# Development History¶

## 2016/02/24¶

• Added save method, to save part of a Signal to a new hdf5 file.
• Assorted bugfixes. In particular, the example signals (available through python -m freqdemod.demodulate), and the IPython notebook example, now work on Python 2 / 3.

## 2015/07/21¶

• Changed Signal object so no hdf5 files are created by default. Previously, Signal objects always created hdf5 files when they were closed; they would also overwrite any previously existing filename by default. In most cases, this eliminates the need delete temporary files. To save the Signal data file to disk, use Signal('filename.h5', backing_store=True).
• Added load_hdf5 and load_hdf5_general methods to Signal to allow importing signal data directly from hdf5 files.

## 2015/07/12¶

• Consider, then decide against, changing the bandpass filter-width definition. As defined, 1 kHz-wide filter will pass a 2 kHz band of frequencies, from 1 kHz below to 1 kHz above. The power spectrum will show noise falling away starting 1 kHz away from the carrier.

• The only bandpass filter available so far has been a brick-wall filter (default order = 50). This filter was introduced to avoid aliasing of high-frequency noise into the low frequency region of a cantilever frequency-noise power spectrum. The brick wall filter, because of its sharp cutoff in the frequency domain, induces sinc-type oscillations in the time domain. To avoid these sinc-type oscillations, we have introduced a cosine filter.

A call like s.freq_filter_bp(1.00) will create a brick wall filter of bandwidth 1 kHz and order 50 (default); this call is unchanged. To create a cosine filter instead, s.freq_filter_bp(bw=1.00, style="cosine").

## 2015/01/28¶

• Typos fixed in documentation.
• Tests can now be easily run on packages installed via pip, using freqdemod.test().

## 2015/01/23¶

• Release the package into the wild.

## 2014/08/09¶

The demodulate.py file has undergone a major rewrite!

• The “quickstart” files are essentially all broken. You may instead see how the new code works by running, at the command line, from the home directory of the package

python freqdemod/demodulate.py --no-LaTeX --testsignal=sine
python freqdemod/demodulate.py --no-LaTeX --testsignal=sinefm
python freqdemod/demodulate.py --no-LaTeX --testsignal=sineexp


Running these commands will bring up a number of windows. You will have to click each window closed before the program will proceed and show you the next window. Each window should have a pretty self-explanatory title I hope. You can try the --LaTeX option to see all the plots in fancy LaTeX typesetting.

• All the data is stored as an HDF5 file. If you have the h5py package installed correctly, you should have available the h5ls command line program which is useful for inspecting the contents of HDF5 files. Each of the above programs saves its data to a hidden HDF5 file. You can see the files by running at the command line

ls -ha | grep h5


or simply looking for the files that start with a dot, ., and end with .h5. To examine the contents of the HDF5 files produced by running

h5ls -rv .temp_sine.h5
h5ls -rv .temp_sine_fm.h5
h5ls -rv .temp_sine_exp.h5

• The code is only lightly documented. To get an idea of how things work, start by looking at the functions

testsignal_sine()
testsignal_sine_fm()
testsignal_sine_exp()


in the demodulate.py program.