MIDIbox SEQ: next attempt

[T.K]

After the last beta-testing days with my MIOS based MIDIbox SEQ

I noticed that so much advances functions have been added, that

the sequences can only be edited ergonomically with a dedicated

control surface - so, here my first try - feel free to add your

comments:

From the left to the right:

Track buttons: I still won’t use more than 4 tracks, since

this an analogue style sequencer + arpeggiator and no groovebox.

Therefore the 4 tracks can be selected directly with these buttons

Layer buttons: to select the “Note number/CC Number”,

“Veclocity/CC Value” and “Gatelength” Layer.

For TB303 fans: accent can be realized by varying the velocity,

slide by varying the gatelength or by assigning a portamento

controller to another track

Mute button: to mute tracks and steps

View button: to select different views on the displays

ChngAll button: when pressed, one rotary encoder changes

the values for all rotary encoders at the same time

Reset button: resets the Layer values

Mode Edit button: the multifunction buttons are used to

play a note/controller

Mode Pattern button: the multifunction buttons are used

to select a pattern on-the-fly (from A-1 to h-8: 128 patterns)

Mode Chain button: to select a chain of patterns (up

to 16)

Menu button: shortcut to jump into a menu directly

2 * 2x40 LCD: since a 2x16 display is really too small,

I want to see all notes/values which are played

rotary encoders: for a simple analogue style sequencer

pots are ok, but the powerfull track/layer function really

demands for endless knobs, otherwise the finetuning of a

pattern doesn’t make fun!

Multifunction buttons: see above

Data wheel: this rotary encoder replaces the Menu

Left/Right buttons, provides faster data entry

Select button: was “Snapshot” button in MB64

Menu button: was “Exit” button in MB64

F1-F4 buttons: free assignable buttons, can trigger

MIDI events or any functions directly (e.g. “Mute all”,

“MIDI Clock Master/Slave”, “Fwd”, “Rwd”)

Start/Stop button: function should be clear

Case: I will try to mount the panel on a common

• flat! - computer keyboard. The case will only contain

the control elements and DIN/DOUT modules, the other

modules (CORE, AOUT, Power supply) will be put into a

seperate box. Advantage: in this way the surface fits

nice in front of my MIDIbox SID

The core unit and the control surface will be connected

via a 6-wire cable.

Ok, thats all for today.

Best Regards, Thorsten.

P.S.: the 64-pot sequencer will still be provided by the

MBSEQ application, also the 2x16 display - for people

who don’t plan to use all features so extensive like me.

P.P.S.: the sequencer will also run without control

surface - with a singe core module. In this configuration

it can be configured from the Java frontend (JSynth).

[d2k]

my oh my oh my… Yes yes TK!

There’s the resta januray sorted :>

[pay_c]

Why buy stuff if you can make it even better with less money??

Can´t believe what I´m seeing here every now and then. Don´t wanna know where this whole page will be in five years!

We´ll need someone for writing instruction manuals for those babies!!!

TK: First gotta try to get everything into my brain. Perhaps then I can come up with some ideas. You´re definitely to fast… ;D

:o :o :o

[pay_c]

OK, got the point ( I hope … …).

Questions & Suggestions:

• Will there be octave up/down buttons (the free assignable ones e.g.)? We talked about that in our emails. Would be cool I think (going the whole range of about 5 octaves up and down with that enc isn´t so much fun)

• The Reset button should react after two seconds (like the old snapshot did). This way you wont delete the track accidentaly (and afterwards bite your own ass)… *if* you had´nt already thought of that.  ;)

Would be cool, this way, this button is freed up for another function if its just pressed

• Idea: What´s with some “Morphing Fun” in this box, too? E.g. just morphing the velocities and CC Values? (Morphing Notes & CC# doesn´t make to much sense, I think) I think that would be … just … … cool again.  ;D

Greets!  ;D

*lookingforwardforthisbaby* *cantwaituntilmytestsareoveragain*

[Captain_Hastings]

Whoa! Another beauty TK!  

Will this run with 2 cores, one  for the app and one for the control surface (driving the lcds and din/dout)?

[uclaros]

No, it’s just a dedicated controll surface for the MB64SEQv2.

It’s just a suggested layout!

[T.K]

Dan: I just have ordered the parts @ Reichelt. Maybe the first release will

be ready in 2 weeks, once I finished the breadboards Btw.: for this project

I will use detented rotary encoders (the China ones which were

organized by Ian Hurlock some time ago), since highest precission is

required for tweaking the notes. They have an integrated push button,

maybe I will use it for an additional function

Pay.c: thanks for this very good input!

transpose buttons: the beta version I sent you one week ago already

features a transpose function for each track. When enabled, 32 notes are

strechted over the whole pot range and can be transposed octave-wise. This

function will also be available in this variant (check the transpose button in

the menu section). Ok, it should be a piece of cake to provide also a

Transpose +/- button for the free assignable F-section (note: in fact every

DIN input can be used as free assignable function button, so its not

limmited to F1-F4)

Reset: yes, important!

