Linear Collider Forum

Home » Analysis and Reconstruction » Tracking & Vertexing » TPC Event Display?
TPC Event Display? [message #1010] Thu, 23 August 2007 04:07 Go to previous message
Messages: 43
Registered: August 2007
Location: DESY Hamburg

I just hacked a very simple a very specific event display for our small TPC protoype. In our group the question came up: why didn't you use CED? Or is there sth. else that works?
So I thought: why not forward this question to a broader audience Wink

Well, here my thoughts to that (also keeping in mind that a display for MarlinTPC and the large EUDET protoype will be needed too):
From my experience, ROOT is not an option because 3D drawing is slow and very limited (works for a quick solution for a very limited geometry.)
CED is nice because it is very fast, but I think it needs some more features:
  • Sth. like Projections (especially on pad plane: XY, but maybe also YZ and XZ)
  • This of course includes handling of pad plane geometry and display.
  • Color code depending on charge deposited charge in a voxel / on a pad (with a palette, e.g. dark to bright). The palette would of course have to be displayed in the display with corresponding charge values.
  • Axis with labels and ticks (incl. values).

Some of these features would probably also be useful for other detector components.

Further features a good TPC event display could/should have are IMHO histograms for interesting values like charge per pulse etc. and maybe even a further reconstruction that results for example in histograms of angular distributions of the tracks etc..
And it would be nice to be able to export the histograms for further analysis...
Of course ROOT pops up in my mind thinking of the histograms. Would that be very bad? (An event display won't run on the GRID).
And if ROOT is bad: is it so bad, that we reinvent the functionality (histograms etc.) and a data format for that?

For the GUI, QT would be my choice (it's also already necessary for other ilcsoft, so no new dependency would be introduced).

Has sbd. else already done some development? Maybe there is sth. else than CED (although I would like not to reinvent the ilcsoft-wheel, but improve and extend existing code)?
I think many people have at least thought about it or even gained some experience while hacking their own specific display... Smile

An last but no least, I think we should keep in mind that it is crucial that the display will be moderatly fast... well, maybe not 25fps fast, but not like 1 event every 30s.

What do you think an display would need, should have, would be nice to have?

I could probably at least start working on that and do some parts needed.
Read Message
Read Message
Previous Topic:LCFIVertex v00-01 released
Next Topic:Evolving the Track Class
Goto Forum:

Current Time: Thu Nov 14 19:14:10 Pacific Standard Time 2019
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.1.
Copyright ©2001-2010 FUDforum Bulletin Board Software