I hear it fairly often in discussions about passives that operate in a frequency range excluding DC. If the geometry defines the frequencies, the upper and lowe

I hear it fairly often in discussions about passives that operate in a frequency range excluding DC. If the geometry defines the frequencies, the upper and lower frequency limits will have a fixed ratio as you scale the device up/down. For instance, textbook directional couplers typically have octave range (1-2GHz, 2-4GHz, 10-20GHz) while "clever" couplers can have decade range (1-10GHz) or more. Similar stories apply to cavity or microstrip filters/resonators. Since frequency and bandwidth are banner specs the terms often leak into the names themselves -- someone might advertise a "multi-decade coupler" and compare it to "octave couplers." You can readily find examples on google. Harmonics are a key concern in RF but I usually hear the terms "fundamental, second, third, etc" used to refer to them and it would sound odd to my ear if someone used "octave" in that context.