371 years LFO

Tim Kleinert · Posted: Sun Jan 26, 2014 4:19 pm Post subject: 371 years LFO

I was thinking about the numerical resolution of the G2 system in relation to its sampling rate, which, among other things, is the reason why the slowest LFO speed is 699 seconds (2^24/24000).

So, as a thought experiment, I made a DIY sawtooth LFO with a cycle time of 371 years.

Why?

Because I can. Laughing

EDIT: The 6.23 billion years LFO available upon request. Just 6 more modules.

371 years LFO.pch2

Description:

Download (listen)

Filename:

371 years LFO.pch2

Filesize:

1.44 KB

Downloaded:

4223 Time(s)

blue hell · Posted: Sun Jan 26, 2014 5:55 pm Post subject:

I think 699 s is a bit short at times .. 371 years would be too much for most practical purposes ... but lets say one hour or a day maybe and then with some different wave forms .. i could actually use that I guess.

Had a quick look at the patch .. could use a couple of lines of how it works ... and how I could change the speed.

In the past I've used another way to make things slow BTW ... using a delay controlled with a saw .. that is .. downsampling an LFO.

For random I'd now use a sample and hold I guess .. but would be nice then if the smoothing could be set to be really long too ...

Anyways .. just some babblings

6 point how many billion years Shocked

you're nuts Laughing

I'd like to see that one too tho ... but again .. with a couple of comment lines please.

Tim Kleinert · Posted: Mon Jan 27, 2014 6:13 am Post subject:

It was just a little thought experiment using two nested counter circuits counting the smallest possible increment within the 24 bit numerical system @24kHz. Counting to 2^48 at 24000 times a second takes 371 years. Laughing

The 6 billion version would simply add another counter stage. (2^72)

I'll have to (happily) eat my words though pertaining the G2 LFOs, because some further testing revealed that their internal counters DO in fact seem to have an even higher resolution than 24 bit after all. Meaning: If you simply apply a negative constant to the LFO mod input, it goes even slower than the slowest base setting of 669 seconds.
The mod inputs can take a maximum of +/- 64 units and track exponentially (unattenuated). So, if the math still is correct at these low rates, a -12 constant (12 semitones or an octave lower) should drop the LFO frequency by half, -24 by four, etc. -60 (5 octaves, or factor 32) should give you 5.94 hours, and -64 even more than that.

I haven't measured it precisely (*), but it's definitely running waaay slower. Try it.

Come to think of it, in this case one could patch the 6 billion years LFO simply with two LFO modules (using the concept of interference and modulating the phase of one with the other). Laughing

(*) EDIT: I tested it and it is robust. The test patch below demonstrates a sawtooth LFO at 669 seconds and another one tuned down 5 octaves and its range boosted by factor 32 -both modulating test oscillators panned L and R. The pitches match up perfectly. Well done, Clavia. thumb up

slow LFO test.pch2

Description:

Testing the tracking of LFOs at slowest rates. See topic for discussion.

Download (listen)

Filename:

slow LFO test.pch2

Filesize:

1.3 KB

Downloaded:

4187 Time(s)

Tim Kleinert · Posted: Mon Jan 27, 2014 10:02 am Post subject:

Addendum to the post above:

Electromagnetic Wave · Posted: Tue Jan 28, 2014 12:04 am Post subject:

Hi Tim! This topic is very interesting. Since few days, I studied in detail the numerical resolution and the sampling rate of the G2.

dorremifasol · Posted: Thu Jan 30, 2014 10:00 am Post subject:

Great! I'm going to start testing it. Excuse me if I die before showing any results! Wink

Tim Kleinert · Posted: Thu Jan 30, 2014 5:37 pm Post subject:

Roland Kuit · Posted: Fri Jan 31, 2014 3:30 am Post subject:

drapdap · Posted: Sat Feb 15, 2014 11:59 am Post subject:

nice patch... mp3 demo? Very Happy

can you estimate the lifetime of a g2? Very Happy