Hotz Demo/Explanation: Can We Record and Filter?

Just watched this:

(There’s a brief glimpse of Scaler ca. 00:28.)

Had never heard of this device. My main reaction is that, while some of the features are indeed missing from current products, the setup sounds too complicated for my tastes.
(To be honest, I even feel Scaler itself could afford some simplification, “progressive disclosure”. Maybe that’s just me.)

The pressure-sensitive membrane controller pads might indeed be fun. The Embody Erae Touch could deliver a similar experience, despite some issues with the hardware.

I’m mostly interested in that one feature on the Hotz which, if I understand things correctly, allows you to record a progression and then filter incoming notes to be in the proper chord/scale relationship. I’ve been trying to build something similar on several occasions, using different tools, and there’s always something missing. It even got me into a wild set of raw notes about what the optimal setup would be.

There are many ways to use scale filters. Some of them have a weird thing where playing a new note will cut the sound (because it ends up on the same note so it sends a note off; I prefer it when it just maintains the note). Few accept notes from external devices to switch the filtering. Fewer still (if any) do chord/scale relationships. Have yet to find/create the proper setup.
Expected behaviour: arbitrary chords are played/fed into a filtering device. That same device takes in the notes from a performance-focused controller (in my case, probably a windcontroller) and filters notes to play within scales associated with each chord.
So, say the current chord is A♭dim7 (implying A♭ WH Diminished scale). If I play a C♮, it filters down to C♭. If I play a B, the C♭ still plays. If I play an E (legato), it glides to E. Next chord is D♭7 (implying D♭ Mixolydian or D♭ Lydian Dominant, etc.). The E keeps playing, creating tension. If I play it again, it filters down to E♭.
I was at least able to prototype this by sending scale notes to a filter, in sync with corresponding chords. It was as much fun as I expected.
I’d like to do this with arbitrary chords, directly from Scaler.
And it’d be nice if it could select different scales within the ones which make sense. So, in the D♭7 example, it could restrict to D♭ Lydian Dominant on one run, D♭ Mixolydian another, and even switch to D♭ Major Pentatonic at some point. Yes, it can be jarring. That’s partly the point. It’s the kind of thing I enjoy as it breaks my expectations.

It’s somewhat similar to the effect of playing rotating chords, like the setup Robby Kilgore created for Michael Brecker and which has been used on several recordings. Playing a note repeatedly creates different chords. Technically, because it’s deterministic, you can tell which chords will play next. In practice, it feels almost like random. Yet it sounds quite good. I really enjoy the tension/release effect.

So, back to the chord/scale filtering. I wish Scaler had such a feature.
Now, maybe I’m completely wrong about the Hotz. Maybe it doesn’t do anything similar to what I have in mind. In fact, come to think of it, it does sound like you’d have to program those relationships in advance, for those to work. Much less fun.
Still, that old device does sound like it can inspire current devs, especially now that some UX principles are better understood in the industry.

1 Like

This was an interesting (and deep) post, which I followed through as best I could to try and improve my fragile grasp of music theory. I was left with several areas I wasn’t quite clear about.

{1} It seems to that Scaler has elements of this filtering with the ‘lock to scale’ capability ? What might be different is that if your description covers the situation where the chord pattern in toto was not know beforehand (as it is in Scaler) but fed in at the same time as the controller notes.

{2} ‘Arbitrary’ chords could cover {a} a deterministic sequence, in that it was constructed from some set of rules (b} random or {c} chaotic but deterministic. [In the latter case it would appear random to the filter logic, but be entirely repeatable, which would beg the question of whether the filter would also act repeatably.

{3} I mused on the the alternative scenario of feeding notes and matching to a scale to produce chords, as opposed to here feeding chords and matching meaningful notes,

{4} I’m not sure about the filter logic. If the chords are fed on the fly, I’m assuming the filter
{a} on receipt of the first chord, identifies the set of scales which might contain it, in order to deduce what an incoming note might map to.
{b} performs the same on the second chord, but updates the set of possible scales on the (possible incorrect) assumption that the first chord was in the target scale.
{c} this is repeated until no chords exist within the set of potentials, in which case the logic has to either go with the notion that it’s a new scale or work on the basis it was a borrowed chord etc.
It sounds quite complicated to work it all out.

{5} I didn’t quite understand the example. Ab WH diminished scale is G# #A B C# D E F G. Abdim7 contains G# B D F, so I couldn’t see why C# would be change to C - probably being thick.

{6} I followed the link to Robby Gilgore. I got the gist of the ‘rotator’ idea, but was puzzled by the reference to Enigma. This machine was (a) totally deterministic and (b) completely reversible (obviously) There was nothing random about it, so any input for any given setting of the wheels would always be mapped to the same thing. It’s just a set of rules for mapping. Indeed, the cipher people at Bletchley Park back engineered the complete logical structure of the Enigma from a code message which was sent twice - an amazing intellectual feat. It led to the development of what was arguably the worlds first external program machine, Colossus.

