The baudline program is a useful tool for analyzing audio. First you get a waterfall presentation of the spectrum, then there is a amplitude diagram and many possibilities for selecting input source, sample frequency etc. You can find it here http://www.baudline.com.
I have used it in many instances, from analyzing the hoover’s sound spectrum (yeah, I know…) to determining different types of noise. For this blog entry I used baudline attached to the line out signal from the Jupiter 538. The line out is accessible through the ACC1 socket on the rear panel. This is the one you would use for receiving digital modes. According to the schematics it is the same signal as the one fed to the AF amplifier though.
A major feature for me with this rig is the almost continually variable receiver bandwidth. There are several bandwidths to choose from namely 8000, 6000, 5700, 5400, 5100, 4800, 4500, 4200, 3900, 3600, 3300, 3000, 2850, 2700, 2550, 2400, 2250, 2100, 1950, 1800, 1650, 1500, 1350, 1200, 1050, 900, 750, 600, 525, 450, 375, 330, 300, 260, 225, 180, 165 and 150 Hz. Quite an assortment!
AM sounds lovely at 8 or 6 kHz, morse code usually around 1 kHz and SSB somewhere in between 2100 and 3600 depending on interfering stations. The perceived effect is a continually variable filter. Very neat!
Baudline reveals the filter’s effect and specifically the shape factor in the transition between pass band and stop band. For a full 8 kHz filter bandwidth this is what it looks like (direct screenshot):
Receive filter at 8 kHz. Radio was set at USB.
In the screenshot above the x-axis is frequency from 0 to 10 kHz (1 kHz/division), the y-axis is graded in decibels (10 dB/div). We see that the passband is virtually flat from 1 to 8 kHz. The transition from pass to stop band is 1 kHz wide and there the amplitude drops a whole 40 dB. At around 500 Hz there is an inexplicable boost in the signal which surprised me as I never heard anything odd from it. On the contrary the audio is always very clean and balanced.
As the bandwidth is reduced all that happens in the picture above is that the transition slope moves to the left. Its actual slope is the same and it drops with 40 db/kHz at every setting. With a width of 1050 Hz:
and 150 Hz:
In the above picture the passband hardly gets started until it is cut down to the stop band, still sometimes it is useful.
While there are several menu controls for the transmit audio there is only one menu item for reception: HIGH BOOST. This control is a treble control of sorts. It act perhaps more like a bass attenuator and baudline reveals what is going on:
HIGH BOOST at 100% and bandwidth of 8 kHz.
The previously entirely flat response now emphasizes the higher frequencies. The bass response is down some 10 dB but the actual amplitude level for the higher frequencies are about the same as in the flat response. I imagine HIGH BOOST’s main purpose is to overcome bad headphone’s frequency response and/or an old ear’s drop in frequency response. For me the flat response (HIGH BOOST 0%) sounds the best, be it through speaker or head phones.
On the front panel is the NR (Noise Reduction) push button. This noise reduction is intended to make SSB audio clearer and easier to read. It succeeds sometimes and sometimes not. Well, it always reduces noise but sometimes at the price of less intelligibility overall. Used with caution it can help, but as interference can be of so many different origins one algorithm would be hard pressed to solve every case.
With white noise it demonstrates favourably though. First the passband (6 kHz to the far right now) with no NR:
The noise floor drops dramatically circa 30 dB. Under the right conditions this is makes wonders for the listening pleasure.
The last two pictures unfortunately also reveal a flaw with the DSP. At certain frequencies and selected bandwidths spurious carriers can appear. The pike at 1.5 kHz above is such a signal. I believe this is not a remnant from the mixer stages of the radio hardware in itself but a side effect of the implementation of the DSP filter algorithms. Or perhaps a combination of the two? As there are three intermediate frequencies in this radio it is a tough job eliminating mixer products as they sometimes appear in-band.
Note: With a band width of 8 kHz and AM mode the maximum theoretical frequency within the passband is 4 kHz as AM has two sidebands.