Original material:
Beethoven’s 5th symphony is possibly one of the most recognizable pieces of music out there. It is scored for 22 instruments. These instruments give the symphony a broad range of frequencies. From the very high piccolo to the low bass. It is also important to note the high dynamic range.
System used:
In order to transmit many phone calls over the same channel, the human voice is transmitted using a little over 2kHz (a bit less than C5 to C7 for you musicians). Now a day these filtered signals are also digitized into a very small size and sampled at under 8kHz (that’s about a 5th the sampling frequency of a typical MP3). At the other end of the line, the sound is regenerated to complete the transmission.
Surplisingly this is enough to understand human speech. However, as anyone who has been put on hold, you know that music does not transmit well at all over the phone. This is because instruments and singers produce frequencies way the above mentioned range.
So, here is the question. What would happen to Beethoven’s 5th if we transmit it over and over through a phone line? How long will it take for the band-pass filter to swallow the high piccolo? Will the dynamic range survive the normalization?
How was this effect done?
An actual phone call was placed between two phones and both sides of the transmission recorded. Using both signals and an adaptive filter created using and FPGA (an intel Cyclone IV) we can recover the coefficients for an infinite impulse response filter that very closely mimics the telephone transmission line. For the last part, we send the signal through this filter (once again using the FPGA) and re-sample it using a microcontroller connected to our computer.
Max 8 was used for normalization and visualization of the resulting signal.
In retrospect, the second recording of the signal sounds better than the phone call signal. This leads me to believe that my method for recreating a phone line is not too realistic.
Google Drive link: https://drive.google.com/drive/folders/1ILeY6phiYcoI7TqMZHV6Ah8xn96e8Qsb?usp=sharing