5.1 Station Setup, Speech Processors & Audio DSP Filters
Key Takeaways
- Speech processors increase the average-to-peak power ratio of Single Sideband (SSB) voice transmissions without exceeding the transmitter's Peak Envelope Power (PEP) rating, improving signal intelligibility at distant receivers.
- Over-adjusting a speech processor causes severe intermodulation distortion (splatter), degrades audio clarity, and amplifies unwanted background acoustic noise such as cooling fans and room reverberation.
- Digital Signal Processing (DSP) implemented at Intermediate Frequencies (IF DSP) filters interfering signals before the receiver AGC detector, preventing strong out-of-passband signals from desensitizing the receiver.
- An Automatic Notch Filter (ANF) uses adaptive filtering to locate and attenuate multiple continuous interfering carrier heterodynes automatically on SSB voice, but must be disabled during CW and digital modes.
- The standard High Frequency (HF) S-meter calibration defines an S9 signal as an input signal level of 50 microvolts (\mu V) RMS into a 50-ohm load (-73 dBm), with each S-unit representing a 6 dB change (a 4:1 power ratio).
5.1 Station Setup, Speech Processors & Audio DSP Filters
Optimizing an amateur radio station requires mastering the operational controls of modern high-frequency (HF) transceivers, understanding audio dynamics in voice modulation, and leveraging digital signal processing (DSP) to extract weak signals from high-noise environments. In HF communications, where signal paths are subject to severe ionospheric fading, atmospheric noise, and crowded adjacent-channel interference, proper station adjustment marks the difference between an unreadable transmission and a crisp, intelligible contact.
This section covers the core principles of transmitter audio processing, Automatic Level Control (ALC) calibration, receiver intermediate frequency (IF) and audio frequency (AF) DSP filtering architectures, notch filtering techniques, and international standards for S-meter calibration.
1. Transceiver Controls & Dynamic Audio Shaping
Operating an HF transceiver effectively requires a thorough understanding of the controls that govern both the transmit and receive audio chains.
+---------------------------------------------------------------------------------------------------+
| TRANSCEIVER AUDIO & GAIN CONTROL STAGES |
| |
| [TRANSMIT CHAIN] |
| Microphone ---> [Mic Gain] ---> [Speech Processor] ---> [ALC System] ---> [Balanced Modulator] |
| ^ ^ ^ |
| | | | |
| Sets baseline Compresses peak- Prevents PA |
| input level to-average dynamic overdrive & |
| range (Boosts avg W) splatter |
| |
| [RECEIVE CHAIN] |
| RF Input ---> [RF Gain] ---> [Mixer / ADC] ---> [IF DSP Filter] ---> [Audio Gain (AF)] |
| ^ ^ ^ |
| | | | |
| Attenuates front- PBT, Shift, Notch, Sets acoustic |
| end overload in Noise Reduction (NR) listening level |
| strong signal areas |
+---------------------------------------------------------------------------------------------------+
Primary Transceiver Gain & Audio Adjustments
- Microphone Gain (Mic Gain): Adjusts the amplification level of the low-level electrical signal generated by the microphone element before it enters the speech amplifier and modulator. If mic gain is set too low, the transmitter will be undermodulated, resulting in weak, quiet audio and wasted transmitter power capacity. If mic gain is set too high, the audio preamplifier will saturate, causing harmonic distortion and driving the transmitter into heavy ALC compression.
- RF Gain vs. AF Gain:
- AF Gain (Audio Gain): Controls the power output of the post-demodulation audio amplifier feeding the station headphones or internal/external speaker. It affects only listening volume.
- RF Gain: Controls the sensitivity of the receiver's front-end RF amplifier and early intermediate frequency (IF) amplifier stages. Reducing RF gain in the presence of extremely strong local signals prevents front-end intermodulation and mixer overload while maintaining full sensitivity on weaker signals.
- Voice-Operated Exchange (VOX): Enables automatic hands-free transmit/receive switching triggered directly by the operator's voice:
- VOX Gain: Sets the acoustic sensitivity threshold at which speech triggers the transmitter into transmit mode.
