Stardust@home Suggestion Thread

Discuss your experiences with and ideas about Stardust@home here.

Moderators: Stardust@home Team, DustMods

Post Reply
Franz
DustMod
Posts: 23
Joined: Fri May 19, 2006 6:27 pm
Location: Netherlands

Image Quality

Post by Franz » Sat May 20, 2006 5:47 am

The technical FAQ states some interesting facts:
- Images are in JPEG format
- A single image is about 12KByte large
- 800x600 screen size is considered a tight squeeze
The tutorial uses 601x451 pixel ~12MByte images.

The tutorial gives a strong signal: find focus in a blurred world. The positive trails can be characterized as being 'in focus' (sharp) on several depths.

However, the used JPEG images can be considered as being lossy low quality. Using low quality JPEG imagery removes 'high-frequency components' resulting in loss of detail/sharpness and artifacts. (as reference: High quality JPEG images at that resolution lie around 100KByte. Lossless TIFF can even be 3x larger).

Artifacts are clearly visible as distortions around text, and I suspect that objects that should be sharp get blurred like other near-out-of-focus objects making recognitions much more difficult.

With this in mind I have a couple of questions:
- What is the resolution of the scanning microscope?
- Are presented images being downsampled?
- What is the rationale behind choosing the current JPEG format?
- Is there an option to examine high-quality (lossless) imaginary?

I can imagine that using small-sized images is a trade-off for the otherwise long download times, especially on slow bandwidth connections. However I imagine that many volunteers have high-speed connections and having such a connection is also suggested by the technical FAQ.

I would like to give a quality answer and am willing to invest time and energy in doing so. However, I feel that I am not able to give such an answer based on lower-quality material.

Please make the quality of presented images configurable so that *we* can choose our trade-offs. Please also present some high-quality imaginary in the tutorial so that we can see what the effects of lower-quality imaginary are.

Sun Tzu II
Posts: 14
Joined: Sun May 21, 2006 3:11 am
Location: Pearl of the Orient

Post by Sun Tzu II » Sun May 21, 2006 4:29 am

Sigh... I don't like JPEG too, prefer JPEG2000.

Unfortunately there are not many browsers (if any) support it.
Amazing how much interest specks of dust can spark

PeteSeeker
Posts: 36
Joined: Wed May 24, 2006 8:51 pm
Location: Upstate, NY
Contact:

Post by PeteSeeker » Wed May 24, 2006 9:16 pm

yes, it would be nice to have an image that was much bigger than 600 X 400 although I do like the interface. A choice between image sizes would be quite nice as those of us with more bandwidth/patience can choose a higher resolution to look at. But then again, this is considered a first pass of sorts and it is important to go over as many of the images as possible to weed out the ones that are not suspected of having a trace or rather, to identify the ones that may have a trace and thus, warrant further scrutiny. Still, it would be nice to be able to choose as Franz said…
Image

Rojer
Posts: 31
Joined: Thu May 18, 2006 7:34 am
Location: China

Translate this project into non-English

Post by Rojer » Sat Jul 29, 2006 6:58 pm

I am one of the members of the translation team of EQUN.com, the biggest distributed computing information website in Chinese. We have translated the Stardust@home website into Chinese.

Our website was built to provide distributed computing information in Chinese. We are mainly voluntary university students and have successfully translated some websites, including the SETI@home classic project organised by Berkeley :)

The Stardust@home project has been reported many times in the local media and obviously many people are interested in it! We hope that a Stardust@home Chinese site will serve Chinese people better(because most of them can not read English well) before and after the official launch of the project!

The Stardust@home Site in Chinese is hosted on:
http://www.equn.com/stardust/
We hope that officials can add a link on Strdust@home website, or make a language choice on it.

I'm expecting for your reply. :D
Wish you can understand my English. :)
Image

renboeverywhere
Posts: 14
Joined: Fri May 19, 2006 11:49 pm
Location: Dalian,China

Post by renboeverywhere » Sat Jul 29, 2006 8:39 pm

Well Rojer,I've been waiting for such a post for so long, and at last!
I've felt what you've felt, and I suggest strongly what you suggest.
Hope the moderators consider it seriously.
And, thank you for your efforts ,Rojer.
When nobody becomes everybody, nothing will become everything.

gcarlsen
Posts: 23
Joined: Thu Jun 01, 2006 10:22 am
Location: Arizona

Please wait!

Post by gcarlsen » Mon Jul 31, 2006 9:03 am

This is an excelent project, and it looks like the first steps to implementation are very simple and quick.
BUT, please wait until the project with the current scope is launched before attempting a enhancement with the potential complexity of an enhancement like this.

