| | 1 | === Mask Analysis for Run 3 with the reconfigured GPC1 : 2009.05.10 === |
| | 2 | |
| | 3 | up to [wiki:IPP_for_PS1 IPP for PS1] |
| | 4 | |
| | 5 | I've generated a mask set for the reconfigured camera. For this mask, I have attempted to use the automatic dark, flat, and CTE masking analyses. There were a couple of challenges. |
| | 6 | |
| | 7 | The first was to set the darkmask analysis parameters in a way which would not overmask images with correlated read noise. I used a set of 10sec dark images, and set the MASK.SUSPECT.SIGMA value to 8.0 to limit the masking to signficant dark current features. This seems to have been quite successful, though it may be worth trying the longer dark exposures rather than the short ones. Also, I used a dark master built from 10, 30, 100, and 300s inputs, with the model fitting trends of exposure time. This is not always successful; it may be more effective in terms of masking to use a single exposure time range. |
| | 8 | |
| | 9 | The second challenge was to choose the CTE mask variance threshold in a reasonable way. I ended up setting it to 0.5, and this generally worked well. The only glaring exception is chip XY11, which actually does have an overall high CTE, but is probably being masked more excessively than is reasonable. It may be worth in the future playing with these levels in more detail. |
| | 10 | |
| | 11 | I combined the DARK, FLAT, and CTE mask bits into a single master mask using a mana script (attached). Two points about the CTE masking: for reasons I don't understand, some cells were completely masked with the BLANK value. After some limited exploration, I concluded that these were bogus, and excluded that bit from the master mask generation. Also, due to floating point round-off in the ppImage output compression, some of the CTE superpixels get shredded into masked and unmasked pixels. I solved this by smoothing the CTE mask with a Gaussian of 4pixel sigma. |
| | 12 | |
| | 13 | Below are the masked-fraction images for the new mask with and without the vignetted region shown. Compare with the Run 2 masks on this page: [wiki:GPC1_Mask_2008.06 GPC1 Mask 2008.06]. |
| | 14 | |
| | 15 | masked fraction (full focal plane): 0.085 |
| | 16 | |
| | 17 | [[Image(Mask-full-focalplane.20090423.jpg)]] |
| | 18 | |
| | 19 | masked fraction (unvignetted region): 0.037 |
| | 20 | |
| | 21 | [[Image(Mask-unvignetted.20090423.jpg)]] |
| | 22 | |