- Anti-VOX: Samples the audio coming from the receiver's speaker and injects a canceling signal into the VOX detector. This prevents audio coming out of the station speaker from accidentally keying the transmitter.
- VOX Delay (Hang Time): Adjusts the duration the transmitter remains keyed during brief pauses between words, preventing rapid, erratic relay "chattering" during normal sentence cadence.
- Automatic Gain Control (AGC): Automatically adjusts receiver IF/RF amplifier gain based on incoming signal strength to maintain a stable, comfortable audio listening volume:
- Fast AGC: Features rapid attack and release time constants; ideal for fast-paced Morse code (CW) and digital modes (such as FT8 or RTTY) where signals start and stop abruptly.
- Slow AGC: Features longer decay times; ideal for Single Sideband (SSB) voice to prevent ambient background noise from rushing up during natural pauses between spoken syllables.
- AGC Off: Used in weak-signal DXing in combination with manual RF gain control to prevent strong nearby noise pulses from desensitizing the receiver.
2. Speech Processors & RF/Audio Compression
In Single Sideband (SSB) phone emissions, there is no continuous RF carrier and no redundant second sideband. The transmitted RF output power is directly proportional to the instantaneous amplitude of the modulating voice waveform.
Human speech has a high dynamic range with a high peak-to-average power ratio (typically 10 dB to 14 dB). Modulating speech consists of brief, high-amplitude vowel peaks and much lower-amplitude consonant sounds. Because the transmitter's power amplifier is strictly limited by its Peak Envelope Power (PEP) ceiling (e.g., 100 watts or 1,500 watts PEP), the average power of an unprocessed SSB voice signal is typically only 10% to 20% of its peak power (e.g., 10 to 20 watts average for a 100-watt PEP transmitter).
+---------------------------------------------------------------------------------------------------+
| UNPROCESSED VS. SPEECH-PROCESSED SSB WAVEFORMS |
| |
| [UNPROCESSED SSB SPEECH WAVEFORM] |
| Peak Limit (100W PEP) ------------------------------------------------------------- |
| | | | | |
| | | | | | | |
| Average Level (~15W) ..|.....|.|...|.......|...|.......|...|...|.|.....|.......... |
| | | | | | | | | | | | | | | | | | | | | | | | | | | |
| Low average power; consonants lost in background noise |
| |
| [PROCESSED / COMPRESSED SSB SPEECH WAVEFORM] |
| Peak Limit (100W PEP) ------------------------------------------------------------- |
| | | | | | | | | | | | | | | | | | | | | | | | | | |
| Average Level (~40-50W) ..|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|.|........ |
| | | | | | | | | | | | | | | | | | | | | | | | | | |
| High average power (+3 to +6 dB); dramatic DX intelligibility |
+---------------------------------------------------------------------------------------------------+
The Operational Goal of Speech Processing
The fundamental purpose of a speech processor (whether implemented as an audio compressor, an RF clipper, or a DSP algorithm) is to increase the average voice power relative to the Peak Envelope Power (PEP) without exceeding the transmitter's linear peak power limit.
By compressing the dynamic peaks and amplifying the lower-amplitude consonant components (which carry the majority of linguistic intelligibility), the speech processor raises the average transmitted power by 3 dB to 6 dB (a twofold to fourfold increase in average power). At a distant receiving station struggling with ionospheric noise and fading, this produces an apparent improvement of 1 to 2 S-units, often making the difference between completing a difficult contact or going unheard.
The Hazards of Over-Compression
While moderate processing provides substantial benefit, excessive speech compression introduces serious operational and technical hazards:
- Intermodulation Distortion (Splatter): Over-driving the processor causes severe nonlinear clipping, generating harmonic and intermodulation products that splatter across adjacent frequencies, causing illegal interference to operators on nearby channels.
- Harsh, Muffled, or Distorted Audio: Over-processed speech loses natural dynamic range, sounding flat, harsh, and difficult to comprehend.
