Getting Started With Ghost Hunters
Ghost Hunters is a tool for capturing and analyzing ghost signals in RF environments. It is not a parlor trick app that works out of the box on your laptop. You need an SDR dongle, a decent antenna, and patience. The software itself is free but occasionally buggy. That is normal. You will troubleshoot it. I used it for about six months after acquiring a RTL-SDR V3 and setting up a directional Yagi on my roof. The first week was mostly frustration. The demo clips online make it look slick. In practice, you are fighting interference from neighbors' Wi-Fi, cell towers, and a lot of garbage RF that has nothing to do with anything you care about.
What The Ghost Hunters Actually Does
The core function is time-domain signal capture with spectral analysis. It records raw IQ samples, then applies a moving average filter and anomaly detection across frequency bands. When it finds a signal that does not match known patterns, it flags it. That flagging is where the name comes from. The signal is "ghosting" through your spectrum, and the software highlights it. Known signal sources include FM radio around 88 to 108 MHz, DAB at 174 to 240 MHz, GSM at 900 and 1800 MHz, LTE bands, and Wi-Fi at 2.4 and 5 GHz. Ghost Hunters will show these normally. What it tries to isolate are transient signals that appear outside these bands or behave unpredictably. A good example is a short burst at an unexpected frequency that lasts less than 50 milliseconds. Those are what the software labels as ghost events.
Installation and Basic Setup
Download the installer from the official site at ghosthunters-tool.net. The Windows build works on 64-bit systems only. I ran it on Windows 11 with no issues, but the Linux version is more unstable. If you are on Linux, use a virtual machine or dual boot to avoid driver conflicts with your SDR hardware. After installation, connect your SDR. Open Ghost Hunters and go to Settings. Set the sample rate to at least 2.4 MS/s. Lower rates will alias and produce false positives. Set the center frequency to somewhere in the 100 to 400 MHz range to start. This is a quiet band in most residential areas and gives you a clean baseline to calibrate against. The first run should be a calibration sweep. Click the Calibrate button and let it run for ten minutes without moving your antenna. The software builds a baseline noise floor. If you skip this step, every minor fluctuation in ambient RF will register as a ghost event. I made that mistake early on. My logs were full of noise that looked like interesting signals until I realized they were just my refrigerator compressor cycling on and off.
Get the Full Details
Configuring Detection Thresholds
This is the part most people skip and then complain the tool does not work. The default sensitivity is too high for real-world use. I recommend setting the threshold to between -85 and -90 dBm depending on your environment. Higher values in noisy urban areas, lower values in rural settings where ambient RF is minimal. You can also adjust the time window parameter. This controls how long a signal must persist before the software flags it. The default is 100 milliseconds. I lowered mine to 50 milliseconds after noticing the software was missing quick bursts that lasted only 30 to 80 ms. But going too low introduces false positives again. You are always balancing between sensitivity and accuracy. There is no perfect setting. Another setting worth adjusting is the frequency bin width. Smaller bins give you finer resolution but increase processing load. On a modern CPU this is not an issue, but if you are running this on older hardware, set the bin width to 50 kHz or higher. Anything below that and the software will choke on multi-hour captures.
Running a Capture Session
Once your thresholds are set, start a capture. I usually run sessions for two to four hours overnight. The software generates log files in CSV format by default. Each entry contains a timestamp, frequency, signal strength, duration, and a confidence score. The confidence score is the most useful field. Values above 0.7 are worth investigating. Values below 0.4 are almost certainly noise and can be filtered out. During the day I review the logs and cross-reference flagged events with known RF sources in the area. This takes practice. You learn to recognize the fingerprint of common emitters. A neighbor's security camera will show up as a repeating 2.4 GHz signal. A nearby cell tower will produce a steady presence around 700 to 800 MHz. These are not ghosts. They are just not interesting to you. The real work is in the events that do not match any pattern. I once logged a signal at 342 MHz that lasted 12 milliseconds with a confidence score of 0.91. No known service operates in that band in my area. It appeared once and never again. I recorded the raw IQ data and analyzed it in a separate spectrogram tool. It looked like a frequency-hopping signal, possibly military or government use, but I could not identify it. That is the honest answer. Some signals remain unexplained. That does not mean anything supernatural. It means your environment has RF sources you have not yet catalogued.
Common Problems and Workarounds
The most frequent issue is sample rate mismatch. If your SDR hardware reports a different sample rate than what Ghost Hunters expects, the timing data in your logs will be wrong. Always verify the sample rate in the SDR driver settings before starting a capture. The RTL-SDR drivers can report incorrect values on Windows. Use SDRto check the actual sample rate independently. Another problem is antenna resonance. Cheap antennas have sharp resonant peaks that can create the appearance of ghost signals at specific frequencies. If you keep seeing the same false ghost at one frequency, your antenna is likely the cause. Move to a different antenna or add a broadband preamplifier. A low-noise amplifier around 20 dB gain helps significantly. I use a Mini-Circuits ZFL-500LN+ and it cut my false positive rate roughly in half. The software also occasionally crashes during long captures. This is not rare. If you are running multi-hour sessions, set up automatic log rotation so you do not lose data when it crashes. The built-in option exists in Settings under Recording. Enable it and set the maximum log size to 50 MB. The software will split large logs into separate files automatically.
Interpreting Your Results
Do not treat every flagged event as significant. Most will be trivial once you understand your local RF environment. Build a reference library over time. Log every known emitter in your area and label them. After a month or two, the unknowns become easier to evaluate. The ones that still resist identification are the ones worth spending more time on. For deeper analysis, export the raw IQ data and process it in a tool like GNU Radio or SignalSpheres. Ghost Hunters gives you a first pass. It is a detector, not a full analysis suite. If you find something that interests you, spend another hour or two examining the raw signal characteristics. Modulation type, bandwidth, and pulse repetition rate can all help narrow down the source. The honest limitation of this tool is that it cannot identify unknown signals. It can only flag anomalies. You still need to do the work of figuring out what those anomalies actually are. That is the real value of using Ghost Hunters. It saves you from having to watch a spectrum display manually for hours. It does the watching for you and then hands you the interesting parts to examine.
