Resources for Hubble’s Tiny Tim Tool
Tiny Tim (Krist & Hook 2011) is a point spread function modeling tool. Details of the code's operation, configurable parameters, and products are given in the Tiny Tim User's Guide while the recent updates to characterize the Wide Field Camera 3, installed in May 2009 are described by Hook & Stoehr 2008.
The Tiny Tim source code presents the following input choices:
- Camera: You may choose to model the PSF specific to a chosen channel for any of the HST imagers past and present. Available options are given in the menu.
- Chip: In some cases you will then be asked to provide the specific detector of interest. Applicable choices are given in the menu. Note that the image file format places ACS/WFC1 in science image 2 and ACS/WFC2 in science image 1. Similarly for WFC3, UVIS2 is in the first image and UVIS1 in the second
- Pixel Position: The PSF is field-dependent so the chip location of interest is input. Appropriate range is offered.
- Filter: PSF is filter dependent. Only applicable filters are given in the menu.
- Spectrum Type / Spectrum Value: Methods for calculating flux over the band are offered. A menu of stellar spectral types, or a field for BB temperatures or power law indices are then offered if applicable.
- PSF diameter: Output PSF image extent. ~3" is recommended.
- Focus: User-supplied focus value in microns at the Secondary Mirror. The best estimated focus for a time period of interest is provided by the focus model. The focus model output can be used directly in this field without unit conversion. Entering zero for focus in this field will result in the generation of a PSF with no wavefront error in focus.
The Tiny Tim source code produces an output web page (example) containing:
- an image illustrating the modeled PSF characteristics
- a summary of the input parameters
- links to the output data files, including the fits images of the distorted and undistorted modeled PSF
- links to the output figures
- a link to the tar.gz file that bundles all the output products for download.
See the Tiny Tim User's Guide for a description of the output products. For ACS and WFC3, PSFs are calculated for both undistorted (tiny2 output) and distorted (tiny3) images. Only the tiny3 distorted image is sampled at the detector resolution. This product corresponds to the images supplied by the pipeline as "_flt" files. For the other instruments, the distortion is much less and has not been considered in Tiny Tim.
We note that at this time Tiny Tim does not model the PSF found in multidrizzled (_drz) files. It is possible that this enhancement could be prioritized in the future, given resources and depending on community demand.
The generated PSFs can be useful for image simulation and observation planning, or to indicate the level of PSF variability over the field and with time, or for small aperture photometry. For PSF subtraction where details of the outer portions of the distribution are important, and for other cases where accurate PSF characterizations are needed, these modeled PSFs may lack suitable accuracy, and empirical PSF techniques should be considered when practical.
Future refinements to the models may enable their more effective use for other data reduction tasks. PSF fidelity will also be found to vary among instruments. For example the ACS Tiny Tim PSF is believed to be better modeled currently than the WFC3. For WFC3, Tiny Tim currently uses aberration coefficients that are derived from the as-built optical ray-trace model, as a function of field position. While the ground test and on-orbit image quality generally matches expectations, no attempt has yet been made to adjust the aberration coefficients used by Tiny Tim to best describe the observed field dependence. This refinement may be pursued in the future using phase retrieval results for the lowest order aberrations from a large ensemble of on-orbit observations.
Tiny Tim also uses blur kernels to describe the effects of charge diffusion and, for the IR channel, inter-pixel capacitance. The IR and wavelength-dependent UVIS kernels have all been updated from those listed by Hook and Stoehr, to better match more recent pre-launch measurements. See Hartig, ISR WFC3-2008-40 and ISR WFC3-2008-41.
The modeling engine and its modifiable calibration parameters are more fully discussed in the Tiny Tim User's Guide.
Tiny Tim author Krist provides some personal science examples of earlier WFPC2 Tiny Tim model application, and summarizes the tool's use over 20 years of HST science operations in Krist & Burrows (2011). Also see Rémy et al. 1997.
As stated elsewhere, however, we are very interested in feedback from the community on the current incarnation of the PSF and focus models, especially regarding any relative performance comparisons between them and earlier (or empirical) attempts.