back to article Open-source hardware group puts out vid system-on-a-chip

A radical tech coalition has produced an open-sourced music visualiser that modifies input video rather than generating patterns - and interestingly the box includes a system-on-a-chip that could one day compete with ARM. If just 80 units sell, the "Milkymist" will have already paid its way. Backer Qi is the company …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Anonymous Coward

    How?

    Did they manage to design AND fab a SoC for less that $40k? That sounds a bit impossible to me. Most SoC devices cost 100 of K to tape out and fab, never mind the cost of design tools.

    1. Anonymous Coward
      Anonymous Coward

      The project is using a FPGA as a central SoC, so fab cost can be skip for now i think

  2. Anonymous Coward
    Thumb Up

    Nice!

    Want one.

  3. Phil Endecott

    FPGA

    The "system on chip" is an FPGA.

  4. Anonymous Coward
    Anonymous Coward

    More pooh-poohing...

    ...from the apparent champion. In fact, aside from patent litigation (mentioned in the piece about the NanoNote), there are several other significant challenges when it comes to realising open hardware or even custom hardware using whichever chipsets you can get your hands on.

    You've got the problem of scale - commit to hundreds of thousands of units or get the brush off from whichever manufacturer is going to put your kit together - which in turn is partly caused by matters of funding and risk management that wouldn't trouble corporations (which is precisely why they incorporated, of course). You've also got the problem of applicable experience and expertise, where people may have a background in one area, but it takes knowledge in several areas to bring a product to market. Again, if you aren't an established organisation with a commitment to continuous product development, you're going to be struggling.

    A bunch of this has nothing to do with whether the hardware is open, as I already pointed out, but if it is then there are a bunch of additional pitfalls like the shitty attitude of chipset vendors towards documentation, let alone freely viewable documentation, even when the customer is itself a major corporation. Nokia spokespeople were always claiming that they weren't important enough to their chipset vendor to get documentation for the community (or even themselves on occasion), so I doubt that a small non-corporate project even gets their inquiries returned by the salesdrones.

    That a bunch of people working through the various issues in public, so that others may benefit from their experience and generally shed some light on how such stuff is put together, should be seen as an opportunity to make wisecracks about open hardware, just goes to show how moronic and docile the author expects even technology enthusiasts to be in the modern era. It may be amusing to the author to quip away (presumably on Twitter as often as not) while waiting for the next goods-drop from the fashionable brand of the day, but it's a sad attitude reflecting wider society's indifferent, impatient, spoilt consumer culture where having the shiny new stuff with the bigger numbers is more important than actually going out and doing something for yourself.

    The Britards may not be interested in anything "sciency" any more if it's not part of some "zany stunt" (where's that jet-powered car with the rocket strap-on these days?) but it's not exactly something to get all smug about. And it's that kind of tone I see seeping through in this article.

    1. James Hughes 1

      Bloodhound SSC

      Currently under construction. But then, you could probably have looked that up yourself....or were you too busy being smug in your own way?

      1. Anonymous Coward
        Anonymous Coward

        Re: Bloodhound SSC

        "Currently under construction. But then, you could probably have looked that up yourself....or were you too busy being smug in your own way?"

        Nope. Was more interested to see whether anyone remembered, or whether shinier things had displaced it from the public eye, necessitating more smoke and mirrors before/during/after, say, the London Olympics or any one of a number of "look here!" sideshows.

    2. Richard 81
      WTF?

      Huh?

      What wisecracks? What pooh-poohing? The author seemed pretty pro-this kind of thing.

      Also, watch what you say about 'Britards'* and sciency stuff. Some of us do just that.

      *Thanks for that by the way... *sigh*

      1. Anonymous Coward
        Anonymous Coward

        Re: Huh?

        "What wisecracks? What pooh-poohing? The author seemed pretty pro-this kind of thing."

        Yes and no. Yes, he didn't just slam the idea of people doing this kind of thing, unlike his remarks about small-scale personal networking from, erm, the same people. No, because the "technically-literate underemployed" is precisely the kind of pooh-poohing that was used against Linux (and still is by the "in your own bedroom!" morons who must therefore believe that IBM has a campus building that is all bedrooms somewhere), suggesting an attitude that independent efforts are by default worthy of derision whereas corporate efforts are by default worthy of praise.

        "Also, watch what you say about 'Britards'* and sciency stuff. Some of us do just that."

        Key point: not all Britons are Britards. Optimistic assessment: most Britons are not Britards. But many people must still learn to ignore the smoke and mirrors and dare to question their society's failings.

