On Thu, March 3, 2016 8:32 am, Ffanci Silvain wrote: Several companies have asked this question over the years and the usual result is a product that ends up costing significantly more than £2500. However it is certainly possible to roll your own custom keyboard synth. Ken Restivo has done it and had some success. http://www.restivo.org/ Boot time may be an issue if you want to load a full desktop environment. However it is also possible to boot directly into X and run your UI there. That also saves a few resources but means you are locked into a specific application. If you don't need a visual environment then you can run without X. I have servers online running this way. They don't even have sound cards and use jack -dummy instead. Everything can be controlled by external hardware and/or command line fu. If you want to save yourself time and hassle buy a prebuilt manufactured solution. If you want to save some upfront costs but are prepared to spend your time and efforts then you can build your own and make it work. You might not get much appreciation from the audience as they tend to get gooey over sponsorship agreements rather than self realised technical expertise but you will solidify your tech cred with the gear head crowd who do appreciate such things. i.e sound engineers, lighting technicians and roadies. Probably a good scene if you are into hairy, dirty, sweaty, boozy men ;-P -- Patrick Shirkey Boost Hardware Ltd

On Fri, 4 Mar 2016 20:20:02 +0000, Ben Bell wrote: Sure, you listen to many kids and at least one crank, they just play and their parents and countries pay for their fun, they think in an academically approach, far away from the wilderness. Buy the Nord! Startup time was mentioned by this thread. That's even not an issue with very old HDDs, upstart and systemd are very fast, much less than 10 seconds for a clean audio setup. Reliability on stage is something completely different, that has abolutely nothing to do with upstaret time. Why do you think everybody on stage is playing a Nord and not some Win, Apple or Linux computer? Everything is in an as small as possible case, without the need to use a display, but with all controls ergonomically provided and even without a flight case already steady as a rock. You pointed out "Stage" ;). "Stage" = Nord and/or similar, but not a computer set-up, with cables and rack and keyboard. That's ridiculous.

On Fri, 4 Mar 2016 16:18:14 -0800 (PST), Len Ovens wrote: A synth, such as the Nord has got everything a stage musician needs. If you instead use a master keyboard, that requires different Linux synth to provide everything, you unlikely will be able to provide a consistent schema for the available controllers, let alone that a synth, such as the Nord has everything in a solid case. If you put your atom board into a 19" case, you need to carry a rack, a master keyboard, a professional external sound card, you need to connect by cables..., nobody who has experiences with real life, who is really living on this planet, can recommend to replace a stage proved environment by an experimental computer based pseudo-solution. What next? Instead of playing a guitar, using analog effects and an amp, playing a sampled guitar with fluidsynth and emulate the analog effects and amp with guitarix? To do that you don't need to buy guitar, effects and amp anymore, you just need an atom board in a 19" case in a rack, a professional sound card, a master keyboard, and tons of cable connections. This ridiculous obsessive approach to replace everything music related by a Linux computer is much more wrong, than loudness war and auto-tune at one go. Linux or any other computer based solution is an extension to stand alone music gear, but it can not replace this gear. And by the way, if the OP insists in replacing real gear, by a computer solution, the OP should consider to buy an Apple based solution. Linux simply can not hold ground for this kind of usage. But again, IMO the computer is the wrong approach, even when using proprietary Apple solutions. There is a good reason that companies are still able to sell stand alone gear such as the Nord. Please stop to encourage somebody to do something stupid.

They say that those who do not know history are doomed to repeat it. I say that those who do not know history should google for "lionstracs keyboard", although they may still end up repeating history anyway. On Sat, Mar 5, 2016 at 3:11 AM, Ben Bell <bjb-linux-audio-user(a)deus.net> wrote:

