Mister Sombrero
Offline

What types of single cycle waveforms are best for use in LGPT's oscillator mode?

1.) Should the SCWF be exactly A440Hz for the tuning to be calculated correctly?

2.) In the LGPT10k sample pack most of the SCWFs are $100 or $FF samples in length..is there an advantage in using this length for oscillator mode? (perhaps with PLOF, LPOF or RTRG?)

3.) What exactly is the difference in the way a SCWF is looped in loop mode and oscillator mode?

4.) What is the advantage of using oscillator mode rather than normal loop mode? Should I just use normal loop mode?

LGPT mailing list:

"[LGPT] makes certain assumptions about the root note of the cycle...assuming that your
oscillator waveform is tuned to A440 or whatever the base assumption is
about that mode (it's been a while since I made a Piggy track with oscil
so I can't remember)"

(sorry for posting all these questions at once)

Last edited by Mister Sombrero (Jul 3, 2013 5:00 pm)

herr_prof
Offline
Unsubscribe

1)  CORRECTION: Osc mode doesn't care for the tuning nor assume anything for the length of the sample. It uses the sample data's length as a way to determine its pitch (assuming only it's a single cycle) and tunes it from there.

2. A single waveform cycle should be about FF in length at 441. Check out the MONORAVE compo for longer waveforms. Read more about the commands in the manual.
http://gorehole.org/lgptWiki/doku.php?i … ce_manual. You get better results with them with longer waveforms in my opinion.
3. Loop mode doesnt make any assumptions about the root pitch.

edit: Everything else I thought i knew was a lie.

Last edited by herr_prof (Jul 8, 2013 2:26 pm)

herr_prof
Offline
Unsubscribe

Of course rules are meant to be broken, so if your song sounds better in one versus the other, who cares.

Mister Sombrero
Offline

Thankyou for the help and the info. Much appreciated. monowave looks cool, many instruments out of only one sample.

By  "a single cycle waveform will be FF in length" does it mean:

oscillator mode is designed to use only SCWFs of length FF? (longer or shorter SCWFs will be out of tune) / I should only use SCWFs of length FF in oscillator mode?

The reason I am asking is that the single cycle waveforms within monowave appear to be 256 long instead of 255 "every 0x100 increment skips to the next waveform"

Thanks for the help

Last edited by Mister Sombrero (Jul 4, 2013 3:23 pm)

herr_prof
Offline
Unsubscribe

the period of one waveform in a440 is about ff in length usually. Sometimes its different, but its a good place to start if you dont know what you imported. You may have to tweak this value based on whatever source audio you use, because the period of a a440 waveform will be shorter than say a period of a lower note on the scale, and way longer than a higher pitched waveform.

http://library.thinkquest.org/19537/Physics3.html

If you say grabbed a piece of random audio and tried to grab waveforms out of that, it may be totally waveform periods.

Mister Sombrero
Offline

Hey, thanks again.

Found this website for making A440 basic waves to be cut into single cycle afterwards:

http://www.audiocheck.net/audiofrequenc … letone.php

jefftheworld
Offline
Toronto, Ontario, Canada

I happen to have just posted a tutorial on using audacity to create SCWFs: http://www.youtube.com/watch?v=kjPR9tEKe_s

If you want to tune to A440, just replace the frequency for middle C that I provided and use (you guessed it) 440Hz instead.

With that tutorial you can generate a single sine, square or saw and the edit it by hand to alter the timbre. You can draw additive or combination waveforms that mimic how certain old soundchips would sound or create entirely new and interesting timbres.

Last edited by jefftheworld (Jul 5, 2013 1:56 pm)

Mister Sombrero
Offline

I made some basic waveform samples at A440 on audiocheck.net (440Hz, 1 sec, -3db, 8Khz).

I cut them down to be single cycle waveforms using renoise, chopping at zero crossing to get exactly 1 cycle.

Loaded them into LGPT. Changed to oscillator mode...loop end was 12, Root note was A... every note was played out of tune...

After I changed the root note to Csharp..every note was played perfectly in tune (notes in the pattern editor matched the pitch of notes played).

LGPTs oscillator mode works in a weird way.

I tried single cycle waves (from the same website) at A440,44.1Khz as well and these were not really tuneable. No matter what the root note was, the pitches did not reflect the notes shown in the pattern editor. weird.

Seems like the oscillator mode requires a particular length of ?A440? single cycle wave, perhaps also a particular sampling frequency to work properly...maybe it matches gameboy waveforms or SID waveforms or something else?

Last edited by Mister Sombrero (Jul 5, 2013 2:30 pm)

jefftheworld
Offline
Toronto, Ontario, Canada
Mister Sombrero wrote:

I made some basic waveform samples at A440 on audiocheck.net (440Hz, 1 sec, -3db, 8Khz).

I cut them down to be single cycle waveforms using renoise, chopping at zero crossing to get exactly 1 cycle.

Loaded them into LGPT. Changed to oscillator mode...loop end was 12, Root note was A... every note was played out of tune...

After I changed the root note to Csharp..every note was played perfectly in tune (notes in the pattern editor matched the pitch of notes played).

LGPTs oscillator mode works in a weird way.

I tried single cycle waves (from the same website) at A440,44.1Khz as well and these were not really tuneable. No matter what the root note was, the pitches did not reflect the notes shown in the pattern editor. weird.

Seems like the oscillator mode requires a particular length of ?A440? single cycle wave, perhaps also a particular sampling frequency to work properly...maybe it matches gameboy waveforms or SID waveforms or something else?

I know that it assumes oscillator waveforms to be at A440 and adjusts accordingly so you're supposed to leave the root note at C.

A single cycle at a given pitch at a given sample rate will be an exact length. Cutting several data points short or long will start to detune the pitch. This will be more apparent at lower sample rates as each data point will represent a longer period of real time.

Last edited by jefftheworld (Jul 5, 2013 3:00 pm)

Mdashdotdashn
Offline
herr_prof wrote:

1) OSC mode makes an assumption that the root pitch is going to be a440, it will scan the waveform and do the pitch for you automatically (You quoted the way it works). More or less. Some waveforms feel pitchier to me so you may have to futz with the base tuning to get it back into the proper tuning as i mentioned in the other thread.
2. A single waveform will be FF in length. Check out the MONORAVE compo for longer waveforms. Read more about the commands in the manual.

Actually this is not at all true big_smile.

Osc mode doesn't care for the tuning nor assume anything for the length of the sample. It uses the sample data's length as a way to determine its pitch (assuming only it's a single cycle) and tunes it from there.

herr_prof
Offline
Unsubscribe

Yea Sorry! Will Correct!

herr_prof
Offline
Unsubscribe

Would you agree about answer 3 and 4? Its what I assumed from usage, but Im realizing I dont know shit.

Mdashdotdashn
Offline
herr_prof wrote:

Would you agree about answer 3 and 4? Its what I assumed from usage, but Im realizing I dont know shit.

Not really. The sample 'rendering' is exactly the same. You could see osc loop simply as a convenience setting so that tuning / root key is automatically set.

herr_prof
Offline
Unsubscribe

!