Content

DVI Mixer: Project to Product

2012

I’ve wanted a simple, compact VGA mixer for as long as I’ve had a laptop. Something I could fit in the same case as the laptop, and something that wouldn’t render my precise computer imagery through the lens of a 1950’s television: a single pixel line should not end up blurred and flickery when projected on a screen. My need is for live video performance, but anybody who has had to put a powerpoint presentation through their laptop’s tv-out socket or seen the chaos caused by swapping cables in and out of computers in front of an audience should have an idea of just how useful such a thing could be.

I wrote this in 2009, when I had cracked how to co-opt a new bit of AV industry kit into a laptop friendly HD mixer. I wanted to get it out to everybody else I knew who was in the same position as I’d been. In 2012, I finally got there: the *spark d-fuser, backed by trading entity Spark Live Ltd: http://sparklive.net/dfuser

For some at-the-time context, there’s a nice piece on Create Digital Motion reacting to the release announcement: The Era of Hardware Mixing for Laptops Cometh: SPARK D-FUSER Available. And rolling back to the beginning, you can see what I set out to do in the original version of this project page. The opening paragraph was taken from it, and over those pre-release years the page had ~30k reads, by far the most popular here.

The development of the project has made a fair few diary posts, notably the joy of getting the project out there: ripples around the internet; rebooting the project two years later: pcbs, parts, plans redux; and after the successful pre-order run, taking it retail: *spark d-fuser: to retail.

I have a talk about the whole journey. It’s pretty special, I think. It’s also long, but it’s exposing the detail that in part makes it. Anyway, I’d have loved to hear something like this when I was younger – a parable of you-can-do.

Kickstarter pitches are now part of maker culture, but when do you hear back from the other side?
How did an Arduino hack turn into £50k in pre-orders? How do you get an assembly line going if all you have is a laptop? Four production runs and a retail partnership later, what were the final accounts?
Join Toby Harris as he talks product and dissects a successful run of a maker business.
https://tobyz.net/project/taking-stock

Finally, as a follow-up to that talk and the discussion around a product being just a bunch of files made on your computer, here is a zip of what you’d need to make a Spark D-Fuser. Having spent the talk trying to give a very concrete answer to ‘what is a product?’, it’s my most tangible version of that.

Files

Diary entries

electrovision: the *spark d-fuser

as a warm-up to d-fuse’s trip to brazil, we performed a test of the new live piece “particle” at electrovision. it definitely felt like a test, as things were plugged in and loaded up to be used in anger in for the first time, and while i wasn’t so happy creatively with the form this first rendition took, that is secondary to what happened there: the dvi-mixer i’ve been dreaming of for years - and that my work turning transforming d-fuse’s live shows has been predicated on - worked as simply and unobtrusively as it should. all the sophistication and craziness lives in the laptop, where we have creative control as far as we’re wishing to configure and code, and we have hardware reliabiliy to ensure we can a) guarantee solid output signal to the projectors no matter what is going on with the laptops and b) mix together and tag team the performance between two visuals laptops.

i am working on getting this out to the vj world for a limited run, and a full announce will follow. it will cost somewhere in the range of a $1000 - £1000 , input and output VGA and DVI, and allow you to do what nothing else will: dualhead at 1600x600, triplehead at 1920x480, HD at 1920x1080@60Hz. it is based on a conference bit of kit that i have developed a controller for, and that i hope to get some extra vj-love put into its firmware. the trick to both those goals is to aggregate our demand into one order big enough for a production run of the controller to be made and for it to be worthwhile for the developers of the conference kit to spend some time enhancing it for our uses. so expect a full announcement once everything is locked down and orders can be taken.

dvi mixer ripples around the internet

well… the word is officially out, and rippling around the internet. never seen so many twitter mentions or positive adjectives next to my name - which is nice - but the real deal is are there enough people out there who want one to make a limited production run from the prototype: its not about interest, it will be about orders. http://tobyz.net/project/dvi-mixer

is amazed from this example of userGeneratedProduct: 1° #DVI #Mixer supporting Matrox/Hd. @tobyspark http://tinyurl.com/yfnt4vh Support!

Nice! Compact DVI mixer. http://sparkav.co.uk/dvimixer (via @tobyspark + @_vade ) Make sure to read the direct impetus for its creation!

the amazing @tobyspark has put together the first affordable DVI mixer. you know you all want one! sign up for it now!

yes you, person who has repeatedly complained about the lack of “affordable” or “digital” mixers, you’re about to get both. So head on over to the *Spark D-Fuser project page, read all of the juicy details, and then hit Toby’s expression of interest form.

So now it is here. Not quite. Its really up to you now to go over to his page and put your money where it counts and order one of these things, custom built just for you. Your support will drive this device into demand.

dvi mixer q&a

here’s an update on the dvi-mixer project; i’ve been through the replies to the expression of interest, am working on some things that have come up, and here are a load of answers to common questions that came up.

CONTROLLER

Tap buttons: This is something a fair few people have asked for, and yes, I’m planning on adding this in.

Some kind of switch to route A or B to the output: Apart from temporary overrides of the tap buttons, the crossfader will be the only control for this. Its the hardware angle of just knowing that whatever the crossfader is set to, is what is actually happening. That said, I understand the concern of guaranteeing a solid output of A or B, not a flickering mostly A and a bit of B, and there’s already some logic in there to only start crossfading after a certain movement away from the extremes.

DJ-style faders: A single crossfader is all the control an over/normal or multiply blend mode needs. When you have additive mixing, you might want the A and B levels to be more than simply what is on either side of the crossfader’s knob. I’m thinking about this, DJ-style faders is probably overkill (expense, signal noise, break-ability and loss of simplicity of control surface), a fader curve setting at the back alongside the resolution setting is what I’m preferring at the moment.

Still button: I’m in the fence but looking into this. Displaying a preloaded still is possible, but then why not just full-screen an image on your laptop? A button to hammer on each channel while mixing live is probably not doable with the hardware as-is. We’ll see.

Single controller, multiple processors: This would fit what I want to do with a v2 controller.

