DIY Sequencer



  • I was pretty excited, then I saw the demo.

    Yes I would like plain white sprinkles on my vanilla ice cream please...
  • the case is horrible, no doubt. looks like a fun little project though...
  • The big gotcha is that this is not a sequencer - this is a custom MIDI/OSC controller for a piece of sequencer software. So you cannot use this without a computer behind the scenes :(

    I don't want to shift the topic, but I am interested on my side on building a sequencer after the Shruthi-1. Some "preliminary discussions": The main question is to find a sweet spot between a MB-Seq (full-featured but large) and the crowd of naive x0x or step sequencers. My original idea was something ableton-live-ish in which MIDI patterns are prebuilt on a PC and stored on a SD-card (one folder per song, one subfolder per track, one tiny MIDI file per pattern) ; with an interface focused on mute/chaining/CC and stacks of algorithmic effects (transpose, mute ever 2nd step, reverse, randomize order). But it seems that people into hardware prefer a full hardware workflow (not just something for live with main editing done on a PC...)

    Anyway, sequencers are a strategic bit of equipment and probably the one for which interface/interaction/flow are more important than ever ; so this is a topic worth discussing!
  • There is also a new midibox mini sequencer in the works that matches the sammich:
  • I've gone back and forth a million times about building a sequencer, but I always wonder whether it will just be an MPC with more knobs, analog stuff, and corny jokes. I think it would be really fun, but as Olivier hinted at, even a really full featured hardware sequencer (while really cool) would still probably get its ass kicked by Ableton in most respects. But I would be really curious to hear what people think as well.

    For my two cents, building a step-only sequencer seems kinda boring. I like the monkey pads on the MPC, with quantize and all that mess. Though the weird matrix-ey step-ey stuff like this:
    seems good for a lot of applications. People seem to like CV stuff as well.

    Yeah, so, what do you think?

  • unless you can come up with some kind of fancy interface or radical novelty take on what has gone before, i think you'll end up rehashing old ground...

    i do like the idea of plodding through a midi sequence in a daw and then mashing it up in a lil box though...
  • edited June 2010
    @Altitude: This 8-steps thing is interesting, it looks like a Gorf ^ 4. Indeed I would be happy to point some of the people asking me for a minimal seq to this project.

    @Todd: The point is to 1/ draw a line and say "I won't go further than this line, in terms of features" (otherwise it'll end up like a MB-Seq, which is probably the most complete hardware thing I've seen, and which clearly shows the huge amount of tedious stuff there can be on the road on such a project), 2/ have a new story to tell.

    The main story I want to tell in a sequencer is the concept of MIDI effects and algorithmic mutations of tracks - the same way you turn knobs on a synth to filter a basic waveform, you tweak objects that make something happen to a sequence. The whole concept of loading stuff on a SD-card is more of a corollary - something to keep me focus on the effects/mutations parts, instead of spending days reimplementing the step editor, x0x editor and preventing people from requesting features and features on the boring part. But maybe hacking some FAT32 lib and getting the SD-card to work + all the code to parse robustly the MIDI files is as boring as coding a basic x0x editor and step sequencer, meh...

    A long time ago, I programmed something like M ( for BeOS, and had fun doing Music in Twelve Parts-ey things with it. What I have in mind would be a hardware version of this, with several layers of algorithmic modifiers (shuffling, transposition, CC or parameter randomization) acting like filters/envelope over raw patterns, and CV inputs or buttons for the "conductor" features. Would anybody want that?
  • edited June 2010

    Ha, thats what I said when I talked to nILS about it, and he was not ammused. The main thing about the kaffeSEQ is everything will be available for real-time tweaking. No time frame yet..
  • Todd, I am with you on the mpc deal ( I would definately own one of these, but the moneyz... I also like to build this stuff myself :) )

    For me an ideal Sequencer should have a way to record events at realtime (i.e. firing a sample at a certain time in a 4 measure groove) and NOT quantizising it. Leave it untouched, exactly at the time I pushed the button.

    Also it should be able to swing like a mad man. Personally I could have heaps of fun with a sequencer like that and a 4 sample slot wtpa *HINTHINT*

    I am still a supernovice at the synthside of things, so I can not contribute on that topic - but for firing samples /drum sounds the above features really would help to get dwon some heavy electrofunk shuffled beats (which is what I want to do with it).

    Maybe a step sequencer with a pot above each step that shifts the triggerpoint foward or after the specific step... do you know what I mean? I think this could be really, REALLY nice!!!
  • I think every music hardware developer guy has considered building a step sequencer at one point; it's just the natural progression of things.

