Gaussian vs. Rectangular RBW filtering

• as a practical example, let's take a look at the analogue television RF spectrum
• in real life analogue transmitters, linearizer circuit alignment is a must
• (modern digital transmitters are capable of automatic alignment)
• for the alignment, vision modulation (pulsing), FM sound modulation and NICAM QPSK modulation is removed
• in Finland we use the PAL B/G VSB with intercarrier FM sound at 5.5 MHz and digital NICAM at 5.85 MHz
• here is a plot using a spectrum analyzer with Gaussian (10k) RBW filter response
• here is a plot using a spectrum analyzer with Rectangular (10k) RBW filter response
• all basic SPA settings are identical (auto), but as you can see, you can sweep faster with Gaussian filtering
• for clarity, here the two plots above are overlayed (4k) to emphasize the difference
• analogue FM and digital sound carrier 1. order IMD 350 kHz sidebands are visible within the DSB area
• this IMD is sometimes more evident in common mode transmitters due to vision amplifier non linearity
• in this case, IMD is generated more severely during the peak sync pulse period
• to analyse this problem, triggering and gating can be used, but that is a different story
• as you can see, Rectangular Filtering yields better resolution with marginally slower sweep rate requirements
  
• Gaussian Filtering is is practically mandatory for noise measurements, including digital spectrum power measurements