stryd_one

No worries mate...take a look at the code for the example apps and start doing some reverse engineering, and grab a few books on assembler code and most importantly I guess, the PIC's datasheets from the microchip website... While youre there grab MPLab so you've got an IDE to write the code in... And dive in! ;D

Heya :) Thanks for your words of encouragement! You sure that sequencer is based on MIOS? There isn't a MIOS sequencer yet, not one started by TK anyway... Although with MIOS being out there, there is absolutely nothing stopping someone from writing their own seq code for it, which is not based on the MB64Seq... It seems that the code for this project uses Thorsten's bootloader to enable software updates... Hard to say more than that, as they haven't made the source codes available.... As for the piano roll thing, the current design works something like what you've suggested, however it runs vertically (Sonar/Logic/Cubase style)... Basically at the moment there are 16x12 buttons, and when in the grid editor, those buttons will be 16 steps of 12 keys (one octave) with octave up/down buttons...FWIW I'd really like to make it 32 steps long but that may not be practical so there might be a thing where i split it into steps 1-16 and 17-32 or something like that). The reason for doing this vertically is to preserve the ability to alter pitch on different notes at once... I guess the hard part about a piano roll interface is setting note's start time and length. The actual note that is played, is easy to set, as the available options are limited (127 note numbers)... However the placement of that note in time, is interesting. When we use a step seq, by slicing up a bar into 16 steps, we limit the available options for note start and end times, so it gets easy, but if you want to work outside of those steps, it gets interesting because the number of times the note could start and end are limited only by the machine's clock (at worst case its a 2MHz clock so thats 2 million places to start or end a note every second hmmm... at best, it's gonna be 24 places to start or end the note, in every quarter note which passes (MIDI clock resolution)... I thought about having an LCD screen with two mousewheel style rollers, one vertical and one horizontal, which could be used to set the current position, so we'll see how that goes... I realised after thinking up this idea that a PC mouse could also be very easily modified to do the same job... What's more the mouse could also be used as an X-Y controller ;) .... A very interesting part of this is going to be what's involved in getting the piano roll info on a LCD.. I must admit that the LCD side of things is something I know very little about :( I'm strongly considering alowing control of the box by PC, and only supporting the piano roll under that interface. This is cool with me despite my hatred for PC's and their love for crashing at inconvenient moments, as if the PC crashed, due to the nature of the machine, it would just keep playing... You'd just have to reboot the PC and reattach to the seq which will still be thrashing along. the software on the PC would be very basic, and could easily be run on the cruddliest old PC with a serial port and a copy of win95. we like that, cause we're broke ;) Of course if you didn't connect a PC then the machine would still work fine, just no piano roll. well, youd get that grid thing of course, just no graphic piano roll. If you don't need pitch info then youre just repeating the same note, why not just loop it? Anyway fortunately for us, the MIDI spec doesn't allow for notes without pitch or velocity, so if the info has to be sent, I figure I might as well make it so you can edit those things, but it will work in the same fashion as fruity, where if you just push the button and turn it on, it'll set itself to velocity 100, Gate time of one step, middle C. And yes there will be accent and slide. Accent will scale the other notes down to 90% of their original velocity, this is to allow for accent on notes that are already at the maximum velocity. I realise that the slide (portamento) feature is more built into the synth itself than the seq, however i wanted a quick way to draw longer notes, so say if you have a step with a C3 note, followed by another step with a C3 note, and you put the porta/slide on the first step, then it will automatically blend the two into one C3 note, as long as 2 steps. If the notes are different pitches, then it will overlap them, so the first step will actually finish shortly after the second begins. This should get the portamento happening on your external synths... Anyway, once again the existing MB64Seq already fills your requirements. The buttons on the Seq can be used to mute on/off the notes, and the knobs can be used to make an accent I think this thread is about the most I've written about this thing in one go heh! I'm starting to scare myself off, seeing all the features listed out ;)