Ethernet / OSC / Midi interface: This would fit what I want to do with a v2 controller.

Fader response time: Currently, this isn’t as I’d like it. Its been fine for D-Fuse or *spark use, but a scratch mixer it isn’t. Talking to the manufacturer turned up a technique that I hope should sort the communication side, and we should be shipping with a crossfader with much better electro-mechanical qualities. Regardless, there’ll be a demonstration video showing exactly the kind of response the shipping models will have before taking anybody’s money.

PROCESSOR

Higher resolutions: The hardware can only go so high. There is a bandwidth limit and a line length limit, so while it can do 1920x1200, it can’t do 2400x600 which is actually fewer pixels. I’d love it to be otherwise, I have a major project that really needs that Triplehead at 800x600, let alone the requests for 3072 x 768 I’ve had! Still, triple 640x480, dual 1024x768 and straight 1920x1080 are such a leap from 720x576. Addressing this seems the obvious next step for a version two of the processor.

Different resolutions: The twelve timings I settled on (six resolutions at 50 or 60Hz) were what seemed the most useful in my experience of AV work. Now bear with me: the processor has stored many more resolutions, but to get the plug’n’play ability I want the EDID info transmitted on the inputs needs to match. The tweaked firmware should increase the number of EDID memories, but this isn’t finite and we might not even get twelve. So as shipped, I haven’t had feedback that changes what the six common resolutions would be. The good news is that they should be reassignable, the bad news is that it won’t be trivial - lots of fiddling at your end.

Dual-link DVI: See above, the unit cannot process dual-link resolutions. There is a bonus here in that single link DVI cables are nicer, they should bend easier and weigh a good chunk less than dual-link ones. The dual-link cables I have are the single thing I really don’t like about the setup.

DVI-I sockets: You get DVI-D and VGA in the same socket, done right. Mix and match DVI or VGA inputs or output. EDID transmitted on the inputs that can be independent of what the output is doing.

Additive blend mode: The good news is that this should be happening! This is supercool, a great win. Also I’d had a blind-spot in not asking for multiply as well, so thanks to the feedback this is on the list as well: in terms of implementation, they’re pretty much equivalent, so the omens are good.

Photoshop-style blend modes: No chance. From my personal perspective, this is where you want to be doing stuff in software, the mixer is there to guarantee your signal to the projector, to give a hardware controlled fade to black, to allow seamless switching between between laptops outputting an image already fully composited in the modern vj app of your choice.

Audio In/Out: I’m not touching audio for a load of reasons, but would like to make a v2 controller that is controllable externally, so you could link an audio mixer with this via Midi / OSC for instance.

Split/Preview output: Its just two in, one-out. If you want to split anything, you’ll need separate DAs. Typically for me this is downstream of my TripleHead anyway, splitting the three VGA projector feeds to have a monitor preview of each.

TripleHead: Any TripleHead is separate to this. Bring your own if you want to use one in conjunction with the mixer. I’ve tested with a TH2Go Digital Edition only.

Latency: As per the fader response time, there’ll be a demonstration video showing exactly the kind of response the shipping models will have before taking anybody’s money. While we’re at it, the processing is in 24bit 4:4:4, so there should be no quality loss because of the mixing.

Internal power supply: It sucks, but its an external power supply for both the processor and controller I’m afraid. I did work through an all-in-one version, but the cons outweigh the pros.

THE SWEDES WON’T BUY A PIG IN A SACK

What a great saying; hat tip to Mikael. Once I have a processor running the tweaked firmware, and a controller representative of what will ship, I’ll make a video to show it all in action, and have the purchasing terms all laid out. And hopefully, you won’t just have my word to take for it… more to come soon.

dvi mixer - sliders, sliders, sliders

its been quite some time since november and nothing visible has been happening. this is a quick post to say that stuff is happening behind the scenes, albeit with lots of delays caused by my spare time being completely out of sync with people i’ve been trying to get things going with. but now the momentum is back, and here shawn bonkowski and i are choosing sliders from the seemingly limitless selection on offer.
being trained in product design and loving this book, i’d have said i have a fair understanding of how much work it can take to transform a prototype into something suitable for manufacture and the real world. but i have to admit, this has taken far far longer than i expected. though now, i think, i can say that the controller is falling into place: still far from having a final, manufacturable, design, but the road to get there is clear and doable.

dvi mixer - pcbs, parts, plans

another work-in-progress update for the *spark d-fuser aka dvi mixer project: here we have the v1 pcb design, parts specified, plans drawn, and – with shawn multimeter in hand – the leg-work translating that into something that works. the real announcement here, though, is this: i will be presenting the project in full in berlin on the 12th june, so expect to know a lot more about what, when and how around then.

http://festival.visualberlin.org/day-program/
http://festival.visualberlin.org/news/spark-dvi-mixer-coming-to-berlin/

dvi mixer - i name thee 'berlin'

a night tweaking firmware, a morning visiting the laser cutter workshop, some black vinyl, and the first ‘manufacturable’ prototype is born. huzzah: and promptly off to berlin…

vbfest » dvi mixer presentation

having spent most of my time since arriving in berlin behind a laptop screen, it was time to unveil effort #1: a dvi mixer project presentation, keynote document with i’s dotted and t’s crossed. more importantly, for the swedes won’t buy a pig in a sack, the presentation was followed by a demo of - and hands on with - the new *spark d-fuser prototype.

thanks to prack for the photo, although i do look like a muppet…

dvi mixer - pcbs, parts, plans redux

it’s been two and a half years since the magic week of going from idea to breaking a working dvi mixer package and the ensuing dreams of getting it out there for everybody. problem is, thats still in dream territory: where’s the manufacturable hardware or website buy-button?

by june 2010 there was a v1 and things were looking good. it had taken far longer than seemed necessary… but that turned out to be just the tip of the iceberg. this was a year ago, and this slightly less. so what has been happening? long story short, the software[1] had been waiting for hardware that was perpetually “almost” there. if you’re time-constrained, don’t work with someone else who is also time-constrained; doubly so if you’re entirely relying on them. if you don’t have the skills to make you the master of your own destiny, get on with getting them.

getting on with getting them, for the past two months i’ve been working away and what you see above is an extract from my first PCB design, a from-scratch reworking of what a d-fuser PCB needs to be. i’ve just sent it out to be manufactured and all-importantly assembled up (data sheets and making up eagle library parts may no longer scare me, but soldering 0.5mm pitch FFCs does). if it doesn’t work its because of the design, and that i can work with – and now have some guru backup.

all of which means it’s as if it’s may 2010 again: we’re in “pcbs, parts, plans” territory, complete with corresponding announcement: i will be presenting the work-in-progress at dorkbot london #78[2]. the difference is, this time i know the whole widget, its entirely down to me, and i’ve even secured a little start-up funding to expedite this prototyping.


