Author Topic: Let's go Supernova Hunting  (Read 68818 times)

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: Let's go Supernova Hunting
« Reply #30 on: August 11, 2009, 01:28:35 pm »
I'm assuming that if the "new" is smaller than the "reference" (and there's no obvious candidate etc), it's unlikely to be a supernova? ;D

I just had one where it looked to me like there was a new bit on the "new", but the target wasn't centred on it. I clicked "no", was I supposed to do that?

It would be really helpful if the instructions could be broken down into a diagram of the basics, like

NEW
= what "new" means,
REFERENCE
what "reference" means,
etc. It would save my eyes ploughing through the very very long paragraphs!

echo-lily-mai

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2125
    • View Profile
Re: Let's go Supernova Hunting
« Reply #31 on: August 11, 2009, 01:29:53 pm »
 ???  i think i need to read it all again !

Art does not reproduce the visible....  Paul Klee

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: Let's go Supernova Hunting
« Reply #32 on: August 11, 2009, 01:31:55 pm »
What on Earth does "Has the candidate subtracted itself clearly?" mean? :o :o Please don't tell me to read the instructions. I have, several times.

Blackprojects

  • Hero Member
  • *****
  • Posts: 9518
    • View Profile
Re: Let's go Supernova Hunting
« Reply #33 on: August 11, 2009, 01:55:28 pm »
Alice

A Scientist probably wrote it and Forgot that Lay People and ZOOITES would have to read it and Make Sense out of it.

So far i have clicked and Clicked and Clicked and Found Nothing that Looks interesting.

Lots of Small Wite Dots in the Centre of the Image but nothing Outstanding Oooh and a few Holes.

zookeeperChris

  • Admin
  • Hero Member
  • *****
  • Posts: 723
    • View Profile
Re: Let's go Supernova Hunting
« Reply #34 on: August 11, 2009, 01:58:10 pm »
What on Earth does "Has the candidate subtracted itself clearly?" mean? :o :o Please don't tell me to read the instructions. I have, several times.

A clean subtraction is one where nearly all of the supernova pixels are positive (or white in these images). Anything else is a sign that something has gone wrong, and you should click 'no' and then chose from the options given to explain what it is that's gone wrong.

For example, instead of a nice white dot, the object at the centre of the subtracted image may show both white and black pixels. These could be randomly scattered, or sometimes arrange themselves into a 'bullseye' pattern. Or, for example, the subtraction routine might just produce a load of nonsense scattered all over the field of view. We only want to keep the very 'clean', all-white candidates.

graham d

  • Hero Member
  • *****
  • Posts: 2408
    • View Profile
Re: Let's go Supernova Hunting
« Reply #35 on: August 11, 2009, 02:29:32 pm »
That's correct: the subtracted image may be irregular, white black stripes, elongated mix of white and black pixels. Sometimes there is a block of several white pixels, square shaped which I regard as rounded, it's not distorted.

The reference image is often much brighter than the new image. What counts is the comparison of brightness of several stars in the reference image that can be compared with the relative brightness of the same star pattern in the new image. Bear in mind that the same star pattern isn't always evident, the field of view may have changed.

I think if you go through the tutorial a couple of times or more and then start classifying you will still feel confused, inevitably. Yet I think the tutorial is actually very well written. Possibly 1 in 50 clicks is a yes. More often than not the subtraction image may be a single white pixel; it's too small. The bulls eye comes up ca. 1 in 10.

Imagine the way that American amateur astronomer discovers new supernovae. He holds the record. He has a 14" Dobsonian he drags out onto his porch and spends several hours a night viewing a restricted patch of the night sky. He concentrates on a rapid look at ca. 100 galaxies per night, but he doesn't have a digital or photograph of those galaxies from day to day for comparison. He went for 2 years without spotting a supernova and then discovered two in three weeks. He notices a sudden increase of brightness at a point against his mental image of the whole galactic pattern. His 14" reflector doesn't redsolve all those stars but he can pick up an increase of brightness by a couple of orders of magnitude. against a background of billions of stars. Back to our samples. Were the new image just 10 fold brighter for one star, the target star, against the relative pattern of brightness , star with star around the target , then the target star of the new image would be very brilliant and after subtraction there would be a very bright circular image, a very positive yes. But that's very unlikely; we are talking about one supernova per galaxy per, you tell me-100years perhaps?

