Why did you (not) sell your MARF?

Discussing some incredible modules that don't quite fit into the other forum categories.

Moderators: Kent, luketeaford, Joe.

User avatar
rklem
Common Wiggler
Posts: 221
Joined: Sat Oct 07, 2017 1:51 pm
Location: Berlin

Why did you (not) sell your MARF?

Post by rklem » Sun Jul 18, 2021 5:33 am

Apologies for starting another MARF thread – I know there are plenty already. But I noticed that there are a lot of MARF clone 2nd hand offers, and I'm wondering why. I would like to hear from those of you who own(ed) a MARF: Why did you sell it? Did you move on to a 250e or 251e? Or why would you never sell it?

I don't have a 250/250e sequencer yet, but I had an OD ER-101/102, which seems similar to the 251e – and I had a love/hate relationship with it and finally sold it. The MARF seems like one of the most unique modules in the 200/200e world, and more hands on than the 251e. I read that most of the bugs are ironed out in the latest clone revisions and I'm tempted to give it a try, but first I like to hear why I might regret it...?

ropesalesman
Learning to Wiggle
Posts: 18
Joined: Sun Mar 14, 2021 12:11 am

Re: Why did you (not) sell your MARF?

Post by ropesalesman » Sun Jul 18, 2021 9:22 am

This is not really answering your question, but from my own experience builders like myself avoided it because there was always news that it was being fixed/worked on. Now that the bugs are being ironed out, many builders took the plunge and many builders offset costs of building a system by building 2 of everything and selling the second one. The first "wave" of units I saw available were all from known builders.
Last edited by ropesalesman on Sun Jul 18, 2021 10:07 am, edited 1 time in total.

Reed
Blame Canada
Posts: 399
Joined: Thu Jan 28, 2010 9:58 pm

Re: Why did you (not) sell your MARF?

Post by Reed » Sun Jul 18, 2021 10:04 am

I got a couple of the pretend Marfs whenever he started selling them, and even after some back-and-forth software revisions with him and his programmer, it never did the things that make the original Marf magic. In fact, it never did the ONE basic thing that made the Marf unique among sequencers.(Here's where 6 people, including the late Mr. Buchla, chime in that the original Marf never worked either. I'll get to that).

The groundbreaking concept of the Marf isn't the row of ARP-like sliders on the right. It's the multiple (get it, multiple?) output modules on the left. In the most musical sense, the sliders and per-step settings on the left are a motive, and the right hand output modules are different staves of music. Suzanne Ciani wrote a paper about doing hocketed patterns on a Marf, which is basically that.

You can extend the concept by having some of the sliders do pitch/duration rows, and some of them do gestures with glide, sustain, etc. Then you can simultaneously have one output module do notes and another output module do envelopes. Or you can have some of the sliders, via one output module, control other sliders and everything programmed with them, including loops, pulses, etc. In fact, with enough output modules you can have one Marf replace ALL of the voltage control modules on an instrument: LFO, envelope (including ADSR), lag processing, external quantizing, pulsing, you name it. I have an old Marf with six output modules, and it's like having a group of people riffing off the same musical phrase.

All this stuff is in Allen Strange's brilliantly thorough Marf manual, btw.

If they really wanted the new Marf to be Marfier, the expander module would have had extra output modules, not just 16 extra sliders. But I bet he sold a ton of those expanders. It's all about looks. (Wow, is that a 32-step ARP sequencer???)

When I switched output modules on the fake Marf, the programmed data went away from the right side. Kinda defeated the purpose. The 249e did the same thing. Basically two completely separate sequencers trying to share one set of controls. Maybe they ironed that out. Clearly wasn't a priority. I know there's a digital copy/paste thing you can do on the 249e and the 248r, but once you have a big patch going and you want to change the sliders to affect everything in real time, it's a dealbreaker.

