NRAO Home  >  Green Bank  |  Wiki Topic:    GB > Software > ModificationRequest11C306

Changes | Index | Contents | Search | Statistics | Go

Add Setting of Spectral Processor Parameter, taper, to the Configuration Tool

Modification Request #11 (C3 2006)

• 1. Introduction
• 2. Background
• 3. Design
• 5. Deployment Checklist
• 6. Test Plan
• Signatures
• CCC Discussion Area

1. Introduction

2. Background

1. Box : a window function that has a value of 1 across all lags. This provides the best frequency resolution, best sensitivity, but at the cost of more 'leakage' compared to the Cosine taper. Since the SP is an FX spectrometer, a channel in the SP has a frequency response of Sinc(x)^2. In contrast, a XF spectrometer like the GBT Spectrometer has a Sinc(x) response. Leakage for the SP manifests itself as a contamination of channels adjacent to an RFI or other strong spectral spike and not the 'ringing' kind of leakage we're familiar with in our Autocorrelators.
2. Cosine : A windowing function that also goes by the more familiar name of Hanning. It further reduces the leakage from RFI or strong, narrow lines but at the cost of sensitivity and a frequency resolution that is twice that of using the Box taper. As we noted last time, "Boxed" data that is Hanning smoothed in an analysis system very closely mimics what the Cosine taper would have done.
3. Half-Box : The SP documentation states that Half-box is only meant for tests. It's a windowing function that is 1 for the middle half of the 'lags' and zero elsewhere. This reduces the sensitivity by 50%, doubles the frequency resolution, but does not improve leakage significantly. To me, it's a very, very bad taper to use for astronomy.

We no longer wanted the default of Hanning windowing when the Aips++ and SDFITS filler processed Spectrometer data.

The observing group decided (ObsIssues25Jan06) the config_tool should set the taper to "box" always, as we do not support any of the other types of tapers. If a users desperately wants to use another mode, he/she can always do so through the "SetValues" command.

Currently, if one user changes the taper, subsequent users will inherit that setting. Only by studying the CLEO screens will someone discover they are using an unintended taper. We've essentially been lucky that no one has ever used anything but "Box". Someday, we'll probably be bitten if config_tool isn't augmented to always set the taper.

3. Design

I believe it is just the SpectralProcessor(taper) that needs to be set to "Box".

5. Deployment Checklist

This change simply needs to be announced in the cycle release notes.

6. Test Plan

Using CLEO, set the taper to any value but box. Run the a configuration for the SP in astrid and insure it chages to "box".

Signatures

APPROVED: I acknowledge that my request is fully contained in this MR, and if the SDD delivers exactly what I specified, I will be happy.

ACCEPTED: I acknowledge that I have validated the completed code according to the acceptance tests, and I am happy with the results.

Written	- KarenONeil - 06 Apr 2006
Checked	- MelindaMello - 06 Apr 2006
Approved by Sponsor	- KarenONeil - 06 Apr 2006
Approved by CCC	- JohnFord - 10 Apr 2006
Accepted/Delivered by Sponsor	- KarenONeil - 06 Apr 2006

Symbols:

• Use %X% if MR is not complete (will display )
• Use %Y% if MR iscomplete (will display )

CCC Discussion Area