Linkwitz-Riley - built it at last

[Ge_Whiz]

Well, two planned pedal projects ground to a halt over the vacation for lack of essential parts, so I knuckled down and built the Elliott Linkwitz-Riley crossover http://sound.westhost.com/project09.htm instead. Using the design program provided, I designed it for a crossover frequency of 184 Hz using 47nF capacitors and 13k resistors (yes, they ARE available), to encompass the first octave of the guitar's range.

The resulting gadget is fantastic. I split off the frequencies above and below the crossover frequency and process them separately. The bass frequencies go through a Digitech BP50 (sorry if I offend anyone's analog propensities) and are voiced, compressed and have the sub-octave added, before going through a separate bass amp. (I do have an original Coloursound Octivider, but it has always sounded crap.) The resulting sound, with higher strings split off and processed separately, is awesome.

To anyone who wishes to try this, here is a further thought: of course, the crossover only separates the fundamental of the low notes, while the harmonics come through on the treble side. This means that the unprocessed bass is muffled and boomy, so it needs some processing. The treble side, however, sounds like a normal guitar sound with the amp bass control turned down. A better approach may be to ignore the treble side of the crossover, and instead build a three-stage 36dB per octave bass filter. With one opamp used as a buffer, as in the original, this would only require one quad opamp and give sharper filtering.

[R.G.]

Congratulations!

I love frequency-splitting effects. It adds some very non-standard kinds of sounds, and gives you a much broader capability than you find in funnel-it-all-in effects.

A couple of notes based on things you mention.
- With none of the bass fundamental left in the signal, you still hear it. This is one consequence of the human ear's nonlinearity. Since all the harmonics of a missing bass frequency are still in the signal, the human ear beats them all together and generates the missing fundamental. In the speaker biz this effect is known as false bass - a speaker too small to really reproduce bass may sound like the bass notes are really there.
- Things really get fun when you distort only part of the musical spectrum. Mark Hammer's "woody" pedal works this way. The extra harmonics can really liven things up, and the expression changes as you move around the fretboard.
- Cellos.  Cello bodies do not reproduce the lowest notes that the strings vibrate at; the bodies do not couple the lowere notes to the air well enough. They *do* couple all the harmonics to the air, so what we hear of a cello is the harmonics (including false bass notes) and the nasal quality of the body resonances.

I've had in mind for quite a while a device that would use 3 to 4 bandpass filters per octave for the four octaves that a guitar can produce for fundamentals plus envelope detection and some gating logic to allow you to fish out of a guitar signal only the lowest note in the sound - kind of a lowest-tone extractor. Distortion done on this would be much cleaner sounding because it would have little cross modulation with other strings and the harmonics on one string.

Again - congrats!

[Ge_Whiz]

Thanks for the feedback, R.G.

This gadget is excellent at demonstrating the psycho-acoustic effects that you describe! It is also interesting to note that the two-stage filter, at a theoretical 24 dB per octave, leaves one with an audible crossover region of at least half an octave, an important consideration if you intend to split off more bands. This is based on the use of 1% resistors and +/- 10% tolerance capacitors, the best I could get. Musically, this 'overlap' region is pleasant, so version no.2 may consist of the low-pass filter only, switchable between 24 and 36 dB per octave.