In the world of rhythmically synchronized music, the original Marf had an issue when being pulsed externally, fixable by a patching workaround using a 281 for a short pulse delay. Because the 21st century copy Marf doesn't output a dual-voltage pulse, you can't do the same trick. Mark Verbos did a clever mod to my original Marfs that made it even easier, and for some reason the C++ guy couldn't, well, didn't do it.

Other workflow things made the original Marf faster to get around, like the sideways switch being able to continuously loop through the steps, or being able to hold down a programmer switch and sweep multiple steps so they all have the same setting.

Donald Buchla scrapped the Marf when his broke during a live performance. It was an ambitious undertaking using one or two hundred CMOS chips instead of a microprocessor from the future. The Buchla 300 did a software-based Marf, which you can hear on the LP "Rosenboom and Buchla, Together in Performance."

The new thing that looks like a Marf (but with slider caps that fall off) should ideally be the thing Buchla was dreaming of. One master microprocessor interfaced with a ProTools Command-like physical surface. People are working on getting the software closer. It is such a deep module that my concerns are probably a tiny corner of the big picture, but from what I've seen so far, there's a long way to go.
Reed and Caroline
"Hello Science" and "Buchla and Singing"
available on Vince Clarke's veryrecords.com

User avatar
shoegazer86
Common Wiggler
Posts: 189
Joined: Sat Oct 20, 2018 9:16 am
Location: Buffalo

Re: Why did you (not) sell your MARF?

Post by shoegazer86 » Sun Jul 18, 2021 2:38 pm

For me, my reasoning behind leaving my 248R in the dust kind of mirrors Reed's response, as in it is quite broken and not quite so similar to the original in terms of function. I've always had a bad taste in my mouth with the "workalike" for other reasons.

I was the first person besides Roman to complete my 248R MARF build. Which means I bought a $700 dollar pcb set and a $100 panel and waited a year and a half before a BOM could even be released. Right out the gate, this was a gigantic red flag. I was shocked that nobody else seemed to be mad about this. I tried and tried to get ahold of the BOM for months, well after the PayPal return window. After this happened we all know that the boards that were initially purchased had so many errors that we had to be sent out replacement boards.

This waiting game amounted to having an expensive module that in short - "worked". I was a bit nonplussed. I had a working unit but it seemed like it was a very bittersweet victory. I think this was the first time I decided that I would work towards discovering Buchla as Buchla, instead of through the eyes of someone who hasn't really played or experienced the ecosystem. Will it do what you think it does? Absolutely. But that nagging question tugs at the back of your brain. "Am I really experiencing the MARF?"

I guess this question was asked a little louder in my brain than others. But with anything, YMMV.

The OG MARF was very elegantly designed (or was it MAFG? :lol:)
Image

The Achilles' heel was not the sum of its' parts, but the A series CMOS ICs that were highly sensitive. Most MARFs in the wild need tons of decoupling caps added to avoid race conditions and lockups.
The MEMS Project
Modular Electronic Music Systems
MEMSProject.info

User avatar
Fuzz
Common Wiggler
Posts: 172
Joined: Mon Aug 01, 2016 5:05 pm
Location: Seattle

Re: Why did you (not) sell your MARF?

Post by Fuzz » Sun Jul 18, 2021 4:27 pm

Over the past year and a half Steve Barsky, Dave Brown and others have done a lot to fix up the v2.1 MARF. Many of the issues that Reed stated have been addressed (external clocking from a single pulse, ability to hold down a program switch and quickly sweep through stages to modify multiple steps), as well as many other noise and programming issues that plagued the module from the start.

I feel that it's a very capable module at this point. I was able to use an original at EMS last year and it was not working as well as my MARF clone at home. I got to experience what Don felt when the thing "died" on him in mid patch.

I agree with Reed, the 248-1602 setup is just the tip of the iceberg, and the current 17-32 step expander without more output modules doesn't make a lot of sense to me. Being able to have 6 more output sections sounds like a dream to me.
darksparkler.bandcamp.com - Music
sourceofuncertainty.audio - Podcast

jimfowler
Ultra Wiggler
Posts: 849
Joined: Mon Feb 22, 2016 1:13 pm
Location: New York
Contact:

Re: Why did you (not) sell your MARF?

