Massive instrument model - is messing with my mind.....

[d.healey]

@Lindon said in Massive instrument model - is messing with my mind.....:

Ok well where the f***k did my round robins all go then?

Gone for good in this model no?

Groups are just groups, ignore the fact that they are labelled as rr groups. They are just one of the methods for organising samples - like note range or velocity.

If you're using the groups for xfading then you lose HISE's automatic round robin facility - but you probably weren't using that anyway, you were probably managing the RR yourself via scripting, so you can carry on with that when using xfades, nothing changes except you have more groups to manage per RR.

[Lindon]

@d-healey said in Massive instrument model - is messing with my mind.....:

but you probably weren't using that anyway,

But I was using it!

So now I must use scripting to turn on specific groups right?

So a note arrives and I point it at RR1 and RR6 using:

Sampler.setMultiGroupIndex(var groupIndex, bool enabled)

EXCEPT!!!! this is what the documentation says:

This function can be used to enabled multiple groups at once. By default, only one RR group is active at the same time - as long as the Group XF property is disabled, then it will play all groups.

-- which confuses the bejesus out of me.... will it work or wont it with Group XF on?

[d.healey]

@Lindon said in Massive instrument model - is messing with my mind.....:

-- which confuses the bejesus out of me.... will it work or wont it with Group XF on?

Oh I wasn't aware of this. I just tested it and no, as soon as group xf is enabled all groups play! So that kills this idea - go with velocity.

But it would be a nice option to be able to control the active groups even when group xf is enabled... @Christoph-Hart

[Lindon]

@Christoph-Hart said in Massive instrument model - is messing with my mind.....:

@Lindon Use RRGroup for the Xfades and encode RRs & dynamics into the velocity data slot. You probably don't have 127 velocity layers, so you can pack that information into one number.

So if you have eg. 4 dynamic layers and 8 RRs distribute them so that the first RR group goes from velocity layer 0 - 15 (0-3, 4-7, 8-11, 12-15) etc.

Then do the "decoding" in the script.

Well I just re-read the spec. and yes (of course) they want velocity to be used to control the volume (in all cases) - as well as optionally set the dynamic layer cross over, so this stacking RRs in the velocity layers wont work will it. I programmatically change the velocity to use the correct RR sounds, and the overall volume of the sampler is changed e.g.

HiseSnippet 1170.3oc6X0saaaCElJNLKwKqnsnEX2L.cwtHAnnvNMMc.8h5XGmhfVmHXkErcU.iDsMQjHEnnRW1vdm1E6AnOR8MX6PIYKJGkTWC2zre7EF9bHOje77G+ncjBOZbrPhrV63KinHq0wtWxUi5Lhv3nC1CYcObORrhJsyT09xHRbL0GYYU60ZEVqsLJ8yGdUaR.g6QKTgPmHXdz2xBYpBsNsdCKHXehO8XVnwr2t0AdBdGQfHAvSMbCTDw6bxP5gD8zVBirVoqOSIjtJhhFirVtsv+R2Qh2wyl+IrX1YATsPSjKrPYp2WD3qQr92nNiXA9NiO2wHjE1ovKTKyK7HbOlOah9Buw8SGvtvBS+g0RkgWsRvq40AuJfjkAjVNCRO.65IYQphQz34qwGvgfy.B31MgR1bQVuG2Q.SfqdZH4b59RPXhAarSiFOwF9ZyWNHg6oXBtsfenPQOhuwl0+s5qU+2qaO8PCFT4X5sQJBBnxJGVGok2jgavSBOiJeh8EjfD5jIBG+x9zUlMepW1o1XhB9Abl5nHZtbQRQyJh.nb2F7qe7f8HJhNnjqClWDUpXZ3XsG8BHCOKDsFdOZ74JQDjiek3Gj4H7SBHpxoS5Zn7A.+QoXnNPwiYpKMqwlNGa4q2eznxRfJOsyHbe.1go7FUMdWpB7BdsO23Muh8avcGLf5oJ.6x38+oar7bgCkUulVnqmVDBRTYyEQCznYtApUs+Q0v7VMY9yZs2r6M++ZuEBT9tLn7PrqRRIgL9PWRXDbcTJdpiyktkq+bjz.Aw2k8qFS68sZm.9KYYs+QqzsZ2PQBuzdki79zHJQAgeCaVsU+9uVJRhl1pO7pzjkik.AJvUXNvQbJ32UlmwNRQb7.33jtVwlC4jHGp8UEZ5SufJiKq6vjPH1y4z.vXrkk9VyXpNVy75QTR1uXN6zcowwjz3MvpKUt4TxaMk7ylRd6oje9Tx6Lk7KJjybn8HQYzaeyOeZaZPfCivEWgtwWMak3vh4PfiuwLayFNYixzTtciwTm3+SwnQ64J58txWfdueBWMLyDWWG2uuMjy3qKQqh45ehsuFVo8fUAdUvSiopSfxKOnq2FM2ZyWdGjsJdQwV8emLT+DRrVx4JOU89XWlt.qK+BHM.znw6CgS3.RRfZr1xc76I3hnQBNyqbOMnG0vgTo44nxC2tJEzRsPyiZ0mFPIwFMk+9VuEX5QjkaUeU+xB80gUF69VbFbs04y12odkQNU453wUvo.9wSDsmbbZ9wQKbWSd2cynGbMkp70T6Q8XmQCJGXfq3y2yz0HsRJ8+ZXQEtVHD7p8edBd2jq6d4z6rgcWpti+H50woDRr07G+RA09hDEvKbL2LLPfyEHs5QMowsj9RlL4FZYMfbob+Tg+B9jOXywb9zC1b7fHuwKEb1x+cVwUMJWWok9tk.fMYfYJcLPMVCInB7VAk2F6QHwSJN0Ki5f1erZpFvSvS+mMWC2SKa2DkRmvzcDBD3N0yq7RcEC2ZdM7YyqgaOuF9740vclWCew7Z3O7wMTyud2DkHLqgMB0yoaVRtU2II40P+MfKihjO