  1. including the gift of the tv-one header for any other projects out there, and doing some reverse engineering to get EDID upload functionality that plug’n’play would need. props to vade for the insane undertaking of reworking the original tv-one header to have every rs232 command and resolution under the sun in there, not to mention bootstrapping the QC plug-in itself. ↩︎

  2. if you’re in town, come down! ↩︎

dvi mixer - square peg, round hole

package was waiting for me, opened the box: two assembled pcbs, all looking as they should.

to my utter amazement, the electronics all work out: once the pinouts were updated in the mBed firmware, its talking RS232 to the video processor without drama or debugging. to my double amazement, after a quick primer on SPI and decyphering the oled’s datasheet, a test program incants the screen into life, and some time after that the random pixels of an uninitialised buffer have turned to a pattern of my coding. some image editing and two processing sketches later, i have the byte sequences to display a full screen image and typeset my choice of pixel font. the test program fleshed out and re-rolled as a library, and we have the above sight. given the delays the screen has caused since 2009, its so satisfying having this work as per the data sheet, as per design, out of the box; and not even relying on other people’s voodoo: my library, from scratch.

of course, things are never quite that simple. the electronics check out, but the physical fit requires some rejigging – note to self: order the enclosure in advance next time, no matter how well dimensioned its spec drawing may seem. a millimetre here, and a millimetre there has now forced a complete re-layout of the PCB.

aaaand: the ultimate irony? the physical fit issues weren’t just components hitting case internals. what you can’t see in the photo are 40 jumper wires coming out of the mBed’s socket on the PCB leading to a displaced mBed sat in a prototyping block. for all of the correct design and manufacture, these assemblies are compromised by a square peg (mBed pins) not fitting into a round hole (a quirk of the socket strip i spec’d). tssssch!

physical fit, fitted

for a few fractions of a millimetre here and there, this is pretty much a whole new pcb design. it also meant no magnetically isolated ethernet jacks (the mag in magjack should really be magic instead/), which puts you on a path of doing that isolation your own circuitry, and that… that can put you on the path of feature creep.

yep, feature creep it has, and with that more delay. but for a very good end: that ethernet jack can now do more than ethernet…

student entrepreneur fund awardee

huzzah: cash to prototype the mixer.

the d-fuser in four minutes

after being one of the forces behind qmedia’s inaugural open studios last year, this year i was playing behind-the-scenes fixer, and with things fixed was able to get a few hours hacking on the dvi mixer before the show wrapped. even better, having established last year the reward of documentation, there was now a film crew looking intrigued and asking me to explain my research…

lpm'12 » *spark's stall: mixers, screenrunners, and making it as post-vj

in which i set out my stall, hawking the wares and ideas that have helped make my ‘post-vj’ life. half an hour, somewhat off the cuff, and probably ill-advised in parts. the bulk of the talk covers some of the backroom work and event life of the dvi mixer hardware and screenrunner software projects, but contextualised by how i. vjs should work towards exploiting the liveness possible in their practice, ii. which should lead to a more rewarding, less clichéd life, iii. and one more profitable and sustainable.

dvi mixer - dmx, check

file under ‘so glad this worked first time, because if it hadn’t i wouldn’t have known where to start’.

as alluded to in the last mixer dev post, the ethernet jack wasn’t just an ethernet jack any more. my first pcb design had the ethernet jack as a ‘magjack’, which has the electrical isolation to protect the circuit from rogue cabling within the jack and was simply a case of leading the signals from the mbed to the jack. however, all the magjacks i could find could not in the end fit within the final tolerances of pcb position and case height, so i had to move to using a discrete isolation chip and a low-profile RJ45 socket. and at that point, the now plain RJ45 socket was suitable for direct connection of DMX - there’s a standard RJ45 pin mapping as well as the XLR type normally seen. so version two of the pcb, in addition to the physical fit issues, had two subcircuits for ethernet and RS485 leading to the socket, carefully electrically isolated and switched.

however, the assember did a snafu and a wrong part was fitted to power the RS485 subcircuit, and it blew on my attaching power. the advantage of it being a subcircuit is that it didn’t take anything down with it, so with an otherwise operational board, the DMX side was sidelined. some months later now the reworking has finally been done, so i made up some RJ45-XLR DMX patch cables, linked in a contributed DMX library from the mbed repository, worked it into the controller’s code, borrowed some DMX gateway hardware… and it just worked! fabulous.

dvi mixer - code released

things are building to a crescendo. as promised the software that runs the controller will be open-source, and so here it is being released.

http://mbed.org/users/tobyspark/code/SPK-DVIMXR/

notably -

i’ve also been corralling the OSC code available for mbed into a library: http://mbed.org/users/tobyspark/code/OSC/

for history’s sake, and perhaps it will help any hackers, attached is a zip of the arduino code i had before the leap to mbed was made, v07 to today’s v18. none of the interface goodness, but has got the fast serial communication technique i came to along with keying etc.

dvi mixer - ceci n'est pas un mixer

drawings do not a mixer make, but in their finality for production and retail they make things that much closer.

dvi mixer - the countdown starts!

i’m super happy to announce that the site through which i’ll be detailing and selling the *spark d-fuser mixer will be going live next friday, then will start taking orders for a week starting the friday after that, and on the first september the manufacture run will start. huzzah!

