About the CM 4476661V1 and its sisters

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

Moderators: Stardust@home Team, DustMods

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

About the CM 4476661V1 and its sisters

Post by McAngus » Tue Mar 30, 2010 11:05 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 clicked on NO TRACK and I obviously got a negative score.
Later on I saw the post of zioriga and all the others replies.

If someone didn't meet the movie CM 4476661V1 I add three images: the first jpg, the median and the last jpg of this movie.

CM 4476661V1 jpg 1
Image
CM 4476661V1 jpg 19
Image
CM 4476661V1 jpg 37
Image


And moreover : Why in the movie CM 90731V1 and CM 1216593V1 the same track assigns a positive score? (See the images below)

CM 90731V1
Image
CM 1216593V1
Image

Below, detail of the three tracks:

CM 4476661V1 jpg 37 detail
Image
CM 90731V1 detail
Image
CM 1216593V1 detail
Image

Comparison of the three tracks
Image

So is there someone who knows the correct coordinates of the true track in the CM 4476661V1 ?

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: About the CM 4476661V1 and its sisters

Post by jsmaje » Tue Mar 30, 2010 11:58 am

McAngus wrote:So is there someone who knows the correct coordinates of the true track in the CM 4476661V1 ?
Well, as explained here, I was scored 'correct' by clicking on the little blob on the very top edge of the frame evident in your CM 4476661V1 jpg 37 & detail, at coordinates 365,1 (sorry I didn't specify these in that posting).

John

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

Re: About the CM 4476661V1 and its sisters

Post by McAngus » Wed Mar 31, 2010 2:51 am

Thanks jsmaje, you are right.

McAngus
Image
ad augusta per angusta

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

Re: About the CM 4476661V1 and its sisters

Post by jsmaje » Thu Apr 01, 2010 1:08 pm

I hope McAngus will forgive me for sharing his personal comments to me via PM regarding 4476661V1, which I think are important, particularly: "Why in the CM 1216593V1, 894906V1, 1479606V1, 90731V1, 8892013V1, 5865706V1 (but I have seen some more of them) we must click on the 'globe' of the same track, even if scaled, and in CM 4476661V1 we must click on 'dust on camera'?"

By 'dust on camera' he is of course referring to our old friend "Larry" the lens-dust stooge (see here and here).
I have to admit having failed to notice that the blob I clicked on at 365,1 is indeed most likely to be Larry (even though I was scored correct). Looking at McAngus's other movies above, at least some of which do appear to contain the same pasted sample, I would otherwise have clicked on the very vague terminal feature arrowed in each here (and in all similar cases so far have been credited as 'correct').

Image

The equivalent in 4476661V1 is indicated below (black arrow), at 353/10, 12/9 pixels from Larry (grey). Meanwhile, the prominent 'globe' feature (white) is centered at 336,23, 13/13 pixels from the former:

Image

Could my click at 356,1 in 4476661V1 therefore have been fortuitous; could in fact the radius of tolerance be 12 pixels?
Hence my request elsewhere to Dan for clarification of the matter.

John

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

Re: About the CM 4476661V1 and its sisters

Post by DanZ » Thu Apr 01, 2010 7:28 pm

jsmaje wrote: Could my click at 356,1 in 4476661V1 therefore have been fortuitous; could in fact the radius of tolerance be 12 pixels?
Hence my request elsewhere to Dan for clarification of the matter. John
I think I was told 30 pixels, but I will confirm and let you know.

Dan

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

Re: About the CM 4476661V1 and its sisters

Post by DanZ » Thu Apr 01, 2010 7:39 pm

Confirmed, it is 30 pixels.

Dan

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

Re: About the CM 4476661V1 and its sisters

Post by McAngus » Sat Apr 03, 2010 12:43 am

Hi jsmaje.

At this point I propose a "summary" of the matter.

Below you find the image.

Image

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: About the CM 4476661V1 and its sisters

Post by jsmaje » Sat Apr 03, 2010 1:27 pm

McAngus - I think your image above explains the situation very well, where my click on Larry seems to have been (only just) close enough to the real terminal particle to be scored correct, while the more obvious 'globe' is too far away (I must correct the x-coordinate disparity between the latter two that I posted as being 13; it is of course significantly greater at 17).

This is consistent with your 30 pixel diameter areas of tolerance, meaning a 15 pixel radius. Dan, did you in fact mean 30 pixel diameter rather than radius, which was what I'd asked for and McAngus has clearly assumed?

