earthsea

  • @[a.d] it's pretty crazy there. lots of experiments!

  • here's a little recording i made with the new thing: https://soundcloud.com/galapagoose/meditations-00-earthsea

  • So, it seems that one could sort of make a clock out of earhsea, when a pattern is "linearized" and played as arp loop, then the gate produced by the EDGE output can feed some other clock module.
    this gets a little confused when root note is switched, but I was glad to stumble on this "solution"...

  • if you "legato" press another note while still holding a note, the pattern will continue playing and shift position without restarting, so you'll have a constant clock source.

  • i'd like to add a "global linearize" shape to linearize the entire bank of patterns to the same time base, so you can dynamically switch patterns and keep time.

  • how do you do the "legato press"??

  • just meant if you push another key while still holding the previous key, instead of fully releasing all keys before pressing the next one.

  • I see, so, if I understand thi scorrectly: that means that in order to not break the tempo during transposition one would need to hold a key down at all times, yes?

  • that would work, but you can also just press the transposition on the beat and it will continue in time?!

  • well, if I run that gate into a SCM or some other clock module, then it extracts the timing from subsequent pulses, wher even slightest shifts cause ripples in the clock structure. So, trying to hit it on the beat makes it even more problematic, because it often generates super fast clocks, producing bursts etc.

    I also tested the legato treatment as suggested by tehn, nd it does not seem to work. Even when holding down the last position key, and then selecting new position, the new position produces its own gate, which is not in time with the arp pattern...

  • i don't have mine in front of me to check, but this might only work in HOLD (default) edge mode??

  • ALSO, is there a map somewhere of actual note/pitch values that are assigned to specific grid buttons?

  • if it's not present, i should make a firmware mod that makes hold mode not reset timing.

    there's no pitch map. that was originally in there as highly configurable, and both trent and i found it completely confusing and not terribly useful. what are you aiming to do?

  • RE: pitch map: just trying to learn :-) I want to see how the pitches are distributed on the grid, thinking about ways of referencing/complementing other scale/pitch producing modules. I guess as a very visual person i find it helpful to see the actual layout of the instrument.
    Crippling as it often is, even after working with sound as a medium for over two decades, I do not have perfect pitch and can not always identify pitch values from hearing only... :-(

  • i think our difficulty in providing a 'layout' is that the whole grid is entirely arbitrary, especially in eurorack where oscillators are generally freely tuned across the full range.

    for reference though, the bottom left key (one in from the control column) is 0volts. octaves (volts) are on 'double diagonals', that is two right and two up.

    we could make a diagram based on that but the better approach is to pick a tone in the middle of the grid and tune your oscillators at that setting (you can always use the same key if that helps). this way melodies can go both up and down.

    after it's in tune i'd suggest it's far easier to spend a few hours playing it to learn the map rather than worry about what note names you're playing. you can learn scale patterns pretty quickly like this. learn within a tight section (say 5x5) then explore how notes are duplicated across the grid / up and down octaves.

  • thank you for your thoughtful reply @galapagoose
    i will definitely pursue your suggestions.
    totally true regarding the fluidity of modular environment and tuning/pitch.
    my question was I think about your last observation about how notes are duplicated across the grid. just trying to form a visual/mental map of it.

  • So, I examined the timing in arp a little closer, and what I see happening is that in HOLD mode, when doing the legato switch, the POS (pitch) value does indeed switch in time (meaning that the pitch defined by the second button press is reflected on next beat in the arp sequence). HOWEVER that is not what the gate/EDGE does. The EDGE outputs a pulse whenever button is pressed... If the EDGE was doing what POS is doing thn the "tempo" on EDGE would be maintained.
    Hope that description makes sense :-)

  • Sorry about all the posts :-) But am spending some quality time with earthsea, so all sorts of questions arise!

    Like this one:

    Would it make sense if the "linearize" function was a toggle?
    I would find it super useful to be able to essentially undo the linearized pattern and/or move between the two ways of organization. It could open up some new possibilities for structuring and restructuring the performance in a very organic way...

    What do you guys think?

  • undoing would be tricky. i could make a shape to "copy this to next pattern" do destructive edits would be ok.

    i'll investigate the hold/legato timing.

  • I suppose save to next pattern would work, although I think being able to just toggle between the saved pattern, and the linearized version would have a much better/easier, more intuitive flow.

    +

    Also, I think I might have encountered a bug:

    When accelerating a pattern speed, there is a threshold beyond which I am unable to decelerate that pattern. So, can go up in speed, and at some point can not go back down: it gets stuck up there. The only way to fix it was to not save the work and cycle power to the case...
    I did not test it super carefully to see exactly how many times I need to double the speed to get it stuck, but I was able to reproduce this problem few times.
    Curious to see if other users can reproduce it as well.

  • when you get it "stuck" does it crash?? can't you just recall the preset?

    definitely there's a point to where you speed it up until all the values become crazy tiny and loose all granularity.

    i'll test this out and should incorporate a fix to prevent anything that'll stop operation, if that's happening.

  • hoy- i made a video of this thing:

  • i've watched that about 10 times trent, awesome work

  • yeah sounds great. those two VCO (?) modules look/sound interesting too...