dvi mixer - preproduction demo done

“The *spark d-fuser lets you cross-fade between laptops. It’s compact, affordable and — after quite some time(!) — I’m now ready to make the manufacturing run. So this is the final prototype, and with this I’m making this video to demonstrate the features and functionality so you can decide whether you want to have one yourself. If you do, you need to order now so that it can go and be made as part of that manufacture run.”

https://vimeo.com/tobyspark/sparkdfuserdemo

the video doesn’t look like the above, but thats how i remember it: a huge amount of faff to get the point of being able to knock out not far off a one-take wonder. you can probably see the time in keynote presenters display, i’m not sure i want to: the afternoon turned to evening and then night, and i ended up returning home when the tube had started up again. note to self: start in a studio instead of trying to transform the office most of the bits and bobs were already in. at least the night allowed me to hijack a reception area, even if it meant ripping up half the floor to find the cabling.

dvi mixer - committed: publicly, financially

with two huge milestones, it’s the point of no return: the site with price, spec and ordering info went live yesterday at sparklive.net, and i’ve just committed real money to get the first manufacturing task primed.

dvi mixer - orders close midnight monday

Just a quick reminder that if you did want to order one and haven’t so far, we’re now in the last 48 hours for ordering.
Given the delays in opening – banks and paypal, urgh! – I’m going to hold orders open for an extra day. Orders close 23:59 Monday 10th, GMT.

It’s now or never, world!

dvi mixer - packaging origami

at some point, you realise that foam lined box isn’t going to cut it. to hold cable, power supply and controller neatly together, and do it without squashing requires some form of carefully shaped insert. hence a night of cardboard, metal rule, craft knife and ever-refining design print-outs.

dvi mixer - i like custom manufacturing

a wonderful moment: the enclosures have arrived, and on opening the box the quality is outstanding. what was in the mind’s eye, now perfectly realised in hand.

dvi mixer - factory production

back to the south coast where an electronics facility has assembled the PCBs and is about to start the box build. people in lab coats walking around with trays of electronic innards with my logo on it… a new one for me, and the first concrete sense of the scale of this.

dvi mixer - factory tests

the assembled PCBs and manufactured cases passed the physical fit test, but thats moot if the damn things don’t actually work: the real business of being at the factory was to finalise the test spec. finish up a self-test firmware, supply a pc with rs232, open sound control, dmx and a processing sketch to tie the different comms together, and spend the rest of the day detailing up the test procedure document, as pedantic a document as something can be.

dvi mixer - blend [ -x---- ] add

this is the first factory build, running the firmware to be shipped, and with that an announcement: the video processors are coming with custom firmware that allows additive mixing, and the controllers have an implementation that morphs the crossfader behaviour from a flat blend to full add.

dvi mixer - the package

there’s a certain terror in thinking that all the time and effort invested into getting a good product could be wiped out by damage in transit: designing the packaging has been just as involved and stressful as the product itself.

dvi mixer - programmed processors

not just the solitary 1T-C2-750 that we, as D-Fuse, bought in the summer of 2009 any more! a shelf full of the things, all programmed up with custom firmware to do additive mixing. what you can’t see is the floor of the industrial unit laid out with a sea of boxes, ready to be assembled up and shipped off.

dvi mixer - first pack

by rights, as the first off the line this should be sean healey’s, but the delivery note at the top of the pile was one for the US. but - this is it! the beginning of the end, the first packaged product.

a project no more: products, shipping.

dvi mixer - with v25 and a video, first run done

the demo video upon which pre-orders were placed had prototype hardware; here now is a manufactured product, out of the shipping packaging, running the final v25 software. i say final, because as the units spread around the world that prototype is no longer the sole test-bed: v25 encodes feedback and testing from real customers. i’m proud of where i’ve got it to.

i never could have anticipated how much work it would take to get here – and it isn’t over, and there’s always scope for more – but right now, it all feels a good place to be, and that video a good point to draw a line. from here, let the videos start coming back, scenes exotic with a small black box with diagonal stripes somewhere in the mix…

dvi mixer - a production runs worth of updates

with the production run now done, for posterity here is what it took, as evidenced by the update emails i sent out to everybody whose future-mixer was being made.

Hello,

Thanks for ordering a *spark d-fuser, you’re one of 80. I realise its quite a thing to put your money down on something sight unseen, doubly so something that has yet to get through its manufacture run. So I’m going to do my best to keep you in loop as I get them manufactured and out to you.

Before getting onto the production status, here’s a little story. While the orders were open, the prototype was out on a field trial in the hands of Jim and Adam of As Described, who were running a stage at the Bestival weekend. Saturday morning, my phone rings, and its Jim on the other end. This makes me worried: I assume he’s ringing because something has gone horribly wrong. But no: its a very excited Jim, reporting a great night VJing with the mixer, guest acts such as Pfadfinderei rocking it, and saying Matt Black of Ninja Tune saw it and asked straight away how could he order one. And later that day, in comes an order from him.

Firmware

  • I’m working on the controller firmware to read in settings from a configuration file, aiming for the resolution menu and keying parameters to be set from a text file. I was hoping to say this was done by the time of writing this email, but I’ve been having grief with all the config file reader libraries I can find. At this rate I might just roll one myself, its only reading in lines from a text file after all.
  • The people at TV One were finally back in the office, and while I still haven’t got a firm answer about the processor firmware with additive blend, they are at least making positive noises.

Summary: developments, but nothing definite yet.