Just my two cents.

George
99 percent of lawyers give the rest a bad name.

simon
Posts: 14
Joined: Tue May 23, 2006 2:59 am
Location: China

Post by simon » Mon Jul 31, 2006 9:34 pm

yeah,quite a good idea.
Image

gr8tfly
Posts: 3
Joined: Wed Aug 02, 2006 3:09 pm
Location: Goleta, CA

VM Interface Suggestion

Post by gr8tfly » Wed Aug 02, 2006 3:48 pm

Just a thought: I would find it very informative to have some feedback on the relative depth while focusing (Z-axis, like 0µm to 140µm). Even just a static vertical scale alongside the focus control would be nice. But, a label for the active focus "slice" would be better.

As I'm writing this, it occured to me a X,Y gauge would be useful, too. Although obviously the team has access to the X,Y,Z coordinates, it might help us when we discuss a feature [with the team].

thanks!
background includes work as an engineer at a nano-scale microscopy company

BIGAL
Posts: 44
Joined: Mon May 22, 2006 8:51 am
Location: SISTERS, OREGON

calibration movies suggestion

Post by BIGAL » Wed Aug 02, 2006 6:44 pm

It would be nice to be able to go back and review a calibration movie that you got wrong.

foobentrem
Posts: 3
Joined: Wed Aug 02, 2006 8:53 pm

Re: calibration movies suggestion

Post by foobentrem » Wed Aug 02, 2006 9:59 pm

I agree totally ... now I'm tempted to second guess myself w/o any idea how I blew it.

Only 1 of 16, but still.

dashadeaux
Posts: 21
Joined: Sat Jun 03, 2006 9:19 am
Location: USA, Peoria Arizona

Post by dashadeaux » Wed Aug 02, 2006 11:03 pm

Prior to registering and using the VM for real, I didn't think it important to be able to selective view calibration movies!!!

My attitude has changed and I'll throw my support in for this suggestion.

I have missed 3 tracks without any idea of why. Each time I was virtually certain the movie was a calibration movie, and I paid more than enough attention to the frames.

I hope I'm not overlooking something basic.

fizzbowen
Posts: 2
Joined: Wed Aug 02, 2006 11:08 pm
Location: Ventura, CA (USA)

Post by fizzbowen » Thu Aug 03, 2006 12:40 am

While on the subject of the interface ...

The submission buttons (No Track, Bad Focus, Examples) are small and close together. Several times I've moved the mouse pointer quickly to the general vicinity of one button, but since my eye-hand coordination is slowly deteriorating with age, I nearly click on a neighboring button by accident before realizing I'm on the wrong one. I'm hoping the buttons can be made slightly larger (10-20% would suffice) and better separated (perhaps by 40-50% of their vertical size). Sooner or later I'm going to click on the wrong one, if they stay as they are.

Also, would it be possible to pre-load more than one on-deck movie? (Perhaps as an option for people like me with dial-up.) Some movies (the dusty ones) take a while to examine; others go quickly. Efficiency would increase if I could have several on-decks load during the longer examinations, so that I wouldn't have to wait for the next download following a "quickie".

Other than that, I'm impressed with the simplicity and efficiency of the system.

marsnut
Posts: 1
Joined: Thu Aug 03, 2006 12:28 am

Interface Suggestion - relocating buttons

Post by marsnut » Thu Aug 03, 2006 12:45 am

I've been using the VM on my notebook with a touchpad, which I think is much easier to "fine tune" than using a mouse. However, the close proximity of the 3 buttons to the bottom of the scroller makes me a bit cautious when scrolling around there, in case I might accidentally press on "No Track". It's quite easy to make erroneous clicks on a touchpad. Perhaps move the buttons to the top of the scrollbar (where it is less used)?

cldrjones2
Posts: 13
Joined: Wed Aug 02, 2006 1:25 pm

Post by cldrjones2 » Thu Aug 03, 2006 12:57 am

Preloading more than one movie would be very nice indeed. Maybe you could give us the option on how many to preload.

Perttu
Posts: 7
Joined: Wed Aug 02, 2006 4:35 pm
Location: Finland 61°30"40' N , 23º39"45' E

Post by Perttu » Thu Aug 03, 2006 2:59 am

This is what my current quick 'n' dirty proxomitron filter for relocating the buttons look like:

Image

Would look a lot better if I moved the No Track and Bad focus buttons to the bottom and just moved the focus adjust buttons a little closer to the aerogel pic from their original position so their remain in one line.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest