Varun's Blogs

Thursday, January 12, 2012

Factual accuracies in Vaaranam Aiyiram

Its been a long long time since I last posted anything in this blog. So its time for one. Vaaranam Aiyiram is the recent one I saw and I decided I should blog about it. If you already know me, you are thinking I am going to b*tch about it, but you are wrong. I was quite surprised at the accuracies in the story. Let me list them for you:

1) Hero (Surya) gets Heroine1 (Meghna)'s number. Here is the clip:

The number is 510 555 3434. While 555 subcode numbers are often fictional, 510 is really area code that includes Berkeley. Great start for a movie trying to be accurate.

2) Hero says that he lived in University Ave for 40$. The story is assumed to be in 1995s. With US inflation at (http://www.usinflationcalculator.com/), I calculated it to be 59$ ~ 60$. From: http://www.law.berkeley.edu/142.htm, I can at least see bunch of them in 60$ range. Close enough.

3) The story is set in 1995 and you can see while he "roams around the city" that there are cut-outs of Jayalalithaa (the current Tamilnadu CM) here:

Note that when this movie was filmed (2007?), Jayalalithaa wasn't the CM. So, at least they must have morphed, or they must have been super lucky to drive through one of Jaya's birthday cutouts in the city that are usually placed around her birthday - sometime in February.

While all of these can be dismissed as just being accidental, I'm giving a benefit of doubt for this movie.

Monday, December 26, 2011

OpenNI and Kinect

OpenNI is a framework for Natural Interfaces. I started to get my kinect work for OpenNI. I followed the instructions in http://davetaz-blog.blogspot.com/2011/03/installing-kinect-on-ubuntu-full-guide.html

and there were issues. Apparently, the stuff in Sensor/Platform/Linux-x86/Build/Common needed to be copied to /usr/include/ni.

Also, I had to grep -R LIB_USED" kinect/Sensor/Platform/Linux-x86/Build and I needed to change LIB_USED to USED_LIBS in those Makefiles (there was only one instance IIRC).

Even after this there were issues. There was an error related to USB (Failed to set USB interface). The solution for it was to sudo rmmod gspca_kinect
(Apparently there is another kernel kinect module that needs to be unloaded).

With both these changes done, I was able to compile and get the sample program running.

I also tried compiling the python bindings for openni (thats called onipy). The guys at onipy already had made Makefile for linux, but it appears that the order of compilation is off in the final .so file.
I had to g++ -fPIC -shared -Wl,--export-dynamic -o OpenNIPythonWrapper.so ../../OpenNIPythonWrapper/Source/wrapper.o ../../OpenNIPythonWrapper/Source/OpenNIContextWrapper.o ../../OpenNIPythonWrapper/Source/OpenNIDepthGeneratorWrapper.o ../../OpenNIPythonWrapper/Source/OpenNIImageGeneratorWrapper.o ../../OpenNIPythonWrapper/Source/OpenNIImageMetaDataWrapper.o ../../OpenNIPythonWrapper/Source/conversionHelpers.o ../../PythonOutputStream/Source/PythonOutputStream.o -lboost_python -lOpenNI -lpython2.7 instead of the original one (where the -lboost_python.*$ would be at the beginning - as suggested here http://stackoverflow.com/questions/3871724/using-boost-python-with-make-instead-of-bjam)