    3. Intractable Potsherd

      @ pooh-poohing AC ...

      ... you clearly read a different article than I did. I detected no sign of any diminishing of the importance of what has been done here*. Overall, it is a good article.

      * I did wonder if there was going to be any analysis of whether there is actually a market for this thing (swirly patterns on walls in pubs - really?), but I suppose there could be 80 people in the world with $500 to throw at this thing.

  5. Uwe Dippel
    Pint

    Then go and buy one ....!

    As one of the proud owners of the nanonote, I'd be happy to try the MilkyMist just as well ... .

    Had a lot of fun with my nanonote and still have - partially. There is one major drawback, and - AFAICS - a blind spot on those who 'need' to sell 3000: If I had to have those numbers, I'd have known that the thingy needs to connect to a network, and cannot be just a piggy-back to another PC.

  6. EddieD

    Hah...

    I was about to say that Milkdrop did roughly the same for me, when I found that MilkyMist was inspired by it (and running winamp + milkdrop + win7 can hardly be called "low latency, but it does work...)

    1. Shades

      If you're running...

      ... a PC with enough oomph to run Win7 then you sure as hell should be able to get WinAmp+Milkdrop to do damn-near-as-realtime-as-makes-no-difference visuals! My XP Netbook has no problems DJing with video (thats two AVI video's playing at the same time!) with a USB hardware controller/soundcard... and of course, DJing with them it HAS to be realtime (well, a buffer of less than 25ms). I suggest you need to check your buffer settings.

      1. EddieD

        Playing digital audio files,no latency, no worries - and has been so for many iterations.

        Using the line-in from external audio is a tad more laggy though

  7. Vic

    Oh come on, Bill, at least read the page you're reporting on

    > Pretty visuals, no copyrighted hardware

    It *is* copyrighted. And released under a CC licence.

    Giving others the rights to copy your work does *not* mean abrogating all copyrights.

    Vic.

    1. Danny 14

      indeed

      else a big corporation can come along, nick it then copyright it.

  8. PyLETS
    Linux

    successful open source h/w projects

    I've heard of various open source hardware projects which are very successful in their own educational terms. Not neccessarily "successful" based on the cynical big-media terms of the author of this piece. Successful enough for very many interesting educational projects, but this area is about learning and fun, and is not currently about sales in the millions of items.

    A couple of examples:

    Ronja optical networing: http://ronja.twibright.com/

    Arduino computer controlled active electronic object prototyping platform: http://www.arduino.cc/

  9. Steve Rivers
    Stop

    Open Source hardware can never be open source in sense that software is. So please find another term for it, as all it does is confuse the people who have no clue about hardware design.

    Open source software works because the marginal cost of copying is negligible to nil. Yes copying the design is cheap, but copying hardware has always had, and always will have a non trivial cost.

    1. Anonymous Coward
      Stop

      Not quite

      Someone earlier commented that an FPGA was used.

      We are seeing that what one person calls "hardware" is really firmware and there's a real feeling now that there is a big crossover in the FPGA realm where what used to be implemented in custom fabs is now more feasible for the home hobbyist using technology like the FPGA.

      I think for a lot of embedded type systems where extreme high performance is not a pre-requisite, we will see an explosion in this kind of thing and I jolly well look forward to it.

    2. Anonymous Coward
      Anonymous Coward

      "Open source software works because the marginal cost of copying is negligible to nil."

      Open source is really all about the collaborative aspect of sharing and improving. Back in the day, even if you were paying a "nominal cost" to get access to the sources, it might well have been non-trivial, especially if a lot of people were paying it. Nevertheless, it didn't mean that no-one was participating in Free Software - the label more likely to be attached at that point. So, no, open source works because people see more value in collaborating to improve something than the cost required to enter into such collaboration (which admittedly is less than it was in some ways, but still significant in certain other respects).

      The costs can be especially problematic with hardware if you're outside the corporate supply and production chains and need access to the facilities that are typically only available within those chains, perhaps because you want to be competitive with off-the-shelf products. In short, there's a significant up-front cost in addition to per-unit costs, making bootstrapping efforts quite difficult (persuading punters to put down cash), complicating the scheduling (you might not be able to commit to sourcing, manufacturing without commitments from punters), and thus making production a risky affair (even if you manage to source stuff and get a shot at manufacturing, something might go wrong, leaving unhappy punters).

      I guess you could distinguish between open hardware design and manufacturing, but then that just illustrates that "open manufacturing" is the real stumbling block.

