format_conversion.effect 8.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343
  1. /******************************************************************************
  2. Copyright (C) 2014 by Hugh Bailey <[email protected]>
  3. This program is free software: you can redistribute it and/or modify
  4. it under the terms of the GNU General Public License as published by
  5. the Free Software Foundation, either version 2 of the License, or
  6. (at your option) any later version.
  7. This program is distributed in the hope that it will be useful,
  8. but WITHOUT ANY WARRANTY; without even the implied warranty of
  9. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  10. GNU General Public License for more details.
  11. You should have received a copy of the GNU General Public License
  12. along with this program. If not, see <http://www.gnu.org/licenses/>.
  13. ******************************************************************************/
  14. //#define DEBUGGING
  15. uniform float4x4 ViewProj;
  16. uniform float u_plane_offset;
  17. uniform float v_plane_offset;
  18. uniform float width;
  19. uniform float height;
  20. uniform float width_i;
  21. uniform float height_i;
  22. uniform float width_d2;
  23. uniform float height_d2;
  24. uniform float width_d2_i;
  25. uniform float height_d2_i;
  26. uniform float input_width;
  27. uniform float input_height;
  28. uniform float input_width_i;
  29. uniform float input_height_i;
  30. uniform float input_width_i_d2;
  31. uniform float input_height_i_d2;
  32. uniform texture2d image;
  33. sampler_state def_sampler {
  34. Filter = Linear;
  35. AddressU = Clamp;
  36. AddressV = Clamp;
  37. };
  38. struct VertInOut {
  39. float4 pos : POSITION;
  40. float2 uv : TEXCOORD0;
  41. };
  42. VertInOut VSDefault(VertInOut vert_in)
  43. {
  44. VertInOut vert_out;
  45. vert_out.pos = mul(float4(vert_in.pos.xyz, 1.0), ViewProj);
  46. vert_out.uv = vert_in.uv;
  47. return vert_out;
  48. }
  49. /* used to prevent internal GPU precision issues width fmod in particular */
  50. #define PRECISION_OFFSET 0.2
  51. float4 PSNV12(VertInOut vert_in) : TARGET
  52. {
  53. #ifdef _OPENGL
  54. float v_mul = floor((1.0 - vert_in.uv.y) * input_height);
  55. #else
  56. float v_mul = floor(vert_in.uv.y * input_height);
  57. #endif
  58. float byte_offset = floor((v_mul + vert_in.uv.x) * width) * 4.0;
  59. byte_offset += PRECISION_OFFSET;
  60. float2 sample_pos[4];
  61. if (byte_offset < u_plane_offset) {
  62. #ifdef DEBUGGING
  63. return float4(1.0, 1.0, 1.0, 1.0);
  64. #endif
  65. float lum_u = floor(fmod(byte_offset, width)) * width_i;
  66. float lum_v = floor(byte_offset * width_i) * height_i;
  67. /* move to texel centers to sample the 4 pixels properly */
  68. lum_u += width_i * 0.5;
  69. lum_v += height_i * 0.5;
  70. sample_pos[0] = float2(lum_u, lum_v);
  71. sample_pos[1] = float2(lum_u += width_i, lum_v);
  72. sample_pos[2] = float2(lum_u += width_i, lum_v);
  73. sample_pos[3] = float2(lum_u + width_i, lum_v);
  74. float4x4 out_val = float4x4(
  75. image.Sample(def_sampler, sample_pos[0]),
  76. image.Sample(def_sampler, sample_pos[1]),
  77. image.Sample(def_sampler, sample_pos[2]),
  78. image.Sample(def_sampler, sample_pos[3])
  79. );
  80. return transpose(out_val)[1];
