Pipeline Analyzer

Pipeline Analyzer measures the response of the active Effect Pipeline without changing the audio you hear. It sits beside the pipeline in wide windows or below the pipeline header in narrow windows, letting you adjust effects while watching the response update.

Open it with the graph button in the Effect Pipeline header or View > Pipeline Analyzer in the desktop app. When Auto is enabled, pipeline edits trigger a new measurement automatically. Clear Auto to remeasure pipeline edits only when you click Refresh measurements; changing the channels, speaker responses, or measurement settings always starts a new measurement.

Choosing channels and speaker responses

Select one input channel from the active pipeline (up to 16 channels) and up to four output channels using + Add Output. At least one output always remains, so the last one cannot be removed. Deleting an output also removes its speaker-response setting.

Each output can use No speaker IR or a saved measurement point for the connected speaker or driver:

Without speaker IRs: Before is an ideal flat reference impulse, and After shows the response of the effect pipeline alone.
With speaker IRs: Before is the combined response of the selected speaker measurements, and After is the combined response after the pipeline processes each output. This lets you verify a Channel Divider or FIR Crossover together with the actual driver measurements.

A saved response that is no longer available stays selected but appears as — until you replace or clear it.

Because separate speaker measurements are aligned at their detected onsets and do not preserve physical arrival-time differences between drivers, configure relative driver delay and polarity in the pipeline before evaluating the summed response.

Measurement settings

Open Measurement settings to adjust how the pipeline is measured:

Signal: Log Sweep (default), MLS, TSP, or Unit Impulse. Use Log Sweep to inspect harmonic distortion as well as the pipeline’s response. Its sweep covers 20 Hz–20 kHz within the sample rate’s limits. MLS and TSP use periodic signals; Unit Impulse captures the impulse response directly.
Level: Peak test-signal level (default -12 dBFS). Linear filters measure the same at any level, whereas compressors, saturators, and other level-dependent effects change with Level.
Requested sweep length (Log Sweep) / Sequence Length (other signals): The default requested length for Log Sweep is 131072 samples. It may use a longer sweep to separate harmonics; the actual Sweep length and Total stimulus time appear after measurement. Its response covers up to 1.5 seconds after the main peak regardless of sweep length. To inspect longer reverb tails, delays, or filter ringing, select MLS, TSP, or Unit Impulse and increase Sequence Length as needed.
Stabilization Periods: Number of periods played before capture (default 12, used by MLS and TSP) so slow-moving effects can settle. This setting is unavailable with Log Sweep and Unit Impulse.
Averages: Number of captures averaged together (default 2, used by Log Sweep, MLS, and TSP) to reduce variation.

Reading the graphs

Use the Graph selector above the plot to switch views without remeasuring:

Frequency: Magnitude response (dB) versus frequency, referenced to 0 dB.
Phase: Phase response (degrees) versus frequency.
Min Group Delay: Group delay implied by the minimum-phase magnitude response.
Excess Group Delay: Remaining group delay after removing the minimum-phase part (the plot is limited to ±100 ms, while cursor readouts show the actual value), making pure delay and linear-phase or all-pass timing easy to spot.
Impulse: Time-domain impulse response from -2 ms to Impulse Range (ms), normalized to its peak. Use the mouse wheel over the graph to zoom the time axis around the point under the cursor.
Distortion: With Log Sweep, shows harmonic distortion as a percentage of the fundamental across frequency. H2, H3, and THD (orders 2–3) are shown initially; H4–H10 can be added without remeasuring. Larger values mean more distortion. Select Log Sweep if this graph is unavailable with another signal. A silent or otherwise invalid measured output cannot provide distortion data. Saved speaker responses contribute their linear response, not any distortion from the speaker itself.
Waterfall: Shows frequency, relative level (dB), and decay time (ms). Choose After (default) or Before under Response to inspect the pipeline output or the selected speaker responses before processing. Each view is normalized to its own peak, so compare the decay shape rather than absolute levels. Drag to rotate the 3D plot; use the mouse wheel or pinch to zoom.

On the frequency, phase, group-delay, and impulse graphs, hover to read Before and After values at the cursor; hover over the Before legend label to temporarily hide After. Use Smoothing (oct) for the frequency, phase, and group-delay plots, and Impulse Range (ms) for the impulse plot.

How the measurement works

Pipeline Analyzer captures the active pipeline, its routing, and any selected speaker responses, and automatically compensates After for latency reported by the pipeline. For nonlinear, modulated, or random effects, treat the plot as a snapshot at the current settings and Level. To confirm the result with a microphone, see Measuring with the Effect Pipeline in Acoustic Measurement.

← Back to README