Post by jimfowler » Sun Jul 18, 2021 8:45 pm

I have two and have no intent to dispense with them any time soon. I'd recommend anybody who was (understandably - as I was) let down by the 248r in its many prior iterations to spend some time with it in its current form and reassess. I feel that it is now at once immediate and very deep and am certain I've only scratched the surface.

Cablebasher
Wiggling with Experience
Posts: 462
Joined: Sat Aug 03, 2013 6:51 am
Location: England, UK.

Re: Why did you (not) sell your MARF?

Post by Cablebasher » Mon Jul 19, 2021 3:17 am

It would be great if someone could do a new marf overview video showing all the features that now work and maybe mention build version / software version.

The source of uncertainly video is great but I believe much has improved in function since then.

I am still confused and on the fence about building and R or S clone version.

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1727
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: Why did you (not) sell your MARF?

Post by cygmu » Mon Jul 19, 2021 4:51 am

I was quite heavily involved in the development of the latest firmware for this last summer and had some conversations with Reed about features to add or fix, some of which we did get done, but this is one that I never noticed until now:
Reed wrote:
Sun Jul 18, 2021 10:04 am
When I switched output modules on the fake Marf, the programmed data went away from the right side. Kinda defeated the purpose.
This had me scratching my head for a while but I think I understand it now. In the Electric Music Store MARF workalike (I can't call it a clone because the internals are entirely different) the "level 2 programming" i.e. quantized vs continuous, voltage range, sloped/stepped, stop/sustain/enable mode per step, is conflated with the "level 3 programming" i.e. the output sections. That is to say, you have a set of level 2 settings per output module. If you "display" programmed output 2, for example, then when you set the step settings, they apply to output 2 only. Reading the MARF manual carefully this morning I realise this is different from the original and I am glad to have learned this! When we picked up the task of refreshing the firmware for the 248r we just ran with what was already there and made the best of it but this seems fundamental.

Would you be interested in an alternative firmware where the level 2 programming applied to all output modules? I think this would be relatively easy to do. It gives strictly less flexibility than currently -- as things are you could have one output module entirely in "stepped" mode and one entirely in "slope" for the same sequence, which would not be possible in this alternative version -- but you save on only having to program each step once, and bring the UI closer to the original.

A few other things I think we did get done:
In the world of rhythmically synchronized music, the original Marf had an issue when being pulsed externally, fixable by a patching workaround using a 281 for a short pulse delay. Because the 21st century copy Marf doesn't output a dual-voltage pulse, you can't do the same trick. Mark Verbos did a clever mod to my original Marfs that made it even easier, and for some reason the C++ guy couldn't, well, didn't do it.
The latest firmware does allow for external clocking. It was harder than we wanted it to be for reasons I can't remember now. But simultaneously pulsing the start and stop jacks will advance one step at a time.
Other workflow things made the original Marf faster to get around, like the sideways switch being able to continuously loop through the steps, or being able to hold down a programmer switch and sweep multiple steps so they all have the same setting.
I am pretty sure this works in the latest firmware. I think it worked in earlier versions too but other issues may have made it harder to use.
It is such a deep module that my concerns are probably a tiny corner of the big picture, but from what I've seen so far, there's a long way to go.
Perhaps so! There were many things that were frustrating about working on the code for this thing after Roman open-sourced it. The hardware architecture has a lot of quirks that we were fighting against and there is certainly not much scope for adding additional output modules, which is a shame. For the versions with the microcontroller on a daughterboard, it might be possible -- but a lot of work -- to create a fresh daughterboard hosting a completely different microcontroller with some spare pins to communicate with expander output sections. But that's definitely a long job for someone...

User avatar
shoegazer86
Common Wiggler
Posts: 189
Joined: Sat Oct 20, 2018 9:16 am
Location: Buffalo

Re: Why did you (not) sell your MARF?