  81. } else {
  82. #ifdef DEBUGGING
  83. return float4(0.5, 0.2, 0.5, 0.2);
  84. #endif
  85. float new_offset = byte_offset - u_plane_offset;
  86. float ch_u = floor(fmod(new_offset, width)) * width_i;
  87. float ch_v = floor(new_offset * width_i) * height_d2_i;
  88. float width_i2 = width_i*2.0;
  89. /* move to the borders of each set of 4 pixels to force it
  90. * to do bilinear averaging */
  91. ch_u += width_i;
  92. ch_v += height_i;
  93. sample_pos[0] = float2(ch_u, ch_v);
  94. sample_pos[1] = float2(ch_u + width_i2, ch_v);
  95. return float4(
  96. image.Sample(def_sampler, sample_pos[0]).rb,
  97. image.Sample(def_sampler, sample_pos[1]).rb
  98. );
  99. }
  100. }
  101. float4 PSPlanar420(VertInOut vert_in) : TARGET
  102. {
  103. #ifdef _OPENGL
  104. float v_mul = floor((1.0 - vert_in.uv.y) * input_height);
  105. #else
  106. float v_mul = floor(vert_in.uv.y * input_height);
  107. #endif
  108. float byte_offset = floor((v_mul + vert_in.uv.x) * width) * 4.0;
  109. byte_offset += PRECISION_OFFSET;
  110. float2 sample_pos[4];
  111. if (byte_offset < u_plane_offset) {
  112. #ifdef DEBUGGING
  113. return float4(1.0, 1.0, 1.0, 1.0);
  114. #endif
  115. float lum_u = floor(fmod(byte_offset, width)) * width_i;
  116. float lum_v = floor(byte_offset * width_i) * height_i;
  117. /* move to texel centers to sample the 4 pixels properly */
  118. lum_u += width_i * 0.5;
  119. lum_v += height_i * 0.5;
  120. sample_pos[0] = float2(lum_u, lum_v);
  121. sample_pos[1] = float2(lum_u += width_i, lum_v);
  122. sample_pos[2] = float2(lum_u += width_i, lum_v);
  123. sample_pos[3] = float2(lum_u + width_i, lum_v);
  124. } else {
  125. #ifdef DEBUGGING
  126. return ((byte_offset < v_plane_offset) ?
  127. float4(0.5, 0.5, 0.5, 0.5) :
  128. float4(0.2, 0.2, 0.2, 0.2));
  129. #endif
  130. float new_offset = byte_offset -
  131. ((byte_offset < v_plane_offset) ?
  132. u_plane_offset : v_plane_offset);
  133. float ch_u = floor(fmod(new_offset, width_d2)) * width_d2_i;
  134. float ch_v = floor(new_offset * width_d2_i) * height_d2_i;
  135. float width_i2 = width_i*2.0;
  136. /* move to the borders of each set of 4 pixels to force it
  137. * to do bilinear averaging */
  138. ch_u += width_i;
  139. ch_v += height_i;
  140. sample_pos[0] = float2(ch_u, ch_v);
  141. sample_pos[1] = float2(ch_u += width_i2, ch_v);
  142. sample_pos[2] = float2(ch_u += width_i2, ch_v);
  143. sample_pos[3] = float2(ch_u + width_i2, ch_v);
  144. }
  145. float4x4 out_val = float4x4(
  146. image.Sample(def_sampler, sample_pos[0]),
  147. image.Sample(def_sampler, sample_pos[1]),
  148. image.Sample(def_sampler, sample_pos[2]),
  149. image.Sample(def_sampler, sample_pos[3])
  150. );
  151. out_val = transpose(out_val);
  152. if (byte_offset < u_plane_offset)
  153. return out_val[1];
  154. else if (byte_offset < v_plane_offset)
  155. return out_val[0];
  156. else
  157. return out_val[2];
  158. }
  159. float4 PSPacked422_Reverse(VertInOut vert_in, int u_pos, int v_pos,
  160. int y0_pos, int y1_pos) : TARGET
  161. {
  162. float y = vert_in.uv.y;
  163. #ifdef _OPENGL
  164. y = 1. - y;
  165. #endif
  166. float odd = floor(fmod(width * vert_in.uv.x + PRECISION_OFFSET, 2.0));
  167. float x = floor(width_d2 * vert_in.uv.x + PRECISION_OFFSET) *
