Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[JimH]

I started a poll yesterday on a proposed Linux player based on the NUC.  There was a little push back, so I thought I might explain more here.

Our goal with this is to create a family of products that compete in the space occupied by Sonos, but also do video and TV.

I believe that most of the beta team is HTPC focused.  Most are quite comfortable with PC hardware.

Most of our customers are still audio centric.  That's understandable because it's where we started.  We also got a big boost from a couple of vocal proponents in the audiophile community.  This led to high end audio manufacturers recommending JRiver.  There must be 50 such companies now, and they each have significant communities.

I'm stating some of this as fact when I don't have good stats to base it on, but it is based on a general observation of users posting on the MC board.

When people purchase, we ask them how they heard about us, and I watch these closely.  Most people don't answer, but today's data has the following:

Various Internet Forums/Apple appstore
Researching via internet
Audio magazine
Audio Shop
Mytek
Friends
Audio Store (AVWORX)

You can see the audio tilt.  It's typical.

When we've asked people how they use MC, video is important, so I know we have significant penetration in that niche.

So back to hardware ... My vision is that MC will be used in a home to organize and consume all types of media, anywhere, any time.  Any device will be able to play media from any other, and any device will be able to control playback.

We can't get there, in my opinion, by relying on customers to configure computers.  There must be parts of the puzzle that are totally plug and play.  That's why I think hardware is part of JRiver's future.

We've dabbled a few times in this and it has been generally encouraging, but we were relying on others to do the hardware, and quality and availability were a problem.  Even relying on the big consumer electronics manufacturers has been disappointing.  They make DLNA Renderers that are rigid and don't work well.

One reason I started the poll was to try to stimulate discussion and to test ideas.  So far, it's been very helpful.  Thank you.

[babgvant]

FWIW, I think it's a great idea.

[csimon]

I think I was put off by the thought of a $400 DLNA renderer with a fan. There would be more weight to calling it a MC Client!

What Sonos does extremely well is user-friendly zone control and synchronisation between zones. At the moment, there's no easy way to doing that with DLNA, USB DACs or even onboard audio. You need to crack that to make a road into an all-house solution.

And dare I mention the central server issue again - with regard to any device being able to play from and control any other, it can be terribly confusing with the myriad of library servers you eventually need to enable and which machine you have to connect to using, say, Gizmo or JRemote, I'm getting really tangled and I only have two PCs! In the ideal all-house scenario, there would be one central library/database server which you always connect to via a remote, it collects details of all players/zones on the network and is able to send commands to each...or each player can be used with its own interface to browse the server. I guess this is similar to how a DLNA network works??  The player (client) does all the audio and video processing, no need for transcoding on the server side when using MC as a renderer.

Another problem you've got with the Synapse is working out what should indeed be plug 'n go, what bits you can configure as default and it'll "just work". Are you thinking that it'll be pre-configured as an MC and DLNA client therefore will just appear on the network and can be used straight away?

[JimH]

Another problem you've got with the Synapse is working out what should indeed be plug 'n go, what bits you can configure as default and it'll "just work". Are you thinking that it'll be pre-configured as an MC and DLNA client therefore will just appear on the network and can be used straight away?

Yes.

[babgvant]

I think I was put off by the thought of a $400 DLNA renderer with a fan.

The fan is almost completely silent on that box even when running prime95 on it. $400 for that system isn't bad as long as they can get HWA video going on it (eventually). It will never have the necessary GPU for madVR style rendering though.

I wouldn't pay $400 for a DLNA renderer personally, but that more because I don't need one (I have JMC already running at my only set of decent speakers ). I do see a market there for this type of device assuming that it Just-Works®.

[InflatableMouse]

We can't get there, in my opinion, by relying on customers to configure computers.  There must be parts of the puzzle that are totally plug and play.  That's why I think hardware is part of JRiver's future.

Your solution is called Glynor. Hire him and all your worries dissapear 

[mwillems]

I thought the concept was a good one overall; the main reason I voted "no" in the main thread (personally) was the price point and the lack of video support.

Admittedly, I'm an HTPC user so I'm obviously not the target demographic 

I do have some audiophile friends, though, who might be in your target demographic.  They do want an "audio-only control point in a box," especially one that you could just plug in and have it work.  But the hitch is they (like many audiophiles) want nice analog audio outputs (nice DAC, balanced outputs) on whatever that box is.  So they're willing to pay 2k for an Oppo player, because it has the nice analog outputs (and plays SACDs, etc.).  