Morph: great idea! Morphing between two patterns can be realized

very easily, the morph value could be controlled either by the data wheel, or

by an external modulation wheel, or by the velocity value of the played note

(-> transpose and arpeggiator mode).

For note values it could be usefull to switch between the note from Pattern

A to Pattern B when a certain morphing level is reached.

Morphing should be switchable (enable/disable) for all 12 layers seperately

Captain Hastings: only a single core module will be required. The design will

be extentable from:

• tiny: core module w/o any control function - especially nice as
  standalone sequencer or arpeggiator with predefined patterns, editable via
  a Java GUI (Win/Mac/Linux)
• minimal: core module, 1 2x16 LCD, 6 buttons and one rotary
  encoder (optional: two additional Inc/Dec buttons) connected via Port J5
  (requires no DIN module)
• vintage: 64 pots, min. 32 buttons and 16 LEDs: the classic MB64
  sequencer
• deluxe: see above
• extravagant: like above, but with 48 rotary encoders to control all 3
  layers at the same time. However, with endless knobs such effort isn’t really
  required, the layer switches are ergonomical enough

Best Regards, Thorsten.

[Duggle]

The different H/W formats is a great thing!

I like the Tiny option because it will allow me to trial the application and see if I like the workflow as a way of making music before I go ballistic and build a dedicated one.

Thorsten, will it be possible to edit in real time using the Tiny plus Java applet?

[Captain_Hastings]

only a single core module will be required.

Nice!  :)

The reason i asked about the number of core modules was that the control surface is only connected with a 6-wire cable and i wondered how all the data for the lcds, din and dout would be transferred?

[T.K]

Duggle: yes, patterns can be edited in realtime via SysEx or CC (there are

4 dedicated MIDI channels for the layers and config data) - nice for testing

Captain Hastings: you are right, this is something I forgot, the LCD requires

some additional lines, so two cables (15 pin connector + power) are required :-/

Btw.: new function

Groove button: to shift the 16 steps globally in positive or negative

direction by +7/-8 MIDI clocks

Morph and Groove will replace the MIDI and SysEx shortcut

button (these menus are reachable with the menu button at the right side

anyhow)

Best Regards, Thorsten.

[Nomical]

Two questions:

1. Will there be enough space (or even be an option at all) to change the code (by myself, i know you don’t need it TK) to a 16 track sequencer? 4 tracks is by far not enough for me.

2. How about pattern/song loading? Is it possible to load/change patterns/songs in sync, in other words without having to stop the sequencer?

This project is looking very good TK!!

[pay_c]

Btw.: new function

Groove button: to shift the 16 steps globally in positive or negative

direction by +7/-8 MIDI clocks

Morph and Groove will replace the MIDI and SysEx shortcut

button (these menus are reachable with the menu button at the right side

anyhow)

Oooooh my oh my oh my…

if it´s even getting better still now, I´ll loose my mind! That will be some encoder flood ordering from our side again!!

For that groove again: I already thought about something like that earlier. My idea was a random function added to the MIDI Clock (or subtracted). Since the PIC doesn´t support random values (naturally), you would have to produce it (which isn´t any problem - just take the noise of a transistor e.g., amplify it and bring it in on an AIN of the core). Just an idea…

Random grooving! I could imagine that would bring alot of natural feeling into some sequencer… what do you think?

[arumblack]

Momical: I too would like more tracks…but perhaps the core can’t handle them? if so perhaps a chain system like on the sid could be used, 4 cores would be 16 tracks, but using one CS.

i also wonder would it be possible to add more layers per track.

note number , gate length, velocity use up three and i could use those plus a couple cc’s? i suppose i could use another track on the same midi channel, but that uses up tracks.

so i guees what i really wonder is how much breathing room is left in the pic for modification/enhancement? Not that i even posses the skill to do it at this point, but theese are things i think about still.

[banneduser]

Looks great! Lot’s of new functionality

You probably know these people with Roland fixation: Well, I’m one of those So two suggestions:

Groove button : I like the idea, but how about Roland TR-909, 7x7 style Shuffle levels?! It sure does make sequences Damn Funky! Is it difficult to implement some kind of selectable groove templates for that?! (And has anyone got a suitable TR, so timings for those shuffles could be extacted?!)

Alternative sequencing control method, kind of like MC202:

ASCII illustration:

Non code:

          12 note "keyboard" [-] [+] octave transpose
 Â  Â  Â  Â  Â  Â  Â [][] Â [][][] Â  Â [S] slide (tie)
 Â  Â  Â  Â  Â  Â  [][][][][][][] Â  [<][>] [X] delete
 Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  Â  step

So, nothing fancy, just plain old step entry.. but that’s how I like to work: very fast way of “scratchpad” sequencing and collecting ideas. Also updating (just pressing keyboard button) live while sequencer is playing could update that step. And this mini keyboard could also be used to transpose sequences live or for driving arpeggiator

Well, I might just be dreaming,

Bye, Moebius

/edit fixing ascii

[T.K]

Pay.c: again - a nice idea!

