1) Live/interactive version of one or two of the algorithms. I'd vote
for IResponse, RelaxRate and ColorMulti as the ones to do live on the
laptop. David is looking at this...
2) The frame skipping issue remains. I looked at this while travelling
today, but couldn't figure it out. We're still saving out meshes at
about half the rate of them being created.
3) The following three analyses do exactly the same thing currently,
and it would be good if they had something approaching a real analysis
happening. Right now it's just a simple z-map in black and white with
some noise on top and lighting... kind of pretty but bit boring
currently:
ShapeFromShadingAnalysis
CamNoiseAnalysis
ShadowScapesAnalysis
* moved the synthesis saving to an abstract function called
saveImageSynthesis()
* allocating the images we're saving out in a renamed
create_dir_allocate_images()
* MAKE SURE TO PUT THE CONFIG FILE IN /data/
* the save functions for image should work on both windows and osx now
(although we might want to move the creation of these images outside
the thread later)
* params for each analysis timing in the XML config
* minor warning / bug fixes in AbstractAnalysis.cpp and
RefractiveIndex.cpp
* cleaned up the use of OF_IMAGE_COLOR/OF_IMAGE_GRAYSCALE in all
analysis classes
the looping structure for RUN_NUM now occurs in AbstractAnalysis, so
that the sequencing is correct - a few related variables were not reset
- so those are fixed now too.
this code awairs:
* add (from david) of RelaxRate analysis synthesise and display
* any learning that Tom brings back from the Middlesbrough trip
tomorrow.
a pretty nice little piece of code, gents!
if(_save_cnt >= NUM_SAVE_PER_RUN) _RUN_DONE = true;
and using max_frame_cnt as the end condition for the run being finished...
not sure if this is working the way it should... need to think through the math of the frameRate/timing calculations in the setup for each class
* removed unused and unneeded classes and added the ones that were
missing
* put the most recent graphics code into David's great structural org!
current issues:
* the config file is still not loading at the start of the code (only
in RefractiveIndex.cpp does it work - not in main.cpp - where the 'full
screen', etc. are set
* the analyses being called in RefractiveIndex.cpp always run the first
listed one twice - weird!
* the saving (NUM_SAVE_PER_RUN) and frame (_frame_cnt_max) conditions
are competing - can we relate them?