On Sun, March 6, 2016 12:54 am, Ben Bell wrote: Another contributor to the difficulty of making a successful keyboard/DAW hardware solution is the amount of effort/time required to build and run the system. It's a good couple of man years of effort to wrap it all up into a simple UI. Add the cost of manufacturing hardware and it quickly become very difficult to compete on price with the established players. It's a minefield to build a simple system that just works. Realistically it requires building out a complete stand alone distribution. Lionstracs got close to that goal but in the end the sales dried up. It's very hard to compete with cheaper solutions made by manufacturers who have economies of scale on their side. The MOD is also struggling from the latter problem too. It's very hard to keep costs down until you are manufacturing large quantities of product and that makes it hard to compete with the other established players. Giving away product or drastically reducing price to compete doesn't do much for the bottom line and doesn't help pay salaries or rent. Trying to sell something that is priced higher than the rest of the market is hard unless you have the brand loyalty of Apple. They need all the support they can get so if you are serious about doing something cool with Linux Audio Hardware invest in the MOD and help them grow so they can compete with the big players. If 50 LAU'ers had bought a lionstracs that might have made the difference between the company surviving the great depression of 2008 and going bankrupt. As far as I know, not one LAU'er actually paid real money for a Lionstracs console. The MOD is significantly cheaper so they should get more traction with this community. -- Patrick Shirkey Boost Hardware Ltd

My 2cents here: On the hardware part, I might be wrong but a powerful ARM platform is more interesting than any of the Intel NUCs. If we are talking multiple cores we have the Cubie8, the Optimus and others based on the AllWinner A80 which is an 8 core, 2MHz, ARM v7 + ARM v15 (BIG.little) CPU If we’re talking performance per core, the Pine64, the DragonBoard 410 and even the Raspberry 3 (though I would not recommend the later) are a great option for the new 64bit ARM cpus. And of course, there are at least a handful of companies offering these SBCs already with a Board Support Package for the Industry: Toradex, Eurotech and Variscite are some. On the software part, I’d divide it into the OS + BSP and your actual software (the actual application) If you're making a business, then outsource all the OS + BSP and focus on your application. The specification of OS + BSP are defined by the application, so I would do it at least in two rounds, doing a first round of prototypes using as much readily available elements as possible (and caring very little to aesthetics, ergonomy and actual performance), build a preliminary/pilot application over this prototype and then, if you decide you’re game on, you can then precisely specify the outsource job to be done. I can assure you it will pay off in the long run :-) if you are using AllWinner CPUs, companies like Free Electrons can do all OS and necessary drivers. Even better, they do it in mainline vanilla kernel, which not only keeps your Open Source assumption but also gives you a much longer life cycle. If you’re doing for fun, try to find the board that offers the smallest effort on the OS + BSP. Magic word here for the tool is Buildroot ;-) hope I have helped Gianfranco Ceccolini MOD Devices +49 160 646 9313 gianfranco(a)moddevices.com <mailto:gianfranco@moddevices.com>

Why paying ten thousand Euro, to get a grand piano for the master class? Any recommendations for a Linux software based solution? Sorry, I couldn't resist.

We used the Intel DN2800MT Great board. By the specs, before we purchased it, seemed to be perfect. When put to action, we had the following drawbacks: 1 - Is it a fanless mother board, but it does not mean the enclosure of the product can be closed. The MOD Quadra required a lateral cooler that was a pain. The only way to have it totally fanless is to have an enclosure that acts as a radiator - like those amplifiers for cars - but that was not a possibility for us. 2 - although Intel makes lots of efforts to have an embedded board, they are far from being successful. One flop in our board is that the EFI BIOS has a “fast boot” mode which, in principle is awesome, but when activated requires a keyboard confirmation when rebooting after an unsuccessful boot. Really frustrating for use in headless (no monitor and keyboard) units and shows that Intel is really a newbie in this embedded field. 3 - the EFI is said to be highly programable and Intel seems to even offer an IDE to that, but in real life, even contacting Intel professionals (we had the Intel Inside Logo) I could find a person who could do me a custom EFI. 4 - This last point has nothing to do with the mother board itself, but with the Intel Architecture. Since you are doing a consumer device and it probably will have USB connection, Intel starts being a pain because it does not offer support for USB-DEVICE. Intel computers, by definition, are only hosts to USB devices and cannot act as devices themselves. I don;t know the status of this now with the Galileo and such, but last word I had still kept like that. 5 - Last, but not least, I would not recommend Atom. Celeron and Core seem to be much better CPUs. Again, in our case, the Atom had a Graphic Chip from PowerVR that was totally unusable in Linux. If we had a Linux Supported GPU, some libraries could make use of OpenCL and move parte of the computing to the GPU. I think that, if you are not moving into a modular CPU like the ARMs and chose to stick with Intel, then go for the horse power. Evenly efficiency shall not be an issue in this kind of product (unless you want it battery operated) If you still insist on the Intel Architecture, then I highly recommend www.logicsupply.com Best Gianfranco The MOD Team