When you look into the source code you will notice that I’ve

already implemented a pseudo random generator which

is used in “Random Direction” mode. The advantage of this

generator (in distance to a “true random” gen) is, that

it works deterministic - it can be reset, e.g. via MIDI Start,

and this makes random sequences reproducible.

It shouldn’t be a problem to use this generator also for

random shifts - since one step takes 6 MIDI clocks

(when the divider is set to 1), the groove feature makes

only sense from -5 to +5, other values could produce

a random value, e.g.:

+6 - random from 0..2,

+7: random from 0..5,

-6: random from -0..-2

-7: random from -0..-5

-8: random from -5..+5

Nomical:

patterns can already be changed within less than

1 mS - so in sync of course - latency was one of my

highest priorities during the implementation of

the sequencer engine.

Songs are nothing else than chains of patterns

(with loops from 1 to 16), it starts immediately,

so no issue.

16 track sequencer: everything is possible

BUT: only with some disadvantages, which

I don’t want to consider in my implementation.

With 16 tracks it wouldn’t be possible to store the whole

pattern in RAM. The sequencer core has to read the

pattern directly from EEPROM instead. This has no

timing disadvantages (remember that sending a

MIDI event takes 1 mS, an EEPROM value can be

read within 10 uS…), but it makes realtime editing

via MIDI impossible. Most parts of the control surface

handler would have to be changed, so that only

the select track is buffered in RAM. Features like

Morhphing wouldn’t be possible anymore, since this

requires a second data set of the layer values.

The whole SysEx store structure would have to be

changed to a new format, it wouldn’t be possible

to store 128 patches in a single BankStick, but only

32 or so… a second JSynth GUI would be required to

take the changed data structure into account (note:

realtime editing with JSynth wouldn’t be possible

anymore), and last but not least: a second documentation

would be required and somebody who answers the

questions to this different implementation.

Just take this into account…

And to repeat it for the guys who didn’t read my other

postings regarding the reason, why I only

need a 4 track sequencer: I’m using it as creativity

tool. Once a nice sequence “works”, I record it with

Logic, and the MBSEQ is free for the next track.

Sooner or later I’ve to rearrange and mixdown the

sequences anyhow, and for this an editor like Logic is

much more efficient (yes, in this situation I really like

to use a mouse).

If you really want to record, play, arrange, mix a whole

song with a standalone device, then you have to program

something which has been done by Emagic or Steinberg

15 years ago - on the Atari ST (a second-hand Atari costs

you less money and less effort)

Arumblack: for every CC’s you have to use a seperate

track. But see above, why more than this doesn’t make

sense. It is possible to chain multiple MBSEQs (synced

via MIDI clock), but it isn’t possible to control them from

a single control surface (it would have to save all

editable parameters in RAM anyhow)

Best Regards, Thorsten.

[T.K]

Shuffle: if somebody could tell me how it is working

exactly? (With examples - .mid files prefered)

Your seq editing idea: I wouldn’t be much effort for me

to integrate such a step recording function, it could

also be controlable from an external keyboard, so that

only the Step Left/Right and Slide/Accent/Delete button

is required. Or by a simple button replacement (just some

additional “special function buttons”, you know: up to 128

inputs are available

Best Regards, Thorsten.

[banneduser]

Too bad I sold TR-707 ages ago (bad sounds and no need for trigger output anymore).. Quick search on the net didn’t reveal anybody willing to share their Shuffled patterns.

SO: Has anybody got: TR-909, 7x7 or 626?!

(Basically it’s kind of a swing feel, notes between beats get shifted, by.. well, maybe we can get those midi files.)

And thanks for your input on control surface.. I was afraid that it could be too far from your original analogue style sequencer idea.

Thanks. Moebius

[sil909]

Reply I might have an archived thread about shuffleing ( tr707- 606-909 specific)  from AH archives.  I believe they are still down but I might have hard copy will look tomorrow morning and post info if i can find.

[banneduser]

Thorsten:

Thanks to sil909 and AnalogueHeaven archives Grooves extracted! No midi file, but

Here’s the table, representing one beat and it’s 16th note values shifted in relation to midi clock:

[pre]

              16ths:

         1      2      3      4    (1)

none:     0      0      0      0    (0)

Shuffle1: 0     -1      0     -1    (0)

Shuffle2: 0     -2      0     -2    (0)

Shuffle3: 0     -3      0     -3    (0)

Shuffle4: 0     -4      0     -4    (0)  

[/pre]

I think you can easily figure out what’s going on..

Bye, Moebius

oh, and link to the original posting:

http://www.retrosynth.com/ah/search/lookit.cgi?-v9705.1068

/edit: Yet another ASCII fix, (Opera not functioning properly with YaBB) I hope I got IT

[sil909]

good catch  ;D