    Anyway, I'm all about sequencers and new ways to be interactive with them, but you guys are right: you shouldn't reinvent the wheel. MBSeq has every feature under the sun, the Gorf is as simple as it gets, and the MPC is known for its realtime capabilities and swing (a.k.a. the magic behind DJ Premier, DJ Muggs, et al.)

    If you want to make something fresh and new, realtime algo stuff sounds very interesting.

    Tell me how this sounds:
    What about a sequencer that doesn't require the user to input individual notes? What if the user presses a "generate" button and, according to various settings, the notes, lengths, velocity, etc. are seeded. In other words, the user sets up the rules, and the pattern is generated.
    The user has realtime access to various parameters via the knobs, such as: variation (re-seeds the algo, which results in a slightly different pattern with each turn of the knob), quantize (timing) amount, quantize pitch (scale) amount, note shuffle/shift, etc. etc.!

    Obviously, it would be a lot deeper than this, but the idea would be to have control over the effect of the algorithm on various elements on a pattern-level, rather than direct control over individual notes.

    I'd call it, "The Instant Inspiration Box".
    Oh wait, what's this? Autechre just called and they've ordered 16 of them.
  • Glitched, that sounds good :) But i'd imagine that would take some very clever algorithms to pull off!? Potentially that seems like it could be more work than the hardware side of things if one is expecting some nice musical results.

    I love quirky sequencers.. But it has to be said, there is quite a number of them to choose from these days as it is. So if you were to produce one, then it has to be different somehow. And therein lay the challenge!?

    But just to throw my ideas out there.. Last year I got a Yamaha Tenori-on (Its a Matrix style sequencer, plus synth, with midi). Silly expensive thing, and the built in sounds are questionable to put it mildly. And although you can load your own samples into it, even that feature is compromised in many ways. So I end up using it more as a midi sequencer.
    Then I wondered why Yamaha didn't make a midi sequencer only version. As the sequencing abilities of these things, although still flawed in some ways, are also equally brilliant in other ways. Which lead me to think about a matrix sequencer in kit form. With open source/user created programming modes. The Tenori-on actually has 6 sequencing types, yet most folk only seem vaguely familiar with 1 of them, score mode. Bounce and Random mode, to name but 2 are also very good and highly useable.. I think clever folk could come up with other cool writing modes too, given the opportunity.

    Before anyone shouts "MONOME!" Yes, it is initially a similar idea. However, the monome is basically a USB controller that relies on a computer and the software on it to work. Im thinking much more of a standalone midi sequencer that can do it's thing without a computer in sight. So yes, rather like a Tenori-on, but minus the expense.. Comes in kit form, and has the possibility to use user created writing modes that can be dumped into it.

    Good idea? Or am I mad as a mad hatter having a particularly mad moment?
  • MONOME! Oops...

    What you describe is awesome, Luap; why not combine both ideas?
    An algorithmic, stand-alone, tactile, open-source, MIDI sequencer that only Dr. Manhattan can program and use.

    You're right, programming the algos and rules would be difficult, but I'm sure they're well-documented somewhere. In fact, isn't Ruin & Wesen's Minicommand open-source? It certainly has a lot of generative features. Oh, and I hope @pichenettes sees this: it loads data from a microSD card. (Wasn't he looking for code that does that? Was that Todd? Ack, who knows.)

    Anyway, I think part of research should involve the question, "Does the world need another sequencer?"
    We were having this discussion over on the mutable-instruments forum, too:
  • Sorry, i don't get the Dr Manhattan thing!? But yes, why not a generative matrix sequencer! Or one capable of running such a mode.. Im all for it, if it is doable :)

    Still, as well as the usual software sequencers, I have a Doepfer MAQ 16/3, and the Tenori-on, which has me pretty well covered for odd ball hardware sequencers. Although I quite like the look of that FutureRetro Orb, with it's remix function. Not exactly generative, but a variation of sorts that seems quite interesting none the less.
    Then there is the Cirklon, and the Analogue solutions Europa and a whole bunch of others besides.. Not so many in kit form though, granted.

    I'll check that thread on the Mutable forum shortly..
  • There's a decent thread at the MB-808 forums about the desire to have a pretty simple, standalone 808/909 style drum sequencer:

    The project seems to have stalled though...
  • chezch this out you guys:
  • Yep, thats pretty much what I was talking about I guess! Nice job he did there too by the looks of it.

    Vid here:

    If it were me though, 64 buttons wouldn't quite cut it. I'd be after 8 X 16 for 128, or go the whole hog and have 16 X 16 for 256

    Still, I like that a lot!
  • @ luap more huh?