Controller Manufacture

  • Case: I’ve approved the production sample and transferred the payment for 85 machined and printed cases to be manufactured. This what I had quoted at 5-6 weeks, and has gone off within the first week, so we’re still on track here.
  • MBED: This is the ‘brain’ of the controller, and is a £40 part a bit like a Arduino on steroids. The minimum order quantity to get them direct is 100, so I thought I was in the bad position of having to find 80 plus spares through places that sell them one by one… and the place I usually use was out of stock. The great news is that they’ve managed to bend their system to allow me to order under the minimum quantity. The not so great news is that their hand over to their sales department got dropped, so while I’ve prompted them and its back on track, I still haven’t been able to commit to the order, and there may well be a few week lead time to get them to me.
  • DC jack: I want the controller power supply to be interchangeable with the processor’s power supply. This meant finding the special locking socket like TV One use on the processor, which I had done and tested its fit. Unfortunately that part has now gone out of stock and won’t be back in stock till the end of the year. I’ve found an alternative, and while I’d have to order 500 the price at bulk is actually quite OK. But that has an eight week lead time. I haven’t made a decision on this yet, but there’s a good chance I might just have to use a generic non-locking jack. Its not the end of the world, but it is a bit of a shame.
  • Electronics: I sent out the design and bill-of-materials to get quotes for manufacture and assembly of the PCB, and while they’ve started to come back I still haven’t had them all back. They’re looking ok so far, but I need to resolve the MBED and DC jack situation before being able to get that going. With the lead time of 5-6 weeks for the case, I thought there wouldn’t be pressure on the PCB assembly, but if the MBED supply means this can’t start for a few weeks, that might change things and push back everything.
  • Assembly: I’ve made up instructions for assembly of the complete controller. We’ll see whether its economic for the PCB assemblers to take on the whole package, or whether its me who slips the PCBs into the slot in the case. Regardless, its a nice moment, it makes it feel a good step ‘more real’.

Summary: waiting on a quote or two, possible delay due to MBED sourcing.

Toby

 

Hello,

This week’s update. At the end of week two the totalling and quoting for the controllers is done, company chosen and monies paid. There’s even an assembly and test document, if not the actual test rig or code setup. The TV One situation is still unresolved.

The bit you care about: when? The controller manufacture is stated at five weeks, which starts Monday. So thats one week over, I’m afraid.

Toby

 

Hello,

First - I’ve done this update mail through the service I’m using to handle future production run enquiries, any issues let me know.

There have been two weekends without updates - apologies. I should have known better than to wait on somebody getting back to me with an update, and definitely known better than to let that happen twice! I have been busy though, updates below.

Firmware

  • I had a less than fun weekend realising there was a crashing bug in the controller code. Without a proper a debugger tracking this kind of thing down is a nightmare, especially when you read a load of internet posts about the library you’re using at that point displaying just the behaviour you’re seeing. Inevitably, the mbed library turns out to be faultless and it my code that was the problem. An edge case in my optimised TV One comms code was quietly setting the stage for a later crash, but at least its fixed now and I’ve learnt a thing or two in the process.
  • The code is in there to work with resolutions and keyer settings read from a config file – much better for you than having to tweak the source code – but I still haven’t got a library running as part of the firmware that actually reads and writes settings to a file. Perhaps more eyes on that will solve it trivially, its probably a compile problem or suchlike.
  • You wouldn’t believe it, but TV One still haven’t given me an answer on the additive firmware. But by the ‘end of the week’, ie. now, they should ‘have something to show me’ and a development cost. It seems they’re doing the work before quoting me a cost to get my approval? Next week I’ll be able to say, surely!

Controller Manufacture

  • Case: They arrived this morning, and the quality is outstanding. Boom! Really stoked. http…
  • MBED: They arrived ahead of schedule and are at the assemblers. Getting them direct means dealing directly with ARM, which felt quite surreal: invoices and letterheads from the architects of the chips that power pretty much all smartphones and tablets!
  • Electronics: Remember that DC-In issue? Well this is where the news gets more annoying. Long story short, I had a supplier who thought they could source that part, and two weeks later found out that they couldn’t, so we’ve had to make a plan B, and we’re delayed. The controllers will now come with a non-locking part, and I’ve ordered that 500 lot of a UK manufacturer’s equivalent part just so this never need happen again. This is doubly annoying, as by this point after much searching around for power supply options, I’d decided to go with the genuine TV One ones, which of course have the locking plug. Quality assured matching part, the plug shaft will will just stick out 1mm from the controller.
  • Assembly: The product assembly and packaging will be done by the PCB assemblers, which is great news for my kitchen table and general sanity. The cardbox box supplier is even making up tooling for the fold-up insert that will have to be made to hold everything in place. Thats yet another example of things being more complicated than you first thought: I had standard foam lining in my costing spreadsheet, but when it comes to making up a test you realise you can’t squash together the power supply and cable along with the controller… hence a night of origami and cardboard as an insert is designed and shaped to hold everything just so. http…

Anyway, the bit you’re really interested in? I’m hoping they’ll ship the week of 5th November.

Toby

 

Hello,

Here’s the bit you’ll like and makes me super happy: TV One are going to develop the custom firmware, have been paid for it, and are getting on with it.

On a personal note, adding additive mixing was pretty much the reason I put this project out to the community back in 2009. I thought with the leverage of many TV One could be persuaded to tweak the 750. Didn’t quite prove that simple and while the journey between then and now has been frankly crazy making, at least there’s the reward of the feature as an artist I felt was missing right from the start.

Here’s this week’s standard bit of production WTF: I’ve just spent the best part of two weeks going back and forth with a cardboard box company to find out that they couldn’t in fact do the basic job I’d asked them about in the first place. When you say to them the sample box is good but can it be made up of a better grade of cardboard, there’s quite a difference in answering “Yes, we have lots of grades of cardboard” instead of “Well, we have lots of cardboard but that box can only be made up in the grade we sent you”. Especially when they’re going to spend the next week working up tooling quotes and samples that are specific to the exact properties of that box. That can’t exist.

Here’s the really horrible side you don’t ever want to know about: despite having got all payments up-front, with the wonders of PayPal I’ve had an oh-so-interesting time actually accessing that money. I’ve had to put in £10k of my own money so far to push things along, and its just occurred to me I’m probably going to have to pay the VAT on all the sales before actually receiving the money for half of the sales. Which could make that figure a lot higher. Gah.

On a lighter note, this week saw some pretty pictures going through the prototype for the opening concert of the Belfast Festival, with D-Fuse providing a visual score to Holst’s The Planets. Plenty of power politics and setup stress at such things, but so good to get back in the mix.

Toby

 

Hello,

Since writing last time, I’ve spent a day in a lab coat at the factory. They’re getting on with the final build and testing, and I have the first off the production line in front of me here. Huzzah!