Hey mate... The MB64SEQ does almost everything this box does so its hardly worth building the project you linked to if you already have an MB64SEQ...I thought the page worthy of bookmarking though :) Thanks.. I do like the idea of a beatbox from one of these things! If you could combine a MB64Seq with a MBSID, you could upload waveforms to the SID's wavetables and fill it with drums and basses :) ... By the time I'm ready for a sound source, the ISA bridge may be ready, and we could use a turtle beach pinnacle card and basically have a Kurzweil K2000 in a box..... But I digress ;D I've mentiond a few times on the forum and have been working extensively with flownez (also on this forum) to find a way to use the MB64 platform to make a step sequencer more along the fruityloops style than the oldschool analog style. I must admit there's some merit in the logic of many people who suggest that I might as well just get an atari or something... But if youre like me, you just dont trust a computer in a live situation... Anyway the flow diagrams and basic panel layouts and the inner workings are laid out and ready, along with some kinda interesting math :-/ (Pages and pages of it), which I thought I might mention because the guy on that page wrote: "Currently, only 16 steps per bar and 6 (24PPQ) clock ticks per step is implemented. (I haven't worked out the logistics of crossfading between two different patterns of different lengths or different time signatures.) " Well thanks to some very late nights, the help of a math guru at work, some technical expertise from flownez, and a HUUUGE helping had from TK of course :D I have worked that out, and I'm just chilling now until the MIOS sequencer is released, so that I can base my code along it's principles and capabilities... But I warn you now, one thing is clearly evident - This thing is going to need a few core modules at least, and heaps of ram, and won't be a small project or one that gets finished superfast.. I (un)fortunately have a job hehehe Seeing as the topic has come up and I've got something more concrete to put forward this time, I might mention a few bits I wanted it to have and are currently included in the design... Quite a few of these don't exist elsewhere... Full duplex - record and edit while playing back without interruption Rrealtime Quantizing - so you can record say drum triggers and it will snap all the hits to a step on the pattern, on the fly, and play away looping that great break you just did, then hit mute buttons and you get the picture... A "Really honest officer, I didn't steal the idea for this interface from fruityloops" interface with rows of buttons and a little screen for editing parameters of the step, and the ability not only to sequence notes into patterns, but to sequence those patterns into songs (actually I wanted to do three layers, so you can sequence steps into patterns into phrases into songs) Polyrhythmic structures - mixing up time signatures and keeping them in sync A side effect of the timing/length/resolution math required for this is that you''ll also be able to do triplet notes or weird things that I dont even know if they exist like fifthlet notes (this is the real reason I'm going to all this trouble) Piano roll - It's more of a Grid edit but it's nicer than faders IMO (But hey I'll make it so you can use knobs/faders too)... We'll see if I can find a way to get a set of piano keys in there... I really can't see it being difficult, without veolicty sensitivy, theyre just a row of switches like any other...Just gotta find a set of plastic keys.. Despite all this stuff it's still very much a step sequencer... I plan on using a Roland MC50Mk2 and/or kawai Q-80EX for the more intricate stuff, however to be honest, because of the way this thing works (I hope;) ), it would be quite happy doing that kind of work itself...but I can't think of an easy way to display and edit it... Try and think of a way to emulate your mouse and PC screen with sonar or logic or something's piano roll.....On MIOS :-/ Last thing for now, Envelope style MIDI CC's - You set the level at desired steps and the way to curve the envelope, and a resolution at which to send the controllers - a nice way to send filter sweeps or autowahs etc which uses hardly any controller info (which eats RAM for breakfast) (No officer I swear I didn't steal this from Sonar and neither did Logic hehehe) Anyway I'm usually a "less talk, more action" kind of guy so I'll go back to being hushed and shy about my little addiction heheheh ;) I hope you like it so far, obviously you're thinking of something kind of similar, so feel free to throw in any thoughts or suggestions, I'd love to hear them :) Geez that's a long post! sorry! :-X!

I see the idea now... Yeh you could do this in the same way anti-skid braking works (at least on airplanes I don't know squat about cars heheh)... You basically watch the speed of rotation, and measure the change, and react according to change (IE if the wheel is turning slower than the plane's speed should make it, its skidding so let the brake off, etc)...would be pretty easy with a turntable because you can use a fixed speed of rotation so you dont even have to monitor the speed...then you just measure the difference between the fixed rotation speed (shall we say 45rpm? heh) and the actual RPM read from the light sensor ... The math needed to do this in MIOS is already there, take a look at the MIOS code for some of the projects, or on the PICList website... Definitely not hard to time a series of pulses and measure a BPM, nor to measure the difference between the two and send a pitchbend message accordingly... But the thing is, there's infinite room for change in speed going faster, and a limited, but still fairly broad range going slower... I mean how could we get it to handle it if you stop the spinning disc, or do a back spin, or a forward spin for that matter? Most MIDI synths and samplers only pitchbend a few octaves either way at best....doh! Food for thought... I'm willing to help with code if ya like but I'm not sure it's really workable... I mean you could make the device, there's 16,383 different pitchbend settings after all (14 bits...), so lots of different speeds could be accounted for, but the thing is that not much will be able to make really practical use of it :( I dunno, maybe I missed the point again heheh