On Fri, Mar 04, 2016 at 07:45:55PM +0000, Ben Bell wrote: Though that said I've just spotted the Banana Pi M3 which boasts eight 2GHz cores and 2GB of RAM. That certainly qualifies as a bit beefier. It wouldn't be able to run anything like PianoTeq, and I wonder about the reliability of low latency kernels and so on on ARM chips, but still..

On Sun, 6 Mar 2016 11:31:11 +0100 Jeremy Jongepier &lt;jeremy(a)autostatic.com&gt; wrote: Another possibility is the Odroid C2. http://www.hardkernel.com/main/products/prdt_info.php?g_code=G145457216438 -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.

On Sun, 6 Mar 2016, Jeremy Jongepier wrote: Right. I was looking at the odroid as the brain for my controller project (think no audio). In the end I chose the RPi because there are so many more out there and support is more current. -- Len Ovens www.ovenwerks.net

On Fri, 4 Mar 2016 19:56:18 +0000 Ben Bell &lt;bjb-linux-audio-user(a)deus.net&gt; wrote: I'm using Komplete Audio KA6 - 4x analog I/O + 2 digital + phones amp + MIDI Front 2 inputs are Mic/Instrument with phantom power on Mic. -- Will J Godfrey http://www.musically.me.uk Say you have a poem and I have a tune. Exchange them and we can both have a poem, a tune, and a song.

On Sun, 6 Mar 2016, Patrick Shirkey wrote: ... I would not pick the i7 for this use for anything other than experimental use. The thing to remember is that these are not desktop processors, they are mobile processors. The i7 mobile... is two core only. The whole thing with mobile intel cpu is heat. Heat limits what can be put on the chip to two cores with ht. Heat limits what performance you can get out of what chips are on the board too. So to pick out the best NUC for a workhorse stage box.... look at power rating of the chips vs. speed. The last .1 or .2 Ghz speed may up the power use by as much as 25% (don't even look at Boost speed because for audio work you will not ever use it) Understand that for live work you might want to turn HT off and loose that as well. (test first though) Do test with a cpu hog what temperature the CPU gets to after 10 minutes and see if you are getting xruns due to speed cuts for heat. In the end you may need to set the CPU Governor to UserSpace rather than performance so highest temperature a CPU hog gets you is right around 60-65C just to be sure stage work is not over heating things. While it is true audio will never use 100% cpu (without a really bad sound :) ) Stages can be warm places... particularely places a little box like the nuc might get stuffed. I would definately run a temperature logger. The NUC has a nice box, but do not be fooled, this is not made for stage work, it is not designed for outside use, or to be thrown into a box when packing up, being stepped on, kicked, or otherwise abused. It's connectors (as discused in other similar threads) are not road worthy. A larger box that allows extending the i/os to better connectors on the box would make sense to me. A box made of metal... or 3/4 ply. Personally, I would tend to a slightly bigger MB at the same time with more options. I would see if I could get a true 4 core CPU like an i5 which generates a lot less heat than an i7. and build it for the road. What no one else has said, building ones own road kit will cost more than a prebuilt solution. One has to be aware of this going into the project. If cheap is the main reason... that says something about how much one cares about their performance too. I think putting together an open source kit is wonderful, but not to save money. It should be about being able to bring a better performance to your audience. -- Len Ovens www.ovenwerks.net

On Fri, Mar 04, 2016 at 10:24:52AM +0100, Atte wrote: Yes, that's important. It's good advice, but it's sort of the other direction from where I'm going at the moemnt. My current live rig's already quite extensive (Hammond XK3c, Yamaha P90, original Nord Stage Compact, Prophet 08 rack and a couple of other old and special S&S racks). I'm really trying to figure out how to upgrade and consolidate at the same time. Replacing both the P90 and Nord Stage Compact with a Nord Stage 2 EX is probably the sensible choice, but the hacker in me is railing against it and trying to figure out whether there's a way to do something a bit different. bjb