Post by shoegazer86 » Mon Jul 19, 2021 5:21 am

I'd like to clarify that I haven't experienced a MARF with the new software and hardware updates, so my experience was a bit skewed. I eventually stopped looking at the thing because Chip and I started to work on the OG MARF for the project. The idea is, as insane as it sounds, to make a real clone of it and see how deep the rabbit hole went. This is still in the works. I'm sure with the help of Dave Brown we will get somewhere with circumventing some of the ICs that were troublesome.
The MEMS Project
Modular Electronic Music Systems
MEMSProject.info

tarandfeathers
Wiggling with Experience
Posts: 398
Joined: Wed May 07, 2014 3:02 pm
Location: Leeds, UK
Contact:

Re: Why did you (not) sell your MARF?

Post by tarandfeathers » Mon Jul 19, 2021 8:41 am

cygmu wrote:
Mon Jul 19, 2021 4:51 am
The hardware architecture has a lot of quirks that we were fighting against and there is certainly not much scope for adding additional output modules, which is a shame. For the versions with the microcontroller on a daughterboard, it might be possible -- but a lot of work -- to create a fresh daughterboard hosting a completely different microcontroller with some spare pins to communicate with expander output sections. But that's definitely a long job for someone...
When I traced it out ( https://github.com/DunningtonAudio/248_2.5 ) it occurred to me that the "to computer" UART connection could maybe be used for an output expander. I don't know whether the current uC/EEPROM have space limitations (or if the UART speed would be limiting - I'm not an STM guy...) which would make this unworkable but since when Roman originally announced the v2 he indicated that an output expander was on the cards I sort of assumed that there would have to have been some hardware provision to make that possible.

I have not done any playing with it yet but I guess that the UART is currently used for debug, though I'm not sure why this is required since there's a dedicated JTAG header?
Dunnington Audio - Various Buchla format accoutrements available
GitHub Repo - Open source Buchla format projects

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1727
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: Why did you (not) sell your MARF?

Post by cygmu » Mon Jul 19, 2021 9:05 am

tarandfeathers wrote:
Mon Jul 19, 2021 8:41 am
I have not done any playing with it yet but I guess that the UART is currently used for debug, though I'm not sure why this is required since there's a dedicated JTAG header?
UART is not used at all. One of the UART pins is repurposed to pick up the setting of one of the DIP switches, to tell the MARF whether there is an expander attached or not. I had forgotten about this -- if we give up on that DIP switch then the UART could perhaps be used to talk to further stages. Hmmm, interesting! And thanks for the schematic too, that's wonderful! We did the coding essentially without one.

I'm conscious that I am derailing the thread a bit now so I won't drone on.

tarandfeathers
Wiggling with Experience
Posts: 398
Joined: Wed May 07, 2014 3:02 pm
Location: Leeds, UK
Contact:

Re: Why did you (not) sell your MARF?

Post by tarandfeathers » Mon Jul 19, 2021 9:21 am

cygmu wrote:
Mon Jul 19, 2021 9:05 am

UART is not used at all. One of the UART pins is repurposed to pick up the setting of one of the DIP switches, to tell the MARF whether there is an expander attached or not. I had forgotten about this -- if we give up on that DIP switch then the UART could perhaps be used to talk to further stages. Hmmm, interesting! And thanks for the schematic too, that's wonderful! We did the coding essentially without one.
Good point, I'd forgotten that that pin was used in two places - more consideration required I think. Coding without access to the hardware design sounds like a total nightmare! Makes the improvements that have been made all the more impressive.

Feel free to give me a shout if you want to discuss further without muddying this thread up.
Dunnington Audio - Various Buchla format accoutrements available
GitHub Repo - Open source Buchla format projects

User avatar
rklem
Common Wiggler
Posts: 221
Joined: Sat Oct 07, 2017 1:51 pm
Location: Berlin

Re: Why did you (not) sell your MARF?

