The dataset consists of wide-band image cubes of the celebrated radio galaxy Cyg A obtained from 7.4 GB of VLA data. The imaged field of view is FoV=2.56 arcmin x 1.536 arcmin, at the spatial resolution 0.06 arcsec, and over the frequency range [4,8] GHz.
We provide wideband imaging results of Faceted HyperSARA (Thouvenin et al. 2020) in comparison with the monochromatic imaging approach SARA (Onose et al. 2017) and the multi-scale multi-frequency deconvolution algorithm implemented in WSClean (Offringa and Smirnov 2017).
We investigate the scalability and precision of the proposed algorithm via two experiments.
Experiment 1: Reconstruction of 15 GB image cube at full spectral resolution (8 MHz) from the full 7.4 GB of VLA data. The cubes obtained by the following methods are provided: Faceted HyperSARA, SARA and the joint deconvolution method in WSClean.
Experiment 2: Reconstruction of 0.94 GB image cube at low spectral resolution (128 MHz) from the reduced 0.22 GB data, leveraging a joint image and data dimensionality reduction technique (DR). The cubes obtained by the following methods are provided: Faceted HyperSARA with DR, averaging over the spectral dimension of Faceted HyperSARA full spectral resolution cubes (Experiment 1) and the joint deconvolution method in WSClean with image dimensionality reduction.
We provide Faceted HyperSARA estimated image cubes (at both spectral resolutions) in FITS format, as well as in AVI format. Imaging results of the benchmark methods are provided in AVI format.