Messages

61

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 22, 2023, 10:54:55 PM »

I hopefully fixed the two issues
- fix for the aliasing in Ping & Diff fx : added extra filtering in those
- fix for the bad noise click when going to the Diff fx

Here is an attached fw file.

There is a pull request as well !
https://github.com/Ixox/preenfm3/pull/41

62

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 14, 2023, 10:27:47 AM »

No problem, thanks for reporting.
i'll have a look at this issue next week.

63

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 13, 2023, 11:46:28 PM »

Hi Xavier,
excellent addition !
it will be very useful to have both fx layer.

about the Diffuser click, i'm not sure what's happening, i need to do some tests,
but I'm currently on vacation so i can't do anything until next week.

I had a look at the PR : nothing to complain on my side.

can't wait to test that firmware !

64

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 12, 2023, 05:43:26 PM »

Sure, here is the commit :
https://github.com/pvig/preenfm3/commit/dbab9d26872e675900df2eb76705185d585e0cde

65

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 07, 2023, 01:47:23 PM »

Hello @superfake,
thank you for your detailled report, i previously did not use spectrum analysers, only my ears, which has now some hf loss i fear.
So here is a version with more lp + notch filtering around 12khz on those fx, i hope it will make it usable, though it remains some aliasing.
Please, let me know if that's ok for you.

While i was at it, i added an optimisation i found lately on all the pitch shifter fx, which save quite some precious cycles !

66

preenfm2 and preenfm3 / Re: paraphonic effects

« on: May 06, 2023, 12:04:38 PM »

Hi superfake,

The delay fx could be done with a sample rate reduction trick, which is x4 factor, so yes, the 12khz noise is awaited..
I did put some lowpass to soften this, but it's far from perfect, let's say it's a feature and call it vintage delay ^^
Though i will try some notch filtering to minimize the noise.

about the delay length, there is not so much ram left... better keep it for other features.

67

preenfm2 and preenfm3 / Re: paraphonic effects

« on: April 24, 2023, 07:10:42 AM »

yes, the menu option is unlocked, no need for CC access.

68

preenfm2 and preenfm3 / Re: paraphonic effects

« on: April 23, 2023, 05:20:55 PM »

Hi,
after some bugfix & enhancements, here is a quite mature version of these fx.
I removed the karplus strong fx, it was not so useful as an effect. Instead, there is a diffuser, to blur transients or add a light reverb "touch" to the sound.
There was improvements in some fx : the pitch shiters were simplified to lower cpu usage, it allowed to add some Eq ing in these to retrieve the lost bass, same for Bode filter.

@Xavier : pull request done in github.

69

preenfm2 and preenfm3 / Re: FM3 MIDI CC list?

« on: April 18, 2023, 11:03:58 PM »

there is also CC_MIXER_SEND ( CC#11 ) for reverb dry/web, for each timbre

70

preenfm2 and preenfm3 / Re: Envelope curve selection

« on: April 15, 2023, 08:56:50 PM »

no, there is no one cycle mode for lfos, but its worth saying that you can do cool things with the matrix : the mtx1-4 destination let you modulate the level of anything in the matrix slot 1 to 4,
so a lfo can last a few cycles when modulated by an envelope for example.
The preenfm matrix is already quite powerful !

71

preenfm2 and preenfm3 / Re: Envelope curve selection

« on: April 03, 2023, 06:52:17 PM »

I reckon it would be nice to add this feature to the aux env,
but these envelopes do not works the same way as the osc ones, and it's not an easy task to adapt the code

Maybe one day, but don't hold your breath !

72

preenfm2 and preenfm3 / Re: paraphonic effects

« on: March 09, 2023, 09:37:35 PM »

I wish there were a way to cure aliasing but i fear we must live with it.
The usual way to lower aliasing is by using the note scaling in the modulation matrix, with modulation index as destination.
For example, full modulation at C4 up to pure sine at C8 (only the carrier left, no more modulation).

73

preenfm2 and preenfm3 / Re: paraphonic effects

« on: March 06, 2023, 08:07:00 AM »

Hi ElVicente,

yes, oversampling is definitively doable, but i'm sure it would be so useful :
it would be good to make a very clean filter, but the aliasing already there is not easily removable.
however, i have to do some tests.

I agree about the harmonizer : as it does not use the scala tuning and not even regular tuning, it's difficult to have a usefull result, works best on drums  . I will try to change that.
I did tuned delay line with the pluck algorithm, it may be usable in the harmonizer.

About the chorus and flanger, i agree it is not ideal, but i'm not sure how to do it : one may like using multiple lfo and other modulation source.

74

preenfm2 and preenfm3 / Re: paraphonic effects

« on: March 04, 2023, 11:55:53 AM »

Still a work in progress...
I attached on the first post of this thread a new fw version, with 2 more fx :
- ping pong delay : very short one, but may be of some use
- Karplus Strong : 2 plucked voices on top of the instrument
 

75

preenfm2 and preenfm3 / Re: paraphonic effects

« on: February 26, 2023, 11:42:55 AM »

Hi Xavier,

these fx can be found in the FILTER_TYPE list, so on the Filter page. (after unlocking access with CC70 = 54 for example)
the code is in Timbre.cpp, is starts here :
https://github.com/pvig/preenfm3/blob/532fcac3ccb117e609c812c4b0ef09173416787e/firmware/Src/synth/Timbre.cpp#L759

the bigger cpu impact is up to 5.6 % for the Harmonizer and the Widener algo (both are double pitch shifters),
the least is for the delay which is like 2%.

Attached file is  a fw iteration of this :
- delay has now some highpass in the feedback path (more feedback, more hp)
- the Bode fx sounds more like it should on pure sine (bug fix). It's also a bit less harsh when feedback : did put some lowpass in there.

About the Bode fx, the frequency shift may be up or down : for values above or below 0.5
Same behaviour for the pitch shifting algo : The pitch is one octave below for value 0.25, one octave above for val = 1.