Opened 13 years ago

Closed 13 years ago

#268 closed defect (fixed)

new_average fails for Mopra broadband data

Reported by: Malte Marquarding Owned by: Kana Sugimoto
Priority: normal Milestone: Unified development
Component: General Version:
Severity: normal Keywords:
Cc: takeshi.nakazato@…

Description 

Hi Kana,

has there been development of  a method in ASAP to merge adjacent IFs into a single IF in your group?
I had a user request here and wasn't sure if you have been working on this?

Cheers,
Malte

Attachments (4)

broadband.asap.tar.bz2​ (206.2 KB ) - added by Malte Marquarding 13 years ago.
Broadband 4 IF data
t.py​ (166 bytes ) - added by Malte Marquarding 13 years ago.
test script
broadband.png​ (53.2 KB ) - added by Kana Sugimoto 13 years ago.
The original input data 'broadband.asap'
broadband_aveall.png​ (57.3 KB ) - added by Kana Sugimoto 13 years ago.
spectra averaged by asapmath.new_average

Download all attachments as: .zip

Change History (9)

by Malte Marquarding, 13 years ago

Attachment: broadband.asap.tar.bz2​ added

Broadband 4 IF data

by Malte Marquarding, 13 years ago

Attachment: t.py​ added

test script

comment:1 by Malte Marquarding, 13 years ago

Attached is the data and the file which fails the method.

comment:2 by Malte Marquarding, 13 years ago

tracking in Ticket now rather than email.

comment:3 by Kana Sugimoto, 13 years ago

Status: new → assigned

I found that new_average does not run properly in case spectra has negative increment in FREQUENCIES table. I started fixing the bugs. It may take a few more days to fix them.
Malte, is the time scale fine for your release?

comment:4 by Malte Marquarding, 13 years ago

Sure, I am on holidays until the 18th.

by Kana Sugimoto, 13 years ago

Attachment: broadband.png​ added

The original input data 'broadband.asap'

by Kana Sugimoto, 13 years ago

Attachment: broadband_aveall.png​ added

spectra averaged by asapmath.new_average

comment:5 by Kana Sugimoto, 13 years ago

Cc: takeshi.nakazato@… added
Resolution: → fixed
Status: assigned → closed

I believe I've fixed Scantable::regridChannel, Scantable::regridSpecChannel and STMath::new_average so that asapmath.new_average works for data which contains IFs both with positive and negative spectral increments. Please find the attached figures of plots of spectra before and after new_average(compel=T).
I've tested new_average by checking numerical values of an averaged spectrum using a simple artifical input spectrum. The result had a good agreement with the analitical solution.

However, there is still a known issue of asapmath.new_average. The method raises array comformance error when the input scantable has multi-channel Tsys columns. Takeshi agreed with taking care of it.

Summary of commits:
Commit dates: 2012/04/12, 13, 16
Revisions: r2462-2463,and r2466-2467
Modified Codes: Scantable.cpp, STMath.cpp

Note: See TracTickets for help on using tickets.