- Background Noise Amplification: Speech compressors increase gain during quiet intervals. If the compression level is set excessively high, the processor amplifies background room acoustic noise—such as station cooling fans, amplifier blowers, power supply hum, and room echo—transmitting this noise at full peak power during pauses in speech.
- ALC Overdrive: Excessive speech processing drives the transceiver's Automatic Level Control (ALC) meter into deep, continuous saturation, causing further signal degradation.
[!IMPORTANT] Proper Speech Processor Adjustment: Always set the speech processor using the transceiver's ALC meter. Adjust the processor compression/level control so that the ALC meter deflects strictly within the manufacturer's recommended "ALC zone" on voice peaks, never pegging the meter or remaining continuously elevated.
3. Receiver Digital Signal Processing (DSP) Filtering
Modern amateur transceivers rely extensively on Digital Signal Processing (DSP) to manipulate signals mathematically in digital form using high-speed analog-to-digital converters (ADCs), digital signal processors, and digital-to-analog converters (DACs).
+---------------------------------------------------------------------------------------------------+
| IF DSP VS. AUDIO (AF) DSP FILTERING |
| |
| [IF DSP ARCHITECTURE (Superior)] |
| Antenna ---> [RF Mixer] ---> [ADC] ---> [IF DSP Engine] ---> [Demodulator / AGC] ---> [Speaker] |
| ^ |
| Filters interference BEFORE |
| AGC detector; prevents receiver |
| desensitization & pumping |
| |
| [AUDIO (AF) DSP ARCHITECTURE (Legacy / Basic)] |
| Antenna ---> [RF Mixer] ---> [IF Filter] ---> [Demod / AGC] ---> [ADC] ---> [AF DSP] ---> [Spkr]|
| ^ ^ |
| Strong adjacent carrier Filters audio |
| triggers AGC, desensitizing after damage is |
| the entire receiver already done |
+---------------------------------------------------------------------------------------------------+
IF DSP vs. Audio (AF) DSP
- Audio DSP (AF DSP): Operates on the demodulated audio signal after the product detector. While AF DSP is effective at post-detection audio filtering and noise reduction, it cannot protect the receiver's front end or intermediate stages from strong adjacent-channel signals. If an interfering carrier falls inside the analog IF passband, it will trigger the receiver's AGC circuit, pumping down receiver gain and desensitizing weak desired signals before the audio DSP can process them.
- Intermediate Frequency DSP (IF DSP): Operates directly on digitized IF signals before the AGC detector and demodulator (or in direct-sampling software-defined radios, directly on down-converted digital baseband data). Because filtering occurs ahead of the AGC loop, strong adjacent-channel signals outside the selected digital passband are eliminated before they can desensitize the receiver or cause AGC pumping.
Advanced DSP Filter Functions & Capabilities
| Filter Feature | Full Name & Domain | Operating Principle | Primary Operational Application |
|---|---|---|---|
| PBT / Twin PBT | Passband Tuning (IF DSP) | Electronically shifts the upper and lower cutoff frequencies of the DSP bandpass filter independently. | Narrows the receiver passband from either the high or low side to eliminate adjacent-channel interference without altering the pitch of the desired signal. |
| IF Shift | Intermediate Frequency Shift | Moves the center frequency of the entire IF filter passband up or down relative to the tuned carrier. | Shifts the passband window away from adjacent interference while maintaining constant overall filter bandwidth. |
| ANF | Automatic Notch Filter (Audio/IF) | Employs adaptive Least Mean Squares (LMS) algorithms to detect and notch out continuous carrier tones automatically. | Eliminates multiple interfering continuous heterodynes or tune-up carriers on SSB voice. Must be turned OFF for CW/Digital. |
| MNF | Manual Notch Filter (IF DSP) | An extremely sharp, operator-adjustable notch filter placed inside the IF passband. | Manually cuts out interfering carriers; operates inside the IF loop to prevent strong carriers from pumping the AGC. |
| NR | Noise Reduction (DSP) | Uses mathematical autocorrelation and spectral subtraction to separate random noise from periodic speech signals. | Suppresses continuous background hiss, white noise, and ionospheric atmospheric noise, enhancing voice readability. |
| NB | Noise Blanker (RF/IF Hardware & DSP) | Detects fast-rise-time, high-amplitude pulses and momentarily mutes (blanks) the signal path for microseconds. | Eliminates repetitive impulse noise such as automotive ignition pulses, electric fence clicks, and power-line arcing. |
[!CAUTION] Automatic Notch Filter (ANF) on CW and Digital Modes: The ANF algorithm cannot distinguish between an unwanted interfering heterodyne tone and a desired Morse code (CW) or digital (FT8, RTTY, PSK31) audio tone. Leaving the ANF active during CW or digital reception will cause the receiver to automatically notch out and silence the desired signals! Always ensure the ANF is disabled when operating CW or digital modes.