So at this point I cant see a way forward without dumping RRs altogether -or being able to control active groups when XFade is enabled.

[d.healey]

@Lindon you can still use velocity to control volume and also assign it to group xf. I'll check out your snippet shortly when I'm at my PC.

[Lindon]

@d-healey said in Massive instrument model - is messing with my mind.....:

@Lindon you can still use velocity to control volume and also assign it to group xf. I'll check out your snippet shortly when I'm at my PC.

yeah --- but...... thats not all Im trying to do here - I'm modifying the velocity so I can trigger the correct "stacked" RR wav file, so for a single note I would have 5 wav files so using Christophs model these would be on velocity 1-24, 25-50,50-75,75-100, 100-127

To get the correct one to play my on note will need to modify the inbound velocity to fit in one of these ranges - and as soon as I modify the velocity in a script then the volume control using velocity fails.

[Lindon]

I think this is very easy to specify.

I have this set of files:

StdGuitar_60_RR1.wav
StdGuitar_60_RR2.wav
StdGuitar_60_RR3.wav
StdGuitar_60_RR4.wav
StdGuitar_60_RR5.wav

and this set of files:

DistGuitar_60_RR1.wav
DistGuitar_60_RR2.wav
DistGuitar_60_RR3.wav
DistGuitar_60_RR4.wav
DistGuitar_60_RR5.wav

These are all mapped to the same note(Middle C) - each set mapped into 5 RR groups - I wish to cross fade between these two sets of 5 RRs, in a single Sampler. Oh and each one is a multi-mic sample set, but Im ignoring that at the moment.

As far as I can see I can either cross fade between these OR use them as round robins - but not both.

Yet the documentation says this(for setMultiGroupIndex):

"However there are a few legitimate edge cases where you need multiple, but not all groups enabled - the most common one might be if you want to implement round robin behaviour in combination with Group XF samples."

[Lindon]

@d-healey said in Massive instrument model - is messing with my mind.....:

@Lindon said in Massive instrument model - is messing with my mind.....:

-- which confuses the bejesus out of me.... will it work or wont it with Group XF on?

Oh I wasn't aware of this. I just tested it and no, as soon as group xf is enabled all groups play!

Did you use setMultiGroupIndex in your experiment or not? Because if not then yes they all will play...

HISE | Docs

(docs.hise.dev)

[d.healey]

@Lindon said in Massive instrument model - is messing with my mind.....:

as soon as I modify the velocity in a script then the volume control using velocity fails.

Use a global velocity mod instead. This will pick up the incoming velocity before you modify it.

@Lindon said in Massive instrument model - is messing with my mind.....:

I wish to cross fade between these two sets of 5 RRs,

Place each set in a different group, using velocity to separate the RRs.

@Lindon said in Massive instrument model - is messing with my mind.....:

Oh and each one is a multi-mic sample set, but Im ignoring that at the moment.

Makes no real difference, we can treat them as a single sample once they are merged in HISE.

