System Voicing & Target Curve Mastery
I. Theoretical Context
1. The "Haas Effect" and Spatial Localization We have modeled the PA, hung it safely, and verified the network. Now we "Voice" it. The first principle is Spatial Localization. Because of the Haas Effect (Precedence Effect), the human brain localizes sound based on the first arrival. If the Side-Hang arrives at a listener 5ms before the Main-Hang, the listener will think the sound is coming from the side, even if the Main-Hang is 10 dB louder. We use Delay Offsets to ensure the "Image" is centered on the stage. We often "over-delay" the Side-Hangs by 2-3ms to ensure the Main PA is always the "First Arrival."
2. Voicing to the Target Curve A PA system should not be "Flat." A flat PA sounds thin and clinical. We voice the system to a Target Curve.
The "Tilt": Most professional curves (Clair, Eighth Day) feature a "Low-End Tilt." This is a gentle rise from 200 Hz down to 30 Hz, usually +6 dB to +12 dB above the high frequencies.
The HF Roll-off: To compensate for the "harshness" of compression drivers at high volumes, we apply a gentle roll-off above 8 kHz.
Phase Linearity: While we EQ the "Magnitude" (frequency), we must monitor the "Phase." If we apply a narrow, high-Q notch to fix a room mode, we create a "Phase Shift" that can smear the transient response of the snare or kick drum.
3. Atmospheric Compensation in Real-Time During a show, as the audience fills the room, the Temperature and Humidity change. Heat rises, creating "Temperature Gradients" (layers of air with different densities). This causes sound to "Refract" (bend) upward. An engineer must use Smaart to track these changes. If the 8 kHz response drops by 4 dB during the first three songs, the engineer doesn't just "EQ" it; they adjust the High-Frequency Compensation in the system processor. This allows the system to remain tonally consistent as the environment shifts.
4. Verification of the Mechanical Model The final step is verifying that our Wednesday model matches Friday's reality. We compare the Calculated Magnitude from the software to the Measured Magnitude in Smaart. If they don't match, it means:
The splay angles are physically incorrect.
A box is not working (Driver failure).
The venue geometry was modeled incorrectly. If the Coherence is high but the Magnitude is 6 dB lower than the model, we look for a mechanical error before touching an EQ fader.
II. Practical Lab: Voicing the Stadium
Tool: Smaart v9 / System Processor (Lake/P1).
Tasks: * Create a "Target Curve" in Smaart.
Voice the virtual PA to match the curve using a combination of "Array Filtering" and "Global EQ."
Haas Alignment: Place two speakers at different distances. Use the "Impulse Response" (IR) to align them within $50 \mu s$.
Air Comp: Simulate a 20% drop in humidity and "fix" the response using only the processor's compensation filters.
III. Daily Assessment (Friday)
Theory: Describe the relationship between the "Tilt" of a target curve and the human ear's "Equal Loudness Contours" (Fletcher-Munson).
Alignment: You are aligning a Front-Fill to a Main PA. At the transition point, the Main PA is 50ms away and the Fill is 10ms away. What is the Haas-effective delay you should apply to the Fill?
Synthesis: If your "Coherence" trace drops significantly during a show, but the "Magnitude" stays the same, what is likely happening in the environment?
WEEK 3 FINAL ASSESSMENT: The Master Exam
(Mechanicals) An array weighs 1,000 kg. You are hanging it on a bridle with a $120^\circ$ apex angle. What is the tension on each shackle?
(RF) Explain "Passive Intermodulation" (PIM) and why it is harder to solve than "Active Intermodulation."
(Fiber) You are running a single-mode fiber trunk. Your power meter shows -20 dBm at the FOH rack, but the SFP requires -12 dBm to lock. Where are the three most likely places to look for the "Loss"?
(PA) What is the difference between "Time Alignment" and "Phase Alignment" at a crossover point?
(Modeling) Why does a "Shadow Zone" in a 3D model often show high coherence but low magnitude in the real world?