I've just discovered a technique for sample based presets that may be of use to some.
I had noticed that hidden away on the orba 2 are very small samples of sine, saw and triangle(? I think). The ones I looked at are just 512 samples long. They don't seem to be used in any of the presets though. It occurred to me that, if the Orba can handle them, it would be possible to create very small sized sample sets using looping.
Making such small samples that loop well is much easier than it is with bigger samples - but you pay for it by losing some of the character of the sounds, eg: the tonguing of wind instrument notes and the vibrato typical of cello and violin sounds. The samples only need to be as long as the loop. The start value needs to be 1 (not 0), and the end, the total number of sample points in the sample. I used mostly lengths of about 1,700, but you can get away with less, or more. The size difference is substantial.
I would also mention that Audacity has some problems with such short samples. You have to get the pitch from a longer area of your source recording - Audacity gets it's wildly wrong otherwise.
I would guess it won't work well with some instruments, like plucked strings for example. But it might offer interesting sounds if anyone did try it.
I have noticed something strange with an experiment I was trying.
I made a preset which used two different sample sets - changing between them using velocity thresholds. (A crazy idea, I know, but it could be done so I thought I would try it.)
My first attempt showed that velocity threshold change registered in different areas of the keys - some near the edge (on the Y axis), others in the center. It was almost like the sensors were physically misaligned in the body of the Orba. So I tried removing seekers not needed and got the change happening only at the rim of the Orba - not quite what I expected, but at least it was consistent.
I don't know what to make of it, but I thought I would share, in case. The samples I used were Artiphon ones so no samples had to be loaded. If anyone wants the preset file to try it just ask, and I will post it.
Curious! Just noticed one UK retailer is offering the Chorda for pre-order for delivery roughly when the Kickstarter campaign should deliver. And another is offering a B-stock Orba 2 for a higher price than the brand new ones they also offer.
It's a strange world!
Ah - thanks for that. Since I don't have an Orba 1 that makes me even more mystified about how I got it! Shame it wasn't what I thought it might have been - may have been useful. Oh, well.
This is a log showing the reset procedure for Orba 1. Orba 2 works differently however. Tech support recently replied to confirm there is no way to access the data partition on a looping Orba 2.
@BJG145 - I can't recall when or where I found this file but I think it may hold some of the information you (and some others) may have been looking for.
An oddball thought here but.......
I was looking at a user manual for a Korg product ( nanoKey Studio) and kept seeing things that reminded me of the Orba. The touchpad uses x and y axis for pitchbend and effects (which seem to be like the Orbas gestures) - perhaps the data it sends gets sent to the same midi destinations and perhaps a similar format.
I also wondered whether the eight pads for percussion sounds were laid out in a similar (or the same) set of midi outputs most of the time , but the Orba presets allows varying that in the parameters for things like ethnic drums etc.. Perhaps there is an industry standard used by most brands - I don't know.
I also saw a reference to what looks like bezier curves.
Perhaps there is technical information to be found that could inform some of the things people have have been trying to understand for the Orbas?
...I can decode note data, understand that bit fairly well, but I haven't looked at expression data. I use a tailored preset on the Orba 1 to filter that out; swamps the note data otherwise. I haven't dared experiment with edited presets on Orba 2 though.
So I did a quick test creating a song with drums - trying to play with the metronome a bit and then making sure I'd hit all the keys. Then I added a scale up and down on a lead preset. (Not at all musical!)
The event data is in base64(?) - much as we see on the Orba1 - which is not that much use to me, but if you want me to post the song to see what you can learn from it ask and I will post it for you.
In a drum preset kitPatch each entry has "note="44" midiNote="39"" - whether the Orba pays any attention to it I don't know, I presume midi output does. I have noticed the same midiNote is sometimes in more than one entry.
In an artisong part of the lead setting, for example we have " <PresetRef name="Madaharp" fileName="Madaharp_8aee6a9869530ed2cc60c5ac045c368f.artipreset"
fileHash="12550"/>"
The fileHash is the 10 base value of the crc number the orba gives it when it is stored on the Orba - that is not easily available to non hacker types.
There is also this " <LoopData writeIndex="0" recordStartTime="0" recordStopTime="0" lastEventTime="0" nBars="0"
eventDataCrc="FFFFFFFF"/>"
In this case the eventDataCrc for all entries is "FFFFFFFF", although the song is empty of events - don't know whether it's used or required in any way. (Perhaps it means keep on playing until the user eventually stops the play/recording with A + key5?)
But at this point in time I've not studied it enough.
"Songs have an inherent problem if you want to mess with the voices - the crc or hash values are included the artisong file but the correct values may not be straightforward for people to access"
The way song data is represented in the XML is very similar to MIDI. The main difference is that instead of having a sequence of note-on/note-off messages, the Orba uses note duration. In terms of voices, the loopData entry simply plays the currently loaded voice; if you load a piano preset, the loopData for Lead will play as a piano. The above demo was based on the "Bedroom" preset where the MIDI note values for the eight pads (and corresponding sounds) are:
36 - Kick
38 - Snare
42 - Closed Hat
49 - Open Hat
43 - Perc1
45 - Perc2
51 - Perc3
70 - Perc4
I expect the eight drum pads always have the same MIDI number assignments, though I haven't checked.
Here's a quick demo of an Orba drum loop converted from a MIDI file.
The same process I've been describing in the ChatGPT topics works for the drum loopData entries too. I can recap the process if anyone wants to try. I'd probbaly recommend testing on Orba 1 before copying to Orba 2.
@BLG145 - actually I dismissed any interest in stems - I listened to some of the samples and realised what they were. I described them previously as the musical equivalent of collage, but using a pre-defined set of pictures (samples). Perhaps painting by numbers would also be a suitable metaphor. I find the concept of them verges on the insulting.
@Ignis32 - When I ordered another Orba for myself I promised that I would stick to using known and proven presets on it - but I confess I am already thinking about ideas for new presets which I will have to curb, but oh, the temptation! There certainly were things I wanted to achieve and may have to give up on.
Generally, I have followed that principle of looking to see what has already worked elsewhere, but I'm not good at proofreading and checking my work so the problem is likely to be simple, stupid mistakes . Never attribute to genius what can more easily be explained by stupidity!
@BLG145 - don't trouble too much on the drum thing - I think I should create a simple song with just a one or two bar pattern and see what that looks like. Songs have an inherent problem if you want to mess with the voices - the crc or hash values are included the artisong file but the correct values may not be straightforward for people to access, we probably can, but .......)
A little help from devs would be great for this, for sure.
We could achieve much more if it was not that scary to brick O2 irreversibly, and there was a known way to make a full reset :(
At leas at some point I felt that I should stick to just using custom samples, and never tried messing with parameters outside of the known territory - I mean before any changes to parameters I always spent time parsing all existing factory presets xml's to check possible known "safe" values, which are already used by stock presets. Does not guarantee anything, as there could be wrong combinations of parameters which are in "safe" boundaries of their own, but increases expected life span of O2 - mine is still alive at least.
Looks like I should have highlighted that in a more explicit way. (While I do not feel like it would stop David from exploration)
Subskybox
This forum is intended to share Orba 2 hacking tips amongst the Orba 2 community. NOTE: Please post facts that are well understood & useful. If you have theories to discuss, please start another forum and link to it here.
2 people like this idea