Post by rklem » Mon Jul 19, 2021 11:10 am

Thanks for all the details and insights! I learned a lot already.

And please feel free to "derail" the thread. I really appreciate that this started a discussion which might lead to a even better "MARF V2.x" in future.

User avatar
dougcl
Number 6
Posts: 4083
Joined: Sat Jul 26, 2008 3:04 am
Location: Portland OR

Re: Why did you (not) sell your MARF?

Post by dougcl » Mon Jul 19, 2021 11:40 am

The functions of the 250e appear to me to be very similar to 248. Wouldn't it be correct to say that they are the same, except that the 250e has one "output module?"

maxl0rd
Common Wiggler
Posts: 250
Joined: Mon Aug 05, 2013 1:51 pm

Re: Why did you (not) sell your MARF?

Post by maxl0rd » Mon Jul 19, 2021 2:14 pm

dougcl wrote:
Mon Jul 19, 2021 11:40 am
The functions of the 250e appear to me to be very similar to 248. Wouldn't it be correct to say that they are the same, except that the 250e has one "output module?"
Yes, true. However, as the 250 also includes looping stages it's actually closer to the digital AFGs in the later systems starting with the 300.

User avatar
Peake
I'm in ur DIY. Filling cases with Buchla
Posts: 8015
Joined: Fri Nov 07, 2008 10:00 pm
Location: Loss Angeles

Re: Why did you (not) sell your MARF?

Post by Peake » Mon Jul 19, 2021 2:31 pm

This is very exciting.
This is not the place I'd imagined it to be.

User avatar
szzdz
Common Wiggler
Posts: 140
Joined: Sat Jun 22, 2019 3:18 pm
Location: chicago, il

Re: Why did you (not) sell your MARF?

Post by szzdz » Mon Jul 19, 2021 5:04 pm

hey roman. can you clone the 249 please :hail:

User avatar
exeterdown
Wiggling with Anxiety
Posts: 711
Joined: Tue Sep 27, 2011 12:58 pm
Location: Australia

Re: Why did you (not) sell your MARF?

Post by exeterdown » Tue Jul 20, 2021 12:45 am

dougcl wrote:
Mon Jul 19, 2021 11:40 am
The functions of the 250e appear to me to be very similar to 248. Wouldn't it be correct to say that they are the same, except that the 250e has one "output module?"
Yea this is largely what I understand - people seem to be going nuts over the MARF and I'm looking at my 250 going 'maybe I'm missing something?' The outer ring controls either output 1 or 2 selected with a toggle button - is what's being described here with the MARF that you should be able to add extra expander modules and add extra outputs - still controlled with that same outer ring?

I don't get how you could replace LFOs with this. Envelopes sure. The two (or three) ring outputs of the 250 are always locked in step too - is the MARF supposed to be able to run these polyrhythmically?

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1727
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: Why did you (not) sell your MARF?

Post by cygmu » Tue Jul 20, 2021 3:26 am

exeterdown wrote:
Tue Jul 20, 2021 12:45 am
I don't get how you could replace LFOs with this. Envelopes sure. The two (or three) ring outputs of the 250 are always locked in step too - is the MARF supposed to be able to run these polyrhythmically?
Yes, each output section of the MARF runs its own clock so it can cycle through the slider voltages at its own pace.

User avatar
ws9848
Super Deluxe Wiggler
Posts: 1201
Joined: Fri Feb 04, 2011 1:20 am
Location: Sacramento, CA

Re: Why did you (not) sell your MARF?

Post by ws9848 » Wed Jul 21, 2021 9:58 pm

szzdz wrote:
Mon Jul 19, 2021 5:04 pm
hey roman. can you clone the 249 please :hail:
I think dougcl should do it because at least it would work properly

User avatar
ryangaston
Wiggling with Experience
Posts: 435
Joined: Fri Feb 22, 2013 5:53 pm
Location: Little Rock, AR

Re: Why did you (not) sell your MARF?

