What's Inside
- Defining the Wall of Sound
- The Challenge of Steep-Sided Valleys
- Analyzing Acoustic and RF Reflections
- Strategic Antenna Placement and Power Tuning
- Turning Interference into an Artistic Asset
- The Landscape as a Synthesizer
Defining the Wall of Sound
What happens to a broadcast when the landscape itself refuses to listen?
Topographical interference occurs when physical landscape features—such as steep granite walls, block, reflect, or scatter radio frequency signals, transforming a clear transmission into a chaotic web of echoes. Across the 88–108 MHz FM band, wavelengths range from approximately 3.4 to 2.8 metres. Cliffs, talus slopes, and the valley floor become electrically large structures rather than minor obstacles. I treat that scale as the first measurement, not a footnote.
The case was a temporary FM transmission project built to push unheard electronic music into an isolated, steep-sided valley. Restricted-service radio transmissions into remote landscapes leave little margin for guesswork. The working signal chain ran transmitter, feed line, antenna, terrain-dependent path, field receiver, then local loudspeaker. Trouble at the RF stage had to be logged separately from any coloration introduced after demodulation. That separation kept the broadcast archives honest when the geology started rewriting the programme.
I walked in expecting standard FM transmission behavior. The valley had other plans.
The Challenge of Steep-Sided Valleys
Sheer rock faces behaved like massive RF mirrors. Multipath arrived hard: the same carrier reached the receiver several times at slight delays, and the avant-garde material on the feed began to fracture.
Near 100 MHz, one wavelength sits at roughly 3 metres. Move a receiver less than a metre and you can step from constructive reinforcement into a deep multipath null. I watched stereo lock collapse beside a rock face while the same set, twenty paces into open ground, held cleanly. Phase cancellation, dropouts, and sudden distortion threatened the integrity of compositions that depended on sustained tones and wide stereo image.
The field assessment separated symptoms by place and cause. A receiver that lost stereo lock against granite pointed to RF multipath. Clean reception followed by a long audible tail from the loudspeaker pointed elsewhere—acoustic behaviour after demodulation. A reflected RF path that travels an additional 300 metres arrives roughly 1 microsecond after the direct path. Sound needs about 0.87 seconds to cover that same distance in air at approximately 343 metres per second. Confusing those two timescales is how you misdiagnose a valley.
Field receivers sat at a repeatable listening height of roughly 1.2–1.5 metres. Ground placement changes antenna coupling and the local reflection geometry, so I refused to treat a set on the scree as a valid sample.
Null Geometry
A high received-signal reading can coexist with poor audio. The meter may report abundant total RF energy while the demodulator is chewing several incompatible paths. Trust the ear and the stereo indicator before you trust a single field-strength number.
Analyzing Acoustic and RF Reflections
Mapping started on the valley floor. Repeatable transects, stops at bends, cliff bases, open clearings, and the transitions between rock and vegetation. At each stop I held the receiver still long enough for the carrier to settle or fail on its own terms.
Survey stops sat 10–20 metres apart where reception flipped quickly, farther apart where the field changed slowly. Each stop needed roughly 30–60 seconds of stationary listening so a persistent null could be told from a momentary fluctuation. Three programme types did different work: a spoken station identifier confirmed continuity, a steady test tone exposed warble or cancellation, and a spectrally dense music excerpt made FM multipath distortion easier to hear.
Two layers, one valley
RF multipath hits before or during demodulation. It can disturb carrier capture, stereo decoding, or high-frequency clarity. Acoustic reverberation arrives after the receiver has already converted the transmission back into audio—how the loudspeaker and the rock walls colour what listeners actually hear. I logged them on separate lines in the broadcast logs. Mixing the columns turns a solvable RF problem into a vague complaint about “the sound of the place.”
Dead zones and hot spots appeared where intersecting paths cancelled or reinforced. Walking the terrain with field receivers made those zones physical rather than theoretical. The map that came out of those passes drove every later antenna decision.
Strategic Antenna Placement and Power Tuning
Adjustment followed a fixed order so interacting variables stayed legible. Aim first. Height second. Power third. Polarization as a check, not an afterthought.