4. S-Meter Calibration Standards & RF Signal Levels
The transceiver S-meter (Signal Strength Meter) provides a standardized measurement of the relative or absolute RF signal power arriving at the receiver's antenna input terminal. The international standard for S-meter calibration was established by the International Amateur Radio Union (IARU) Region 1 and is widely adopted across HF amateur equipment design.
+---------------------------------------------------------------------------------------------------+
| IARU HF S-METER CALIBRATION SCALE (50 OHMS) |
| |
| S-Unit: S1 S3 S5 S7 S9 +10dB +20dB +30dB +40dB |
| dBm: -121 -109 -97 -85 -73 -63 -53 -43 -33 |
| Voltage: 0.2 0.8 3.1 12.5 50 158 500 1580 5000 (microvolts) |
| |-------|-------|-------|-------|--------|-------|-------|-------| |
| <--- 6 dB per S-unit (4x Power / 2x Voltage) ---> |
+---------------------------------------------------------------------------------------------------+
The Standard HF S9 Calibration Point
- HF Calibration Standard (Below 30 MHz): An S9 reading is defined as an RF input signal voltage of exactly 50 microvolts ($\mu\text{V}$) RMS into a 50-ohm receiver input impedance.
- Equivalent Power in dBm: Calculated using Ohm's Law and power relationships:
- VHF/UHF Calibration Standard (Above 30 MHz): Because ambient noise levels on VHF/UHF are lower, the IARU standard specifies S9 as 5 microvolts ($\mu\text{V}$) into 50 ohms ($-93\text{ dBm}$). This is 20 dB lower than the HF standard.
S-Unit Scaling: The 6 dB Rule
- Each S-unit from S1 to S9 represents a 6 dB change in signal level.
- In power terms, 6 dB represents a 4-to-1 power ratio ($10^{6/10} \approx 3.981 \approx 4$).
- In voltage terms, 6 dB represents a 2-to-1 voltage ratio ($20 \log_{10}(2) \approx 6.02\text{ dB}$).
- For example, if a signal increases from S7 to S8, the power delivered to the receiver has increased by a factor of 4, and the input voltage has doubled (from $12.5\ \mu\text{V}$ to $25\ \mu\text{V}$).
Decibel Readings Above S9
Signals exceeding S9 are calibrated directly in decibels above the S9 baseline (e.g., "S9 + 10 dB", "S9 + 20 dB", "S9 + 40 dB"):
- S9 + 10 dB: Represents 10 times the power of S9 ($-63\text{ dBm}$, or approx. $158\ \mu\text{V}$).
- S9 + 20 dB: Represents 100 times the power of S9 ($-53\text{ dBm}$, or $500\ \mu\text{V}$ / $0.5\text{ mV}$).
- S9 + 40 dB: Represents 10,000 times the power of S9 ($-33\text{ dBm}$, or $5,000\ \mu\text{V}$ / $5.0\text{ mV}$).
- S9 + 60 dB: Represents 1,000,000 times the power of S9 ($-13\text{ dBm}$, or $50,000\ \mu\text{V}$ / $50.0\text{ mV}$).