I presume that the images we look at have already been processed as likely candidates from a huge data array from more than a billion stars. When we do click yes, very occasionally, the target star may have less than doubled in brightness relative to the other pattern of brightness in the reference image, rather than an order of magnitude change.

Click away and you may get some feel for it. I compare the star patterns against the new image pattern before glimpsing to the subtracted image. Often a possible yes is negated when you realise that the new image hasn't really changed its brightness against a close star adjacent to it and in the same disposition.


zookeeperChris

  • Admin
  • Hero Member
  • *****
  • Posts: 723
    • View Profile
Re: Let's go Supernova Hunting
« Reply #36 on: August 11, 2009, 02:34:32 pm »
Just a couple of things to add :

1. We've refined the tutorial a bit based on your feedback - keep it coming.
2. The 'candidate' the computer has selected is always at the centre - that doesn't mean there won't be something worth seeing elsewhere, but that is very unlikely.
3. These are indeed the best candidates from a search of many, many galaxies (we target galaxies, not stars - that way we get to monitor billions of stars in a single image).

Keep up the good work, and thanks.

Chris

echo-lily-mai

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2125
    • View Profile
Re: Let's go Supernova Hunting
« Reply #37 on: August 11, 2009, 02:42:17 pm »
I would like a little test the same as zoo 1 to give me some confidence that my answers are going to be correct. (This could be optional for people who are ready to go straight ahead)

i have read the instructions a few times now, ready to try, wish me luck.

Art does not reproduce the visible....  Paul Klee

graham d

  • Hero Member
  • *****
  • Posts: 2408
    • View Profile
Re: Let's go Supernova Hunting
« Reply #38 on: August 11, 2009, 02:45:05 pm »
My last click was a positive. On the penultimate query was the star centred in the galaxy ?; infact the target star in the new image was alone, it's companions were absent so I clicked No.
It didn't like the response and froze me out. I've tried to log back to no avail , it directs me to the Supernova hunting blog every time. It was that last click- must think I'm maliscious..

Good luck Lily- go for it: it evolves as I realised that if you click no a lot, then you get a feel for it.
« Last Edit: August 11, 2009, 02:47:33 pm by graham d »

graham d

  • Hero Member
  • *****
  • Posts: 2408
    • View Profile
Re: Let's go Supernova Hunting
« Reply #39 on: August 11, 2009, 02:51:32 pm »
Back on but?- is the server overwhelmed at the moment by zillions clicking?

echo-lily-mai

  • OotD posters
  • Hero Member
  • *****
  • Posts: 2125
    • View Profile
Re: Let's go Supernova Hunting
« Reply #40 on: August 11, 2009, 03:07:21 pm »
"Has the candidate visibily moved position from the reference image?" How would we know?

Hmmmmmmmm

Art does not reproduce the visible....  Paul Klee

jules

  • Hero Member
  • *****
  • Posts: 8402
    • View Profile
Re: Let's go Supernova Hunting
« Reply #41 on: August 11, 2009, 03:08:41 pm »
On the basis that I think I am finding too many "yes" candidates I am returning to the tutorial. :-\

JohnF

  • Hero Member
  • *****
  • Posts: 3596
  • (Software Engineer, MBCS, MIET, FRAS, FBIS)
    • View Profile
Re: Let's go Supernova Hunting
« Reply #42 on: August 11, 2009, 03:37:19 pm »
I would suggest those, on this thread, who have queries, and who are going to Greenwich on Sunday (with their laptops), is to make a list of these queries and ask the ZooKeepers at Greenwich - then post the answers back on the forum.
John C. Fairweather - http://www.jcfwebsite.co.uk

Alice

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 31782
    • View Profile
Re: Let's go Supernova Hunting
« Reply #43 on: August 11, 2009, 03:40:36 pm »
Won't it be over by then?

For the record, I don't mean to suggest the tutorial isn't well written - it's just set out so that my eyes get lost and frantic when I'm trying to find the answer I want!

jules

  • Hero Member
  • *****
  • Posts: 8402
    • View Profile
Re: Let's go Supernova Hunting
« Reply #44 on: August 11, 2009, 03:44:09 pm »
OK - finding fewer candidates now - but still a fair few. My tip of the day is to zoom the screen - helps get your eye in.

Is there a time limit to this?