    3. A J Stiles
      FAIL

      Open Source hardware

      Open Source hardware has been going on at least since the early days of radio broadcasting; when world+cat would build crystal sets and one- or two-valve TRFs using whatever parts could be obtained, working to schematics published by the broadcasters.

      The wiring diagram is the Source Code, which you "compile" with a soldering iron.

      Yes, you had to pay for your own valve(s), tuning capacitor, headphones, the wire you used to wind the coils with, HT and LT batteries, resistors, capacitors and so on. What you *didn't* have to pay for was permission to build the set described by the published wiring diagram, or to re-draw it for someone else to work from.

  10. DF118

    The demo videos are pretty poor. Low latency maybe, but is it really that jerky?

    1. sb0

      (disclosure: I founded this project)

      It is not, and yes our demo videos could/should be a lot better. We have a new one in preparation (yeah, bad timing), but you'll have to trust my word until we get it out, or someone makes a better video.

      1. DF118
        Thumb Up

        Thumbs up to that!

  11. Richard 12 Silver badge

    I'd have more respect if they complied with the standards

    All published versions of the DMX 512 standard specifically state that 3-pin XLR connectors are never to be used.

    Given that they used the wrong connector, how likely is it that they did any other part of the ANSI E1.11 properly?

    Shame really, one would expect geeks to get that kind of detail right - it's not like it's secret!

    That said, kudos for getting a design together that breaks even at 80 units.

    1. sb0

      (disclosure: I founded this project)

      We chose the 3-pin connector on purpose, because no matter what ANSI says it is what is present on most DMX equipment (except maybe the high-end professional one) so it is convenient for the majority of our users.

      If you have the 5-pin connector on your equipment, you will be able to use it with the Milkymist One using a simple passive adapter.

      1. Richard 12 Silver badge

        I am surprised at that decision.

        There are very good reasons for the standard prohibiting that connector. It wasn't written by ANSI, it was originally written by the USITT, then updated by ESTA - all theatre technicians who really do go out and light things ranging from a folk singer in a pub up to the biggest stadiums. It has to work in that whole range.

        I'll acknowledge that you do see 3-pin XLRs on most of the sub $100 fixtures and controllers - the ones nobody expects to work very well or last very long, bought by the Disco Daves of the world.

        Disco Dave doesn't spend $500 on a single VJ gadget no matter how cool, he spends $500 on his entire lighting rig and controller because he's not getting paid enough to spend much more.

        The professionals use the 3-pin connector as a sign that the designer of the kit didn't read the DMX512 standard. Almost all of those have made very serious mistakes like using a framing error as 'Break' instead of timing it and the fixture can't cope above 15-20 packets-per-second.

        Which basically means that it falls over and flickers like crazy the moment it's in a poor environment and/or being controlled (or controlling) a device that it wasn't originally tested with.

        I really hope you didn't fall down that trap.

        The Milkymist One is the price of small-scale professional devices.

        It's only slightly cheaper than professional lighting consoles like the Zero 88 Jester, the same price as a Betapack3 and is much more expensive than the Alphapack2. All of which have 5-pin XLRs.

        That's really why I was so disappointed - the Milkymist One looks and is priced at the small-scale professional level, yet uses disco dave connectors - with the implication of disco dave quality.

        Sorry about that, but it's what it looks like.

        1. Dave Lawton
          FAIL

          @ Richard 12

          i take exception to that, 500 quid wouldn't even bought the Pular dimmer packs (driven by the Zero 88 Mercury) that powers my non-DMX kit, let alone the DMX stuff.

      2. Robin Szemeti
        Thumb Down

        No no ...

        You are selling to geeks here. Standards and compatibility are key.

        The correct answer was "the standard defines a 5 pin XLR, if you have a 3 pin connector on your equipment, you will be able to use MilkyMist one with a simple adaptor"

        Secondly, DMX 512 uses EIA-485, a balanced signalling protocol ... since you have chosen to implement it with an unbalanced connector, you can't really transform from 3 to 5 pin with a "simple passive adaptor" ... if you had used 5 pin, balanced signals you can go to 3 pin, unbalanced (and non standard) fairly easily, but you need an active adaptor to provide a standards compliant interface.

        Sorry, bad and wrong design decision. Follow the standards and then let users break them if they want to, not the other way around.

        1. sb0

          Both the 5-pin and 3-pin DMX connectors use balanced signaling. The two extra connections on the 5-pin plug are not used (see http://pinouts.ru/Home/dmx-512_pinout.shtml). So, yes, I maintain you can use a simple passive adapter to go to 5-pin if you need to.