Here are three diary posts from the factory, and a shot of the real controller. A drawing no more!

Controller firmware is up to v21, there’s a lot of code and time in there. I’m quite proud of the way it handles additive mixing, and you no longer need to hack the code to change things like the resolutions in the menu system: its reading settings from a text file. But really all the effort is about smoothing the edges of the TV One unit. Hopefully it does, more and more.

Away from the digital anvil, I’ve sanded 100 menu knobs down to fit (https://twitter.com/tobyspark/status/263411649485094912), the packing tape came in handy sooner than expected (https://twitter.com/tobyspark/status/263412216122970113), and I’ve got there with the packaging (https://twitter.com/tobyspark/status/264152349923237888). Although since that photo, I’ve chucked it down the stairs and realised that with the way the laser cutter works, I need to cut from the bottom not the top for strength… so just when you’re done, you find yourself doing the fiddly business of changing all the cut marks so valley folds become mountain folds and vice versa.

By now though, you know there’s always a spanner thrown in the works somewhere. Here follows this week’s. The crossfader comes plain or with a black dust cover over the slot. I wanted that black version, but it was never in stock. By the time of the production run, it was, so with some minor joy I ordered them. Of course, getting the assembled PCBs back now, I find out that the shank is somehow ever so slightly different on them, my 100 slider knobs don’t fit any more, and I’m not sure if there’s an alternative I can order that will. Gah!

Shipping: close but not quite! Still a week.

Toby

 

Hello,

Processors are waiting to be shipped to me. Custom firmware for the processors is in my hands. Controllers don’t have their slider knobs, but a plan is in action. Controller firmware is done. Packaging is designed, my hall full of 100 cardboard boxes, and a day with a laser cutter awaits to make up the inserts.

All in all, pretty much there. It will take a few days of sitting in a warehouse upgrading processors, folding boxes and whatnot. The place I had lined up turns out to not be free next week, which is kinda annoying, however my main worry is actually sending them to the right addresses: looking at the list from PayPal, some don’t look quite right.

Important: you should receive an individual email from me at some point today confirming address. Please make sure to reply to it.

Toby

 

Hello,

Packages are shipping!

I spent last week setting things up so that this week the kitting company can package up the boxes and send everything out. They’ve got through most of the orders now and will be done by Tuesday. I’ve just got back from London peers The Light Surgeons and seen theirs arrive. If I wasn’t writing this through the fug of a cold, I think I’d be thrilled =]

Talking of being ill, this week has been a write-off and I haven’t made the ‘take it out the box and plug everything in’ demo video I wanted you all to have by the time you received your mixers. It will come, but in the meantime, feel free to hit me up to help you get started: email operator@tobyz.net and we’ll take it from there, I’m rarely far away from the email.

Two notes on cables:
DVI: The processor is only single-link DVI, so there is no benefit in using dual-link cables. Back in 2009 I bought the highest spec DVI cables I could find and have been lugging around these heavy cables that barely bend ever since! If you need to go and buy yourself some DVI cables, I’d recommend thinner, flexible, cheaper single link ones over ‘premium’ ones.
Controller software and power-on settings: Its open to you to tweak and upgrade, to do so you’ll need a right-angle mini-USB lead to get to the socket inside.

A note I hate to write:
Something may not be right with the additive mixing bonus feature. I will see what is what, and what can be done, in the next week.

A final ask:
Tell me how it works out for you! It would be great to hear about your unboxing experience and first use. What information are you missing? What doesn’t work as you’d expect it? Also I need to know how reliable the kitting company has been in doing the final step of the process with me not there. A photo or two of the state of the mailing bag / box / insides would be really helpful.

Toby

 

Hello everybody,

Thanks to everybody who has sent feedback. I’ve learnt a few things, some of which are for me to do better if there is a next time (packaging - I’m looking at you), but some of which need addressing here and now. I’ve had a hunch that something wasn’t as I expected with the shipped units, and have now verified this.

In short, the fully tested hardware has gone out with the software/settings not applied correctly. I am working on a fix. It should be simple. Apologies to all.

I’m pretty clear now on the sequence of events that led to this, and while my manufacturing partner has let me down bad on their final steps, I can’t push all the blame to them: there’s also a problem in the controller software that is my fault alone. Obviously, this is annoying for so many reasons.

In the next few days, I’ll have videos detailing everything. You’ll need the mini-usb lead I’ve mentioned previously. If you are eager to get things going today, follow the software available online (http…).

Toby

Links to usb leads –

I’ve used these: http…

They can be easily found online, and probably at your local computer shop.
http…

I also found one of these in my house, if you go this route you can leave the adapter plugged in and close the case back around it, and from then on use whatever mini-usb lead you have to hand.
http…

 

Hello everybody,

In the last mail I wrote: I’ve had a hunch that something wasn’t as I expected with the shipped units. The fully tested hardware has gone out with the software/settings not applied correctly. I am working on a fix. It should be simple. Apologies to all.

I’ve got the fix and documented its application.

Note if you bought the controller alone, I’ll send a follow-up with the specific instructions and downloads for you tomorrow - which device and TVOne firmware you’re running complicates things.

Without further ado -
Instructions: http…
Download: http…

The instructional videos are coming, this has consumed my time so far.

Toby

ps. For the hackers, this week included reverse engineering the missing factory step that required a PC so that it can now be applied directly from the controller. Rock!

 

Hello,

A quick mail to say the v24 software for the controller is developing further in response to feedback from some of you. Now its not just me with a controller, I’m having a few people beta test the updates, which is proving really helpful.

If you want to get the fixes, check the latest and put in any feedback before I push out what should be the final version, check http… and get in touch. I’ll email again when its done.

Toby

 

Hello,

These emails were to keep you informed during the production process. That process turned out not to finish with them shipping, but it is done now: may I present v25 of the controller software, along with the getting started video.

v25 has a clearer, more consistent user experience. Behind the scenes, it keeps a much closer eye on the processor and has strategies to handle things not being as they should. It handles processors with different versions of TVOne firmware. It even has a reworked blend fade level algorithm: it took pages of diagrams and equations to get past my intuition to something that was 100% correct instead of mostly correct. Who knew crossfading could be so tricky!

