aleph-proto

monome, the iconic grid controller that launched them all, has always been a device tethered to a computer. Without a USB connection to your machine, it is an attractive but functionless box. The latest monome project, the result of a collaboration between Brian Crabtree and musician Ezra Buchla (yes, there’s a relation) is different. It is a computer, with all the functions that entails, but in a box designed for sound.

It has:
A brain: Two of them, in fact – a DSP chip (BF533 blackfin, 533 mHz with 64 MB SDRAM) and an AVR32 for control.
Audio connections: 4 in, 4 out balanced 192k/24-bit. The first two inputs can be switched to instrument gain.
A grayscale display: OLED, “beautifully low-res” (and thus sidestepping the problems that can come from supporting GUIs on Linux).
Analog control voltage: 4 in, 4 out CV on 3.55mm 0-10V connections, compatible with eurorack.
USB, expansion: Connect to a computer via USB, or host other USB devices – like the monome, natch, or any other class-compliant device (for HID or MIDI). There’s also an SD memory card slot and 1/4″ foot jacks. Curiously, there isn’t MIDI DIN, which seems unfortunate, though you can host one via the USB port. There’s also custom support planned for the Madrona Labs soundplane.
Controls: “Very high resolution” optical encoders.

And with this, you can do whatever you like. You can assign the encoders. You can assign and program the CV controls. You can run software to turn the box into a synth, or a sequencer, or an effects box, or a generative sound machine.

aleph prototype looper and drum synth from tehn on Vimeo.

But you know that – you’ve used computers before. aleph is notable because it’s the latest glimpse of how a “computer” in the near future may not be the general purpose device. In that sense, Steve Jobs’ term “post-PC” is absolutely fitting. And while the concept is nothing new, accessibility of these sorts of devices has gone up substantially, performance has gone up substantially, as costs have plummeted. This isn’t the first device to put a music computer in a box, but we’re now in an age when you would most want to use the result.

aleph-modular

aleph-proto-rear

aleph itself is US$1400. That’s pricey by embedded computer standards, and would seem to leave the door open to a lower-cost competitor. As a dedicated music machine, though, it seems about right, in line with drum machines and so on that have vastly more limited functions and no re-programmability. (Well, not unless you’re Elektron.) It also seems a better investment of your $1400 than one of the larger monome controllers for the same price, in that the return is a self-contained box that can make sound. I also expect the aleph to follow in the footsteps of other monome projects, in both uncompromising hardware quality (made in no small part in the USA), and growing value from user-contributed software.

If you are a developer, you can make your own software in C; monome promises extensive documentation and even a ready-made disk image of the Linux toolchain so you don’t have to set it up yourself.

But you don’t need to be a programmer. Out of the box, aleph will ship with “a complex layered looper, a peculiar monosynth, and a sharp-enveloped drum synth to start, and our highly-flexible patchable control environment called bees.” (Peculiar, eh?)

bees is perhaps the most interesting of these, a “modular control environment” that will let you map the controls on the aleph and create your own instrumental, sequencing, effect, and control patches using modules, as well as manage these presets in performance.

aleph apparently isn’t open source hardware, but the bees software and USB interface are. And custom support is planned, too, for the strange and wonderful shnth, a similar DSP-in-a-box sound creation.

I’ll be interested to see how the Bluefin DSP chip performs; there are, of course, many other possible embedded approaches that could do this. But the approach, in case it isn’t already clear, is to build a high-performance box to replace a computer. From the FAQ:

we wanted a device that could provide a focused and customized dynamic control system for a complex live audio set-up; a way to integrate and manipulate devices quickly and dynamically without the hazards of an overburdened laptop running many software applications often at the mercy of a greedy operating system.

Shipping by the end of the year. Santa…?

http://monome.org/aleph/

Got questions about this? I hope we’ll interview Brian and Ezra (we’ve talked to Brian before, most recently about the arc). So we can pass those questions along.

  • http://fzero.ca/ Fabio Neves

    I would consider it if there was a way to see bees in action. Saying “hey, we have this awesome patching environment!” without showing it and expecting anybody to just throw money at them is a bit unrealistic. Having said that, a small stand-alone box that can be controlled by, say, a launchpad mini is really interesting.

  • nayseven

    And if you want listen thoses guys talking about this new toy ( and livid and soundplanes..): here the last night Seatlle conference video : http://decibel.hosted.panopto.com/Panopto/Pages/Viewer/Default.aspx?id=a9cb0dce-9a70-497e-b24e-cfe728427000

  • Mutis Mayfield

    Expensive…

    • Ethan Vosburgh

      Agreed. I understand boutique devices, and the prices that come with that, but Monome to me crosses the line into the needlessly hipster instead of wanting to see their instruments in as many people’s hands as possible.

    • Bob Rawkz

      There are all manners of price points for every product out there. Get over it. Too much with the hipster name calling bullshit. I wouldn’t even buy the aleph, but people buy ferrari’s and 4000 dollar handbags. Sometimes for status, sometimes for the quality. In this case, it’s to support things not made of plastic from china. More power to them. Buy yourself a dell and a launchpad, and stop complaining.

    • Ethan Vosburgh

      That’s a fair enough point, and I’ll even admit that I’m being hypocritical as someone that only buys expensive Apple laptops. And I buy boutique devices like the DSI Tempest or Elektron Octatrack. But I’ll stand my ground that to me $800 for a grid of buttons (admittedly a device I’d really like to own) is overpriced for a product that has been in production since 2006. The limited, hand made production runs seemed cool back then. Now it just comes across as a bit silly. But this is really just me whining about not having the disposable income to justify the purchase ;)

      Anyway, this product looks interesting and I welcome any devices that try to bring electronic instruments into new territory. I don’t want to come across like I’m bashing Monome too much.

    • heinrichz

      No way should we get over unreasonable price points, they shoudl be fact checked and challenged at all times, because otherwise it will send the wrong message to the uninformed kid that just wants to make music. not everybody is as well informed as you ! That’s exactly what we don’t need hipster shit to show off as if you were driving a Ferrari. Make music please and stop showing off your money.

    • Bob Ro(ck)ss

      What message? People can make music with whatever they want to. Buchla has 30,000 dollar systems. They’re amazing. Is someone an asshole for owning one? Is someone showing off by playing an American strat vs. a Japanese one? Is a 2000 dollar dsi tempest against your dumb musical ethics? The crabtrees aren’t making millions of dollars from Monome, they’re making something quality, so what constitutes unreasonable? What we don’t need is more silly people jumping on the internet hipster rally cry.

    • chompy

      This (and the OP-1) are actually cheap compared to embedded systems using essentially the same hardware, but in other industries. I’ve seen almost the exact same hardware for hospital systems costing over $8000.

    • Mutis Mayfield

      In less time than anyone could expect we are going to see an ipad with full iOSX on it for 500$…

      But even it doesn’t occurs, you could buy TODAY (not at december) an ipad 4, mac mini and iconnectivity 4+ (and if you want a cv2midi unit) for less than a 1400 $, and start patching or just make music.

      Pd: check samplr app it have some things from op-1 and even improvements in the concept.

    • chompy

      Sure, but if you are doing custom hardware based on the blackfin $1400 isn’t expensive. This is more like buying a nord or other handmade electronic instrument which all cost a lot more than mass produced general consumer electronics. My upright bass cost $14,000 and my cousin showed me his $40,000 viola last night. None of this electronic stuff is expensive compared to actual orchestral instruments.

    • Mutis Mayfield

      Tme where become musician was something elitish had gone…
      Soon you could put your expensive bass inside a 3d cloner/printer and take 1 for the right price and even make some customitzation and improvements.

      In other hand I could understand the price of a tool which is being this… A tool. I will never going to buy an electric screwdriver for putting off 4 screws… If someone spent 10 years in a music career then 14000 $ bass instrument is probably the fruit/tool of/for these career…
      The point is: An ipad with alesis/behringer dock plus a keyweighted keyboard and some apps (like ik or alchemy) could do the same job as a clavia or kronos but these stagr units are more expensive and limited. Stability? Don’t lie yourself, inside them there are a computer (sometimes with bugs which the brand didn’t repair like alesis fusion).

      For analogic gear we have another discussion over the table but… Is aleph an analog bass? I lost myself in some point…

    • chompy

      You won’t be able to make an upright bass with additive 3-D printing anytime soon, if ever. More likely would be a robot that did the carving and shaping of the wood and the assembly. But it would just be a wank, at that point it’s easier and cheaper to just have a human do it.

    • Mutis Mayfield

      Why additive?
      First human:
      http://www.instructables.com/id/How-To-Make-A-Plywood-Upright-Bass/
      Then man with cnc machine:
      http://www.technocnc.com/cnc-router-customers/cnc-router-application-articles/Modern-Luthier.htm

      Soon or later 3D cnc printing…
      If you want to discuss it with another musicians or so…
      http://www.talkbass.com/forum/f3/cnc-bass-lutherie-619646/
      ;)

    • Greg Lőrincz

      They already made 3D printed Stradivari.

    • gli

      +1
      samplr IS a great app

      i love my ipad and realize the hardware will improve with newer iterations

      but can the ipad + mac mini + iconnect4 combination connect my devices if i have no programming knowledge?

      according to the description
      aleph would let my ipad, simmons drum module, monome 64, shnth, and tx7 “talk” to each other

      i honestly dont know how to get my laptop to do that

      there might be a way but i am unaware of it

    • Mutis Mayfield

      Lingon + automator.
      I remeber an early article in cdm about scratching setup with mac mini and without screen… But again you could run touchosc/lemur in your ipad or vnc desktop (splashtop) to tweak your mini realtime from the ipad… And even you pay yourself a course in how to do all of this before reach the cost of one aleph!

      In otjer hand if they came with a “noob box” with maxmsp inside and any kind of frontend for easy use… Why not go the apple mainstage way?

      Sorry, I respect your opinion but I still think is expensive and overpriced.

    • gli

      i’m googling those right now

      for clarity…i can use the monome grid and shnth and ipad somehow with em?

      i’m not just talking about audio

      i’d like to send data between them

    • Mutis Mayfield

      http://www.iconnectivity.com/iConnectMIDI4plus

      I will link you some other solutions for the other gear listed below in the other post.

    • gli

      if i only needed midi data sent i think i could agree

      check this out http://shbobo.net

      aleph works will work with that device and that is one of the many surprises that i was not anticipating when i heard that there is a new monome coming

      unless i’m mistaken iconnectivity would not let me use the data coming out of the shnth

      nor would i be able to send and receive data from my monome 64

      that might not matter to anyone else but it matters to me

    • http://nutation.net/ rseymour

      When Ezra B plays live, he was generally using a Mac Mini, a small monitor and some supercollider patches. This is next level stuff, imho.

    • leolodreamland

      hospitals wtf?

    • heinrichz

      It’s mainly gear for rich kids that want to appear hip an tech savy, same, was the case with monome. I really find this kind of hardware elitism rather cheesy and having a negative effect on democratizing music making, when people can really make more music with generic laptops and software.

    • gli

      every artist chooses their instruments & tools

      those choices rarely have to do with money…its purely a matter of personal taste

      what you find beautiful and how you’d like to control and manipulate audio is different from me

      not better
      not worse

      aesthetics and open software aside, i buy from monome because they’re local and modern

      i could get instruments from across the globe or fill my studio with vintage equipment that might make similar sounds

      but i’d rather not

    • active

      please keep in mind that for some people, music equipment is a business investment, not merely a toy. democratizing music creation/production does not mean the elimination of high end market products. instead, it means opening up options to all price points. an API console is a wonderful example. yes, a legacy model is close to 500,000, however the company has also opened up cheaper, modular components of that console. other companies have decided to stay fully in the high end market. does that make their product better or worse? BURL might only sell 5-10 motherships a year, and even their 2 channel DAC/ADC are 2,500.00 each, but that makes them neither cheesy or create a negative effect on the creation of sound.

      we would have a very boring sonic world if everything was made by m-audio/avid.

  • Elder

    I’m hoping they can work with Cycling74 so I can turn my Max patch into a hardware instrument.

    • Miquel Dangla

      Considering that most existing monome apps are written in Max, this is quite likely :)

    • Mutis Mayfield

      If it will be possible, then the best patch to port will be ms.pinky…

    • http://brunoafonso.com Bruno Afonso

      This ability would make the 1400$us price tag appear more justified.

  • experimentaldog

    It looks interesting, but here’s what I don’t really understand. Isn’t this still a computer in a box? A laptop costs this much or less and you can get peripherals to do the same things, write patches in Max, PD,Csound, SuperCollider etc. Laptops are much more resilient than they had been, so I don’t really get the statement about software at the mercy of the OS. If I used one of these and then used a laptop or a small sampler/sequencer to do the same things i.e. make the same music, could I tell the difference between them “acoustically”? Does it matter that the sound would possibly be indiscernible from all processes or is it more a matter of live performance issues? That said, if it’s more portable, wouldn’t you still have to truck around a bunch of gear in order to interface with the USB and CV ports? These are just questions. I have had some of these same reservations with the OP-1 and its moderately hefty price. I bought a monome 64, but that seems a bit more open, affordable and flexible. If this was in the $600 range, I might get one of these an MS-20 mini and a Volca Keys. But now I think I’m fetitshizing too much. Do I need more gear or can I make the most of what I have to make the same music I would with this?

    • beardo manboy IDM postmaster

      all hardware is wank, if you can’t do it in audiomulch or renoise then it’s not worth doing

    • DPrty

      Can you tell the difference between them “acoustically … the answer I think is yes. Most modern keyboard workstations are a computers .. can you tell the difference between the Korg sound and lets say Roland? I can. The Korg M3 Oasys and Kronos are linux systems and they have a very distinct sound. I can tell the difference between recording software also … Cubase sounds different from Ableton. Sure its all zero’s and ones but it really depends on how the software DSP chain is programmed, that’s the magic sauce.

    • assbreath cornbeans

      imagine living your life being wrong about everything

    • DPrty

      Imagine having your screen name assbreath cornbeans.

      assbreath cornbeans

    • minus

      Digital is a symbolic representation of data :
      identical algorithm (whatever the hardware and programming environment are) + same converter = same sound, don’t be silly…

    • DPrty

      Yes .. But what you do with the sound before the converter defines the instruments sound. If you know what the code is you can recreate the sound but that doesn’t mean the code has any less of a signature sound then any other form of instrument.

    • Samuele Cornell

      In theory yes the code affect the sound , but in reality the differences ( if the instrument/program is well made and aimed to accurate reproduction obviously ) are negligible and the difference could be seen only with a signal analyzer . I don’t believe you can hear the difference from a signal run into ableton and one on Cubase as i don’t believe people that buy 1000 $ audiophile gold cables and then pretend to hear a difference .
      The difference is there , sure , but it is so subtle that i strongly doubt the human hear can’t hear it .

    • eon1

      *yawn*

      null test.

      what is this, gearslutz?

    • Samuele Cornell

      the null test is an expedient that let you hear the difference ( and most of the time you can’t say where this difference come from , hardware ? , software ? )
      what i meant is : can you tell the difference without any expedient , just straight out sound ?

      i can go back to gearslutz now .

    • active

      fun fact: the human brain is capable of remembering a sound for roughly 1.17 seconds. after, we merely remember the emotion we experienced with said sound.

    • heinrichz

      i would like to challenge you on some of that with some blind tests

    • Andy Cartridge

      it DOES look interesting i give it that, i’ll be curious to find out more about what it does with some practical demonstrations i can see

    • experimentaldog

      I guess it would be good to clarify what I meant by sounding the same “acoustically”. I wasn’t referring to sound synthesis and whether or not this machine sounds the same or different to other hardware, that’s a different discussion. Also I wasn’t comparing the sound of DAWs either, the DAW is a totally different kettle of fish and arguably is related to you AD/DA conversion / speaker output setup. I was referring to the method of making the “music”. As this box is capable of achieving similar things that a laptop does, what would be the difference between hearing instruments sequenced and played live by this machine as opposed to a laptop doing the same thing? Also at a higher level almost all samplers can achieve the same quality, Akai and Emu were just doing the same thing at some point, sound in / sound out. A sampled sine wave as identical as the other. Later computers with decent DA/AD conversion took their spot. Why is this any different if the live “music” being made is done with the same tools and with samples of the same quality? The music being performed live seems to be achievable with what is already available, or what has been deemed obsolete (i.e. hardware sequencers), and you still need the outboard gear (controllers/synths) that you would with this machine. If it’s not being used live (i.e. aleph + synths into P.A.), how would you discern that the the sound was made by this system and not a laptop based one? How would a recorded performance of the aleph be discernible from another system that is doing the exact same thing? Don’t get me wrong, I’m quite interested in this box and device computers in general. I think we might see an Ableton Push or the like with a computer in it at sometime. It’s early days, this box is probably independently made in the US, therefore the price is steep unlike outsourced mass produced hardware. I just don’t know if I’d be able to justify another purchase to achieve the things that I can with what I already have. It is always tempting, but I keep thinking of the sustainability ethics of buying more gear to do the same…

    • gli

      excellent excellent post

      you raise several great questions and to your main one i’d answer no, i dont think you’d readily be able to tell the difference acoustically

      its an idea i’ve grappled with for a while because there are a million ways to create the same sound. with such wide open possibilities, you can choose which gesture is most meaningful or gratifying and, since the result is identical, the audience may never know (or care).

      so this all boils down to one question: do we feel like using the gestures aleph requires/allows or would we rather use the gestures allowed by other systems?

  • Sean Costello

    Why Blackfin? These are fast DSPs, but their native data format is 16 bit ints. For pretty much any audio purposes, you need to program these in double precision. Is there a decent free toolchain for the Blackfin nowadays, or are the Analog Devices compilers still needed?

    • chompy

      the Blackfin is also in the OP-1 so it seems like at least one other boutique vendor has figured out how to make a synth out of it.

    • Sean Costello

      No doubt that you can make stuff out of it, given enough time and resources. In general, though, the Blackfin is a good DSP for companies where the time spent programming it isn’t as important as the low cost of the chip itself. There is some breakpoint at which low silicon price + higher volumes offsets higher development costs. For an “open” platform, it seems like a really difficult DSP to program for, compared to an ARM or the like.

    • Mutis Mayfield

      Agree. If you are go to invest time in develop… Why not improving android to make it better for audio applications?
      A lot of people will get lucky qnd your fanbase will be big since day one… But I supose it is not so cool…

    • chompleston

      What else is there, though? Aside from general purpose chips like ARM as you mention, I don’t know of any custom DSP environments that are as accessible to music devices as Blackfin.

    • Sean Costello

      Why not an ARM? Even some of the cheaper ARMs can run floating point C. Slightly higher end ARMs have NEON SIMD, which is very close to the SSE2 SIMD used by Intel and AMD CPUs. Compilers can usually optimize for this SIMD, which means that your standard floating point C/C++ audio code will run.

      Compare this to the Blackfin, where standard C code will NOT be efficient (I programmed these for a few years while working at Analog Devices). If you use intrinsic operations, you can get very efficient code on the Blackfin, but I’m not sure if these intrinsics are available in gcc for the Blackfin. You would still have to “think” like the Blackfin, which means always worrying about what fractional data format you are using, single precision versus double precision, 16×16 bit multiples versus 16×32 versus 32×31 versus 32×32, that sort of stuff.

      Don’t get me wrong. The Blackfin is fast, and has great performance for a low cost DSP. However, it is obtuse. I don’t know if “obtuse” is the best choice for an open platform.

    • chompy

      Your last paragraph raises an interesting point. I would guess that one reason the TE guys used the Blackfin for the OP-1 is that it has security features making it next to impossible to dump the OS and reverse engineer it. If this aleph device is an “open system” those features wouldn’t seem to matter. As to why not ARM, I’m not sure. Blackfin is used in many audio devices, from hobbyist projects to mass produced stuff from Roland, Onkyo, Korg and the Roku boxes. It would be interesting to learn “why Blackfin” from the aleph engineers. Maybe Peter could ask them about it in a future article?

    • Sean Costello

      “Why Blackfin?” would be a great question for a future article.

      The ARM is still somewhat of an unproven entity in MI signal processing. The recent Audio Damage modules use a small ARM (STMF4), and sound pretty great, but I would love to see a comparison of ARMs at a certain clockrate versus the DSPs in this range.

      What Roland/Korg boxes use the Blackfin, BTW? I know that a few Electro-Harmonix pedals use the Blackfin.

    • zebra

      sorry for not expanding this thread earlier. needless to say, i disagree with a lot of your statements, but perhaps its not the best time or place. here’s something: most major audio projects i’ve worked on have defined a little macro language for DSP:
      instead of (x * y) , MUL(x, y) … and suddenly,. who cares what the intrinsics look like.

      thanks for the interest! -eb

    • just passing

      Maybe it’s simply a question of power consumption? The kind of volume a $1400 hardware box is going to sell in doesn’t really make component cost a limiting factor, but given that the OP-1 uses the same DSP, maybe it’s simply the most power-conservative way to get a certain MMACS count?

    • zebra

      please check your facts. its 32-bit with packed 16-bit intrinsics. gcc yes.

    • Sean Costello

      I worked at Analog Devices from 2001 to 2006, developing audio applications for the SHARC and Blackfin (VisualAudio, as well as lots of one-off code blocks). Within ADI, the Blackfin was always considered to be a 16-bit processor, at least with regards to signal processing. It has enough speed to comfortably work in double precision, but it doesn’t have a hardware 32×32 bit multiplier. So, consider those facts checked!

      I agree that the Blackfin is more OSS-friendly than the 56K series, but it isn’t friendly to people coming from a straight C/C++ world, IMO. It is good to know that the gcc toolsets work for the Blackfin.

      BTW, if you can dig up the VisualAudio modules, there was some good stuff in there (including a granular pitch shifter I worked on and was fond of). I’d rather see this stuff used in an open source project, versus being relegated to the warehouse at the end of Raiders of the Lost Ark, which is where all the VisualAudio code seems to have been filed.

    • zebra

      the data registers in bf 5xx and 6xx are 32 bits wide. i don’t know what else you could possibly mean. perhaps you were using a differenet chip family since this architecture was created in 2009.

      sorry, i just think your original post is pretty misleading.

    • zebra

      er sorry maybe 2008. anyways here is the architecture spec:
      http://www.analog.com/static/imported-files/processor_manuals/blackfin_pgr.ref.man.rev1.3.pdf
      are you referring to this project?
      http://www.aes.org/e-lib/browse.cfm?elib=13344
      i know its authors very well.

    • zebra

      anyways, thanks for bringing it back to my attention, that was a good project with a lot of useful ideas.

    • zebra

      and unfortunately i think you’re right about its source code being permanently locked away in a dusty vault.

    • Sean Costello

      Here’s a document I found about writing VisualAudio modules, which gives a decent overview of how things worked at a lower level:

      http://www.analog.com/static/imported-files/tech_docs/DevelopingVisualAudioModules.pdf

      Ignore all the .xml metamodule stuff, which was for the visual editor. I wish I could ignore the hilariously bad stock photos used for the “cover” of the PDF. Making fun of those photos gave us a bit of joy before we got sh*tcanned.

    • zebra

      very interesting, thanks

    • Sean Costello

      The BF533 dates to 2005 at least. I know I was working with the BF533 EZ-KIT for a few years before 2006.

      Yes, the AES paper discusses the VisualAudio project I was referring to. I worked with David, Tim and Paul on the VisualAudio modules (Scott was the head of ARTC and wasn’t programming at that time, and Britt worked on the visual editor). The entire department was axed at the end of 2006. I worked with Paul as a consultant through mid-2008, before starting on my own plugin business.

      As far as the VisualAudio source code, I think that Analog Devices removed it from the website, destroyed all backup copies, and salted the earth under the former ARTC division. I honestly don’t know why, as there was some good audio code in there (there were some issues with the editor, but the audio code could be used separately from the editor).

      Don’t get me wrong – the Blackfin is a blazing fast DSP for audio. I just consider it hard to work with, compared with processors that can run floating point C/C++. The mindset required to work with fixed point fractional DSP is far different than floating point C/C++, and there is a far bigger base of developers that has experience with the latter, as “floating point C/C++” describes Csound, Pd, Max/MSP, Supercollider, and most of the VSTs and AUs out there. Something like an ARM seems like a much more friendly platform for these programmers.

      On the other hand, I have worked with Pure Data running on an embedded ARM system before (BeagleBoard), and it started choking pretty quickly on a patch I consider pretty lightweight (i.e. around 1 to 2% CPU on my MacBook Pro). So it may be the case that the clock speed of the ARM doesn’t really translate to real-word DSP performance, and that OS, cache, or memory access issues might be the big bottleneck.

    • zebra

      that is all very interesting! well i know jaffe and stilson from their CCRMA affiliations, especially david. they are both superlative engineers for whom i have deep respect.

      anyways, look: you can say that the bf53x is “primarily” a 16-bit machine, i suppose. maybe “internally” at ADI this is done, of course i couldn’t say. but as an engineer on the outside, i see that it has 32-bit data registers, a pair of 40-bit accumulators, hardware ALUs for 32-bit operands (yes indeed, it really does, 40-bit too, check the ALU block diagram), and fast fract32 intrinsics, and i say, it looks and sounds and acts like a 32-bit duck. the fact that the part *also* functions as a 16-bit processor at double speed (with caveats, because 16-bit operands need to be packed into 32 bit words for advantage! ), and indeed as a sort of vectorized 8-bit processor if you want, is icing to me.

      i do apologize for the curtness of my earlier comments, but i *still* think its terribly misleading to the casual reader to say, why on earth did they use a 16-bit processor? – the part is very clearly intended for 32-bit fixed-point audio applications as much as anything else.

      as far as obtuseness, i guess i can see that, but not much. blackfin ASM is a beautiful instruction set, IMO. and as i said before, i don’t think calling intrinsics for arithmetic is a big deal. for example, SuperCollider plugins use macros for arithmetic on signals, and those macros expand differently on different architectures. i’m pretty sure that under the hood it is the same for Reaktor, MSP and probably most other major digital audio projects.

      as for ARM, well you have pointed out some of the reasons already. (i too have not found the beagleboard etc. terribly useful for real musical applications.) i’ll additionally point out that in the aleph we deliberately separated the controller/host and the DSP into two physical processors. from there, the reasons behind our part-selection decisions get kind of involved (and sure, maybe idiosyncratic!) – definitely a topic for a different discussion… but one nice result of the separation, for example, is that we don’t need the same kind of deep multithreading on the blackfin that you see in the VisualAudio framework.

      as for floating-point, it is absolutely true that people will be put off by its lack. but i don’t need to tell you there are advantages. most obviously, there is speed/cost. additionally: the extra few bits of precision and the built-in saturation means that certain topologies are actually more stable. denormalization is a non-issue. in any case, i have had no trouble developing algorithms first in standard C with floats, then replacing floats with fracts (and a few workarounds), on linux, then running the same code on the blackfin. because we are such a very small team (in fact so far its just me on this side of things) i haven’t done the kind of intensive ASM work that surely went into the op-1, for example. but as you say, c with instrinsics is quite fast.

      you probably also know that ADI has published a quite fast but non-IEEE floating-point implementation for blackfin, so that functionality is there at a less-than-prohibitive cost, if you truly need it.

      in any case, we don’t expect a lot of users to actually want do DSP programming on the aleph (though i certainly hope for at least one or two!); we have a substantial library of DSP building blocks and full-fledged firmwares, and control-side customization is far more accesible and immediately useful as far as “rolling your own” performance machine. i hope to have time once everything is released, to go back and really immerse myself in ASM development to increase efficiency and extend the depth of possibilities.

      and to wrap things up i will say – this really is designed to be an instrument. it is less intended to replace your laptop as a central sound processor, than to allow the types of processing and the depth/flexibility of control currently only available on laptops, to be available in other setups as well. as a touring musician who uses acoustic instruments with idiosyncratic processing, i have wanted something like this for a long time – a small rugged box with builtin preamps and lots of I/O, that turns on and instantly starts doing what i have told it to do. we are putting every possible effort into making it easily adaptable to many specific purposes within that general use-case.

      anyways, thanks for the discussion! maybe we can get more in-depth in the CDM interview, or over email (i’m easy to find.)
      -ezra b

    • Sean Costello

      I think the term 16-bit was used for the BF53x, as it can perform 16×16 bit multiplications at the full clock rate. Once you start calculating higher precision multiplies, your performance goes down to 1/2 or less for vectorized code. You also can’t use SIMD for 32 bit precision multiplications, IIRC.

      On the plus side, you can mix and match the multiplication precision as needed in your code. For example, multitap delays work great with 32×16 bit multiplies, where the 16 bit number is the tap coefficient. The older Lexicons had 16×6 bit multipliers (or even less precision in some cases) and sounded great. 32×16 might work fine for various FIR architectures. 32×31 bit multiplies are almost indistinguishable from 32×32 bit multiplies, and can be pipelined to be VERY efficient on the Blackfin, compared to the full-on 32×32 bit multiply.

      As far as the macro language you mentioned earlier, if you supplied a few examples of this for the aleph, people could probably run with it. I know that Pure Data has a version designed for fixed point processors, so some of this DSP code might run on the Blackfin without much of a problem.

      Anyway, I will be following this project with interest. I’ve been in the native DSP world for a long time now, but I like to keep up to date with the embedded stuff.

    • zebra

      ok, so we’ve agreed that this is a pretty silly semantic distinction, no? bf53x in fact supports two 32-bit reads and two 32-bit writes on each clock cycle, ALUs are up to 40 bits, addresses are 32-bits and it can access a 4GB address space. 16- and 8-bit operations can be SIMD, which of course is nice… but it also has “double precision” intrinsics for 64-bit operands. it behaves exactly like a 32-bit core, or like a 16-bit core that is twice as fast.

      when you say “16-bit processor” most anyone would take it to mean a processor with a 16-bit data bus, like motorola 68k or 8086, even if such a processor supports 32-bit operations at some cost. and this summons a very different picture to mind, specifically a picture of the late 80′s and early 90′s era of microprocessors. its simply not a fair casual description of the blackfin family. really, i can say nothing more about that.

      indeed, it is super-cool that BF can mix operand depths for speed! and that is a great point about pipelining 32×31, thanks

    • zebra

      anyways the hardware multiply register is actually 40 bits wide, if thats what you’re talking about.

    • zebra

      someone deleted my reply? well anyways. this is incorrect. the blackfin is a 32-bit processor. it also has intrinsics for fast processing of packed 16-bit fracts if that’s what you want. there is a venerable free cross-compiler for baremetal ( bfin-elf-gcc ) and two for the widely-used blackfin linux (bfin-linux-gcc , bfin-uclibc-gcc ). of course this part is eminently suitable for audio applications, and it is far more OSS-friendly than comparable platforms from freescale or motorola. hope this answers your question. – eb

  • Charlie Lesoine

    Most musicians are not programmers. We want instruments. Not multipurpose boxes.

    • lokey

      speak for yourself.

    • Charlie Lesoine

      I’m speaking the truth. “We” as in most musicians would rather have a musical instrument rather than a box with buttons and knobs. Otherwise we wouldn’t be musicians.

    • lokey

      oh please. Speak for yourself. You dont get to define what constitutes a musician.

    • Charlie Lesoine

      I think you are missing my point. I’m not trying to define what constitute a musician. I’m just saying that the majority of musicians would rather play a real musical instruments instead of “a box that does anything you want”. Be they guitars, pianos, MPCs or 303s. These are what I consider real musical instruments. They are designed with specific musical purpose(s) in mind, and the musicians who play them far outnumber the musicians who would rather play a monome/launchpad/controller/etc.

    • lokey

      i hear what you’re saying about your personal opinion about what constitutes a ‘real musical instrument’, im merely laughing at the idea that you believe you represent ‘the majority of musicians’. Off the high horse.

    • Charlie Lesoine

      It’s all good. I do need to get off my high horse sometimes. This thing actually looks like a pretty cool replacement for a laptop+audio interface.

    • gli

      i’m ok with you representing the majority of musicians. i’m also ok with being excluded from being called a musician at all.

      i’m not a programmer either. But according to your description of “real musical instruments”…there are currently 13 instruments lying around my room that aleph will allow me to connect and use to make & warp sound. I was unable to connect them before today.

      I’m always looking for new gear so if you can find a more elegant solution to serve as the hub for my (admittedly) crazy collection of tools I’m all ears

      right now aleph is perfect for what i need to do

      peace

    • Mutis Mayfield

      Out there are solutions for connecting anything more or less… What are these unconectable devices? I’m intrigued and want to help.
      ;)

    • gli

      i’m someone who primarily deals with audio and recently have been exploring whats possible with midi and cv. i tend not to use tools exactly how others do but here’s an idea of what i usually have around:

      various acoustic drums & percussion (wood, metal, and clay)
      roland sp 404
      monome 64
      arduino w/ repatcher shield
      ds lite
      yamaha tx7
      simmons 800
      shbobo shnth
      moogerfooger lowpass + ring mod
      crumar performer
      iphone
      ipad
      cracklebox (kraakdoos)
      mixtape alpha
      koma kommander
      kmi qunexus

      also a variety of iOS and OSX applications but those are the physical objects i can remember

    • Mutis Mayfield
    • gli

      thanks

      de donde eres? españa?

      anyway
      i checked all these links and some of those i have already (like touchosc)

      the ones that are REALLY interesting are eurorack like the befaco stuff. the other solutions appear to only translate midi in various ways

      just a quirk but i currently have no interest in eurorack format

      i may dive into that realm in the future but until then i’m “stuck” in the world of standalone tabletop synthesis

      (truthfully i’m not engaging in this dialogue to convince you or change your mind…i respect your opinion and maybe we’ll just have to agree to disagree)

      my desire is not just to connect the pieces in my collection but to do so in a very particular way that is flexible and dynamic

      unfortunately none the solutions you posted allow me to use my gear in the manner that i need to

    • Mutis Mayfield

      Soy de Barcelona compañero!

      I wasn’t tried to convince you, only help and fill my POV with arguments, not only arguing against GAS (hehe)

      I was looking carefully allnyour stuff including shbobo and I believe iconnectivity 4+ host usb and even ethernet unit (usefulnfor your monome) but maybe I’m wrong and Aleph reminds unique in its class (and still justify its price for special setups like yours).

      In other hand if you (or someone) wants info about Befaco I will be very glad to help you.

    • eshefer

      this is not a tool for “most musicians”, whatever that means.

    • Axel Rigaud

      “These are what I consider “real” musical instruments” Sounds familiar…

    • Mutis Mayfield

      Most musicians grow with the necessity of tweak, bend and explore or improve they tool (even when they were not multiporpose boxes) and that’s the best point of Aleph but I agree it has a price of not multipurpose box…

      Maybe this is the point for all the discussion. With the actual gear, price and context (we are not talking about mathlab medical sound experiments, isn’t?) this unit fits in the “not multipurpose box” by its price and in the “multipurpose experimental box” by its concept…

      Monome was a great idea and comprensible price for its age… I can’t see the correlation here, sorry…

    • gli

      not sure what you mean…

      are you saying its affordable? it does more than similarly priced “not multipurpose boxes”?

      price is irrelevant unless you actually see it as useful and plan on buying one. monome pricing has been discussed to death already

      it has nothing to do with age, or what was/is normal for other companies to charge

    • Mutis Mayfield

      When monome was released it was the first of its own then the price was “expensive but comprensive”. When oth grid controllers come then monome become expensive without excuse (and they missed the rgb improvement due to “minimalism concept” (check the forum site for arduinome clones and rgb post and you will understand why I’m arguing))
      In this new one Aleph, I still not understand what make it so useful to raise 1400$ but maybe it is not about usefulness rather than coolness…

    • gli

      mabye i need a refresher course about their methods

      i dont understand what the monome having been in production for many years has to do with the price. its cool if you think it doesnt do enough to warrant the amount of money they charge. but they have been clear: we are paying for hand assembled, locally sourced parts.

      they are are a “company” of four hands working together in a barn.

      if i knew enough about code and electronics i would build my own devices. until i develop those skills monome is a pretty good deal actually

    • Mutis Mayfield

      It is about cost and price, offer and demand, use and specs.