  • hah interesting yes! not at liberty to say anymore...

  • the osc sounds amazing!

    everything about this clip is beautiful

  • this should be front page for bit :)

    great example of all the devices involved

  • read, mangrove, on the front panel of one?

    "Adults of the mangrove cricket show a clear and persistent circatidal rhythm in ... a circatidal oscillator, which has characteristics similar to a circadian oscillator."

    interesting times...

    great clip.

  • Well, I think using earth sea with the new Mutable Instruments "Elements" will be amazing... can not wait to try this combination!!!

  • So, am really enjoying digging deeper with earthsea.

    I am not sure if that would be useful to others, but something I would find useful would be the ability to save a kind of visual map for the grid. Users can decide how they might use this feature: someone might (after tuning their oscillator) want to mark all the "black keys" on the grid. Or somone might want to mark the buttons that are "in scale" that they are using in particular piece. Or someone might just want to highlight a certain octave etc. etc.
    The way that would work is that these "marked" buttons could be just lit at a very faint setting, just barely lit up, so that all other functions can still deliver what they are offering "on top" of the marked buttons.
    Upper right corner button (like ALT in WW) could be the access to the screen wher user gets to select the buttons they want to mark?...

    Anyway: I think it would make the instrument even richer in experience. Thinking of this as custom fretboard markers of sorts.

  • ... nobody thinks this is a good idea?

  • i do think this is a good idea!

  • If nothing else, I think this would make learning the instrument much easier.

    So, tehn, do you think this could get implemented eventually?

    I think this could just be a kind of "global" setting for the instrument as a whole.
    Or these can be saved with each pattern?

  • i'll be thinking about it-- it seems it'd make sense at a "preset" level.

  • You are absolutely right: makes a lot of sense for each "layout" to be saved per preset. Then the patterns and settings within that preset become a sort of explorations of this topology.

  • The more time I spend with earthsea the more I wish to be able to externally sync it. It is so great to work out, and mangle/modify themes and sequences on the fly... but integrating these effectively with other modules or in a more rhythmically cohesive compositions is challenging.

    As is, it is only possible to use the edge out as a clock to slave everything else to it, effectively making earthsea the "central" clock for th epiece. But as a clock it has such incredibly limited scope: once the pattern is commited, you can only double or half the time it plays at, so no fine tuning the tempo (unless I missed that somehow?). If one tries to match things tempo wise in any other way: figuring out the exact tempo earthsea plays at is nearly impossible. etc. etc.

    +

    Also, are not many people using this module?
    Seems sort of quiet here.

  • i've been considering a proper "sync" implementation and it's not at all straightforward unfortunately. if you have an idea of a method that would suit your needs yet have some versatility in the scope of the module, i'd be very interested.

  • How would the external signal be integrated?
    There are no inputs on the module... could the EDGE have dual function? With some user setting it can be the output of the gate, OR a gate input for sync?...

  • what i mean more, is how would the clock integrate with the earthsea user experience?

    ie, just "quantize" start/stop times when recording gestures?

  • My first thoughts on that is that the external clock ONLY controls the playback, not the recording process.
    So, when one wants to use the earthsea as it is now: free-flow improvisation etc. they can, and that is how their pattern is recorded and played without external clock.
    However playback of these paterns get quantized when the external clock is present.

    There could be settings similar to the three EDGE behaviors that are there now, that could define how patterns get quantized. For example whether the durations are preserved or not (quantize only attack or both attack and release)...

  • I think the clock should only trigger the beginning of loop playback. When the clock/loop is running, it will cut off recording and the loop will restart if a note is held. if the loop is not running, recording occurs as before, then the entire length of the recording is scaled within the loop time given by the clock (once the loop is running). since the clock will determine the length of the loop, the quantize grid will be automatically synced with the clock. no need for a clock tick for each note like w.w. and no need to worry further about quantizing. and if the pattern is only one note long then you can still get one note / tick.

    this preserves free-form recording if quantize isn't used but still allows integration with clock sources. also some cool really fast loops.

    this would also solve the other earthsea drawback I agree with and that is the lack of more control over the loop speed, independent from a clock input - double up and down leaves me wishing I had a knob to better control the loop speed - which would just be the loop clock in question.

    the only question would be whether shortening the loop length shortens all the steps so the entire pattern gets played, or preserves the step length and only plays as much of the pattern as the loop length allows. the latter option would kind of derail the convenience of the loop length determining the quantize grid while the former would seem more consistent with the current earthsea functionality. another interesting design choice (hypothetically speaking) if there were any way to include both...don't know that it'd be worth it.

