"Blackboxing" a chorus

[Ichad.c]

Firstly, http://en.wikipedia.org/wiki/Black_box_testing

I have this chorus in my rack that I've always wondered about for ages, so now it's time to throw some test signals at it - > it only has 3 parameters - Freq/Depth/Mix.

So how would I go about it efficiently?

Mix 100% - Freq Min(0.1Hz is the lowest it goes) ?

1) throw a pos only peak at it = this would tell me how many delays there are, what the time differences are, and if there is any level differences.

2) Pure Sine = THD measurement.

3)Slowly sweeping Square wave = check phase.

4)White noise = min/max excursion of notches should tell me what the depth parameter does.

5) Modulator(s) waveshape ?

Any other test signals I can throw at it? And more importantly - is there anything I should look out for that might not be obvious? No.4&5 seems like it might be hard to determine...

Andrew

[BertKoor]

A chorus is a modulated delay. My guess is it's heart is a BBD (bucket brigade delay) with the playback sampling rate being modulated by a LFO, giving variations in pitch.

If this one is not that different from the usual ones:

* Freq & Depth knobs are responsible for the modulation. If Depth can be zero, then Freq makes no difference any more.
* Mix knob only controls dry/wet signal ratio.

First set Freq & Depth to zero. In this setting it should act closest as possible as a Delay effect. Send impulses through it to measure delay time (mix knob set at 50%) and measure the frequency response (sine sweep, convultion) with Mix knob set to completely wet AND completely dry. You could also use the RMAA test suite for a complete analysis of the sound character.

To test the LFO that modulates the delay, send a simple steady sine tone (1 kHz should do) through it. Set mix to complete wet and use a spectrum analyser to see visually what it does.

Just some ideas...

[Ichad.c]

A chorus is a modulated delay. My guess is it's heart is a BBD (bucket brigade delay) with the playback sampling rate being modulated by a LFO, giving variations in pitch.

Actually this device under test is a 20+ year old digital. So I'm going in blind.

RMAA test suite for a complete analysis of the sound character.

Thanks, this will come in handy.

To test the LFO that modulates the delay, send a simple steady sine tone (1 kHz should do) through it. Set mix to complete wet and use a spectrum analyser to see visually what it does.

This I think is going to be the hard/fun part, not doing the test per se, but actually interpreting the data.

Andrew

[avasopht]

To test LFO delay modulation I would most likely test for impulse response. I would then transform the impulse response into events, which would consist of: { sample, amplitude }. I could then compare that data with my expected results, such as:

> expect (sample: approx(30), amplitude: approx(2517))

If the transformed impulse has the following event, "sample(29), amplitude(2518)", you can consider it close enough, and in your test suite you can allow for varying levels of accuracy. But because of the resampling that will occurs with the delay LFO it might be best take the first derivative (x[n] - x[n-1]) of the output signal for the impulse response.

Just a thought.