You are here

Yoshimi extras

For all other computers and operating systems, including Atari, Linux and mobile apps.

Moderator: Moderators

Re: Yoshimi extras

Postby Martin Walker » Sat Aug 22, 2020 7:52 pm

Indeed - I'd love to find out more about the community and all the different things they are doing with Yoshimi.


Martin
User avatar
Martin Walker
Moderator
Posts: 17207
Joined: Wed Jan 13, 2010 9:44 am
Location: Cornwall, UK

Re: Yoshimi extras

Postby Folderol » Mon Aug 24, 2020 1:45 pm

Well, I must say I haven't thought of that, but surely it still needs someone to manage the forum, dispose of spammers etc. and that's not going to be me. Indeed, I'm so saturated with this stuff, that from Wednesday through to the following Monday, I will be taking a sabbatical from all things technological, meandering around the New Forest.
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby blinddrew » Mon Aug 24, 2020 10:11 pm

Depends how you set it up and what you want it to do really. The world is very much your mollusc.
User avatar
blinddrew
Jedi Poster
Posts: 13650
Joined: Sun Jul 05, 2015 12:00 am
Location: York
Ignore the post count, I have no idea what I'm doing...

Re: Yoshimi extras

Postby Folderol » Thu Sep 03, 2020 7:48 pm

Well, after my break from all things tech and a mental reset, I've decided that this is yet another thing I simply can't get involved in. I've put a post on the Yoshimi email list, asking if people are interested, and for someone to to take it on. If someone does, then I'll probably contribute from time to time but that's all.

Currently, this is what I'm already involved in:
Yoshimi development (the biggie)
Yoshimi Advanced User Manual
Yoshimi Website
Yoshimi Users list
Yoshimi Pi development
New Yoshimi Pi website
Preparation for November Linux Audio Conference (assuming it takes place).
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby blinddrew » Thu Sep 03, 2020 7:57 pm

And what do you do in your spare time?
;)

I would suggest that with the blog the first thing is to keep it simple, the second is to encourage your contributors to pace themselves.
:thumbup:
User avatar
blinddrew
Jedi Poster
Posts: 13650
Joined: Sun Jul 05, 2015 12:00 am
Location: York
Ignore the post count, I have no idea what I'm doing...

Re: Yoshimi extras

Postby Folderol » Fri Oct 09, 2020 8:19 am

Unfortunately a bug has been discovered concerning detune. This gives the wrong detune degrees for both fine and coarse detuning. However it only affects AddSynth Global. The individual voices, modulators, SubSynth and PadSynth are fine.

This has been present for over a year, so affects all Yoshimi Pi units out there. It is corrected in the current 'master' for those wanting to fetch it.
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby Folderol » Fri Nov 20, 2020 4:30 pm

For those diving into the instrument editors...

I had someone rather puzzled by the fact that he didn't seem to get much response from the frequency envelope controls. The reason was that he had a very short amplitude envelope.

An excerpt from the notes!

"
The first thing to keep in mind is that amplitude envelopes (particularly release time) set the point at which a note ceases. Frequency/filter envelopes can be shorter, so their effect stops part way, but if they are longer, the last part will be ineffective.

Across all three engines, and kits (if kit mode is active) it is whichever is the longest that sets the overall time of the note, and you may well hear other sections stop if the times are sufficiently different.

Also, within AddSynth itself, it is which ever voice has the longest envelope that sets the overall voice time, and if you set voices with very different characteristics you can hear the shorter ones finish before the overall sound stops. Bear in mind, that each voice can also have a start delay set, so you can get a late sound pickup that is then the last bit you hear, even if it's quite short. However, if the start time of one voice is after all the others have finished it will never sound.

This sort of idea works best with 'Forced Release' disabled.

An unexpected twist to this, is that taking the combined voice envelope time against the main (global) AddSynth envelope, although attack and decay times follow the above pattern, it is which ever (the voice group or the global value) has the shortest release that sets the AddSynth time as a whole. This can really catch you out!

With regard to the modulator amplitude envelopes. They don't change the overall time, but if they are shorter than their voice length (or any voice that the modulator is slaved to) the modulation may end a bit strangely. If they are longer, then part of their action will be missed. Also, when using the Morph type, you must have an envelope set, as this is what controls the way the sound morphs between the voice oscillator and the modulator one.

Finally, there is what I believe to be a bug that goes back to Zyn 2.2.1 - original Zyn :)

If an AddSynth voice is enabled, it's amplitude envelope time is active, even if the envelope is apparently deactivated and not editable. Oh, and by default all the voice times are quite long, so again you could be puzzled as to why a sound is longer than you expected. Because this has always been there I don't propose to change it. To do so would quite likely alter many existing instrument patches, but do keep it in mind.
"
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby Martin Walker » Fri Nov 20, 2020 10:47 pm

Thanks Will - there's a huge amount to take in with Yoshimi Pi, so your gentle remainders and clarifications are most welcome!


Martin
User avatar
Martin Walker
Moderator
Posts: 17207
Joined: Wed Jan 13, 2010 9:44 am
Location: Cornwall, UK

Re: Yoshimi extras

Postby Folderol » Wed Jan 13, 2021 2:21 pm

I've been looking at aftertouch again (although I can only test channel A/T here).