    How about:
  • That is very nice too! Good find.. It is rather like a Tenori-on. Which is no bad thing..
    I'll have to look into this Midibox stuff some more. But I suspect it could be pushing my DIY abilities a bit further than I am capable of??

    Anyway, I think you chaps see what I was getting at with the kit sequencer idea now. I'd be all over something like this as a more or less complete kit.
  • you'll have no problem with your diy abilities there.
    even I got some stuff running, albeit no SEQ finished yet, bummer...
  • edited June 2010
    I'm always looking for a few things in sequencers. i have a monome but not crazy about it as the layers of software you have to load to get it running is simply dumb and not very inspiring, and switching software is also a pain. If I get around to it it goes I think.
    I'd dearly love a stupid scratch pad pcb I could but in an arp to record cv/gate directly off the keyboard that has a button to record/stop and one to play with a knob to turn for tempo +/- from realtime. Sync & memory & midi dump is nice but I could live without that even. Should power off a standard cv synth 15V power supply. Could even have fewer knobs if it was controleld by a jack to footpedal I suppose.

    Next I'd love an 808 style step sequencer DIY. Nice if it was a long very thin linear box that did cv/gate or midi i/o. It should sit on top of anything, a dx-7 or a minimoog, narrow enough to not get in the way and an easy reach above the keyboard. Put the cables on the end, run off batteries?

    Bigger? Well a friend saw a Moon seq at the PNW show and what he really liked was that it could do rachecting, hocketing and all fthose clever T-Dream things becasue its clock could divide or offbeat predictably for different lines of cv, made a lot of syncopated bass lines possible and would allow for weirdo autechre percussion events. Tons of shuffle& swing and a $2000 price tag without actually generating anything but control voltages. A bit steep. Demo:

    I also favour compact and density of features, still use my oberheim step seq all the time, but again I wishe it had a counter and clock divide.

    Them's my thoughts. Back on the road tomorrow, getting tired and cranky.
  • Oh well, haters gonna hate. Beatseqr is a hell of a lot of fun to play music with. It's not designed to be a MIDI hardware purist dream, it's specifically a tool to use with computer music software.

    That's a pretty old video. Before you take your good-manners hat off, at least get up to date on the status of the project.

    If you can make hundreds of instances of a better sequencer, more power to you. And Olivier, I'm looking at you, man. Live and let live. The market of ideas is big enough for everyone.
  • good manners?
    i couldn't find a reason to complain about manners in the thread. you know, some people may not like the design (like myself also) but that seems a matter of taste. and it wouldn't have been a reason for me not to try it. i just don't play audio with a pc (matter of taste also) therefore i'd be in for something that works in the other direction (sw sequencer) or even better, all hardware.
    i really like all kinds of sequencers from different persons, be they small or big solutions, they still have their own character and way to work with, ie different inspiration.
    for that matter, if i could, i'd try to convince dnigrin to transfer his Klee to MCU land ;-D

    and i really couldn't find any really rude or "hater" comments here. after all it's a public discussion, with different opinions and people showing machines they like. maybe look at it in a more positive way, inspiration for the next (standalone) upgrade?
  • edited November 2011
    I'm puzzling over some high end sequencers which are out of my talent league, a nemo and a P3, and yet I play with the gorf as well. The different styles are interesting. Especially coming from old dos land like sound globs. (yeah I Live 8 etc, but its so boring for midi sequencing I'm afraid).

    Sound globs was a matrix of 16 boxes to click on. In each box were midi outs, and a probability calcuation for things like, speed of note generation (tempo-ish) probability of 1 to n notes playing at once, probability of note length (not teh same for every note), probaility of pitch of note/octave/scale/tuning, probability of velocity, or other controllers, probability of lfo modulation shape to pitch, and things like that. You could select any combo of channels 1-16 for each note out, shifting it to a different channel note by note, using 1 meant it all went to one instrument. You could copy any box to another and then randomize evrything, a few things or manually tweak variables in each probability. You played by mouse clicking. You coudl even assign anything to midi commands from say the 1-32 buttons on a dx-7, or even a stage lan of keyboards so several people could control different parts of the program live. Lovely bunch of chaos. And it did scales and weird tunings.

    Re MONOME, no velocity or touch volume. BIG weakness compared to an MPC style.

    I agree with the idea of all sequencers having their own character. I'd like a balance that is tactile but will musically randomize your input on each pass, to a degree you can control. I'd like it to evolve as a process and auto record what's going on so you could save if you were happy with it. No computer. sd card to and from computer, or usb dump.
Sign In or Register to comment.