Here are 2 RTA I built in 1977 and have been in a box since 1983. I got them out and fired them up as you can see in the setup in the first picture before the video. The 1st video is a log frequency sweep from 40Hz to 20KHz and is the most revealing of the difference between the 2 RTA. By pausing the video it can be seen that the top one is badly “bleeding” the signal into the adjacent octaves. At the end of the video are shots of the RTAs.
There were very few inexpensive RTA in 1977 and these were prototypes that never got farther. The bottom one is the first one I built and its 10 octave filters are 6th order meaning 6 capacitors per filter, giving steeper side skirts for each octave “box”. Q = 10 and BW (at -3dB) is 25% either side of the center frequencies. Each filter has its own dB indicator ladder, which is why there are so many wires. I had not discovered multiplexing yet. Individual LEDs were used as in 1977 bar LEDs were not generally available.
The top one is 9 octave filters which are 2nd order with only 2 capacitors per filter. Q=1.7 and BW is 23%. There is one dB indicator ladder and the filters are multiplexed to it at 0.001 second intervals.
For 2nd order filters, larger Q attenuates the signal occurring between the center frequency and the edges (skirts) of the octave ”box” more than does a smaller Q. On the other hand, a smaller Q allows the signal to more “bleed” across the skirt into the next octave “box”, which can be seen in the top RTA when compared to the bottom one. This causes the adjacent octave to be held up and not collapse as quickly as a higher order filter and therefore seem slower responding. Which is worse, to lose (attenuate) signal inside the octave “box” or gain signal from one “box” into the next one? Both cause inaccurate readings. Best to use 6th order filters (which are more costly) !
Unfortunately, the Q and order of the RTA on UT are not generally given. It appears some home made ones are 2nd order.