@Lindon said in Massive instrument model - is messing with my mind.....:

Did you use setMultiGroupIndex in your experiment or not? Because if not then yes they all will play...

Yeah I tested with that too

@Lindon said in Massive instrument model - is messing with my mind.....:

"However there are a few legitimate edge cases where you need multiple, but not all groups enabled - the most common one might be if you want to implement round robin behaviour in combination with Group XF samples."

Looks like Christoph intended for this behaviour but forgot to implement it then. But still this is way more complicated than using the velocity separation method because you're going to be dealing with crossfade tables across a massive amount of groups instead of just two.

[aaronventure]

@Lindon said in Massive instrument model - is messing with my mind.....:

and as soon as I modify the velocity in a script then the volume control using velocity fails.

You can still save the event velocity in the script. You then use eventData and the event data node in the voice start script modulator (or the event data modulator itself if you don't need anything extra) for the samoler's gain modulation, and for each event where you're playing the artificial note, just send the correct gain using eventData to the desired slot, which will be picked up, per-event, by the eventData modulator, removing the need to have the samples spread out to use the velocity linearly as a gain control.

This turns velocity into a simple y axis value for mapping the samples, letting you have 127x128 samples per group.

[d.healey]

@aaronventure Or use a global velocity mod

[Lindon]

@aaronventure said in Massive instrument model - is messing with my mind.....:

You can still save the event velocity in the script. You then use eventData and the event data node in the voice start script modulator (or the event data modulator itself if you don't need anything extra) for the samoler's gain modulation, and for each event where you're playing the artificial note, just send the correct gain using eventData to the desired slot, which will be picked up, per-event, by the eventData modulator, removing the need to have the samples spread out to use the velocity linearly as a gain control.

This turns velocity into a simple y axis value for mapping the samples, letting you have 127x128 samples per group.

wow this looks like the most complicated way to get per-note velocity modulation ever...

[aaronventure]

@d-healey also true, but here you can do any kind of math to get your volume value, even use data from the interface script without first passing it to the real time script (if using globals), so creating volume curves per-articulation in the UI becomes possible without worrying where the modulator is, what it's called or how many of them are there.

[Lindon]

@d-healey said in Massive instrument model - is messing with my mind.....:

@aaronventure Or use a global velocity mod

and this one isnt per-note - is it?

[aaronventure]

@Lindon how is it complicated?

Its just a different modulator (it's already built in, just use it instead of the velocity modulator) and one additional method call right under your playNote call.

The gain expression can be as complicated or as simple as you need it to be.

The idea is that you don't need specific modulator names or counts in the script. You just target a slot index for eventData, and in the modulator you set the slot that it'll read from. Then for every event, it'll check if there's a value stored for the event's id and apply it. It's ridiculously simple and scalable.

[d.healey]

@Lindon Yes, I use a global velocity mod in all my instruments. Another advantage is you get one overall velocity curve you can control, plus an additional table at each of the destinations.

[Lindon]

@aaronventure can you post a snippet?

[Lindon]

@Lindon well it would seem Dave is right and

Sampler.setMultiGroupIndex(var groupIndex, bool enabled)

isnt working, if it was all this would just go away...