1 Like

Hello @panda !

Thanks a whole lot for your reply! Really appreciated.

Just to make sure we’re on the same page…
When I talk about a scale filter, I mean the kind of device which takes notes from an external controller and maps them to a scale. It’s actually called “scaler” in some cases. Or “scale quantizer”. Or “mapper”.
AFAICT, the Scaler 2 plugin doesn’t quantize incoming notes. If I play a C3 in a track which has Scaler on it, it’ll trigger the first chord in the current trigger mode. Maybe I’m completely wrong and there’s a way to use Scaler as a scaler? That’d be great!

EDIT: Ha! Now I get it.
If I select a scale and I don’t select a trigger mode and I set things up for « scale notes only », it does act as a scaler! Didn’t realize this at all. Not precisely discoverable, I’d say.
Now, how do I switch scales based on chords, though? That’s the main part. (And I don’t really mind if it’s a 1-to-1 relationship, that it always switches to C Lydian when I reach a CMaj7 in the progression. I mostly want the scale to follow the chord instead of the opposite.)

Also, when I talk about chord/scale relationships, I mean something like those in Marc Sabatella’s Jazz Improvisation Primer or those displayed in iReal Pro. In fact, if you consider F to be an “avoid note” on CMaj7, you privilege C Lydian instead of C Major, even though CMaj7 “comes from” C Major.
In the case I’m describing, each new chord implies a scale on its own, without reference to the previous chord. There could be additional logic which does take into account what the previous chord was, allowing for voice leading (which is almost inexistant in “theory plugins” and not-that-intuitive in Scaler 2). I’m just talking about: Scaler 2 produces a set of chords and each transition between chords switches the filtering to a scale “implied by” the chord itself.

The whole notion of chord/scale improv is controversial. I personally enjoy what it can do with electronic instruments for precisely the reason Scaler 2 exists: it produces cool results without having to think through principles from a given music theory. In fact, it’s more of a discovery tool than training to be a Berklee-like player.
(Disclaimer: I was trained in classical saxophone before doing graduate research in ethnomusicology.)

Agreed. I enjoy all three forms for different reasons. In this case, I was mostly thinking about option {c} because that’s what 4Pockets’s Progressions does.
I want the filter to choose semi-randomly between a small set of options.

Transpose down a half-step. :wink:

(In other words, I’d be going from C to B, in your enharmonic version.)

Ditto. What does help, from the Engima analogy, is when you loop through sequences of different lengths. Completely deterministic and predictable. It’s just that, when you listen casually, it sounds like the chords are “random”.
I typically have very short sequences for bass notes, alternating between dissonant and consonant intervals. I don’t keep the idea of maintaining a parallel fifth. I switch the other voices around in jumps.

Interested in those musings. I’ve had some of the same, thoughts which is pretty much the reason I enjoy rotating chords. There have been some attempts at “auto-accompaniment” which are similar to that. I’d like something which automatically does species counterpoint. In the 1990s, a Slovene gave me a floppy which contained code to do just that. Never tried it.

So… I’ll investigate things more on my end. And I hope things are a little bit clearer. I do realize the way my brain works is awkward and weird. Let’s call it my “brand”. :wink:

Maybe this is why I encounter note-offs, sometimes, using keys-lock feature for soloing… :thinking:

If it is the case, I hope the issue will be fixed, a day or another

In the case I’m describing, each new chord implies a scale on its own, without reference to the previous chord.

If I understand well, this could be useful to play jazz solos

There have been some attempts at “auto-accompaniment” which are similar to that.

AIR Ignite maybe?

Interesting video
that device still need a good keyboard player like the one in the video, but I think it can help poor ones like me, so I hope some of that will fall into next releases of Scaler as well

Quite possible. Sounds like there are different approaches to filtering so they don’t always match our expectations.

Right. Jazz-style improv on chord progressions. In this case, the progressions need not come from diatonic patterns like ii-V-I.

Thought it might have been in AIEP3 and it’s not. Sounds like it’s been discontinued, actually. How does/did it work?

The kind of auto-accompaniment I mean is a longterm dream of mine (from music school in the late 1980s). You play something and the computer follows along. A little bit like this:

(I do hope this won’t trigger a whole thing about AI. And I do realize that saying so decreases the odds that I’ll avoid it. So… erm… There’s probably another thread for AI chat. If you want to share there, all good with me. kthxbai!)

1 Like

Well, you can do a look here then

and to this post of mine also :grin:

Thanks but no thanks. :wink:

1 Like