Radio Galaxy Zoo Talk

Why does the radio noise have that lattice-like structure?

  • vrooje by vrooje admin, scientist

    I think there will be a blog post about this soon, but in the meantime, any radio experts want to weigh in?

    (Note: you may have to click on the image to get the sliders to appear on the original subject page.)

    Posted

  • DocR by DocR scientist

    Great question! These data come from radio interferometers, which can be thought of as a giant mirror that has been cracked into a bunch of pieces. This creates all kinds of distortions, like looking through a fine mesh window screen at a distant street light. Radio astronomers do their best to remove these distortions, but often some remain. The FIRST Survey (Faint Images of the Radio Sky) has some of the strongest lattice-like distortions remaining.

    Posted

  • KWillett by KWillett scientist, admin, translator

    The structure in any image is also partly determined by the shape of the telescope that took the images. In the Very Large Array (VLA) telescope, where the radio data are taken, the dishes are laid out in a big Y shape. As a result, distortions in VLA images tend to have structure that looks like 6-pointed stars (the Y-shape mirrored on top of itself).

    Other telescopes like the Australia Telescope Compact Array (ATCA), which will appear in future RGZ data sets, are arranged in a straight line. As a result, their noise patterns tend to look different. See if you can spot the difference!

    VLA from above (Y-shaped): http://goo.gl/maps/LGPir
    ATCA from above (straight line): http://goo.gl/maps/3R2Md

    Posted

  • enno.middelberg by enno.middelberg scientist, translator

    One more comment on the strange noise structure of the radio images: we're looking at images which are VERY different from CCD images. One can think of them as the superposition of many, many waves with different frequencies, amplitudes, and orientations across the image - think of throwing a handful of pebbles with various sizes into a still pond, and watching the resulting wave pattern. Because we still need to make images which consist of pixels, we have to "grid" these waves, and the pixels need to be much smaller than the wave with the highest frequency (Nyquist sampling theorem). Therefore adjacent pixels in the image are not independent from one another, something which is known as "correlated noise". In a CCD image, individual pixels are more or less independent from one another, and that's why they look "noisier".

    Posted

  • jbroekma by jbroekma

    I have to say that the radio noise takes me back to my undergraduate days, trying to pick out compact sources for David Helfand's galactic plane survey. Good times, good times. 😃

    Posted

  • smo by smo

    Radio images probably show emissions as log(Power).scaled, so anything between the noise (instrument and other kind) and a signal with cojones will be there. Artifacts should not keep one from noting the essential - if it is there.

    smo

    Posted

  • JeanTate by JeanTate in response to enno.middelberg's comment.

    What, then, does a true radio point source look like? Assume it's moderately strong, and in an otherwise blank part of the sky (I know, I know, highly unrealistic assumptions!). In particular, does its apparent shape change, depending on ... ?

    Posted

  • KWillett by KWillett scientist, admin, translator

    If you were able to do a perfect job removing all of the emission away from your point source (which are called sidelobes), you retrieve what we call the synthesized beam shape. This is usually an elliptical shape - its exact size and shape will depend on several factors: the spacing of the telescopes, the elevation of the object in the sky, and how long you were able to integrate on it.

    Many images of objects in the radio will actually have a little inset showing what their synthesized beam was for this observation; that lets readers know what kind of structure in the image is physical, and which are likely to be noise artifacts. Here's an example from Chyzy et al. (2007) - that small circle in the bottom left is a measure of what a true point source would look like.
    NGC 4254, as viewed in 21-cm HI gas

    Interferometers are definitely tricky to grasp intuitively; the basic principles rely heavily on understanding Fourier transforms. There's some nice explanations in these online notes, though, if anyone's looking for a good resource. Essential Radio Astronomy

    Posted

  • NGC3314 by NGC3314

    There's a quick comparison of a snapshot VLA observation with its computed point-source response and the deconvolved result here:
    http://astronomy.ua.edu/keel/telescopes/nrao.html
    (scrolling past images of the VLA itself). The point-source response ("dirt beam") can be computed from the distribution of the antennae and the direction of observation; an echo of this pattern will therefore appear due to such extraneous things as interference (which will be attributed to a specific baseline) or changes in detector properties. Flagging such bad data can be a major part of processing radio-interferometry results. By and large, by the time you see the data a "pipeline" deconvolution has been applied to remove the effects of the point response, but low-level artifacts can be tougher to track down in batch mode (and the eye is really good at spotting patterns down in the noise level).

    Posted

  • ivywong by ivywong scientist, admin

    Thanks heaps @KWillett and @NGC3314 for the nice pointers to radio interferometry. @enno.middelberg has now written a 3-part blog on the guts of radio interferometry for those of you wanting more details. Check out Part 1 in today's blog:

    http://blog.galaxyzoo.org/

    cheers,
    Ivy

    Posted