  168. width_d2_i;
  169. x += input_width_i_d2;
  170. float4 texel = image.Sample(def_sampler, float2(x, y));
  171. return float4(odd > 0.5 ? texel[y1_pos] : texel[y0_pos],
  172. texel[u_pos], texel[v_pos], 1.0);
  173. }
  174. float GetOffsetColor(float offset)
  175. {
  176. float2 uv;
  177. offset += PRECISION_OFFSET;
  178. uv.x = floor(fmod(offset, input_width)) * input_width_i;
  179. uv.y = floor(offset * input_width_i) * input_height_i;
  180. uv.xy += float2(input_width_i_d2, input_height_i_d2);
  181. return image.Sample(def_sampler, uv).r;
  182. }
  183. float4 PSPlanar420_Reverse(VertInOut vert_in) : TARGET
  184. {
  185. float x = vert_in.uv.x;
  186. float y = vert_in.uv.y;
  187. #ifdef _OPENGL
  188. y = 1. - y;
  189. #endif
  190. float x_offset = floor(x * width + PRECISION_OFFSET);
  191. float y_offset = floor(y * height + PRECISION_OFFSET);
  192. float lum_offset = y_offset * width + x_offset + PRECISION_OFFSET;
  193. lum_offset = floor(lum_offset);
  194. float ch_offset = floor(y_offset * 0.5 + PRECISION_OFFSET) * width_d2 +
  195. (x_offset * 0.5) + PRECISION_OFFSET;
  196. ch_offset = floor(ch_offset);
  197. return float4(
  198. GetOffsetColor(lum_offset),
  199. GetOffsetColor(u_plane_offset + ch_offset),
  200. GetOffsetColor(v_plane_offset + ch_offset),
  201. 1.0
  202. );
  203. }
  204. float4 PSNV12_Reverse(VertInOut vert_in) : TARGET
  205. {
  206. float x = vert_in.uv.x;
  207. float y = vert_in.uv.y;
  208. #ifdef _OPENGL
  209. y = 1. - y;
  210. #endif
  211. float x_offset = floor(x * width + PRECISION_OFFSET);
  212. float y_offset = floor(y * height + PRECISION_OFFSET);
  213. float lum_offset = y_offset * width + x_offset + PRECISION_OFFSET;
  214. lum_offset = floor(lum_offset);
  215. float ch_offset = floor(y_offset * 0.5 + PRECISION_OFFSET) * width_d2 +
  216. (x_offset * 0.5);
  217. ch_offset = floor(ch_offset * 2.0 + PRECISION_OFFSET);
  218. return float4(
  219. GetOffsetColor(lum_offset),
  220. GetOffsetColor(u_plane_offset + ch_offset),
  221. GetOffsetColor(u_plane_offset + ch_offset + 1.0),
  222. 1.0
  223. );
  224. }
  225. technique Planar420
  226. {
  227. pass
  228. {
  229. vertex_shader = VSDefault(vert_in);
  230. pixel_shader = PSPlanar420(vert_in);
  231. }
  232. }
  233. technique NV12
  234. {
  235. pass
  236. {
  237. vertex_shader = VSDefault(vert_in);
  238. pixel_shader = PSNV12(vert_in);
  239. }
  240. }
  241. technique UYVY_Reverse
  242. {
  243. pass
  244. {
  245. vertex_shader = VSDefault(vert_in);
  246. pixel_shader = PSPacked422_Reverse(vert_in, 2, 0, 1, 3);
  247. }
  248. }
  249. technique YUY2_Reverse
  250. {
  251. pass
  252. {
  253. vertex_shader = VSDefault(vert_in);
  254. pixel_shader = PSPacked422_Reverse(vert_in, 1, 3, 2, 0);
  255. }
  256. }
  257. technique YVYU_Reverse
  258. {
  259. pass
  260. {
  261. vertex_shader = VSDefault(vert_in);
  262. pixel_shader = PSPacked422_Reverse(vert_in, 3, 1, 2, 0);
  263. }
  264. }
  265. technique I420_Reverse
  266. {
  267. pass
  268. {
  269. vertex_shader = VSDefault(vert_in);
  270. pixel_shader = PSPlanar420_Reverse(vert_in);
  271. }
  272. }
  273. technique NV12_Reverse
  274. {
  275. pass
  276. {
  277. vertex_shader = VSDefault(vert_in);
  278. pixel_shader = PSNV12_Reverse(vert_in);
  279. }
  280. }