S-Meter Technical Calibration Reference Table
| Meter Indication | Signal Voltage ($50\ \Omega$) | Input Power (Watts) | Input Power (dBm) | Delta from S9 (dB) | Power Ratio vs. S9 |
|---|---|---|---|---|---|
| S1 | $0.20\ \mu\text{V}$ | $0.0008\text{ pW}$ | $-121\text{ dBm}$ | $-48\text{ dB}$ | $1 / 65,536$ |
| S3 | $0.78\ \mu\text{V}$ | $0.0122\text{ pW}$ | $-109\text{ dBm}$ | $-36\text{ dB}$ | $1 / 4,096$ |
| S5 | $3.13\ \mu\text{V}$ | $0.195\text{ pW}$ | $-97\text{ dBm}$ | $-24\text{ dB}$ | $1 / 256$ |
| S7 | $12.50\ \mu\text{V}$ | $3.125\text{ pW}$ | $-85\text{ dBm}$ | $-12\text{ dB}$ | $1 / 16$ |
| S8 | $25.00\ \mu\text{V}$ | $12.50\text{ pW}$ | $-79\text{ dBm}$ | $-6\text{ dB}$ | $1 / 4$ |
| S9 (HF Standard) | $50.00\ \mu\text{V}$ | $50.00\text{ pW}$ | $-73\text{ dBm}$ | $0\text{ dB}$ | $1.0$ (Baseline) |
| S9 + 10 dB | $158.0\ \mu\text{V}$ | $500.0\text{ pW}$ | $-63\text{ dBm}$ | $+10\text{ dB}$ | $10\times$ |
| S9 + 20 dB | $500.0\ \mu\text{V}$ | $5.00\text{ nW}$ | $-53\text{ dBm}$ | $+20\text{ dB}$ | $100\times$ |
| S9 + 40 dB | $5.00\text{ mV}$ | $500.0\text{ nW}$ | $-33\text{ dBm}$ | $+40\text{ dB}$ | $10,000\times$ |
| S9 + 60 dB | $50.00\text{ mV}$ | $50.00\ \mu\text{W}$ | $-13\text{ dBm}$ | $+60\text{ dB}$ | $1,000,000\times$ |
5. Sideband Operation Near Band Edges
Because a Single Sideband signal occupies spectrum on only ONE side of the displayed (suppressed-carrier) dial frequency, operating too close to a band or subband edge can push your transmitted energy outside the authorized segment — a Part 97 violation even when the dial reading itself looks legal.
- USB (Upper Sideband): The transmitted signal occupies the range from the dial frequency UP to the dial frequency plus the audio bandwidth (approximately 3 kHz). When operating near the UPPER edge of a phone segment, the displayed carrier frequency must be set at least 3 kHz BELOW the edge. Example: with the dial set to 14.347 MHz on 20 meters, a 3 kHz wide USB signal occupies 14.347 to 14.350 MHz — its top edge lands exactly on the band edge.
- LSB (Lower Sideband): The transmitted signal occupies the range from the dial frequency MINUS the audio bandwidth up to the dial frequency. When operating near the LOWER edge of a phone segment, the displayed carrier frequency must be set at least 3 kHz ABOVE the edge. Example: with the dial set to 7.178 MHz on 40 meters, a 3 kHz wide LSB signal occupies 7.175 to 7.178 MHz — its bottom edge lands exactly on the lower boundary of the General phone segment.
[!IMPORTANT] Band-Edge Rule of Thumb: On USB, keep the dial at least 3 kHz below the upper band edge; on LSB, keep the dial at least 3 kHz above the lower band edge. Careful operators leave an extra safety margin for filter skirt energy and transceiver frequency-calibration error.
What is the primary operational objective of utilizing a speech processor in a Single Sideband (SSB) phone transmitter?
Which of the following is an adverse operational consequence of setting a speech processor to an excessively high compression level?
Why must an Automatic Notch Filter (ANF) be switched OFF when receiving Morse code (CW) or digital data modes such as FT8 and RTTY?
According to the standard IARU calibration for High Frequency (HF) amateur transceivers, what signal input level corresponds to a reading of S9, and what power change is represented by each S-unit below S9?
A General Class operator sets the transceiver dial to 14.347 MHz and transmits a 3 kHz wide USB signal on the 20-meter band. What frequency range does the transmitted signal occupy?