I abandoned the omnidirectional pattern and moved to a directional Yagi aimed down the valley’s usable axis. Energy went along the listening corridor instead of painting the reflective granite sides. At 100 MHz a quarter wavelength is approximately 0.75 metre; height changes on that scale shifted the balance between the direct ray and ground- or wall-reflected rays. Small vertical moves mattered more than another few watts on the meter.
Power tuning ran in 1–3 dB steps against the same reference audio and the same receiver positions. A 3 dB reduction halves transmitter power, though received field strength does not fall by that same numerical proportion. Lowering output reduced the strength of secondary reflections bouncing off the walls. That felt wrong until the multipath eased and the carrier held.
Power Catch
Reducing power helps when strong reflected energy dominates. It can worsen reception at locations already limited by weak direct signal or external noise. Check the weak end of the corridor before you call the cut a win.
Polarization got a hard test: rotate the Yagi through 90 degrees while the field receivers stay in their normal orientation. A mismatch there can cost more than a modest increase in transmitter output ever buys back. The useful height, polarization, and power setting still shift with receiver orientation, foliage moisture, temporary kit placement, and the geometry of the particular valley. A setting that clears one listening corridor can deepen a null elsewhere. I accepted that trade and locked the configuration that served the principal route.
| Observed symptom | Field check | Likely mechanism | Adjustment |
|---|---|---|---|
| Stereo indicator flickers at one location | Hold the receiver still, switch to mono, and rotate its antenna | RF multipath or polarization mismatch | Change antenna aim, height, polarization, or power |
| Clean RF lock, long audible tail from speakers | Compare headphone feed at the receiver with loudspeaker output | Acoustic reverb after demodulation | Reposition loudspeakers; leave RF chain alone |
| High meter reading, poor audio | Stationary listen with test tone and dense music excerpt | Multiple incompatible paths at the demodulator | Reduce illumination of reflective walls; retune aim |
| Dropout after small body movement | Move less than one metre at 1.2–1.5 m height | Spatial multipath null near 100 MHz | Map null, then shift antenna height or axis |
Turning Interference into an Artistic Asset
Once the carrier stayed usable along the principal listening route, evaluation left pure stability and entered composition. I revisited mapped hot spots and marginal zones with identical programme material and asked what the residual paths were doing to the music.
Comparisons used 45–90-second excerpts built from sustained tones, sharp transients, quiet passages, and wide stereo content. Location-dependent change had to be audible without leaning on meters alone. Stationary checks showed stable cancellation; slow walking passes exposed the rapid transitions between hot spots and dead zones. Residual RF multipath stayed described separately from literal echoes off the loudspeakers, so two physically different effects never got credited to one mechanism.
What remained after the Yagi and the power cut could not be erased without killing coverage. Those unavoidable signal echoes became site-specific texture on the electronic material. Listeners in the remote landscape heard a transmission that felt tied to the geology—phase smear and stereo drift as part of the piece rather than a fault report. Sound art & experimental work thrives on that kind of honesty when the special project is temporary and the pilgrimage is the point.
The Landscape as a Synthesizer
Transmitter settings establish the initial signal. Path length and reflection geometry keep shaping it before any listener hears the demodulated audio. That is the field map’s real conclusion: the valley participates in the synthesis.
Radio waves travel at approximately 299,792 kilometres per second. Reflection from a wall 500 feet away adds an out-and-back path of roughly 1,000 feet, so the reflected component arrives about 1.02 microseconds after the direct component. A 1-microsecond delay equals roughly 100 carrier cycles at 100 MHz; the exact phase relationship still shifts with carrier frequency, modulation, and geometry. The receiver combines those paths before it recovers the audio. The delay does not normally present as a separately audible echo. It alters phase relationships among the FM carrier and sidebands, and after demodulation that can surface as distortion, spectral coloration, stereo instability, or dropout.
Because radio waves travel at the speed of light, a signal bouncing off a rock wall 500 feet away arrives roughly one microsecond later than the direct signal—and that microsecond is the valley writing itself into the music before the sound ever reaches a human ear.