I see that we're also now told that the ubiquitous central 'globe'-like blob in these movies is in fact a "blemish in the aerogel"! Would that be on the surface, or at what depth, and on what evidence? It seems to regularly focus at the same sub-surface level as the pasted calibration tracks, the terminal track/particles assuming a slightly divergent trajectory after this feature, as also seems to be the case with the analogous prominent feature in all the other differently-proportioned CMs I've so far come across.

John

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

Re: About the CM 4476661V1 and its sisters

Post by DanZ » Wed Apr 14, 2010 11:20 am

The true coordinates of 4476661V1 are H365,V5. Our tech didn't have time to confirm, but he thinks it's a 30 pixel radius. Nonetheless, when he has time he'll double check and get back to me. We'll just have to be patient.

I was told the "big blob" is some messed up aerogel, but I've asked again. Regardless, whatever it is, it's not the terminal particle, but because it was there to begin with, it was included for every cut and paste used for the CMs.

In 4476661V1, it appears the terminal particle ended up underneath, or possibly just beyond, "Larry." It's only a coincidence and wasn't done on purpose.

If I get more info, I'll report it here. But please understand, the scientists and technicians on this project are very busy analyzing particles and trying to get more tiles scanned. So as long as dusters now know where to click, they're not too concerned with blemishes or how much leeway are in the CMs. So, answers to that level of detail may take awhile to get.

Thanks,

Dan

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

Re: About the CM 4476661V1 and its sisters

Post by DanZ » Wed Apr 14, 2010 5:34 pm

DanZ wrote:I was told the "big blob" is some messed up aerogel, but I've asked again.
Here's your answer John!:

"This duster is correct: the condensed aerogel blob is in the track. It is part of, and formed with, the track. It's not a blemish, which sounds more like a defect. We know it's condensed aerogel because we analyzed it at the STXM beamline at ALS 11.0.2, LBL. The details of the analysis are in Andrew's Update of narrated slides [see number six at the bottom]. In the slide-set is a composite of x-ray absorption images making up a picture of the entire track, including the blob. We see condensed aerogel (i.e. same composition, but more dense than surrounding aerogel) all the time in the cometary tracks. Every track has these condensed aerogel blobs in the (much larger) track walls. In the case of Orion, we don't know the impact speed or the impactor's mass, so it's difficult to model the physics that causes the mid-track aerogel re-deposit that we've seen in both Track 29 & 30. We don't yet have any lab analogs of midnight tracks."

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

Re: About the CM 4476661V1 and its sisters

Post by DanZ » Fri Apr 16, 2010 10:34 am

jsmaje wrote:This is consistent with your 30 pixel diameter areas of tolerance, meaning a 15 pixel radius. Dan, did you in fact mean 30 pixel diameter rather than radius, which was what I'd asked for and McAngus has clearly assumed? John
According to our programmer, it has to be radius. The code is MIN_DIST from the true X,Y coordinates, and MIN_DIST is set to 30. Now, if you're still getting calibration movies wrong, let me know and I'll try and help. But if you still think it's diameter, then so be it, I can't dig any deeper than this - the programmer considers the matter closed.

Dan

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

Re: About the CM 4476661V1 and its sisters

Post by jsmaje » Fri Apr 16, 2010 12:08 pm

Thanks for your 'digging' Dan, and if I'd done my own a bit more thoroughly I'd have come up with this statement by Brian Mendez on Aug 16 2007: "You must click within a 30 pixel radius of the simulated tracks in order for the click to count as correct." This generous radius has obviously remained unchanged.

If the terminal particle in 4476661V1 was placed in a similar position to 'Larry', as you say above, this would obviously explain why I was scored correct (and would also have been for clicking the more subtle 'blob' between Larry and the central 'globe'), while McAngus was scored wrong for clicking on the 'globe', which happens to be just a couple of pixels further than 30 radially away from Larry.

OK, matter clarified and closed!
John

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

Re: About the CM 4476661V1 and its sisters

Post by McAngus » Sat Apr 17, 2010 7:23 am

"jsmaje wrote: ... and closed!"

Hi jsmaje,
you have "burned" the reply I was going to post, but I accept it.

I think that, at this point, we have to celebrate the closing of this topic
and then I have the pleasure to invite you (obviously I will pay the bill) to my favorite restaurant
(if you don't know it, you can find "here").
If you agree we can invite also Daniel to join us. But just under one condition.
He must not put a cherry inside. I don't approve this.
I only admit this: Image shaken not stirred.


" ....... what should be invented to make you laugh a little." Eddie Murphy
:lol:
ad augusta per angusta
Image
ad augusta per angusta

Post Reply