Problems with a positive calibratio

Post here if you are having any kind of problem with the Stardust@home website.

Moderators: Stardust@home Team, DustMods

Post Reply
zioriga
Posts: 53
Joined: Sat Jun 10, 2006 1:00 pm
Location: Moncalieri, Italy

Problems with a positive calibratio

Post by zioriga » Wed Mar 24, 2010 7:01 am

I found a problem with one or two positive calibration.

For the last one (the 1552881V1) I'm 99% sure that the right coordinates are near 470,220. But I got an error !!!
The previous (may be the 4476661V1!! I'm not sure this is the correct number) where the track is as big as a "house" (near 330,20) I got another error.

DanZ
Site Admin
Posts: 776
Joined: Fri Feb 27, 2009 3:44 pm
Location: Berkeley, CA
Contact:

Re: Problems with a positive calibratio

Post by DanZ » Wed Mar 24, 2010 10:24 am

zioriga wrote:I found a problem with one or two positive calibration. For the last one (the 1552881V1) I'm 99% sure that the right coordinates are near 470,220. But I got an error !!! The previous (may be the 4476661V1!! I'm not sure this is the correct number) where the track is as big as a "house" (near 330,20) I got another error.
Here's my assessment of 4476661V1. At first glance, there appears to be an obvious track headed off to the left (towards where you clicked). But I don't think that's it! Look very carefully in other directions, and I think eventually you'll find the actual path. By the way, 5 other people have clicked on this movie so far, and no one else has gotten it right yet either.

I can't seem to pull up a movie for 1552881V1 - are you sure the number is right?

Dan

zioriga
Posts: 53
Joined: Sat Jun 10, 2006 1:00 pm
Location: Moncalieri, Italy

Re: Problems with a positive calibratio

Post by zioriga » Wed Mar 24, 2010 10:27 am

Sorry the right number is 1552821V1 !!!

Vacuity
Posts: 2
Joined: Tue Mar 23, 2010 8:47 am

Re: Problems with a positive calibratio

Post by Vacuity » Fri Mar 26, 2010 9:21 am

DanZ wrote:By the way, 5 other people have clicked on this movie so far, and no one else has gotten it right yet either.
4476661V1 is a bit off. People say they have tried both the left end of the track and the blob and both give wrong answers.
The track appears to be this one identified in this thread and I think the target is the deflected particle at the end of the track.
Image

The problem is the the movie demands too much precision concidering how out of focus the track is. The bad focus gives two judgements:

A. The particle (at 352,12 in the calibration movie) never becomes clearer than a faint halo. As there is also a feature on the blob that makes a similar halo appear sligtly above it (at 336,16) this gives the appearance that the two halos are just ghost images of the blob so people mark the blob (at 336,25).

B. The further you focus down the more of the left part of the track comes into focus, giving the illusion it is deeper there, making some (including me) interpret it as a carrot track leading left with the blob at the thicker part of the track and mark the left end (at 280,45).

No one can tell the target particle is there unless you know it beforehand.


Now for a different issue concerning the design of this particular CM: Why does the track just pop up 120+ microns into the tile? Wouldn't a real track start at the surface and lead all the way down there, or can a particle really make it that far into the aerogel before it leaves a visible track? This in reagards to question I asked here.

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: Problems with a positive calibratio

Post by jsmaje » Fri Mar 26, 2010 5:19 pm

Vacuity wrote:
DanZ wrote:By the way, 5 other people have clicked on this movie so far, and no one else has gotten it right yet either.
4476661V1 is a bit off. People say they have tried both the left end of the track and the blob and both give wrong answers.
The track appears to be this one identified in this thread and I think the target is the deflected particle at the end of the track.
Image
Here is the relevant portion of 4476661V1 at deepest focus and same magnification as A & B:

Image

I agree with Vacuity that 4476661V1 is a "bit off", i.e. poorly focussed. Not only is the putative track never fully focussed, the surface of the aerogel in fact never appears throughout the whole depth of the movie (we seem to be sub-surface all the way the down). For that reason, when I first came across the same movie I nearly clicked on 'Bad Focus' without further thought, but on noticing the top-right feature and its resemblance to the track in A & B (having now found it to be ubiquitous), took a closer look. Whilst this feature is blurrier, larger and more extended (whether having been artificially manipulated or being a different sample), I eventually decided to click on the vague small, highest blob right on the very edge of the frame (arrowed), being the closest equivalent to that which I've found is deemed 'correct' for A/B (the top-left in B), and conceivably the end of the track. And I was scored correct! Only later did I find DanZ had hinted at this.

