)  # (batch, lightnum, 3)

        # F
        tmp = torch.clamp(1.0 - torch.sum(wi * vhalf, dim=2, keepdim=True), 0, 1)
        F0 = 0.04
        Fresnel = F0 + (1 - F0) * tmp * tmp * tmp * tmp * tmp

        # D
        axayaz = torch.unsqueeze(torch.cat([ax, ay, torch.ones_like(ax)], dim=1),
                                dim=1)  # (batch, 1, 3)
        vhalf = vhalf / (axayaz + 1e-6)  # (batch, lightnum, 3)
        vhalf_norm = torch.norm(vhalf, dim=2, keepdim=True)
        length = vhalf_norm * vhalf_norm  # (batch, lightnum, 1)
        D = 1.0 / (math.pi * torch.unsqueeze(ax, dim=1) *
                torch.unsqueeze(ay, dim=1) * length * length)

        # G
        G = cls.ggx_G1_aniso(wi, ax, ay, wi_z) * cls.ggx_G1_aniso(wo, ax, ay, wo_z)

        tmp = torch.ones_like(denom)
        if "D" in specular_component:
            tmp = tmp * D
        if "F" in specular_component:
            tmp = tmp * Fresnel
        if "G" in specular_component:
            tmp = tmp * G
        if "B" in specular_component:
            tmp = tmp / (denom + 1e-6)

        # some samples' wi_z/wo_z may less or equal than 0, should be 
        # set to zero. Maybe this step is not necessary, because G is
        # already zero.
        tmp_zeros = torch.zeros_like(tmp)
        static_zero = torch.zeros(1, device=wi.device, dtype=torch.float32)
        res = torch.where(torch.le(wi_z, static_zero), tmp_zeros, tmp)
        res = torch.where(torch.le(wo_z, static_zero), tmp_zeros, res)

        return res

    @classmethod
    def ggx_G1_aniso(
        cls,
        v: torch.Tensor,
        ax: torch.Tensor,
        ay: torch.Tensor,
        vz: torch.Tensor
    ) -> torch.Tensor:
        """
        If vz <= 0, return 0

        Args:
            v: shape (batch, lightnum, 3)
            ax: shape (batch, 1)
            ay: shape (batch, 1)
            vz: shape (batch, lightnum, 1)
        
        Returns:
            G1: shape (batch, lightnum, 1)
        """
        axayaz = torch.cat([ax, ay, torch.ones_like(ax)], dim=1)    # (batch, 3)
        vv = v * torch.unsqueeze(axayaz, dim=1)                     # (batch, lightnum, 3)
        G1 = 2.0 * vz / (vz + torch.norm(vv, dim=2, keepdim=True) + 1e-6)

        # If vz < 0, G1 will be zero.
        G1 = torch.where(
            torch.le(vz, torch.zeros_like(vz)),
            torch.zeros_like(vz),
            G1
        )
        return G1