Who also knew how much the software would grow. The arduino code for the first gig had 68 lines of program and 128 lines of TV One library, and it did us proud. v25 has 1834 lines of program, 1047 lines of TV One library, around a thousand lines of program supporting bits and bobs, plus things like the OLED library I wrote from scratch. Crazy!

Even crazier, the MBED version control system tells me 2124 lines of code have changed between the version of software I launched with – ie. the version against which you placed your orders – and the version I’m releasing to you now. If you took into consideration the evolution between those snapshots, that number would be even higher. Thing is – and just like the original arduino code – its not as if that version didn’t work: its what was running for the demo video. Just that user-friendly features, refinement, robustness… these add complexity, and seem to do it exponentially.

Enough! Lets get mixing!

Please download the v25 updating procedure instructions here:
http…

If you just bought the controller only, please also download this PDF and start there.
http…

The getting started video is here:
http…

Toby

ps. Caveats below.

  1. Network modes and troubleshooting are still undocumented at the moment.
  2. Deselecting a network mode now seems to cause the controller to crash - not something you do outside of studio tests but annoying nonetheless. This has worked before and the network code is unchanged from then. I think the developments since have increased the memory footprint of the controller and dynamically creating / deleting the network stacks can now cause a stack/heap collision.
  3. There seems to be a bug in the OSC float receive, and right now its beyond me to fix it. In the open source tradition, this should prove trivial given more eyes looking at it.

*spark d-fuser » to retail

turning the dvi-mixer project into the *spark d-fuser product was very much an act of faith that peer demand meant something. but in no small part too, it was an act of hubris: i said i was going to do this thing, and i found i wasn’t at all happy looking at it fizzling away before the 2012 reboot. put those two things together, and you will likely decide to offer a pre-ordering window, and pledge to make as many are as ordered. it could be two, twenty, or two hundred; a loss at the low-end, a profit at the high end, but one way or another they were going to get out there.

but this is not how people expect to buy a product, nor is it how the world is set up to let you sell products. with daily enquiries from people too late to get in on that first run, december and january saw me getting to the point where the *spark d-fuser could become a retail operation. the face of this you can see in the completely reworked website pictured above, but the big development is partnering with a company to take on the logistics of the d-fuser as a retail operation. it’s a company i used for some parts of the original production run, that worked out well so they’re now my manufacturing partner with orders going directly to them.

all which is quite a preamble to a simple fact: there’s now an order button. go use it!

*spark d-fuser » factory, take two

back at the factory to check on more d-fusers being made. don’t think i’ll ever quite get over seeing other people making my things in bulk.

b-seite » *spark's stall out again

b-seite charge one: provoke vjs into thinking about the future of their practice, via showing the toys and talk about how i made it as a post-vj. this was a tidied up version of the talk i improvised at LPM last year, except that then the mixer was a final prototype, and here I was at a festival with two d-fusers as part of their tech setup. mixer’s i’d sold but had never seen, the boxes opened up by hands other than mine, in a country far away.

*spark d-fuser » vidvox feature

“VDMX has been a big thing for D-Fuse. It’s the engine that drives our most innovative shows, laying out and compositing footage across multiple screens, hosting patch upon patch of custom D-Fuse development work. However, for our kind of theatrical shows, plugging a laptop directly into the projection setup just isn’t an option. We need video hardware to keep a solid feed to the projectors from early soundcheck to us walking on stage, we need to have a master fade control for our output, and we need to crossfade between our laptops to mix collaboratively. For year after year there was no solution to this, as with each laptop rendering all screens simultaneously the video format is outside of what the SD and then HD hardware can handle. What we needed was a DVI mixer. There wasn’t one, and so I made one.”

http://vdmx.vidvox.net/blog/history-of-spark-d-fuser

*spark d-fuser » osc and a/v tutorial

prompted by the vidvox feature, a guest tutorial video showing vj use of the d-fuser, complete with a demonstration of how you can link the d-fuser controller with your laptops for better audio-visual integration when performing as a group.

http://vdmx.vidvox.net/tutorials/using-the-spark-d-fuser-with-vdmx-with-toby-harris

*spark d-fuser » desert engines

happiness is seeing photos like this: prototypes and factories far behind, a d-fuser out in the wild. truly the wilds here, courtesy of sean healey and his audio-visual performance ‘desert engines’. was it almost a year ago he was interviewing me as this whole endeavour was announced, prototype to pre-order?

*spark d-fuser » enter stage at edc

leaving the desert sun behind desert sun we’re now deep into the night: ali demirel tripping the light fantastic for richie hawtin’s enter stage at EDC las vegas. photo taken by the perfectly portraited barbara klein.

*spark d-fuser » v29 and more

a milestone in post-retail *spark d-fuser life - there are compelling new firmware revisions for the controller, and a support website that backs that and more.

notably, i’ve assembled a guide based on people’s experience beyond the getting started video, so now there is an extensive document that details use, what you need to know around mixing with DVI/HDMI/VGA, and what to try if things aren’t working out.

headline features from the controller firmware v25 to v29 are
keying - tweaked keys are now remembered, you can key right-over-left as well as left-over-right, and the menu / user experience is much better.
resolutions - better user experience; fit/fill/1:1 implemented for mismatching sources and output; 24 and 25fps HD modes are listed by default.
network - better implementation of OSC / ArtNet / DMX modes, plus detailed OSC how-to document.

all at http://sparklive.net/dfuser/support

*spark d-fuser » resolume review

If you’ve been keeping your finger on the VJ pulse as close as we do, you couldn’t have missed the arrival of the Spark D-Fuser. However, we appreciate the fact that some of you are busy touring or designing wicked content. Or even may actually have a social life. Either way, you could have missed it getting that elusive ‘Buy Now’ button earlier this year.
Or it could be that you’ve seen the button and the hype, but are wondering if it’s really all that it’s cranked up to be. It could be that you want to know more about this mysterious magic box that will solve all your problems, before you part with your hard-earned VJ cash and actually press that ‘Buy Now’ button.
Either way, we’ve had the pleasure of working with that little bad boy on various occasions, as well as seen it in use by quite a few touring VJs. So we figured it was high time to give a first hand experience of what the D-Fuser actually defuses.

it’s a fun and informative review, and who doesn’t like being told they rock.

above the detail and any quibbles, there’s two meta-comments about the mixer that i really appreciate resolume making. the first is that “the *spark d-fuser is a new mixer for a new age”, with the review leading the reader from the act of vjing where the hardware mixer is the central instrument to where we are now: the sophistication and breadth of software is the prime driver, and so everything should be built around that. second, is that the *spark d-fuser is in many ways a hardware equivalent of resolume - it is a carefully designed product with a honed rather than expansive feature set, a product entirely built around the user experience for the user.

