If your timing is off then you need to work out the best way to correct this if the arpeggio triggering timing is negatively impacting the result.  As you say it's probably easiest to keep the recording of the trigger notes and work with those.  If you have a chance to record and quantize simultaneously then this may reduce your steps since you wouldn't necessarily need to "fix" the trigger notes as the timing would be automatically corrected.

Also if you have the arpeggio timing set to "measure" then this will automatically align the start of the arpeggio which means you can trigger the arp (almost) an entire measure early.  In other words, quantize at the measure level is built in.  This may or may not be available to you depending on when you need the arp to change.

Do you use this process that involve 3 successive records?

I don't think there is just a single answer to the best way to record an 'arpeggio'. 

First you have to figure out what the goal of the recording is. Are you just wanting to capture the AUDIO that the arp produces? Or are you wanting to capture the MIDI behind the arp?

Is the goal to be able to reproduce the arp yourself?

You can record midi to daw a performance which contain arpeggio in different way.

Ok - but recording a performance is different than recording an arp. So again the question is about the goal - what is the purpose?

An example - a performance could contain 8 scenes, 8 arps and multiple sequences. If you record that performance you won't really capture the entire capability of the performance unless you actually execute all 8 scenes, all 8 arps and all the sequences.

There could also be other controllers used during a performance (assign switches, faders, knobs).

So just what does it mean to 'record a performance'?  

The one is to record directly the arpeggio in midi but then it could be more time consuming in editing , sometimes chords which trigger the arpeggio are not fully in time and can make undesirable notes, then you have to work on each notes of the arpeggio

The second process is to not record arpeggio to midi and then it should be easier to correct only the position and eventually velocity of the chord.

A third process is the second process + record again with arpeggio to midi and giving so the possibility to edit in details the notes of the arpeggio.

Are you wanting to capture the MIDI that the arp is using? Or is it a control arp and you want to capture the control changes?

Bad Mister wrote an article about creating arps

https://yamahasynth.com/learn/montage-series-synthesizers/mastering-montage-arpeggio-making-101-part-i/?_ga=2.32638521.666580034.1666042081-1762977864.1666042081

In that part 2 he thoroughly examines how 4 part arps are created, why they are created that way and the details of how they act differently if they are chord-intelligent.

It isn't simple and straightforward to capture all the details of a chord-intelligent since the output they produce depends on how you trigger them.

If you can provide a more detailed example of the use case you have in mind it will be a lot easier to answer your questions.

My understanding of the problem statement is that imperfect timing when recording an arpeggio can make an arpeggio spit out unwanted notes.  I think we've all encountered this if we land on the wrong side of the perfect timing to trigger the arpeggio.   And there can be lots of triggers if you have the arpeggio holding and press different notes to change the arpeggio (increasing the number of chances for timing issues).

If you're recording MIDI somewhere then recording the MIDI output of the arpeggio itself (and not the trigger notes) gives you a bigger task to correct than if you instead only recorded the MIDI trigger notes and corrected those.   This isn't always true - but would tend to be true (depends on how many notes the arpeggio outputs for given trigger note(s) ).

It seems Daniel has a pretty good grasp of the landscape.  

If one was able to perfect their arpeggio trigger timing then this would not present itself as a problem.    There may be ways to manage this (quantize, arp timing, etc).

But, yes.  I think if you record trigger notes with wrong timing then there are cleanup steps.  And I agree that generally cleaning up the arpeggio output would be more difficult than cleaning up the trigger notes then taking a 2nd pass to pass the cleaned up trigger notes back to your instrument to get a clean recording of the result.  In terms of MIDI there's probably no need to ever record the MIDI of the arpeggio output if you're not going to want to manipulate it.  You can just always record the trigger notes and not worry about documenting the arpeggio's MIDI note output.

Knowing the actual use case is the first step.

My understanding of the problem statement is that imperfect timing when recording an arpeggio can make an arpeggio spit out unwanted notes. 

That right 'understanding' is the key and that is the piece I am missing. At first I thought he said 'performance' and 'arpeggio' as referring to presets that he wanted to record. I took it in the context of users that ask why they can't play a performance and sound like the audition.

You might be right if your take is that of recording a live performance or session and getting issues in the result that need to be corrected.

This Bad Mister reply in a thread from 9 years ago is pretty on-point for the timing/trigger issues you mentioned

https://yamahasynth.com/community/moxf-series-music-production-synthesizers/best-methodssettings-for-syncing-arps-to-daw-of-choice-in-live-performance-scenarios/#post-3036

I stand by my statement that the best way to 'fix' an issue is to first know just what is causing that issue.

The trigger method, and timing, controls what an arp produces. So if you aren't triggering it properly you need to fix the trigger. Trying to alter what the arp produces doesn't make sense if it didn't produce anything close to the right content to begin with.

If you record to AUDIO, rather than MIDI, you can't fix the trigger. So for triggering problems I've found it more effective to record to MIDI and ONLY record the performance and arp trigger notes to DAW. Then if the triggering was wrong you can adjust the MIDI that does the triggering. Then you can play that MIDI result and record it to audio since the arp will now produce what is needed.

Bad Mister makes the KEY distinction that once you play an arp it isn't an arp anymore - it is just MIDI like all other MIDI. Anything you 'fix' at that point is pure midi editing. If an arp produced the wrong midi then trying to 'edit' that midi to be correct is, essentially the same as just manually adding the arpeggio yourself rather than having the instrument play it.