HiseSnippet 1541.3oc6X01aaaCDVxNJnwqqnsnCneXePnnEvAyv0Nu0MjML2D6zl03XCorrATLDvHQESTIRAJJ25UTr+R6mv.1ul8s8OX6njrkjiyK1y0qCq9CB5NdG0Cui74N5tblENHfwUTW4nA9XE0apYNfJ5saODgpreSE0ao0FEHvb8XU6LvGEDfsUTUK9LoB0UVRI52e9s6fbQTKbpJEkiYDK7ADOhHUa2Fuf35tGxFeDwKi0azXeKFcWlKKDvSQsZJ9HqWgNCeHRZVAME0kaYSDLto.IvAJpKsCydfYO1qow1eLIfbpKVJTWwDlnX06wbskHV9txt8Ht1cGttCTTT05lFEJFGEtmVahMYj9znwsiFPO0irwC0B4gWwbvq9EAuI.I0LPZoXHcGMSKNwWjNhDOeh19TH43ffvdVnDaqh5uqsKCLfJp5gdEdONHLxgxaUqVEc3wpaWpTImPpkfvn5L5gLAtCs7pkdaoUJ8tR5iOjiyDGS9k3LWWLehCKS17KywxzPuSw7J58Qtg3QFBQf7g0kudgUq3EdFCYz8oDQGebhb59h5SHInjD4f2998ahDHYdIQGXmOlKHR3n1D2G1jGmkVQqIN3UBlOrM+boPXyCyNzEIxuiRdLJY.HdjKMJyUz.hXP1iYiuMaoKNdTahmBl3p8ZB26n0kHr5MY7VXB3EhZuuwaxg1OUqkiC1RjB1kz16GuzSnycnbiKfE8lQmCAILu97fC0+ZygpV7+TblScF4eJ+ogwyQTafzZR7mE9CUM8R5O9w5ABDWne5.cj6qQCBzEgbJgdFvdoK5g0cH7.gNmqCyEXhajRF7fGnybbpVsZIcHw.1zGwkC1g1ghCz+F8WVuxF+z1iOpiS7vubsJqWYyJaIsHqMlHOe.y0gIHZmV0yvhnvPjT4GLb7GrZjiCEqhBRds7pUwTDDYMXgTaC1oDZYGja.v6peAkBFMIAXQ6PWA4YbVn+9Ta7aJOZEUQvAt6suZqiWgUh+la+AXYFs4UYlOVZ4ikVlCP4yigxc0LEbLxCHeRNhEgmRZCOvsXKuzkicYHaSxOmwreqwNgP7hmW6u1H5S8TOfwI22JA4FXeLR.o+L9biFFFQDGi608aDsY4HNbEAHTjcMB7JPbWjcMtKmED3.Kmn4JHq0cC4mIiUoFaf6Cz140cXnGj6oTrqjYPUUdxM.Ky0Dq1HAm7lrVG8UpcDJJe+Hs01n5nee0tvi9cfGNewHk86D6R8o2k0ldWVOwE3BUQxaLl7liIu0XxOIUVUK+l9UtdblsQ9cQPrLik6PR1NCiEqIO2UFSGkLivQlALwBArWHHS2M+.pO1gw8xnZBsKr7+BcyTetye9dkteJ5MrP2y8uJbaMShL41h1GnK.MR7dWnflCBZMXn17TWsYTleOFkXk+vIbX6ryv7rqiIt3dpP.bCoZtWCCrKFEjgc4gMN.5HGwyy4b93xb8h7SL2cesX3pK4a0+fpjcxUZJocLjjrjecIf+rQh5iVN0uZzBjlIrGYyd.eqHOeaSrE4Tra9DCTqJ4aFMGQMNEwBMuRWykNUJ9+9NUtrP2sR5SQ2TdgpOv6H8lZQE7zkMBs.f5UxfLhMUVBJ8dPIroQ0kR0NSmFUTlzowzI8vn6NkcNL8Xv8by2lHoQBy9wxqXcYGCl8ipv8VkE6G17kFfLSnqTKb19zJHAerbMorDBlXpcjveA+RFr9vl5jCVe3fJVCmJHBl7dLoSw36NGcwDWncQ2rqw.n2WIj.lIkfnM6dPOMPWFwB.ipAS1dpl5WCVe.6E3ACEdNIivArikSbLpeNIQ3Wjq83dhGB58HtI+G4OT6scM57cs18nS1qyAMaY7NSg8NtD+S1p1IFF0q9ZTeklg9tDKH9K2Wn79.TKmCTO5bfpIIPjhpsVPnRapP0lKHTszTgpMVPnp3zrsZ8EDnJLMfZsI.pEAmwh3a3gr3rSrhYkkrS2HRCvKQiBMqn0VJqWWo+3DvdPQhSrrxOUmyw0lUGWeVcbiY0wMmUG2ZVc7Iypie4U6nrF9SCELu31JUTZ2sUbIG0ViJ4TT4uAfmAfuG

@Christoph-Hart any chance of fixing this?

[Lindon]

@d-healey
@Lindon Yes, I use a global velocity mod in all my instruments. Another advantage is you get one overall velocity curve you can control, plus an additional table at each of the destinations.

yeah I understand - but its a work around Im not keen on pushing at the customer....