          The DMX signal is generated by a custom FPGA core (not a hacked UART), which produces a correct break and not a "framing error". Feel free to check our Verilog source (online at Github) if you have the slightest doubt about this. If you can find bugs, we'd be happy to fix them and push out a FPGAs bitstream update.

  12. Anonymous Coward
    Anonymous Coward

    Open hardware milky goodness

    Seems to me that this kind of platform needs to be reused by a number of end application open source projects.

    It has video in / out, audio, USB host, Ethernet, etc... Displaying milkdrop style visuals is probably the least interesting project you could do with this!

    How about porting/starting the following projects to it;

    - atari / amiga / whatever old hw revival project clone (maybe minimig)

    - open source games console (on top of a Linux port)

    - open pvr / stb / etc.

    Getting more interest in the shared hw by expanding its use can only bring the cost down :-)

    Before anyone says that this can't be done, it has a strapping great FPGA, so I think it probably can!

    I want one!

  13. Anonymous Coward
    Anonymous Coward

    I think they have a winner...

    ... provided it works at least as well as advertised and they manage to reach out to enough pubs and clubs to shift a couple hundred. But yes, this is indeed a large step for hardware design even if this box itself doesn't succeed, but at those numbers it's hard to see how it couldn't. Sure it's a spendy appliance, but destined for being a line item on the things-to-write-off list at a high-margin fashion-sensitive venue. I'd expect there's a few more than 80 of those around. In fact, with a veejay not being free either, it might turn out to be entirely justifyable for far more than that.

    As to the nanonote, I'm tech-literate, have plenty of time on my hands, but no money. So I can't really justify spending on toys, even moreso after the last such purchase (a 2nd hand illiad) turned out to be mostly a dud and a right pain to use--though the concept, had it been usefully executed, would've been exactly what I needed. So they won't be selling me one of those 800 left. But now that they get a bit of spotlight time with this thing, I'd guess they'll shift a few to others anyhow.

  14. Robin Szemeti
    FAIL

    Marketing fail

    Massive geek and tech-head marketing fail.

    I failed to see the unique features demonstrated, either the graphical camera + patterns feature, or the "its open source" angle. No snazzy video content of the cobination of live vid and patterns, no access to anyting like the source.

    MilkyMist seems to be nothing more than the usual Windows/whatever music to pattern algoritm, in a box. The alternative would seem to be a laptop with a VGA output, which will allow you to play videos. have a GUI, store stuff etc as well. The only vaguely unique thing about it seems to be that it can (apparrently) mix up incoming video into the display ... if it can, they failed to demo it in there videos. I watched each one hoping to catch a glimpse of

    The one unique selling point, and they fail to even show it in action ... lame.

    You are trying to persuade people to not use what they would probably use (a laptop) and take a bit of a leap of faith ... which they may do IF the piece of cheese was tasty enough ... and you don't even show the unique thing it can do??? FAIL.

    A classic example of why engineers shoudl be kept well away from marketing anything!

    In desperation, and following up the nerdy "han I hack on it" angle, I followed the "developer" link on their page ... its supposed to be "open source" so one of the kewl things would be getting into the code and building other projects on top of the hardware platform ... its got some useful IO and with open source of the hardware design, you could have some fun ... and even just the board layouts woudl be fun to look at in KiCAD ... I wanted to see easy access to tools I could use, and documentation, to modify the unit, that would make it an interesting tech toy.

    There are secions in the Wiki for Milky mist, almost all seem to be blank, no pages, no downloads, no source code, so nerdy FAIL

    I did vaguely look at the NanoNote, being such as low-volume unit, my primary concern was "will it be compatible with anything" ... you really need to feel comfortable thats its going to be something other than an interesting curiosity, and that the designers really understand how compatibility is key. The web page for the "B*n" NanoNote, and indeed even the name itself contains some UTF8 character not rendered on any of my various Linux boxes or my XP boxes, clearly compatibilty didn't even cross their minds when deiciding on the name, so it leaves me little confidence the device itself will be compatible with anything.

    Again a marketing fail. Do these guys even have a marketing guy on the team? If so he needs quietly taking out the back and shooting!

    Nice try, and technically, an interestign challenge exectuted very nicely ... amazing what a team of just a few people can achieve. But, sorry, you really need to work on the marketing BIG TIME.

This topic is closed for new posts.

Other stories you might like