Artiphon

I would love if we could have Kalimba presets as good as the piano ones.  Anyone good enough to make this happen?  

I would have if this kind of experiment doesn't risk destroying it. Not worth the risk unfortunately.

??? What are you talking about? there is a whole community making new presets and there is no risk. its not a hand granda 

Curious about this "whole community making new presets" - perhaps for the Orba 1 using Orbasynth but I've not found one for the Orba 2. (Mind you I might not look in the right places.)

By the way, when you say Kalimba do you mean a thumb piano type or a xylophone type? I've seen the name used for both.

I thought you must have been talking about people posting somewhere else. As far as I'm aware there has been very few posting presets here - and I think I am about the only person left making new presets. And I've curtailed my own efforts, partly because I think I've about got as far as I can go, but  also because I managed to brick an Orba 2, so I'm reluctant to push my luck.

I answered your previous posts about this a while ago. Thumb pianos create overtones especially at the start of the notes, and the root frequency changes through the duration of the note. This makes getting tuning right a nightmare and the overtones at the start sound horrible on the Orba. (Even assuming you can get samples that are in tune.)

(The xylophone type of instrument is not so bad but my attempts sounded so characterless that I felt I might as well use typical synth sounds instead.)

The best I managed to make plays poorly for tuning (I suppose I could improve that but it would be rather tiresome to so and I'm not inclined to risk it - if the attempts don't work well when testing you have to get into the dark arts to remove them from the Orba.) The bottom octave also sounds bad because of high overtone frequencies being too prominent. It's not so bad through external speakers with more bass though.

I'll post it later perhaps, but I wouldn't actually recommend it to anyone.

Are you talking about Orba 1 presets...? I'm only aware of about three people on this forum making a small number of Orba 2 presets and one of them killed it in the process.

As far as I understand it, the problem is with creating a set of samples and assigning them to pads without automatic retuning. I think it's fairly straightforward to take a simple sound and let the Orba map it to the different pads, but you won't get a high-quality result this way, and it's pretty difficult to get round this limitation with a set of different samples like the piano preset. After I also bricked an Orba 2 while investigating it, I gave up. Artiphon have told me that experimenting in this way is unsupported, there's no way to reset it, and you're not covered if you break it.

My understanding (although I didn't get this at first) is that you "tell" the Orba what the frequency of the sample is and it adjusts the pitch according to what it is told. So, samples that are midway between two semitones can still be used. I can't recall exactly what I tried but I think you can "lie" to it within limit. I think I tried to use velocity thresholds (how hard you hit the key) and lying to it about pitches to make it possible to play different pitches depending on the touch (to get semitones or octaves). But it didn't work.

A problem with thumb pianos is that when I try to ascertain the pitch of a sample (using Audacity) I get very different results in different parts of each sample. Test at the very start and you can get numbers an octave higher or even more, much more. But even later in the samples there are still significant differences as you go along. I pretty much had to guess and hope what would work. There is a frequency variation at the start of notes on a whole range of "real" musical instruments, but the Kalimba is extreme!

In the preset I referred to above some samples are playing flat - but it's rather hard to be sure which ones are responsible, you don't want to be adjusting something that isn't - those with more analytical minds might be better at working that out!

By the way, I think my Orba crashed because firmware updates were persistently failing, but I tend to the view that my experiments didn't help the situation.

>>??? What are you talking about? there is a whole community making new presets and there is no risk. its not a hand granda 

 
There are about four people making custom presets, including me, that's the whole community. Two of them already had broken their Orbas 2 during experiments, one of them is David, who made the biggest amount of presets.  That's the reality.    

I did come across a couple of things that might be of interest:

https://www.orbalibrary.net _ couldn't actually register or log into it so I don't know what's there - but I suspect it's for the Orba 1 if anything.

And this:

https://www.youtube.com/watch?v=HPxSS77VYt8

I'll pass on commenting on this because it went so fast I couldn't be sure whether it was good information or not.

The comments there suggest there may be others trying to make presets but I don't know if any have shown up here.

And the presets and I have shared (and those you haven't seen yet) were inspired and informed by the other three.

From what I've read and seen so far it looks like Artiphon are missing some tricks for the Chorda which they could have learnt from here. Missing some opportunities, I think.

I have  done a little fixing to the best of my Mbira preset attempts. I have improved the tuning but I can't do anything to get a better sound on the lower octave.

However: I have not tested it for anything but how it plays, on it's own, in the default key of C major.

I don't know if it's in A440 pitch and it's almost certainly not in equal tempremant. This means it may not play nicely with other instruments which I assume are in A440, and it will likely sound dodgy in at least some other keys.

It would be possible to adjust the pitch values for each sample in the preset to correct these things but I reckon it could only be done by careful guesswork - not by simple calculation.

As far as making it as good as the Grand Piano - well, it would need at least 25 samples the more the better (I only had 11), and to catch the effects of harder touches -  that would need as many more samples and they would be the ones harder to work with.