    I know this is a stretch, but could say the 3rd parameter output & pot be retasked, the pot as loop speed and the output to clock input? I really wouldn't miss that 3rd cv out if I could trade it for a clock. but you'd really want an endless fader to adjust loop speed with a delta - know anyone who makes a high resolution endless rotary fader?

    how about a module to integrate an arc?

    imagine a module that had a cv and trigger out (and maybe delta cv out) for each encoder on the arc4, cv stationary or oscillates +/- lfo style (or back and forth within a range - might be tough to execute without pushbutton), and triggers on each negative to positive, edw-style. essentially make each encoder an L+FO with sine and square+ outputs. You could vary 'friction/sensitivity' for each encoder (basically a multiplier for delta), lemur-style, maybe with a pot for each encoder - ooo, such good mental masturbation.

    sorry, I let my imagination get the best of me late at night, I should refrain from spouting such outlandish crackpot ideas

    still I think that trying to sync individual notes to a clock on earthsea is overcomplicated, overlaps too much with the functionality of white whale, and is therefore misguided. I think that syncing a clock to the loop would provide the missing functionality to integrate earthsea into a clock-based setup, while maintaining the ability to create crazy polyrhythms within that environment.

  • This sounds good. I like the idea of only quantizing the loop as a whole as well, but to me this should be one of the options, rather than the only quantizing scenario.

    For example: if you imagine the three EDGE settings (hold, fixed, drone): there could be similar page for quantize options:
    -- first one: quantize only attack points, with all notes equal length determined by clock, similar to "linearized" setting;
    -- second one: quantize attacks and ends, preserving (approximate)note durations as recorded in pattern;
    -- third: quantize the begining of entire loop (as described by @sphiralstudios above)

    I don't see this as duplicating white whale functionality at all.
    the manner in which the patterns are recorded is so totally different on both modules. earthsea is so instantaneous and performance oriented. Now, if there was a way to use earthsea to record patterns INTO white whale, that would be a different story :-) !

  • ok-- i'm considering a tiny extension module for inputs. keep these ideas coming and i'll work on the possibilities.

  • That is fantastic news @tehn!

  • bah - comment posted twice, see below

  • All right @laborcamp, I'm with you on your quantize settings - although I still kinda feel like the loop clock would be enough (with some extra workaround). ** edit - my feelings have changed, see notes re: mode 2

    so,
    mode 1 - an incoming clock tick plays the next note in the sequence regardless of sequence timing.
    mode 2 - an incoming clock tick counts through the sequence based on 16 steps?(or however many steps are set on top row of grid)? and the sequence timing is preserved
    mode 3 - an incoming clock tick begins playback at the beginning of the loop. timing of the loop stays the same (step length preserved) and the loop is simply cut off if it doesn't finish playing before the next incoming clock tick
    mode 4 - an incoming clock tick begins playback at the beginning of the loop. 2 consecutive clock ticks determine the loop length time. i.e. the entire pattern plays between each clock tick

    I'm including mode 3 just for the sake of not leaving any possibilities out at this stage. I do think this mode could be sonically interesting, however. mode 2 would accomplish what you're saying I think, but by simply slicing the pattern into even chunks, which I imagine would be easier to execute than having a midi-style map of noteons and noteoffs. now that I think about mode 2 this way I wonder how much you would use mode 1. if your sequence was quantized, mode 2 would still give you notes on the clock to sync, but I guess you still might not want notes in between ticks or ticks without notes.

    actually, mode 2 could be a winner-take-all answer - ignore when the notes start as far as the clock is concerned, just sync the clock with the steps on the top row of the grid, and let the quantize put the notes on the clock if you want them there (side note - an undo might be a nice feature, for a quantize that kills part of the rhythm or when you accidentally hit a note and lose an awesome sequence. maybe a diamond shape?). If you're going to sync the clock with something more frequent than the loop beginning, the steps at the top of the grid seem like a much better choice than the note events. I'm not sure I'd want to part with mode 4 though - I REALLY like the idea of treating the sequence as a whole like a single note to trigger, and feel like it's consistent with earthsea's dharma (if my impressions are correct).

    what would be useful would be a clock divider - haven't looked yet, I'd be amazed if they aren't out there in some form - that sends a trigger every 2nd, 3rd, 4th, etc. trigger in, for example. maybe a clock in, a clock out, and a pot to set 1/2, 1/3, 1/4, etc. then you could have say a 4 note sequence with earthsea, but trigger it every fourth clock tick from the w.w., and the two should sync. or even cooler if you had a 5 note sequence on earthsea but triggered every fourth clock tick from w.w. you should get a 5 against 4 polyrhythm (using mode 4). I guess meadowlands kind of operates this way? - haven't played with that one yet.

    it's awesome that you're considering this. ever grateful for your listening, @tehn

    I've been thinking more about how a module that integrates with the arc could work, let me know if you ever want to hear ideas (I won't push them - but they're killer). I know the arc seems almost passe at this point, but it is such an awesome interface, just so badass. A module that turned it into 4 lfos / triggers could breathe new life into it...just sayin'

  • (deleted) It occurred to me that this post was better emailed directly to monome.

  • +1 for the addition of scale hints. Also what are the limitations on pattern recording? From a reading of the code it looks like 128 events? What about time wise?

    I note there are debug print statements in the code, how do you debug this code?

  • on the back of the circuit there's a header with rx/tx uart that's pin compatible with this guy: https://www.sparkfun.com/products/9716 (other manufacturers make the same thing).

    use minicom (or whatever terminal) to connect and see output.

    scale hints, made an issue. https://github.com/tehn/mod/issues/11

  • Excited for the "scale hints"! :-)