First there is a bug which I need to sort out.
If you change the A/T settings while it is active (note presses reasonably hard) it will mess up the settings until you do a reset.

Moving on from that, currently the settings aren't saved anywhere, and I actually wondering if they should be. It's very much a performance issue, and dependent on both having an appropriate keyboard as well as the selected instrument patch. Presumably if it was saved it should be to the patch rather than the more general settings.

Also, as A/T is not currently part of anything it won't be cleared by changing to another instrument patch - which has some interesting consequences! I should probably alter that :lol:

Finally, a couple of points about the Mod wheel generally. This has no effect on the SubSynth engine as there are no LFOs for it to change. On AddSynth it is the global frequency LFO it changes, not the individual voice ones. It also works between zero and the current LFO level setting - so set at zero (the default for a basic sound) = no modulation.

P.S.
Thoughts and suggestions welcome.
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby Folderol » Tue Jan 19, 2021 9:51 pm

Yoshimi's little band of helpers (TM) have been busy again.
One of them has just completed a very nice extension to LFOs.
Within a few hours of us notifying the users list that it was in the current 'master' another 'helper' ran with it and produced a youtube vid of his first experiments with it and Yoshimi being controlled by Pure Data.
I recommend watching it full screen. The image is still unfortunately a bit fuzzy, but at lest you can see what controls he is playing with.
It's here:
https://youtu.be/BKbQ6QgQogg

Oh, and he's also been using the new window resize feature :tongue:
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby Martin Walker » Wed Jan 20, 2021 1:29 am

Very pleasing demo Will :thumbup:

Tempo syncing of parameters via MIDI Clock can add a lot of real-world realism, particularly with vibrato-based parameters including LFO freq and filter freq, as this is what string players tend to do anyway.


Martin
User avatar
Martin Walker
Moderator
Posts: 17207
Joined: Wed Jan 13, 2010 9:44 am
Location: Cornwall, UK

Re: Yoshimi extras

Postby Folderol » Tue Jan 26, 2021 11:23 pm

A question came up today, variations of which have been frequently asked.
"Can you get Yoshimi to use an incoming waveform?"
The answer is, basically, No.

To a degree it could be done for the PadSynth engine, as that is a wavetable generator. However, the result would not be as good as the internal oscillator. For AddSynth and SubSynth it's quite impossible.

A simplification of the reason why, is that these sounds are directly calculated in real time sample by sample. Hardly any of the controls directly touch the sound output. They adjust the formulae that are used to generate the sound.

It is this that gives Yoshimi its character. If you go right into the waveform editor of Addsynth you will find that not only can you change the harmonics in real time, but also various filters and shaping controls - including real-time randomisation. You can't do that with a wavetable based engine. The best you can do is period based adjustment, with interpolation to take out the jaggies.

Please bear in mind that this is not the same as morphing between two waveshapes. In that situation the actual forms are fixed, so the mathematics (while way beyond me) are comparatively simple.
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Yoshimi extras

Postby blinddrew » Tue Jan 26, 2021 11:52 pm

For a moment there I almost thought I understood something...


...and it's gone.
;)
User avatar
blinddrew
Jedi Poster
Posts: 13650
Joined: Sun Jul 05, 2015 12:00 am
Location: York
Ignore the post count, I have no idea what I'm doing...

Re: Yoshimi extras

Postby Martin Walker » Wed Jan 27, 2021 1:52 am

Thanks Will - I can confirm that this made perfect sense to me, and that I will at some point be exploring the possibilities.


Martin
User avatar
Martin Walker
Moderator
Posts: 17207
Joined: Wed Jan 13, 2010 9:44 am
Location: Cornwall, UK

Re: Yoshimi extras

Postby Folderol » Tue Feb 02, 2021 11:08 pm

Something that is far from obvious is the way you can build up and then actively modify multi-layered sounds. This is something I discovered myself some years ago, but then became so 'but of course!' that I didn't think to mention it before.

To see this in action, from my 'Will_Godfrey_Collection' bank, select 'Master Synth High'.
Don't play it yet :)

First wind the part volume down to about -10dB

Click on the edit button, and in the new window click on Kit Edit.

On the square blue surrounded 'Mute' buttons, check all of them except the top one 'Breathy Pad'

Hold down a chord, then uncheck 'Ice Field'. Nothing seems to happen... until you press more keys. Then do the same with 'Smooth', and finally 'Full Strings' to hear the full sound - You'll hear why I said to wind the volume down :)

But there's more!
Any check box that has a blue surround is MIDI-learnable, and these particular ones are next-note so you can silently switch them while playing, and if you have a controller with enough buttons, you can learn them and dynamically change these multi-layered patches while playing.

Notice you can also switch the individual engines of and on. However, switching on a padsynth one for the first time will briefly cancel all sound on that part - this is because it has to generate the wavetable.

P.S.
That particular patch is something like 10 years old now :)
User avatar
Folderol
Jedi Poster
Posts: 12387
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

PreviousNext