Measurements
The numbers.
The numbers on this page come from Level Box's own offline test harness, run against the shipping code and regenerated for every release. This is version 0.3.0.
At a glance
Latency
0samples
no oversampling, nothing to report
CPU
200+instances
on one M1 with CONSOLE on
Unity
0.00dB
the committed level is the level out
Bench
±2dB
second harmonic vs a real console channel at 20 Hz, +18 dBu and up
Aliasing
−120dBc
worst spur, 15 kHz at +24 dBu
Console
Distortion as you raise the level
Total harmonic distortion with CONSOLE on, against the level leaving the plugin, sine at the stated output level, 48 kHz, bridging load. The line is the model. The dots are a real console channel measured on our bench at the same levels, and the thin rule beneath them is the measurement chain's own floor: dots on the rule mean the unit is at or below it. One scale for every panel.
1 kHz
100 Hz
50 Hz
20 Hz
Also measured
- CONSOLE off, the output is bit-identical to the input times the committed gain, and the stage costs nothing.
- CPU with CONSOLE on: 0.47 % of one M1 core per stereo instance at 48 kHz, so 200 instances is under one core.
- Second harmonic leads at every frequency, from the standing current through the core; at 20 Hz the model is within 2 dB of the bench from +18 dBu up and reads a little high below that.
- The bench unit's 20 Hz figures include a steady third harmonic of about 0.05 % that does not change with level; it is at the measurement chain's floor and is not in the model.
- Response flat from 20 Hz to 20 kHz, +0.3 dB at 10 Hz; clips above +30 dBu into a bridging load.
Method
Every number here is produced by the plugin's own offline test harness against the shipping code, on an Apple M1, and regenerated for each release. The transformer's constants were set against a real console channel measured on our bench; the dots on the charts are that unit, measured at the same levels through a Focusrite interface whose own floor is about 0.0015 % at 1 kHz and 0.05 % at 20 Hz. This is version 0.3.0.