Spitzer Documentation & Tools
Muxbleed Corrector

Purpose: Identifies and masks pixels potentially affected by muxbleed, attempts correction.

The Muxbleed Corrector presently only works well for deep survey images like the GOODS data

Author: Dan Stern (GOODS team); Sean Carey (IRAC IST)
Date Contributed: 21 Jan 2005
System Requirements: IDL

This software is not considered part of the pipeline and no installation support from IRSA should be expected. We are, however, very interested in your feedback in order to improve the algorithm and are glad to work with you on test cases and analysis. Questions and comments should be directed to the Helpdesk.

Information and Download

Download the Muxbleed Corrector (fmuxbleed_bcd.pro) (PRO, 8 KB). You will also need literstat.pro (PRO, 3 KB), muxbleed_lookup.pro (PRO, 1 KB) and muxbleed_sf.pro (PRO, 1 KB).

The IRAC muxbleed corrector was written by Dan Stern (JPL, GOODS) with help from Sean Carey of the IRAC Instrument Support Team and is provided "as is" for testing and usage by IRAC observers. In short, the code implements a slightly modified version of the pre-launch algorithm. Performance is better for channel 1 as the channel 2 coefficients are less tuned. The corrector does a good job at identifying and masking pixels potentially affected by muxbleed, but the actual corrections are not accurate enough to suppress the effect below the noise level. The algorithm is still under development and improvements are expected. In particular, the code currently fails in the presence of diffuse emission. Corrected images should be visually inspected. To correct an image, start IDL, compile the modules and execute the fmuxbleed function:

IDL> .run muxbleed_sf
IDL> .run muxbleed_lookup
IDL> .run literstat
IDL> .run fmuxbleed_bcd
IDL> fim = fmuxbleed_bcd('image.fits')

within IDL. The corrected image will be in variable fim'. Optionally you can output the corrected image using

IDL> fim = fmuxbleed_bcd('image.fits', TITLE=title, /WRITE, /VERBOSE)

within IDL, where title is a string to prepend to the input filename (default is "mux_"), and /VERBOSE will provide some more information. Yet more information is available by looking at the comments in the code.

The muxbleed corrector uses a model fit coded in function muxbleed_lookup' to correct affected pixels. For triggering pixels between 10,000 and MUXLEV (default=30,000), the scale factor for the correction is calculated by muxbleed_sf,' using a scale factor derived from the counts of the triggering pixel. For triggering pixels greater than MUXLEV counts, the scale factor is determined from the data.