What a real-time spectrum analyzer can reveal depends on its software. The RTSI software running on QIBO TORI's WaveEye series packs 16 measurement functions across three layers: first make the signal visible (real-time spectrum, wide-span sweep), then make it persist (continuous DPX, block DPX, waterfall), and finally make it distinguishable (time-domain power + spectrum, adjacent channel power, phase noise, cellular signal analysis, direction finding). This article walks through all 16 functions by these three layers and ends with a scenario-to-function lookup table.
1. The big picture: one menu, 16 entries
On startup the software opens a function-selection menu. All 16 measurement functions have their own button, plus three system-maintenance buttons (reset ADC master clock, reboot, cancel). For field engineers this menu is the index to the whole capability set — no command lines to memorize, no external PC required.
Fig. 1 RTSI software function-selection menu — all 16 measurement functions at a glance
2. Layer 1 — Make it visible: real-time spectrum and wide-span sweep
2.1 Real-time spectrum analysis: computed all at once within 100 MHz
Real-time spectrum analysis completes an FFT in a single shot across a 100 MHz span — no sweeping. That determines what it can catch: a swept analyzer only observes a narrow slice of spectrum at any instant, so fast-changing, hopping, pulsed, or extremely short signals are easily missed in the gaps between sweeps; real-time analysis computes the entire 100 MHz span simultaneously, so every flash of a signal lands in the same frame.
So it fits situations that demand both fine detail and zero misses — high-resolution, high-accuracy detailed spectrum analysis, especially for those elusive short-duration signals.
Fig. 2 Real-time spectrum analysis — one-shot FFT across a 100 MHz span; short pulses are never missed
2.2 Wide-span spectrum analysis: sweep once you go beyond 100 MHz
When the span exceeds 100 MHz the software switches to sweeping, stepping through the band in 50 MHz or 100 MHz increments. Its job is not fine analysis but a fast wide-area survey: first find out "where are the signals" across a very wide frequency range, then drop back into real-time mode for the details.
They are a relay team, not substitutes — this is also the most common field workflow: sweep wide to find the suspect frequency, then lock onto that frequency in real-time mode to see the detail.
Fig. 3 Wide-span spectrum analysis — spans beyond 100 MHz stepped at 50/100 MHz for a wide-area survey
3. Layer 2 — Make it persist: three afterglow and history displays
The hard part of real-time is not the computation but "remembering that instant a moment ago." RTSI solves this in three ways, which many users blur together, but the division of labor is clear:
| Function | How it displays | What it's good for |
|---|
| Continuous DPXStandard | Digital-phosphor results are overlaid and continuously decayed/refreshed; the more often a signal appears, the brighter the phosphor | Intermittent signals: the kind that pops occasionally but keeps popping |
| Block DPXStandard | Shows phosphor results block by block; block length is user-settable | TDD signals such as 4G/5G/Wi-Fi; RF AI model training; synchronized capture of image and IQ data |
| Spectrogram (waterfall)Standard | Current spectrum trace at the bottom; above it, historical traces stacked as color-coded lines | Complete history of fast-changing signals — see at a glance "when it appeared" |
One-line distinction: DPX answers "how many times has this signal appeared," the waterfall answers "when did it appear." The former accumulates probability through phosphor brightness; the latter stacks history along a time axis. Block DPX takes "block" as the unit — a settable block length means you can slice a TDD frame structure into aligned blocks and compare them block by block, which is exactly why it suits 4G/5G and AI training data acquisition.
Fig. 4 Continuous DPX — phosphor continuously decays and refreshes; intermittent signals accumulate
Fig. 5 Block DPX — settable block length; ideal for TDD signals and synchronized IQ/image capture
Fig. 6 Spectrogram (waterfall) — current spectrum at the bottom, color-coded history stacked above
4. Layer 3 — Make it distinguishable: synchronized time-domain power and spectrum
The first three functions answer "is there a signal"; this one answers "in which segment, and what does it look like." It plots the time-domain power of the entire 10–40 ms signal; the user clicks any segment on the curve, and the spectrum for that segment appears below — time and frequency aligned on the same screen, purpose-built for locating TDD uplink interference.