      In the beggining they were the first and if you want something like it, it was the only source. Today we have more handmade or care-made solutions with better specs and better price, diy and comercial but they still maintain the price and user community support them for the whole concept. It is okay but not for everyone and today there are other options… Maybe they should put a “touch id” or “champagne faceplate” to keep it pricey… ;)

    • gli

      i agree that there are some awesome and affordable grid based devices on the market today. i’m actually not sure i would have ever gotten a monome if these cheaper gadgets had existed 7 years ago

      however, i wouldnt look down on monome for the way they run their business

      dont get me wrong…i CERTAINLY wish they were cheaper

      but i cannot say that the price is unreasonable

      in the past it had nothing to do with the scarcity or the fact that they were, as you put it “the only source” for that type of product. the price was based on the process and parts

      and that is why the price hasnt gone down over the years: the process hasnt changed much (nor has the cost of buying local components and parts for the assembly of the monomes).

      that is also why the user community supports them

    • Mutis Mayfield

      Sure but they (including the community) could improve some specs without manufacturing in China and be a bit competitive/attractive.

      Most of these other units are midi native making your problem about linking them easy to solve than with a monome… Or directy control other midi stuff and be controlled by…

  • Andy Cartridge

    i watched the video.. and well, what can it do that something like an ipad,an audio interface and a few controllers can’t?
    when did the idea of ten inch multitouch screens become outdated so we desired a box with a few knows and led ok OLED screen?
    it just looks like something from the past..

    • Scott

      I don’t think that it can really do anything an ipad, an audio interface, and a few controllers can’t. Although, it’s kind of taking the role of two of those three items with some of the 3rd. Though, in a world where options as a musician are more or less limitless, isn’t it nice to have some constraints? Some of the most interesting and sought after instruments/devices (old MPCs, analog synths and drum machines) are those with somewhat limiting interfaces. I’m not saying there’s nowhere to improve when it comes to interfacing with sound, but it seems like it would be nice to have something that does its job and it does it well.

    • Mutis Mayfield

      For 1400 $ and you still must to patch it in some way?

  • Thopa

    I really like the idea. Spot on. If its true that you can save patches, settings of the modular synthesizer, its a seller. Very buchla-ish :) Also being able to plug the monome is amazing too. Id like to find out if you could load monome patches on it. Using those cv outs to sequence the modular from the monome could be pretty awesome.
    Probably not every one’s cup of tea, BUT there is many people that can see a lot of potential in it.

    • gli

      to your question about monome patches: no

      the current stable of max/msp patches that we use will not be natively support on aleph

      however, the functionality of some apps is not only possible but currently in development (tehn is working on a new version of the popular looper MLR)

      porting a patch directly to aleph is doable but i’d rather people explore the unique abilities of the new system

      like the example you described

  • Dan Browne

    I love it and quite frankly expected this to come out sooner from someone in the Monome community. It’s too expensive and I’m priced out, however I am the target demographic for this device. I respect and agree with the criticism, but this is a useful bridge considering the type of gear i compose with. It could prove interesting to cut the cord to the billion plug-ins and stereo tracks that we’re all used to, and require a different compositional approach. Each instance of it’s minimality that I could complain about, makes me think of exciting ideas that might not have materialized if I were in my “extremely flexible studio setup”. (which I love and am not discarding) If Brian Crabtree wants to ship one to me to inspect, I’ll convince a few kids on the internet to sell their Macbook Pros ;) Well done.

    • gli

      well said

      hopefully you can get your hands on one eventually ;)

  • DPrty

    I find this very interesting. I have been working on a granular synth that is purpose built to run on an embedded system. I am using Flowstone by DSP Robotics http://www.dsprobotics.com/ I look forward to testing the Aleph.

  • M

    I’m watching this space very closely. Having worked with hardware and software for years, I’m ready for flexible, dedicated micro-controllers and this is a step in that direction. If folks aren’t grasping the novelty, it’s probably due to an antique perspective on hard vs soft. This is both.

  • syntheticjuice

    Will it be possible to make/use a looper that streams directly to and from the sd card, so ur looper lengths are only restricted by the size of the sd card?

    • gli

      good question

      to my knowledge yes

      i will double-check but i’m pretty sure ezra commented on this when someone else asked

    • pat

      why not just get an octatrack?

  • Greg Lőrincz

    Why do dudes who demo synths ALWAYS have tattoos?

  • hellojeffreyjames

    This looks like something I’ll never be smart/patient enough to use.