How come then that greuti has has found clicking on the most prominent blob(s) in A & B is acceptable; would this in fact be true in the case of the more extended 4476661V1?

Regarding zioriga's other similar-looking movie, 1552881V1, from the reasoning above I'd have clicked on the small(er) blob at coordinates 480,214 (470,220 seems to lie lower than and between both visible blobs, but hardly a long way away. Can we be told the 'correct' coordinates?)

It's for all these reasons that I've asked what radius of tolerance in pixels the team take to be acceptable (and around what exactly).

Meanwhile, as Vacuity points out, there seems to be no evidence in 4476661V1 that the supposed track has extended from the surface, (a) because it peters out quite quickly into featureless areogel at all levels, (b) has been pasted so low as to be legitimately considered artefactual, and (c) because we never in fact see the surface :roll: !

John

McAngus
Posts: 229
Joined: Sun Nov 05, 2006 8:38 am
Location: Italy (Palermo)

Re: Problems with a positive calibratio

Post by McAngus » Sun Mar 28, 2010 11:35 am

I meet the naughty CM 4476661V1 (where appears a really very big track) since the testing phase in the "jake" site.
I clicked on it and me too got a negative score.
I encountered again the movie in phase 3 and I (really shrewd) clicked NO TRACK and I obviously got a negative score.
Later on I saw the post of zioriga and all the others replies.

Since I were a kid I'm using only one eye, the left is defined "oculus piger" (lazy eye). (Don't cry: the right is very very good!).
So I ask all of you, dear colleagues that have in use both eyes, to tell me where is the track.

If someone didn't meet this movie I add three images: the first jpg, the median and the last jpg of this movie.
4476661V1_jpg1
Image
4476661V1_jpg19
Image
4476661V1_jpg37
Image


And moreover : Why in the movie 1216593V1 the same track assigns a positive score? (See the images below)
1216593V1
Image
4476661V1_detail
Image
1216593V1_detail
Image
4476661V1_1216593V1
Image
Thanks.
McAngus

ad augusta per angusta
Image
ad augusta per angusta

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: Problems with a positive calibratio

Post by jsmaje » Tue Mar 30, 2010 12:15 pm

McAngus - see my post above yours, and response to your virtual cross-posting as a separate subject here. The coordinates of the arrowed feature in the former post, which was scored 'correct', are 365,1.
John

jsmaje
Posts: 616
Joined: Tue Aug 15, 2006 8:39 am
Location: Manchester UK

Re: Problems with a positive calibratio

Post by jsmaje » Thu Apr 01, 2010 1:37 pm

Please see my further (? wiser) thoughts about 4476661V1 on McAngus's sister site here.
John

DanZ
Site Admin
Posts: 776
Joined: Fri Feb 27, 2009 3:44 pm
Location: Berkeley, CA
Contact:

Re: Problems with a positive calibratio

Post by DanZ » Thu Apr 01, 2010 7:19 pm

Seems you are all catching on :D

If it helps, I've been informed that the confusing "big blob" is just a blemish in the aerogel.

Otherwise, recent posts to note include:

1)
jsmaje: I eventually decided to click on the vague small, highest blob right on the very edge of the frame (arrowed), being the closest equivalent to that which I've found is deemed 'correct' for A/B (the top-left in B), and conceivably the end of the track. And I was scored correct!
2) Reference to new calibration movies here.

The important thing, however, is to keep clicking on what appear to be tracks. If the movies are "real," and enough people have clicked on them (regardless of where), we'll take a look!

Dan

Post Reply