We have a separate hands-on guide for this function, with the rationale for each parameter, real measurement cases, and FAQs, so we won't repeat it here: "QIBO TORI WaveEye TD-power+SA: Frame-Synchronized Time & Frequency Domain Analysis" (insert internal link after publishing).
Fig. 7 Synchronized time-domain power and spectrum (TD-power+SA) — 10–40 ms power trace on top, spectrum of the selected segment below
5. Transmitter health check: adjacent channel power and phase noise
Adjacent channel power: The software performs a wide-span spectrum analysis, integrates and displays the power of multiple adjacent channels, computes the power difference between channels, and estimates occupied bandwidth. This is not just "looking" — it is part of the 3GPP standard test requirement, used to confirm that an active base-station transmitter does not leak excessive out-of-band signals into adjacent channels. In other words, it gives a conclusion that can be checked against the standard, not a subjective judgment.
Fig. 8 Adjacent channel power — multi-channel power integration, power-difference and occupied-bandwidth estimation
Phase noise test: Measures the phase-noise curve of a signal source. The use is specific: characterizing the purity of single-tone sources such as signal generators and oscillators. Debugging a local oscillator in R&D, accepting a signal source, or troubleshooting "why won't my noise floor come down" all start with this curve.
Fig. 9 Phase noise test — phase-noise curve of a single-tone source
6. Cellular signal analysis: 4G-LTE / 5G-NR / NB-IoT / GSM-EDGE
All four are optional, with the same logic: not just measuring power, but decoding the signal.
4G-LTE analysis performs detailed analysis of all control signals embedded in the carrier (PSS, SSS, RS, PBCH) — these may come from each sector of the same base station and/or all neighboring base stations, so a single measurement reveals the distribution of surrounding cells. When the CRC check passes, the software decodes the MIB data.
A frequently overlooked but very practical gauge: the decodability of the MIB (whether the CRC passes) is the best coverage indicator at the observation point. High power does not equal good coverage; being able to fully decode the broadcast information is what proves the signal quality at that point is truly adequate. This is closer to the user's real experience than looking at level values.
Fig. 10 4G-LTE analysis — PSS/SSS/RS/PBCH decode and MIB result
Fig. 11 4G-LTE analysis (cont.) — neighbor-cell list and per-cell measurement summary
5G-NR analysis works the same way, except the targets are the SSB beams from all nearby base stations and sectors; NB-IoT follows the same principle as 4G-LTE but targets NB-IoT signals; GSM/EDGE analyzes all 8 bursts within one time slot one by one.
Fig. 12 5G-NR analysis — SSB beam decode from all base stations and sectors
Fig. 13 5G-NR analysis (cont.) — constellation plots of the SSB beams
Fig. 14 NB-IoT analysis — narrowband IoT signal decode
Fig. 15 GSM/EDGE analysis — all 8 bursts within one time slot decoded individually
7. Coverage survey and direction finding: multi-channel scan, blind scan, multi-antenna DF
Multi-channel scanning performs detailed analysis of multiple 4G or 5G channels the user selects "simultaneously" — essentially a drive-test scanner — used to survey the downlink coverage quality of multiple signals across an area, measuring as you walk, without manually switching frequency points one by one.
Fig. 16 Multi-channel scanner — multiple 4G/5G channels analyzed simultaneously (drive-test scanner)
4G-LTE blind scanning solves the "don't know which cells exist" case: it searches for every possible 4G-LTE signal within the selected frequency range and finds all the 4G signals currently active in that area. Ideal for a first entry into a new scene, or for mapping who's around before interference troubleshooting.
Fig. 17 4G-LTE blind scanner — all active 4G cells found within the selected band
Multi-antenna direction finding answers "which direction is the signal coming from." This is the most critical step in interference hunting — knowing there's interference and what it is, but not where the source is, means you haven't solved it. The direction given by multi-antenna DF is the starting point for walking toward it with a directional antenna.
