Can an oscillator start at the same point every note on?

G2Psy

Simple bass notes need to start at the same point every note to sound consistent. There's nothing more annoying than a bass saw wave which sometimes has a click at the start and sometimes doesn't because the phase of the cycle is changing every note. Oscillators don't seem to let you set the cycle start and some LFO's do, but even an LFO for bass notes does not start at the same point everytime - I've never gotten it to work.

Can I do this somehow?

Regards,
B

Wout Blommers · Posted: Fri May 29, 2009 5:26 am Post subject:

In a spontaneous reaction couldn't work the Gate output into the OscSync input?

Wout

blue hell · Posted: Fri May 29, 2009 5:26 am Post subject:

Use the sync input for audio rate oscillators and the reset input for LFOs, and connect these to the note-on trigger (usually the same thing that would start your envelope).

Edit : oops, Wout beat me Smile

Wout Blommers · Posted: Fri May 29, 2009 5:29 am Post subject:

YES! I'm the winner!

That's what spontaneous does for you Wink

Wout

blue hell · Posted: Fri May 29, 2009 5:30 am Post subject:

G2Psy · Posted: Fri May 29, 2009 4:37 pm Post subject:

Thanks for both replying so fast. So I'm sorry for taking so long to reply because of the time difference.

Sync for the audio oscillators sounds good. BUT ... it doesn't work in this example I attach. The Keyboard module gate output goes to the osc sync input. It should reset the saw wave at the start of each note but you'll see it doesn't if you record a few notes.

B

Saw Bass Should Start at 0.pch2

Description:

Simple saw bass which should always start at phase 0, but doesn't.

Download (listen)

Filename:

Saw Bass Should Start at 0.pch2

Filesize:

1.21 KB

Downloaded:

1543 Time(s)

G2Psy · Posted: Fri May 29, 2009 4:47 pm Post subject:

Bloody hell, it DOES work now. It worked after saving the pch2 to attach to you and loading it back into the G2. That is totally weird, adding the gate to osc sync did not make a difference while I had the editor open and before saving the patch.

Please ban me from further posting for your own sanity's sake Very Happy

B

blue hell · Posted: Fri May 29, 2009 5:28 pm Post subject:

G2Psy · Posted: Fri May 29, 2009 6:05 pm Post subject:

I'm baaack....

The attached audio file shows that in fact they don't always start at the same point. These are unquantized notes and the audio shows some difference in the start point. Now for the freaky part: if I quantize those notes and record the audio - they DO all start at the same point.

OK, it is slight but you can hear it with the bass isolated.

B

Bad Saw.wav

Description:

I need to get out more

Download (listen)

Filename:

Bad Saw.wav

Filesize:

743.45 KB

Downloaded:

1081 Time(s)

Antimon · Posted: Sat May 30, 2009 12:30 am Post subject:

I think it's the filter that clicks here. I've edited your patch a bit, changed the saw to a sine and trigger by an LFO. I've you tweak the speed of the LFO you'll here clicks coming in different ways. But if you turn off the filter the clicking is consistent - the same whatever speed you tweak the LFO to.

I you move the filter so that it's after the ADSR I think it works better.

/Stefan

saw_bass_should_sbt_edit.pch2

Description:

Download (listen)

Filename:

saw_bass_should_sbt_edit.pch2

Filesize:

1.28 KB

Downloaded:

1984 Time(s)

iPassenger · Posted: Sat May 30, 2009 2:46 am Post subject:

Looking at the patch I imagine the difference is caused by the release phase of the envelopes, as the envelopes are set to "N", they do not restart on note on at beginning but from wherever the release has got to at the next note on.. switching them to "\/R" should solve the starts of each envelope.

BTW: this is just a guess based on looking at the patch not testing it.. so sorry if i am off the mark. Smile

G2Psy · Posted: Sat May 30, 2009 5:04 am Post subject:

These are great ideas, and I'm learning some fine details along the way. Sorry iPassenger, the reset for the envelope doesn't make a difference but I will definitely use this setting for most envelopes from now on. The blip is coming from the filter, as you correctly spotted Antimon, but I haven't found a way to remove the blips yet. I even inverted the envelope and set a longer attack thinking that the fast opening of the filter caused the blip but that is no different either.