Now go press that ‘Buy Now’ button, everyone.

http://resolume.com/blog/review-hands-on-with-the-spark-d-fuser

d-fuser and 4x1024x768

back in 2009, the d-fuser was conceived to mix ‘triplehead’ dvi. it was a direct need for d-fuse’s live shows, along with HD. that story is long told.

then in 2011, the datapath X4 was released. it’s a triplehead2go on steroids, coming from the pro-av market rather than gamers. still, surprisingly affordable for what you get, which is a 1-in, 4-out dual-link dvi box where each output is completely configurable. before even getting to running quad-head, you could take triplehead input, split that to three outputs, and then use the forth to scale that non-standard 12x3 input back to something you could display on a monitor on your desk. i like, a lot[1].

it was a while until we got our hands on one at d-fuse HQ, but we did, and now the d-fuser can work with one too. of course, this has all happened since the firmware was commissioned and plans went to the factory, so the retail units don’t ship with this support. having just stepped through the process for the guildhall school of music and drama’s purchase of two d-fusers and an X4 for their scenography work[2], i’ve packaged up and made good the resources necessary. it’s not quite ready for http://sparklive.net/dfuser/support yet — it involves a windows pc, amongst other things — but the process works, and i’ve even got a nice align grid and 4x1 to 2x2 pixel map and test movie rolled in there.

so - attached is a zip of files with the resources you need to configure 2048x1536@30Hz through your d-fuser and 2048x1536@30Hz to 4x 1024x768@60Hz through the X4.

SparkDFuser-DatapathX4Support-v1.zip
  Align Grids
    SPK Align 2048x1536.png
    SPK Align 4096x768.png
  D-Fuser Controller Setup
    _See support site for firmware v30 and above_
  D-Fuser Processor Setup
    Windows CorioTools SPK-DF-750 v423 with X4.zip
  Datapath X4 Setup
    spark dfuser - datapath X4 configure script
  Pixel Map
    Spark D-Fuser X4 4x1 to 2x2 Pixel Map.qtz
    SPK Align 4096x768 30fps Bar.mov 

i didn’t see many config scripts for the X4 when i wrote my first, so here’s the configure script from the zip in plain, google-searchable, text.

#!/bin/sh

echo '### Datapath X4 Configuration for *spark d-fuser'
echo '###'
echo '### http://sparklive.net/dfuser/support'
echo 

if [ ! -x '/usr/local/bin/vqscmd' ]
then
  echo 'Error: Missing Mac OSX software for X4'
  echo '       Download and install from www.datapath.co.uk'
  exit
fi

echo 'Setting input format to 2048x1536 30Hz'
/usr/local/bin/vqscmd -DeviceNumber=0 -InputEdid=1,2048,1536,30000

echo 'Setting output format to 2x2 grid of 1024x768 60Hz'
/usr/local/bin/vqscmd -DeviceNumber=0 -OutputMode=1,1024,768,60000,1
/usr/local/bin/vqscmd -DeviceNumber=0 -OutputSource=1
/usr/local/bin/vqscmd -DeviceNumber=0 -CaptureRegionDefault=0

echo 'Saving settings in X4'
/usr/local/bin/vqscmd -DeviceNumber=0 -Commit

  1. its noisy fan and bulky power supply, less so. ↩︎

  2. first use of which was this: a bright spring day spent in a dark concrete cellar… ↩︎

taking stock: from arduino hack to shipping product

it’s common to hear people talk about the beginnings of projects, but rare to witness a post-mortem. doubly so in the world of kickstarter: the pitches are now part of maker culture, but where’s the venue for a summing-up at the end?

i wanted to do that. with what could be called the modest success of the *spark d-fuser, i had something i could be proud about. moreover, and this probably gets to the heart of the issue, without the disincentives of stratospheric success shrouding the business in secrecy or abject failure making me want to crawl into a hole or quietly move on.

and where’s the venue? well, elephant and castle mini-maker faire seemed a good place to get a draft together and punt it out.

it’s still a work-in-progress, to be done when the last mixer is sold and accounts reconciled, but nonetheless it’s great to have the bones together.

bristech » taking stock: what made a maker business

how to network in a new town? i schemed going to talks is good, but giving a talk is better. plus it would be a good prompt to develop the definitive version of the ‘taking stock’ talk, the nascent version of which i’d made for the 2014 Mini-MakerFaire we effectively launched South London Makerspace at.

Kickstarter pitches are now part of maker culture, but when do you hear back from the other side?

How did an Arduino hack turn into £50k in pre-orders? How do you get an assembly line going if all you have is a laptop? Four production runs and a retail partnership later, what were the final accounts?

Join Toby Harris as he talks product and dissects a successful run of a maker business.

// About Toby

A product-focussed generalist with a soft spot for python and a CV that includes getting a robot to do stand-up comedy.

Bristech Meetup, 5th September 2019

there is a recording, though as with makerfaire there was a hitch; this time the audio starts a few minutes in and the video a few minutes after that.

enjoyed the talk before me, too –

dorkbot / quantifying progress

gave a lightning talk at dorkbot – we all know that a hack ain’t production-grade, well here’s me quantifying that.

taking stock: what made a maker business’ is far too big a talk to give at dorkbot, but this little excerpt made for a perfect bite.