Index page for CLEAN results from simulations using ALMA SIL


CLEANED images

All results for 10,000 iterations.

A array results

Posted 14th September 2000
Added 29th January 2001 Results for the -23 declination for the DOT image provided by John Conway. These results show the same fidelity and difference measures calculated in the same way as the 18th January B and C resolution results
Amended 6th February 2001 DOT results from 29th January have fidelity histograms amended to show both Kand C candidates on the same graph, and add a graph showing histograms of cumulative fidelity (i.e. number of pixels with fidelity greater than fidelity x) for both K and C candidates

B array results

Updated 2nd January 2001 These updated results have been obtained as described for the 18th December C array results below, i.e. Fidelity images, stats, and occasional negative values due to their presence in the dividing models...
Updated 18th January 2001 These results replace the 18th December results. The main changes are that the M51, CYG and SDC model images have had any negative pixels truncated to zero, so that they and their smoothed counterparts as used in the Difference and Fidelity image calculations are wholly positive (or zero). These results should now be complete, i.e. the links for the MAR and M51 cases should now work.
Amended 6th February 2001 Results from 18th January have fidelity histograms amended to show both Kand C candidates on the same graph, and add a graph showing histograms of cumulative fidelity (i.e. number of pixels with fidelity greater than fidelity x) for both K and C candidates
Updated 17th February 2001 Two new histograms have been added which show scaled rms in the difference image as a function of source intensity, and also dynamic range as a function of source intensity. This is described more fully in this note.

"H" array results- hybrid of B array long track and D array snapshot

Posted 17th February 2001 These results are obtained by reweighting the D snaphot UV data by a factor of 18.63 using WTMOD in AIPS, and combining with the corresponding B resolution UV data obtained over a 4 hour track, using DBCON. CLEANing, differencing, fidelity calculation and the new results reported for the B, C and D resolutions elesewehere are calculated.

C array results

Posted 14th November 2000 These results have been obtained as described in this note.
Updated 18th December 2000 These updated results include the results for the +25 declination, and show the CLEAN and difference results at the same scale to enable direct comparison. More importantly, they include `Fidelity' images as described by Stephane (more details on how they were calculated appear on the pages of results), which are the first attempt at a fractional-difference-type metric. The difference and fidelity images are now accompanied by histograms of pixel values, and textual max, min, mean and rms data. As this represents a large increase in the data presented, the pages each show a single declination, to keep their size sensible.
Updated 18th January 2001 These results replace the 18th December results. The main changes are that the M51, CYG and SDC model images have had any negative pixels truncated to zero, so that they and their smoothed counterparts as used in the Difference and Fidelity image calculations are wholly positive (or zero).
Amended 6th February 2001Results from 18th January have fidelity histograms amended to show both Kand C candidates on the same graph, and add a graph showing histograms of cumulative fidelity (i.e. number of pixels with fidelity greater than fidelity x) for both K and C candidates
Updated 17th February 2001 Two new histograms have been added which show scaled rms in the difference image as a function of source intensity, and also dynamic range as a function of source intensity. This is described more fully in this note.

D array results

Posted 29th January 2001 These results are a by product of the requirement to produce snapshot UV data for hybrid B and D imaging. Only M51 and CYG results are posted
Amended 6th February 2001Results from 29th January have fidelity histograms amended to show both Kand C candidates on the same graph, and add a graph showing histograms of cumulative fidelity (i.e. number of pixels with fidelity greater than fidelity x) for both K and C candidates
Updated 17th February 2001 Two new histograms have been added which show scaled rms in the difference image as a function of source intensity, and also dynamic range as a function of source intensity. This is described more fully in this note.
Page updated by Steven Heddle on 17th February 2001