Audio is actually music and you have limited ability to make changes. MIDI is pseudo-music which is essentially just a recipe for creating the actual music. There is much more flexibility in manipulating MIDI thank with audio.

I prefer to keep everything I can in MIDI until the very end. If I had a choice between a library of AUDIO master material and the equivalent MIDI master material I would choose MIDI every time.

Posted by: Toby

↑

Ok - but recording a performance is different than recording an arp. So again the question is about the goal - what is the purpose?

Yes it need some details:  the goal is to record in one time a whole performance ( the same I  use in live )witch contains 8 parts and many arpeggios embedded in. The music type is orchestral « semi classical ». There is 20 pieces of music about 5 minutes each. It will be the first time I use this 3 steps process recording that came in my mind, a basic thinking through.

Posted by: Toby

↑

If you can provide a more detailed example of the use case you have in mind it will be a lot easier to answer your questions.

understand that it depends on the type of arpeggio and its setting and I should increase my knowledge . what I can say about my main issue is that for triggering in the best way the arpeggio you need to play the whole chord in one time but when you have also to play in the same time with the same hand another instrument witch is not triggering arpeggio, it can confuse the arpeggio system about the harmony and producing unwanted notes  and also sometime my timing is not perfect, in live situation it could be acceptable in almost case as I am the only player but for a recording it is surely not.

Posted by: Jason

↑

Also if you have the arpeggio timing set to "measure" then this will automatically align the start of the arpeggio

That is the mystery point for me. I test it just now and I don’t heard differences on the triggering between real time and measure, in both I can trigger anytime in the measure the arpeggio, I means the first time of the arpeggio structure doesn’t match with the first time of the measure. if I start the arpeggio in the third time, arpeggio will start the beginning of his composition always in the third time, just sync quantize ( about I didn’t pay attention till now and set to 480 for all my performances! And I wish this was good enough for me at this point) make a dramatic change.

Just a general thinking; in the future ( but i don’t know which one) the arpeggio function should reach new fields. For exemple, you could teach the IA the entire series of the chord you wish to trigger timely and then even if you don’t play the whole chord during the performance,  the arpeggio will follow the intended harmonies but may be this could be done without IA too. surely the IA should bring more staff, may be just you play your piece of music one time and then the neuronal circuit will trigger the exact arpeggio harmonies you expect and if you are not satisfy enough, you could say to her and why, of coarse the time you will speak to your montage is not for tomorrow but maybe for after tomorrow. Arpeggio in montage is a powerful function that make pre recorded sequences sounding natural or at least tend to and I use them a lot for live.

You could also ask the IA to clone for you arpeggio in existing repertory like classical music or whatever, from scores or records and etc . Of coarse come this question at the end: Is the IA will clone you too and making you useless? The answer for me is that depend on your background and skills and how you use the tool but human tend to follow the easiest path and that can bring to human obsolescence and that is general for all high technology, I know this is only philosophy, sorry for that.

Arps do not have measures or a reference to beat 1 until you start the arp.  If the arp isn't running then it will start as soon as triggered.   To use this timing, you should start the arpeggio and have it hold "forever".  Do not turn off the arp using arp on/off or using the individual arps as "off" (or "none"? ).  Instead when you want the arpeggio to stop sounding switch to an individual arp called "Mute X/Y" where X/Y is your time signature.  Mute arps silently keep the time so measures still will be relative to your first arpeggio start.   

MIDI generally offers more flexibility to 'fix' problems.

 the goal is to record in one time a whole performance ( the same I  use in live )witch contains 8 parts and many arpeggios embedded in. The music type is orchestral « semi classical »

That suggests to me that you should record the MIDI of the performance and only the trigger notes of the arps. That should provide a master recording that you can use to both find, and fix, any note and note timing issues you find.

Step one would be to find/fix note and note timing issues. Those are the issues that can cause the wrong arp to play or the arp to play differently than what you expect.

For example if an arp isn't being triggered properly you need to fix the trigger and not try to replace/adjust the midi that the arp produced. When you fix the trigger the arp will produce the correct midi. The easiest way to do that is to NOT record the arp output but only the arp trigger notes.

Once the core notes and note timing is correct you can change course and record the actual arp output instead of the trigger notes.

Having the arp output in midi form is useful when the midi needs to be played, or otherwise manipulated, on an instrument/controller that won't actually have the arpeggios themselves.

As Bad Mister has said many times in the past, once you record the arp output that midi is no longer an arp - it is just midi like all other midi.

What I would suggest is that you test your process by:

1. use the simplest test cases possible

2. use a simple performance with just 1 or 2 parts and 1 or 2 arps

3. make an INTENTIONAL timing or note mistake that needs corrected

4. use your process to record and try to fix that simple mistake

A simple scheme eliminates possible side-effects from parts/configs that have nothing to do with proving your process. It also lets you focus on ONE problem at a time while working out the kinks in your recording/mixing/tuning process. 

Posted by: Jason

↑

Arps do not have measures or a reference to beat 1 until you start the arp.  If the arp isn't running then it will start as soon as triggered.   To use this timing, you should start the arpeggio and have it hold "forever".  Do not turn off the arp using arp on/off or using the individual arps as "off" (or "none"? ).  Instead when you want the arpeggio to stop sounding switch to an individual arp called "Mute X/Y" where X/Y is your time signature.  Mute arps silently keep the time so measures still will be relative to your first arpeggio start.   

 

it is the case, my arps are always on except for some breaks and endings.  I use use the SK for mixing the parts. And I tested with arp on and with hold too. But I don’t heard any differences between real time and measure .that is strange for sure, I surely miss something.