Measure/Software

From OLPC

Jump to: navigation, search

A page for software developers to get involved in the software development efforts on the Measure Activity.


Brief overview of the code

  1. audiograb.py
  2. drawwaveform.py
  3. config.py


audiograb.py is the file that takes care of grabbing input audio samples. It uses a gstreamer pipeline in which the handoff signal of the fakesink element is used to intercept the data in the pipeline. Data logging is done in this class since we can use the sampling rate as a time reference; i.e. we know the sampling rate and also the number of elements in each buffer that is intercepted from the pipeline. So each atomic time interval is (1/Fs * bufferSize). This is the time interval after which we get a handoff signal.


drawWaveform.py is the class that handles the waveform drawing part. It creates and uses a cairo context to draw the background, the grid and the waveform. It uses a combination of pango and cairo for displaying the text as it was found out that cairo is itself not good at rendering text. All drawing occurs in the expose function which is the gtk expose event handler. It does drawing from the real time samples or from a file. In real time, it can either draw time domain or display the frequency domain samples. self.integer_buffer is one buffer that is obtained from audiograb.py and self.main_bufers is the buffer that accumulates all these buffers. The self.main_buffers is sampled at varying intervals of time, depending upon the frequency range selected, to form self.buffers which is the buffer that is scaled and drawn. self.main_buffers is sampled at a larger interval (set by self.spacing) for lower frequencies and at a shorter interval for higher frequencies. An example calculation is as follows. If we form 1 self.buffers by taking all consecutive elements, we would be sampling at 48khz. I have found that to draw a reasonable representation of a sine wave, one needs at least 8 -12 points. Hence the largest frequency that can be represented accurately is given by 48/12 = 4Khz, or less.Text is updated after every self.update_text expose events.


config.py contains all global variables.


Signals and timeouts used in the software

The handoff signal in audiograb.py is emitted whenever a new buffer is available. The signal handler is on_buffer which adds a gobject timeout. On timeout, _new_buffer is called which emits a custom signal new-buffer. The signal handler for this custom signal is _new_buffer in drawWaveform. Within _new_buffer, the new buffer is read and saved /stored. The expose event is queued in the main measure.py which causes the expose function to be called.

Personal tools
  • Log in
  • Login with OpenID
About OLPC
About the laptop
About the tablet
Projects
OLPC wiki
Toolbox
In other languages