Frequency Stability Analyzer
Measure the stability of any frequency source in real time. Analyze jitter (standard deviation), Allan deviation, PPM drift, and settling time with professional-grade plots and CSV export — all processed locally in your browser. No data is ever recorded or uploaded.
Frequency Stability Analyzer Tool
| Samples | 0 |
| Mean | — Hz |
| Min | — Hz |
| Max | — Hz |
| Range | — Hz |
| Duration | 0.0 s |
| Time | Hz | Δ Hz | PPM |
|---|
How to Use the Frequency Stability Analyzer
-
Select Microphone & Interval
Choose your audio input from the Microphone dropdown. Set the Measurement Interval — 10 ms for rapid sampling or 500 ms for longer-term stability analysis. Adjust Sensitivity to suppress background noise.
-
Start Measuring
Click "Start Measuring" and grant microphone permission. Play a steady tone from a signal generator, tuning fork, or instrument. The tool begins collecting frequency samples immediately at your chosen interval.
-
Monitor Jitter & Stability Grade
Watch the Jitter value (standard deviation in Hz) and PPM reading update in real time. The Stability Grade rates your source as Excellent (<1 PPM), Good (1-10 PPM), Fair (10-100 PPM), or Poor (>100 PPM).
-
Analyze the Plots
The Allan Deviation plot shows frequency stability at different averaging times on a log-log scale. The Histogram reveals the distribution of measured frequencies. The Frequency vs Time chart shows drift and settling behavior.
-
Export Your Data
Click "Export CSV" to download all measurements with timestamps, Hz values, deviations, and PPM readings. Use the data log to review individual readings or import the CSV into a spreadsheet for further analysis.
Understanding Your Results
Jitter (Standard Deviation)
Jitter is the standard deviation of all measured frequency values. A lower jitter indicates a more stable frequency source. For a perfect 440 Hz tone, jitter should ideally be below 0.5 Hz. High jitter may indicate a noisy signal, unstable oscillator, or environmental interference.
PPM (Parts Per Million)
PPM expresses frequency deviation relative to the mean frequency: PPM = (stddev / mean) × 106. This normalized metric lets you compare stability across different frequencies. A crystal oscillator typically achieves <1 PPM, while a software-generated tone might show 5-50 PPM depending on system load.
Stability Grade
The grade provides a quick assessment: Excellent (<1 PPM) is laboratory-grade stability, Good (1-10 PPM) is suitable for most applications, Fair (10-100 PPM) indicates moderate drift, and Poor (>100 PPM) suggests significant instability that may need attention.
Allan Deviation (ADEV)
The Allan deviation plot shows how stability changes with averaging time (tau). Short tau values reveal high-frequency noise, while long tau values expose slow drift. A slope of -1 on the log-log plot indicates white frequency noise, -0.5 indicates flicker frequency noise, and +0.5 indicates frequency drift.
Settling Time
Settling time measures how long it takes for the frequency to stabilize within a ±threshold band and remain there for at least 5 consecutive seconds. This is critical for evaluating oscillator warm-up behavior and transient response after power-on or frequency changes.
Frequency Histogram
The histogram shows the distribution of all measured frequency values. A narrow, tall peak indicates excellent stability. A wide or multi-modal distribution suggests drift, instability, or mode-hopping. The shape can reveal whether noise is Gaussian or contains systematic patterns.
Technical Background
Allan Deviation & Frequency Stability
The Allan deviation (ADEV), introduced by David W. Allan in 1966, is the gold standard for characterizing frequency stability. Unlike standard deviation, which diverges for certain noise types common in oscillators, Allan deviation converges for all power-law noise processes encountered in practice. It is defined as the square root of the two-sample variance: for a series of frequency measurements averaged over time tau, ADEV equals the RMS of consecutive frequency differences divided by √2.
This tool computes the overlapping Allan deviation, which uses all possible overlapping samples at each tau value. This provides better statistical confidence than the non-overlapping estimator, especially important when the total measurement duration is limited. The log-log ADEV plot reveals the dominant noise process: white phase noise (slope -1), flicker phase noise (-1), white frequency noise (-1/2), flicker frequency noise (0), and random walk frequency (+1/2). Identifying the noise type is essential for predicting long-term stability and selecting appropriate filtering strategies.
Jitter & Short-Term Stability
Jitter quantifies cycle-to-cycle frequency variation and is computed here as the standard deviation of all frequency readings. While simpler than Allan deviation, jitter provides an intuitive measure of short-term stability. For audio applications, jitter manifests as pitch wobble or warble. In telecommunications and digital systems, excessive jitter causes bit errors, clock skew, and synchronization failures. This tool measures jitter at the configured sampling interval, so faster intervals (10 ms) capture higher-frequency jitter components.
PPM & Normalized Deviation
Parts per million (PPM) normalizes the frequency deviation to the carrier frequency, enabling meaningful comparison across different frequencies. A 1 Hz deviation at 1 MHz is 1 PPM, while the same 1 Hz deviation at 100 Hz is 10,000 PPM. Crystal oscillators are typically specified in PPM (e.g., ±20 PPM for a standard crystal, ±0.5 PPM for a TCXO). The PPM reading in this tool uses the session mean frequency as the reference, and the standard deviation as the numerator, providing a real-time measure of relative stability that updates as more samples are collected.
Frequently Asked Questions
What is frequency stability and why does it matter?
Frequency stability describes how consistently a signal maintains its intended frequency over time. It matters in telecommunications (carrier accuracy), music (pitch consistency), electronics (clock precision), and scientific instrumentation (measurement accuracy). Poor stability causes signal drift, timing errors, and degraded performance in any frequency-dependent system.
How long should I measure for accurate results?
For basic jitter and PPM readings, 10-30 seconds provides reasonable results. For meaningful Allan deviation plots, measure for at least 60 seconds — longer measurements reveal lower-frequency noise processes and drift. The ADEV plot requires at least 3 samples at each tau value, so the maximum useful tau is roughly one-third of your total measurement duration.
What measurement interval should I use?
Use 10 ms for capturing fast jitter in electronic signals. 50-100 ms is ideal for most audio and musical applications. 500 ms works well for long-term drift monitoring where you want to observe trends over minutes or hours. Faster intervals generate more data points but may show more measurement noise from the FFT process itself.
What do the stability grades mean?
Excellent (<1 PPM) indicates laboratory-grade stability typical of precision oscillators and high-quality signal generators. Good (1-10 PPM) is suitable for most practical applications including music and audio. Fair (10-100 PPM) shows moderate instability that may be acceptable for casual use. Poor (>100 PPM) indicates significant drift or jitter that likely needs correction.
How is Allan deviation different from standard deviation?
Standard deviation measures the spread of all values from the mean, but it assumes stationary (non-drifting) data — it diverges for random-walk and flicker noise common in oscillators. Allan deviation uses consecutive differences, making it insensitive to linear drift and convergent for all noise types encountered in real oscillators. The ADEV plot also reveals the type of noise process through its slope, providing diagnostic information that standard deviation cannot.
Is my audio data private?
Yes — 100% private. All frequency detection and stability analysis runs entirely in your browser using the Web Audio API. No audio is recorded, stored, or transmitted to any server. The tool works completely offline once loaded. Only numerical frequency values are stored in memory for analysis, and all data is cleared when you close the page or click Reset.
Related Tools
Frequency Detector
Real-time Hz detection with multi-frequency separation and pitch accuracy meter.
Try it →Drift Monitor
Track frequency drift over time with trend analysis and deviation alerts.
Try it →Tone Generator
Generate precise sine, square, triangle, and sawtooth waveforms at any frequency.
Try it →