Back to index

Source-aware Encoder / Bitterli The Grey And White Room

Source-aware encoders enable straightforward adaptation of a trained model to new content. We train new source-aware encoders for the Tungsten Renderer on a training set built upon publicly available scenes.

The interactive viewer below compares results from several networks:

  1. one trained from scratch (random initialization) on the full set of 1200 frames,
  2. one trained from scratch (random initialization) on a small subset of 75 frames, and
  3. one for which we just trained a new source-aware (from scratch) encoder for a network previously trained on Moana and Cars, using the 75-frame subset for training.

The results are compared against the NFOR denoiser (Bitterli et al. 2016). In most scenes, just training a new frontend using the small dataset yields similar performance to training from scratch on the full dataset.

Images

Use mouse wheel to zoom in/out, click and drag to pan. Press keys [1], [2], ... to switch between individual images. [f] toggles full screen viewing.

32 spp
256 spp
Color
DSSIM
1: Input
2: From scratch (75)
3: Frontend (75)
4: From scratch (1200)
5: NFOR
6: Reference
Input
From scratch (75)
Frontend (75)
From scratch (1200)
NFOR
Reference
1: Input
2: From scratch (75)
3: Frontend (75)
4: From scratch (1200)
5: NFOR
Input
From scratch (75)
Frontend (75)
From scratch (1200)
NFOR
Color
DSSIM
1: Input
2: From scratch (75)
3: Frontend (75)
4: From scratch (1200)
5: NFOR
6: Reference
Input
From scratch (75)
Frontend (75)
From scratch (1200)
NFOR
Reference
1: Input
2: From scratch (75)
3: Frontend (75)
4: From scratch (1200)
5: NFOR
Input
From scratch (75)
Frontend (75)
From scratch (1200)
NFOR

Charts

Relative error, obtained by dividing the reconstruction error by the default path traced rendering error. For each method, we show, from left to right, the improvement at 32 and 128 samples per pixel. For all metrics, lower values are better.

MrSE
DSSIM

Error metrics

MrSE

32 spp256 spp
Input0.233030.03361
From scratch (75)0.004650.00125
Frontend (75)0.003040.00091
From scratch (1200)0.002700.00085
NFOR0.006920.00197

DSSIM

32 spp256 spp
Input0.397250.25114
From scratch (75)0.029530.01425
Frontend (75)0.022380.01184
From scratch (1200)0.019740.01144
NFOR0.037870.01993