This basic patch shows a lot about any synth engine, its my favourite "get to know your synth" patch. In this case it has taught me a lot about the G2 and revealed a mystery. IMO quite a few synths fail this test, but the G2 has this unexpected behaviour.

Regards to all,
B

Antimon · Posted: Sat May 30, 2009 5:28 am Post subject:

Well, there will always be a sharp edge at the start of the envelope, so some kind of tick should be expected.

You could approach the problem from the other side - if you can't set the phase at will, you can wait until the oscillator has the correct phase, and trigger the envelope at that moment. Attaching a patch that tries to do that, unfortunately it only works reliably for one pitch, if you tweak the oscillator it will start clicking again. :/

/Stefan

phase_matcher.pch2

Description:

Download (listen)

Filename:

phase_matcher.pch2

Filesize:

1.16 KB

Downloaded:

1878 Time(s)

Antimon · Posted: Sat May 30, 2009 5:30 am Post subject:

BTW, have you looked at the PM oscillator, maybe you can use that? Unfortunately it only has sine and triangle.

/Stefan

iPassenger · Posted: Sat May 30, 2009 5:47 am Post subject:

Yeah its definately the filter.. take that out and it is the same every time.. how odd... Must be part of the filter algo. Tried a few of the filters and they all exhibit the same behaviour.

Tried a few different modules too (sat module etc) to see if it was caused by the delay in processing the signal through more modules etc.. and they did not intoduce this problem.

The filters must introduce some kind of fixed delay.

iPassenger · Posted: Sat May 30, 2009 5:53 am Post subject:

Tried it with the perc osc and it doesn't seem to do it, so i tried as Antimon said with the env before the filter and that also seems fine..

saw_bass_should_sbt_edit_3.pch2

Description:

Download (listen)

Filename:

saw_bass_should_sbt_edit_3.pch2

Filesize:

1.2 KB

Downloaded:

1550 Time(s)

G2Psy · Posted: Sat May 30, 2009 7:02 pm Post subject:

Thanks for all your help guys. I have the perfect saw bass by putting the filter after the amp. Its been an interesting ride and demonstrates that you don't really know anything unless you verify it is actually doing what you assume it is doing.

For completeness, I am putting the final patch in this attachment.

Regards,
B

Saw Bass 1.pch2

Description:

The perfect saw bass.

Download (listen)

Filename:

Saw Bass 1.pch2

Filesize:

1.2 KB

Downloaded:

1457 Time(s)

xav · Posted: Tue Jun 02, 2009 7:07 am Post subject:

Hello,

This threat interests me very much since I already noticed this "problem":
http://electro-music.com/forum/topic-31044.html
I would really like to understand if the osc doesn't want to start at 0 when synchronized to gate, or if the noise comes from the EG.

I repatched a lot of sounds with EG before filters but I don't like that, because I like to patch several OSC to parallel filters... For instance a saw to a LPF and a Square to a BPF... and if the EG isn't after, this type of patch requires as many EG as parallel voice, and 4 morph groups if one desires the same setting for all the EG.

Tim Kleinert · Posted: Tue Jun 02, 2009 8:25 am Post subject:

xav · Posted: Tue Jun 02, 2009 8:53 am Post subject:

Yes Tim... it seems evident now. The G3 control rate will certainly be 96K. Laughing

Tim Kleinert · Posted: Tue Jun 02, 2009 1:41 pm Post subject:

xav · Posted: Tue Jun 02, 2009 2:05 pm Post subject:

blue hell · Posted: Tue Jun 02, 2009 2:20 pm Post subject:

I was thinking that when the envelope would be triggered from the same yellow signal that syncs the osc ... it should work ok then.

BTW Xavier, you don't need multiple envelopes, you can use Level Multipliers that you control from the envelope module (the Env out, as it labeled on most env. modules), that would at least be a bit cheaper than multiple envelopes.

xav · Posted: Tue Jun 02, 2009 2:35 pm Post subject:

G2Psy · Posted: Tue Jun 02, 2009 5:00 pm Post subject:

As OP of this thread (although it is basically the same issue as xav's thread) I can't help but think that the G2 architecture simply fails with envelopes and control signals. It is not normal or expected to get a pulse or click when a fast envelope modulates a filter or amp. Its just wrong. Sure you can find a work around, but you really shouldn't have to. The only good side to this issue is that many digital synths have this flaw. But the good ones (Virus TI for example) never do this.

Cheers,
B