Post by ryangaston » Thu Jul 22, 2021 8:19 am

Reed wrote:
Sun Jul 18, 2021 10:04 am
The groundbreaking concept of the Marf isn't the row of ARP-like sliders on the right. It's the multiple (get it, multiple?) output modules on the left. In the most musical sense, the sliders and per-step settings on the left are a motive, and the right hand output modules are different staves of music.
. . .
You can extend the concept by having some of the sliders do pitch/duration rows, and some of them do gestures with glide, sustain, etc. Then you can simultaneously have one output module do notes and another output module do envelopes. Or you can have some of the sliders, via one output module, control other sliders and everything programmed with them, including loops, pulses, etc. In fact, with enough output modules you can have one Marf replace ALL of the voltage control modules on an instrument: LFO, envelope (including ADSR), lag processing, external quantizing, pulsing, you name it. I have an old Marf with six output modules, and it's like having a group of people riffing off the same musical phrase.
I just wanted to chime in here to say thanks to Reed for explaining all of this. As someone who has never played either an original or a modern MARF, I have to admit that, despite having read the Strange manual a number of times, it never clicked for me that the sliders/switch settings on the original are more like a single data set/array for each of the output modules to pull from. (Seems I'm probably not the only one who fell into this admittedly weird default assumption?)

Something about working mostly on modern gear made me default to assuming that slider settings were independent for each output module, which TBH always made the MARF seem like it would be...I dunno, kind of annoying to use? So I never really gave it much thought.

But the idea of several outputs sharing the same set of step values/settings is super super interesting (and as you point out, much more similar to the "sequence table" idea in the 300 Series / PATCH than I had realized before).

Anyway, just wanted to say thanks for sharing your thoughts, gives me a lot to think about! :guinness:
"I have always been outside and I've chosen to remain there."
--DB

User avatar
thee ghost ov n_phay
Common Wiggler
Posts: 200
Joined: Sun Jan 25, 2015 11:59 am

Re: Why did you (not) sell your MARF?

Post by thee ghost ov n_phay » Thu Jul 22, 2021 1:48 pm

It would be very interesting (and AMAZING) to see the talented & knowledgeable people here putting together a ground-up interpretation of the MARF that works like the original one is supposed to and also has the extra output modules & ability to add a block of steps too. The "r" version was always a bit of a scary prospect!

User avatar
szzdz
Common Wiggler
Posts: 140
Joined: Sat Jun 22, 2019 3:18 pm
Location: chicago, il

Re: Why did you (not) sell your MARF?

Post by szzdz » Thu Jul 22, 2021 2:35 pm

ws9848 wrote:
Wed Jul 21, 2021 9:58 pm
szzdz wrote:
Mon Jul 19, 2021 5:04 pm
hey roman. can you clone the 249 please :hail:
I think dougcl should do it because at least it would work properly
even better :love: would be the perfect compromise of space vs output modules imo

User avatar
sendepause
Wiggling with Experience
Posts: 268
Joined: Mon Jan 20, 2014 3:54 am
Location: Amsterdam

Re: Why did you (not) sell your MARF?

Post by sendepause » Fri Jul 23, 2021 3:49 am

Interesting thread.
Just to be clear: if we talk about the fake MARF, we talk about the Electric Music Store / Roman AND samodular one?

User avatar
shoegazer86
Common Wiggler
Posts: 189
Joined: Sat Oct 20, 2018 9:16 am
Location: Buffalo

Re: Why did you (not) sell your MARF?

Post by shoegazer86 » Fri Jul 23, 2021 4:44 am

sendepause wrote:
Fri Jul 23, 2021 3:49 am
Interesting thread.
Just to be clear: if we talk about the fake MARF, we talk about the Electric Music Store / Roman AND samodular one?
Yep
The MEMS Project
Modular Electronic Music Systems
MEMSProject.info

Post Reply

Return to “Buchla, EMS & Serge”