LFO

[jpcima]

This adds SFZv2 LFO on top of the mod matrix PR.

There are the basic targets for now (amplitude, pitch, etc)

Simple usage example

<region>
sample=*sine
lfo1_freq=1.0
//lfo1_amplitude=100
lfo1_pitch=100

[paulfd]

I added some tests, trying to automate them also on travis. It may fail, bear with me :)

[jpcima]

On reference data points, may we possibly flac them up to reduce their size?

[paulfd]

So if you're certain about the phase I guess we need to amend sfzformat.com, and if my solution for the reference sampling point is OK, I'm good with the rest 🙂

[jpcima]

Last commit introduces a terrible hack for Garritan instrument compatibility.

• if lfoN_phase is 0<=x<=1, it's considered with the ARIA meaning which is 0-1 normalized
• otherwise, it's taken as degrees, as standard SFZ suggests
• if the user adds trailing text to the value, this forces a particular choice of unit, eg. 0.5d → 0.5 degrees

[paulfd]

This seems like a reasonable solution to a messy problem!

[alcomposer]

lfoN_phase_onccX would then follow the SFZv2 standard, or use both as well?

[jpcima]

I have done some changes as follows to help make the work more mergeable

• control all of plot_lfo settings by arguments
• squashed the test commit and moved on top
• postponed the ARIA compatibility work, kept in a side branch; there is discussion to have on this topic, as every proposed solution has its advantages and inconvenients, and I'm not interested to take the decision here, just advance LFO for coming work

On the tests, I wish to have them preferably included in the regular test suite, I will add some work towards this direction.

[jpcima]

Regression tests are moved into the main test suite. Good for merge.
Do we keep these numpy files?

[paulfd]

I like the plotting one but it can go somewhere else. I'm not used to whipping out gnuplots on the fly :)

We still need something like it to validate changes visually and possibly update the reference.