[aaronventure]

@Lindon

HiseSnippet 1283.3oc4XstSabDEdWCKJ1ospoMUzeN+n+vthZ75ajHTUH.FBpwfElPpTTTzvtGaOh0yrZ1YchUUen5aPeD5iPeD5aP6Y10W1EbLFqjPkhQB44bYluyk4bNiaIENPPfPZXl6rg9fg4WX0dHW0audTF23n8ML+Jqlz.EHIwj1cnOMH.bMLMW4PMAyrqZD84edxtTOJ2AlRxv3bAyAdNqOSMkZqc9Elm2ATW3LV+DRWcmibD78DdhPDOqXUxvm5bIsKbLUKVFKimQC5YX9iVUdbmxtUgZUqW8hZ0cqUoZ4NtvVa4Tysds5a4hTsebkspXXtVCWlRHaqnJH.2zcEtCa2S7Vd7AbNKfcgGnWXazFO4XxF60i441ZryIvvvzp0TW0JwtpGZ0j4xlPepK6qiXPlpQRmlYl4AI6aAjLS.oUigzCrZ6HY9pobz349VGwwHXGJFaRBkXYML+Sq8Dn.bUw9zKgCj3hIJjudoRaPv+UX6NgbGESvIB9wBEbBOegb+Vtr498bjqxpSmYxSeLRgmGHmIac5fbdJlmG1+BPtAY.0KDlHHZ9o8oqsX9TmXqNgfB9Qbl5DeXz5CDdtZek96WOBXLxsge6EGsOUQ0AkQzP47Aohogi49v.7ZPbHJq09PvkJgOdQ3ZwOLyQ3F5QUoSmzWzFw.8Gohg5.EOfoFl7h3sHGqzbywVTH9.qVLkSuYiwLy.inm5iAFGcy7KsZzoC3nlBvUsN3WW1qgK9wud7weeqWRG.Q0QiN7uKZcGgrO4PfCRsuydNUR+qEsRp+BWI8DGEBgyjTdfuH.rStysg9ryDbHHE0qnQ4YpQJp6CpPd5sNlTJozthCPWQB4t2NiIVNIwVTdpMCWmZmZxd2zk+8SZvoX.qMfVt6IANniQ6mSdsnUnW.7RlqpmcREmRtbRxOiJcwXnSpj1URm0r5hUnI882DxOip5qcGznozGnFMe6HpSQj8rZ3j4OxXIgtDX.Vn7HWxOSd0q2ljazxhRH.jCf71keTgsykayMICEgDN.tDUOV.QIHA.2MVchKV2MGFyCTjC8DWP8NUDpX7tMobbzAIt4M3cYbnXWPMKAx+dar0Ds.ThhrtbgDZnOs7JI12YarsC92H79pwxgGfV8ii5QkuvqwiNpFPQeO5PMm7uGI2fToVgsIngJjj21CCWCPfqnRTNx.vS3f0N0lsdM5Kjj.evg0g4PBn888.xETzCivOlKdbnItAQJINhPzGAJmbYmkwWL.TQFlt6U9axf1ffiBjzHNeD1xW3mrKroc45nqYNSCLOGZrmhGK9tCaDCkaDRW6.+TMhg0GpQL9bbrByLIvX1wXrMSmK2fqy3QJZL9MnU0gF5oFSMc8jlBtvumfyRUj9TPIYcwr6jXelFzSUJ7EFSo7vcNE7.7tzTR+vNOGKdPkneBVRewseL9YFu9dqX3Rz4vj67wAWabuoI0Olf4H3tdLChlCYBK6aFss8D3yObg2kdPhnzfy0WJWxoHs+3Oo6Je1Lo67KrfSlQ9e86UV2BmPrLIJTRtahkSZDiEqvTcKrkVar4sCfnjyAOcOFyL5tAwqKoWGO7M2MZw+heFwzVu1bDS6wL+jbF8oNRwabh6dpqmcuHJncyi9YaxZ0TulXaL3pWd6iCu9FGmza00Tr7xpXkkUwpKqh0VVEqurJt0xp3itYE0us3ogJQ+36F3CtZ0HZbDSy32ZEcMw3+.rT4alD

I lied, it's two additional lines of code, with one of the being the GlobalRoutingManager reference.

Here's my ultra flexible continuous per event modulation snippet
https://forum.hise.audio/topic/10635/snippet-per-event-modulation-from-ui-with-global-storage

If you read the thread you can follow the entire history of that how or why that came to be in HISE.

For continuous modulation of volume per event, you can make it simpler as you don't need a network to just control the volume, the modulator itself will do:

1. create EventData Envelope (it's important that it's an envelope)
2. now you get two options. In both cases you'll be calling setEventData from the control callbacks, as well as the initial call for each event right after it's played with playNote
   1. publicize your eventIds by making the array a global instead of reg, then write into them from your interface script using tables or whatever
   2. link your Dynamics slider to another one in your realtime script, then use that slider's value to deduce your correct volume using tables or whatever