55:
was the dominant anti-aliasing technique. MSAA samples (renders) each pixel multiple times at different locations within the frame and averages the samples to produce the final pixel value. In contrast, TAA samples each pixel only once per frame, but it samples the pixels at a different locations in
34:
in the current frame. In TAA, each pixel is sampled once per frame but in each frame the sample is at a different location within the pixel. Pixels sampled in past frames are blended with pixels sampled in the current frame to produce an anti-aliased image. Although this method makes TAA achieve a
111:
operates on similar principles to TAA. Like TAA, it uses information from past frames to produce the current frame. Unlike TAA, DLSS does not sample every pixel in every frame. Instead, it samples different pixels in different frames and uses pixels sampled in past frames to fill in the unsampled
76:
Sampling the pixels at a different position in each frame can be achieved by adding a per-frame "jitter" when rendering the frames. The "jitter" is a 2D offset that shifts the pixel grid, and its X and Y magnitude are between 0 and 1.
56:
different frames. This makes TAA faster than MSAA. In parts of the picture without motion, TAA effectively computes MSAA over multiple frames and achieves the same quality as MSAA with lower computational cost.
68:
both sample each pixel only once per frame, but FXAA does not take into account pixels sampled in past frames, so FXAA is simpler and faster but can not achieve the same image quality as TAA or MSAA.
80:
When combining pixels sampled in past frames with pixels sampled in the current frame, care needs to be taken to avoid blending pixels that contain different objects, which would produce
112:
pixels in the current frame. DLSS uses machine learning to combine samples in the current frame and past frames, and it can be thought of as an advanced TAA implementation.
273:
84:
or motion-blurring artifacts. Different implementation of TAA have different ways of achieving this. Possible methods include:
126:
65:
131:
108:
136:
229:
216:
30:
technique for computer-generated video that combines information from past frames and the current frame to remove
121:
52:
254:
81:
40:
27:
89:
195:
187:
179:
242:
167:
95:
Limiting (clamping) the final value of a pixel by the values of pixels surrounding it.
267:
199:
146:
141:
36:
191:
243:"DLSS 2.0 - Image Reconstruction for Real-time Rendering with Deep Learning"
31:
183:
104:
88:
Using motion vectors from the game engine to perform
166:Yang, Lei; Liu, Shiqiu; Salvi, Marco (2020-06-13).
168:"A Survey of Temporal Antialiasing Techniques"
8:
230:"Temporal Anti Aliasing – Step by Step"
158:
211:
209:
217:"High Quality Temporal Supersampling"
7:
39:, the technique inevitably causes
14:
255:"How DLSS 2.0 works (for gamers)"
51:Prior to the development of TAA,
16:Spatial anti-aliasing technique
127:Fast approximate anti-aliasing
1:
43:and blurriness to the image.
132:Deep learning super sampling
137:Deep learning anti-aliasing
290:
122:Multisample anti-aliasing
274:Anti-aliasing algorithms
215:Brian Kari, Epic Games
172:Computer Graphics Forum
20:Temporal anti-aliasing
35:result comparable to
28:spatial anti-aliasing
100:TAA compared to DLSS
60:TAA compared to FXAA
47:TAA compared to MSAA
241:Edward Liu, NVIDIA
90:motion compensation
202:– via Wiley.
184:10.1111/cgf.14018
281:
258:
251:
245:
239:
233:
226:
220:
213:
204:
203:
163:
92:before blending.
289:
288:
284:
283:
282:
280:
279:
278:
264:
263:
262:
261:
252:
248:
240:
236:
227:
223:
214:
207:
165:
164:
160:
155:
118:
102:
74:
62:
49:
17:
12:
11:
5:
287:
285:
277:
276:
266:
265:
260:
259:
246:
234:
228:Ziyad Barakat
221:
205:
178:(2): 607–621.
157:
156:
154:
151:
150:
149:
144:
139:
134:
129:
124:
117:
114:
101:
98:
97:
96:
93:
73:
72:Implementation
70:
61:
58:
48:
45:
15:
13:
10:
9:
6:
4:
3:
2:
286:
275:
272:
271:
269:
256:
253:yellowstone6
250:
247:
244:
238:
235:
231:
225:
222:
218:
212:
210:
206:
201:
197:
193:
189:
185:
181:
177:
173:
169:
162:
159:
152:
148:
147:Deinterlacing
145:
143:
142:Supersampling
140:
138:
135:
133:
130:
128:
125:
123:
120:
119:
115:
113:
110:
106:
99:
94:
91:
87:
86:
85:
83:
78:
71:
69:
67:
59:
57:
54:
46:
44:
42:
38:
37:supersampling
33:
29:
25:
21:
249:
237:
224:
175:
171:
161:
103:
79:
75:
63:
50:
23:
19:
18:
153:References
200:220514131
192:0167-7055
268:Category
116:See also
82:ghosting
64:TAA and
41:ghosting
32:jaggies
26:) is a
198:
190:
105:Nvidia
196:S2CID
188:ISSN
109:DLSS
66:FXAA
53:MSAA
180:doi
107:'s
24:TAA
270::
208:^
194:.
186:.
176:39
174:.
170:.
257:.
232:.
219:.
182::
22:(
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.