difx:amplitudescaling

This shows you the differences between two versions of the page.

difx:amplitudescaling [2010/11/19 06:06] adamdeller |
difx:amplitudescaling [2015/10/21 10:08] |
||
---|---|---|---|

Line 1: | Line 1: | ||

- | ===== Amplitude scaling in General ===== | ||

- | At the end of the day, a properly calibrated visibility dataset should be an accurate representation of the received signal strength on a given baseline, in Janskys. | ||

- | |||

- | - Accumulating the unnormalised correlation between a pair of antennas | ||

- | - Scaling the unnormalised correlation to correct for the length of integration | ||

- | - Correcting for the nominal occupation of sampler levels | ||

- | - Correcting for the actual occupation of sampler levels | ||

- | - Correcting for quantisation effects (the Van Vleck correction) | ||

- | - Scaling the corrected correlation coefficients by system temperature | ||

- | - Scaling the corrected correlation coefficients by antenna gain to attain actual flux densities | ||

- | |||

- | The first part **has** to be done by the correlator. | ||

- | |||

- | The level of correction done online by DiFX is controlled by the TSYS entry in the DATASTREAM table entries of the .input file. If TSYS < = 0, only the first two steps are done. If TSYS=1.0, all steps excepting the last two are done. If TSYS = (some nominal tsys value), then all steps except the last are performed. | ||

- | |||

- | Each step will now be considered in more detail | ||

- | |||

- | ==== 1. Forming unnormalised correlation counts ==== | ||

- | The correlator unpacks quantised data to a floating point representation, | ||

- | |||

- | Some useful extra info can be found in Leonia Kogan' | ||

- | |||

- | ==== 2. Scaling the unnormalised correlation to correct for the length of integration ==== | ||

- | At the end of one accumulation period the correlator calculates the number of valid samples accumulated and divides the visibilities by this number. | ||

- | |||

- | This step is skipped in DiFX when TSYS>0 - as we shall see, the division by autocorrelations later on makes it redundant, since this same correction is applied to both autocorrelations and cross correlations | ||

- | |||

- | ==== 3. Correcting for the nominal occupation of sampler levels, and for the unpack values chosen ==== | ||

- | This is where things start to get muddled. | ||

- | |||

- | This step is skipped in DiFX when TSYS>0 - as we shall see, the division by autocorrelations later on makes it redundant, since this same correction is applied to both autocorrelations and cross correlations | ||

- | |||

- | ==== 4. Correcting for the actual occupation of sampler levels ==== | ||

- | So, at this stage we have autocorrelations that are equal to 1, and cross correlations that are equal to the correlation coefficients - **IF AND ONLY IF THE SAMPLER STATS WERE PERFECT**. | ||

- | |||

- | This step is the saving grace that sweeps any early mistakes under the carpet. | ||

- | |||

- | ==== 5. The Van Vleck correction ==== | ||

- | This is an upscaling which takes care of the mean decorrelation due to the coarse sampling. | ||

- | |||

- | ==== 6./7. Scaling by measured system temperature/ | ||

- | This is always done offline using [ANTAB and] APCAL in AIPS. If a priori values were applied at the correlator (TSYS > 1, equal to some nominal value for the telescope) then the measured TSYS values must be divided by the a priori values. | ||

- | |||

- | ===== GOTCHAS ===== | ||

- | DIGICOR in AIPS may actually check the array name and refuse to function correctly for non-VLBA data. I'm going to look into this. | ||

- | |||

- | Walter A.: \\ | ||

- | Just checked FITLD and it checks if the array name is VLBA. If it isn't, it will set DELCOR = .FALSE. and DOCORR = -1. This means it switches the corrections off! So DiFX has to declare every array to be VLBA unless we change FITLD. \\ | ||

- | I suggest to change FITLD' | ||

- | |||

- | Walter B.: \\ | ||

- | FITLD does this now. Just use 31Dec09 or later to get this behavior |

difx/amplitudescaling.txt · Last modified: 2015/10/21 10:08 (external edit)