RTI File output problem

[jwegner]

Hi CHI folks,

 

I've been trying RTI recently following your guidelines for image capture and with RTIBuilder.

Everything works perfectly with image capture and processing up through final generation of RTI file. Final stage of Builder goes through to the "fitting complete" window (using HSHfitter option) and generates RTI file in the "finished-files" folder.

 

BUT, the RTI file that results is usually tiny (4kb), obviously containing no data. In a couple execute attempts it created what appeared to be a working RTI file of 2GB +. In either case though trying to open file in RTIViewer causes program to crash.

 

I can't see what's wrong (file names have no spaces, assembly files and dialogue box in Builder show no errors). Maybe a memory issue, it's a Mac with OS 10.6.8.

 

Thanks for any insight!

-Joe

 

[marlin]

hi Joe,

 

I'll chime in with my hunch, tho Carla the goddess jedi of processing and software might have additional input and details in re: to memory management and your OS.

 

A 2gig file size is rather large. Too large in fact for the current version of the RTIViewer to handle. The result, (as you experienced), is that the app crashes and goes down in flames or it just wont respond with any useful information or even error message.

 

Knowing that you're file size is giant, I'm assuming that your acquisition device is a larger / medium format camera? What are you using? Hasselblad? What are your pixel WxH size(s)?

 

The Work Around.

Until the RTIViewer gets rebuilt with the ability to open and view files of that dimension, you should reduce the pixel width and height of the jpegs that you export to the 'jpeg-exports' folder. You will likely have a better user experience. Reduce to what size? Not sure. You might want to try and re process at a few different sizes to see what your machine/ system can handle.

 

The good news is that you have the original data 'in the can', and when the RTIViewer can handle those larger files, you'll be ready to reprocess and view at the larger dimensions.

 

Hope that this helps. Send a flare after reprocessing.

 

Marlin.

[cdschroer]

Joe -

 

What Marlin says about file size the viewer is true.  I think you are running into something else though.  We have seen it a few times where when building an hsh (doesn't happen with ptm) you get a tiny file that is not a valid file.  I don't have a reproducible test case, so we haven't been able to track it down (the developers can't do anything if they can't see it for themselves)  It seems to be specific to machines, not to data sets.

 

I will say that the first order hsh doesn't work.  2nd order is the default - and that is what I recommend. There is also a bug in the 2.0.2 version of RTiBuilder where if you choose an order, then change it to a different order, it doesn't actually change, it builds another file at the same order you previously chose.  You have to reload the data set (i.e. reprocess an existing data set per the user guide) to get it to do the different order.

 

We are working on the 2.1 release of RTIBuilder, and this bug is fixed - I mean the one where changing the order in the interface doesn't change the order of what is built.  We are awaiting some final fixes and then a round of testing and we need to do user guide updates. So, I'm hoping to get this out by the end of May.

 

At any rate, maybe thee notes are useful for working with 2.0.2.

 

Carla

[jwegner]

Thanks for the suggestions,

 

Problem does seem to be with the HSHfitter. Ran it with PTMfitter and works perfectly.

 

-Joe

[cdschroer]

Can you try making an hsh with the order set to 2 or 3 and post what happens.  If it doesn't work, please note specifics about your machine and environment.

 

Carla