Fig. 18 Multi-antenna direction finding — bearing of the signal of interest
8. Unattended: multi-channel spectrum analysis and automatic trigger-based recording
All the previous functions require a person watching the screen on site; this one doesn't. It autonomously performs spectrum analysis on multiple non-contiguous preset channels — relying only on the device's built-in DSP, no external PC. Once the measured spectrum power exceeds a preset trigger level, it automatically records the spectrum data for that band; afterwards it can be retrieved and uploaded to a PC, or, if a network connection exists, sent to a cloud server in real time.
Two typical uses: a unit fixed on site, watching a few key frequencies long-term for signal-intrusion detection; or mounted on a drone, flying patrols along sensitive borders or other sensitive areas. The former replaces manual watch; the latter solves areas "where people can't enter and vehicles can't drive."
Fig. 19 Multi-channel spectrum analysis and automatic trigger-based recording — preset channels, trigger threshold, auto evidence
9. Pick a function by scenario
| What you need to do | Use this function |
|---|
| Quickly tell whether there's a signal at a frequency and what it looks like | Real-time spectrum analysis (≤100 MHz span) |
| Survey a very wide frequency range first | Wide-span spectrum analysis |
| Catch an intermittent signal that pops occasionally | Continuous DPX |
| See each block in a TDD frame structure, or collect AI training data | Block DPX (+ IQ synchronized capture) |
| Review the history of when a signal appeared | Spectrogram (waterfall) |
| Find which slot 4G/5G uplink interference falls in | Synchronized time-domain power and spectrum co-display |
| Verify a base-station transmitter for out-of-limit emissions | Adjacent channel power |
| Evaluate signal-source purity | Phase noise test |
| Measure 4G/5G downlink coverage quality in an area | 4G-LTE/5G-NR/NB-IoT analysis |
| Measure multi-signal coverage while driving | Multi-channel scanner |
| Don't know which cells are around | 4G-LTE blind scanner |
| Know there's interference, need the source direction | Multi-antenna direction finding |
| Unattended, auto-record on threshold exceed | Multi-channel spectrum analysis and automatic trigger-based recording |
10. Frequently asked questions
1. What's the difference between real-time and swept spectrum analysis?
Real-time computes the whole thing in one shot within a 100 MHz span, no sweep, so short, hopping, pulsed signals are never missed; swept analysis steps through segment by segment, the span can be very large, but each instant only sees a small slice. When the span exceeds 100 MHz the software automatically switches to sweeping; the two are a relay relationship.
2. How to choose between continuous DPX and block DPX?
It depends on whether you need "block alignment." Continuous DPX is continuous refresh accumulation, suited to finding the pattern of intermittent signals; block DPX has a settable block length, can slice TDD signals into aligned blocks for block-by-block comparison, and can synchronously grab image and IQ data, suited to 4G/5G/Wi-Fi analysis and RF AI model training.
3. What's the difference between waterfall and DPX?
DPX accumulates "probability of occurrence," using phosphor brightness to show how many times a frequency appeared; the waterfall accumulates "time order," stacking history from bottom to top with color-coded lines. The former helps judge whether a signal is regular, the latter when it appeared.
4. In 4G/5G analysis, if the MIB can't be decoded, what does it mean?
It means the signal quality at that point is not yet sufficient to stably carry broadcast information. The decodability of the MIB (whether CRC passes) is the most direct coverage indicator at the observation point — closer to real usage experience than level values alone. When it can't be decoded, combine with the neighbor-cell list to judge whether it's insufficient coverage or interference suppression.
5. Are these functions standard or optional?
Real-time spectrum, wide-span sweep, continuous and block DPX, waterfall, time-domain power + spectrum, adjacent channel power, phase noise, and multi-channel auto-recording are standard; 4G-LTE, 5G-NR, NB-IoT, GSM/EDGE analysis, multi-channel scanning, and 4G-LTE blind scanning are optional. Multi-antenna direction finding requires the corresponding multi-antenna hardware. Exact configuration depends on the selected model; contact us to confirm.
6. Does the software require a PC to work?
No. Multi-channel spectrum analysis and automatic trigger-based recording are performed entirely by the device's built-in DSP, no external PC; both fixed on-site and drone-mounted scenarios work this way.
