Pictures can say more than words

Foundations of Amateur Radio Recently I've spoken about measuring the frequency response of your radio and what the benefits of doing so might be. Today I've got some progress to report and some initial discoveries. Again, this is preliminary, but then all of this hobby is experimentation, so that should come as no surprise. Let's start with the mechanics of what I'm doing and a "duh" moment I need to confess. The aim of this process is to transmit a known audio signal, receive it, record it and create a spectrogram from it. This allows us to compare the original spectrogram against the received one and show just how the audio path has been affected by getting the audio into the transmitter, the processing by the transmitter, the propagation between the transmitter and receiver, the artefacts introduced in the receiver and any recording device. To begin this process I started off with an audio file of my voice. That wasn't very helpful, since it's a complex signal and comparing my voice before and after is a non-trivial process. At some point I intend to come back to voice before and after comparison, but that's on the shelf for now. The audio that I'm using is a frequency sweep, lasting 5 seconds. That is, there's a tone that changes frequency from DC to 5 kHz. When I looked at the spectrogram of that, it shows as a curve with time against frequency. It occurred to me that I could make two of those sweeps at the same time to measure distortion, so I added a reverse frequency sweep from 5 kHz down to DC. Now I've got two crossing lines showing in my spectrogram. To transmit this audio, I'm using the same tool I use to automatically call CQ during a contest. Every so many seconds I transmit this audio into a dummy load and at this point I should mention that my "duh" moment was that I was attempting to transmit into an antenna and record from a dummy load, rather than transmit into a dummy load and record from an antenna. I still cannot believe that I did that. Moving on. The recording is done using an RTLSDR dongle. In the current initial version I'm using a tool called rtl_fm to tune the dongle to the same frequency as my transmitter. I send the audio from there to the same tool I used to generate the original audio, SoX, that's Sierra, Oscar, X-Ray, and have it detect the silence between each transmission and record each into a new file. If I leave it running, every time I transmit something, SoX will create a new audio file. I'm saying that quite quickly, but getting the squelch and silence detection working in my noisy environment took most of a day and it's specific to my station, today. I'll have to figure out how to make this smarter, but for now I have some data. A spectrogram is generated for each audio file and then we can compare pictures. What was sent, audio wise, and what was received, audio wise. To be clear, I'm not sending images, I'm sending audio and comparing the spectrograms of this audio. I will also note that I'm currently using FM as the mode. I intended to do this with SSB, but the amount of effort to get the squelch right has left me with a future project to achieve that. The code itself is pretty rudimentary, but I've uploaded it to my github page. I've also added the pictures to my project website, which you can find at vk6flab.com. One initial observation, one that I don't yet understand, is that what I sent and what I received don't look the same. My pretty curves in the original audio come back with spectacular harmonics all over the place, very pretty to be sure, but not quite what I was expecting, let's call it an educational challenge. Before I forget, just because I'm using a Yaesu FT-857d, a Raspberry Pi, an RTLSDR dongle, an antenna and a dummy load, doesn't mean that you need to. Essentially, what this does is generate a special audio file, transmit it, receive it, record it and generate a spectrogram. You can play the audio from your own computer if you have digital modes set-up, or from your mobile phone if not. Recording can be something sophisticated with off-air monitoring, or it can be a recorder held in front of your receiver. One final note. You can change settings on both the transmitter and the receiver to see what they do in relation to the audio, so experiment. I'm Onno VK6FLAB