Error propagation on PVDs? and on spectrum coming from PVD?

[renovatio]

Dear all,

I was wondering can anyone share their opinion with me about the error propagation on PVD. I think the way of creating PVD is important here.

I first rotated the original (cleaned) cube using 'regrid' with the 'rotate' keyword that the emission lies in the E-W direction on the sky. I then defined a rectangular box where the whole emission is located. The rectangle has 11 pixels in N-S direction. After summing all the pixels inside the box (using IDL total function), you end up with a nice PVD diagram. However, the question is what is the RMS on this PVD? How does the error propagate? Is it sqrt(11)*RMS? -RMS is the rms in the original cube obtained by calculating the standard deviation of emission free pixels in the cube- 

In the literature people generally use 2*RMS or 3*RMS when they define the contour levels on PVD. Maybe this is ok for only showing PVD contours. However, I am doing something additional, obtaining spectra from the PVD. I take an average of 5 pixels (through the X-axis) per velocity channel (through the Y-axis) on PVD. When you do this for all velocity range you end up with a spectrum of that particular slice on PVD. If you do this through the X-axis for each 5 pixels (it is one beam in my case), you will have all the spectra over the galaxy major axis. SO, the error of the spectrum will be different too, won't it? 

Thanks a lot for any comment/suggestion in advance!


Best,

[Peter Teuben]

you first need to understand the sampling of your image (which comes out
of invert,
ie how many points per beam), as well as how well the velocity axis is
sampled.
If you haven't done any regridding yourself in velocity , most miriad
data will
be hanning smoothed. After that, doesn't matter if you take idl::sum() or
miriad::moment() , you will know how many pixels are involved in the summing
and basic stats should give you the RMS in the resulting image/plot and you
can adjust your contours to that. It's just a matter of taste if you
contour at
2RMS or 3RMS. But it would also be good to predict what your RMS
should be,
and compare that with what you measure.

peter

> --
> You received this message because you are subscribed to the Google
> Groups "miriad" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to miriad+un...@googlegroups.com.
> To post to this group, send email to mir...@googlegroups.com.
> Visit this group at http://groups.google.com/group/miriad?hl=en.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>