Didn't you read that edrum link? Use peizo buzzers.

oh yeh and...yes, you can make smaller triggers with peizo stuff, and the velocity is usually controlled by the voltage received from the trigger/peizo pad/battery connected to wire/whatever analog in.

i really cant see it being that hard to use the ADC on the PIC to do this... Just a matter of whether you want it badly enough to write your own code... I'm in the same situation with a sequencer, and after 5 years of waiting for someone else to build it, I'm gonna use MIOS as a basis and do some extra code, maybe you'll feel like doing the same... FWIW it's pretty easy to pick up cheap trigger to midi convertor boxes and theres many drum machines that incorporate the feature into their design... Honestly it's probably cheaper and definitely faster to buy something like that and build your own triggers... Another thing I did hehe

why not use a combination of lfo's and pitchbend and modwheels modulating pitch?...

That explains all the extra attention to the site lately....

Not to mention a whole bunch of new documentation and images!!! :o You just never stop do you mate :D I think you should copyright the phrase "Thanks Thorsten!" So people have to pay you royalties when they say it... You'll make a fortune out of me alone ;) Thanks Thorsten! [quote name=Thorsten Klose link=board=troubleshooting;num=1054487915;start=0#5 date=06/05/03 at 00:59:40]alright, the 3tracks variant is now in the v101 archive under MBHP->MIDIbox64seq->Download hope it works ;-) Best Regards, Thorsten.

Hi all, I'm doing some planning work for the sequencer I've spoken of building, and I'm trying to find out how the MB64SEQ (and eventually the MIOSSEQ) is triggered to send events, both when using internal timing, and when synced to MIDI Clock... Is is a real world clock (eg counted in seconds etc), or a musical clock like midi clock, or maybe something with the PIC's clock speed, or a bit of both, maybe none of the above ;) As much info as you have, would be a great help. Thanks!

Sounds interesting :) I'm sure it would be possible to whip up some code to do this What does the DB50XG do that the SID can't?

Excellent news... Myself and my good friend Ben will be working on MIOS sequencer's too - hopefully we will all be able to help each other along. I don't know what you have in mind, however I actually stumbled accross uCApps while searching for a way to implement some rather "weird" functions myself - things that are not available anywhere else at all, commercially, free, software, hardware or otherwise... They're ideas I've been nurturing for some years now, so they're pretty specific, and as such I'm not too sure if this is the place for us to discuss it, we might be talking too much ??? But I'd definately like to talk with you and we can bounce ideas off each other and blend them into one box perhaps? Anyway to make a very long story very short, my idea would be like a step sequence-sequencer a-la fruityloops (step sequences are stored as patterns which can also be sequenced into songs) and the major difference is the existence of individual clock multipliers _and_ dividers, both of which can be even or odd numbers, including primes. Simple idea, but nobody has ever done it as far as I can tell. (In case anyone wonders why I want these features, it's for sequencing polyrhythms) Yeh, I'm pretty interested in this too. I originally thought that my ideas would take major adjustments to the source in order to be implemented, but after MIOS was announced, I was not so sure - and now, with the MIOS Code Reference and the 'code snippets' which we've all drooled over, I mean, seen ;) ... Well it looks like I may not have to make any alterations at all to the source code, and might get away with just adding some code snippets to MIOS. Too True! ;D

To be honest, even though it means an increased learning curve, I do kinda like the fact that MIOS is assembler only... However it looks like there is some level of 'scripting' (Refferred to elsewhere as "code snippets") available in MIOS which also make other funtions more accessible... I guess what I'm trying to say is, it's just fine how it is :)

I'd definately be interested... If it's on EfNet I can probably get at least a few nodes of a pretty decent botpack in there too... On the flipside, decentralising the point of reference (currently this forum I guess) could be detrimental.... Any other thoughts anyone?

Man... I've been lurking my way around here, ordering parts, and studying circuit diagrams, waiting patiently for the day I would see this post :) I just had to take this opportunity to state my complete awe and amazement in your work Thorsten - not to mention the selfless sharing of it. You truly are, as we say in .au, a bloody bonza bloke! ;D All the best mate... I'm sure we will see each other around here, as no sooner had I read this post, than I had received my new PIC18F452's in the mail (3 of the little beauties! woooo) Thorsten.... YOU RULE!!!