I've tried to convert a few of these folks to JRiver, and the main obstacle has not been the lack of a "pre-fab" JRiver box; it's been the quality of the analog outs on most computers. They don't want to buy a dedicated computer and an external DAC, they just want a box (preferably a rack mountable box) that works and has balanced outputs and a nice DAC in it.  They also want integrated optical disk support.  One of my friends would probably buy JRiver/a JRiver box tomorrow if it could accept an SPDIF input and had nice analog outs (he would be using it to replace a DSP processor and a few other boxes in his audio chain).

So this product wouldn't be reaching them; it would either need to be much more expensive (with corresponding hardware), or much less expensive (to take the sting out of having to buy a DAC and external CD/SACD player).  

That's my two cents, but, obviously, I'm not in the business. The audiophile community is pretty heterogenous, and I'm sure there are folks out there for whom this product would scratch just the right itch.  I just thought I'd share my experience trying to sell to my audio-only friends  

[cncb]

Can't some of those SOCs run Linux such that you could put MC on there and it would be much cheaper?  What about the Raspberry Pi?

[JimH]

We could do cheaper, but I don't think we could do better.  CPU power would be limiting on some of them.

We've tested a few cheap boards.  Some worked.  The problem with some of them is that we'd probably need to start over with a new board periodically.  Case, power supply, etc.  The NUC is an appealing package to build on.

[jmone]

My vision is that MC will be used in a home to organize and consume all types of media, anywhere, any time.

I've always loved this vision and for me it ticks all the boxs:
- It is enduring
- It can not be done overnight
- It allows me to bang on about all the stuff I want (and have for years)!

Back to the HW.  I see three different pieces of JRiver HW:

Sub $100 "Sticks" - Market is for the Cheap, easy and just works over "quality"

Based on Android & Gizmo, and needs to support:
- HDMI
- Wireless N
- Analogue Out
- USB Ports
- Micro SD Cards
- Ability to optionally support a mouse/keyboard
- Ability to optionally support wired Ethernet (through a USB dongle)

Gizmo Development Required:
- Local File Playback
- Gizmo can be "pushed to" (eg DLNA Renderer) so you can see the Gizmo instance from an MC / Other Gizmos and push content to it



approx $500 "MC HTPC in a Box" - Market is for those that want a preconfigured full MC environment who don't want to build their own but offer full "quality"

Based on MC and Windows / Linux (when it does Video) support:
- Single SSD / msata
- HDMI
- Analogue Out (quality DAC?)
- USB Ports
- IR Port
- Optionally comes with a RC/mouse/keyboard
- Optionally comes with a BD Drive
- Wireless N and Ethernet

MC Development Required:
- Windows base : would need to get and supply OEM Windows Licences
- Linux base : would need to support Video

The NUC is a good candidate as you can get it with various CPU specs and the the rest of the requirements are built in.  It would be worth testing to see how good/bad/indifferent the DAC in these boxs are.  They may be pretty good???


approx $500+ - pending config "MC Server in a Box" - Market is for those that want a preconfigured full MC environment to serve media who don't want to build their own but offer full "quality"

Based on MC and Windows / Linux (when it does Video) support:
- One Drive std but with support for adding multiple drives
- HDMI
- BD Drive
- Analogue Out
- USB Ports
- Optionally comes with a mouse/keyboard
- Wireless N and Ethernet

MC Development Required:
- Windows base : would need to get and supply OEM Windows Licences
- Linux base : I don't know if the current version supports the streaming/transcoding of Video

[jmone]

$? - Audiophile "ID" - Market is for the Audiophile that just wants the highest quality Audio.

Based on MC and Linux, and needs to support:
- Single SSD/msata
- High Quality DAC with fancy gold plated connections
- Some special isolated PS
- Must look incredible (high quality Aluminium, Unusually Heavy)
- Ability to support various other std connections such as HDMI, Wireless N and Ethernet, RC/mouse/keyboard etc

Development Required:
- None unless you want to invent some new special cables!

Pricing: Use the old pricing method.  Stand in front of a mirror and keep saying a price, go higher, repeat until you laugh.  Take the previous price!

[Hendrik]

Pricing: Use the old pricing method.  Stand in front of a mirror and keep saying a price, go higher, repeat until you laugh.  Take the previous price!

I thought the last step would be "Double this value!"

[jmone]

the trouble is when describing this new breakthrough device you will start laughing so you are stuffed (unless you can learn to fake sincerity)!

Here is the link to the build of the pics I posted above.  It is a bit old but they even recommend MC!

[JimH]

the trouble is when describing this new breakthrough device you will start laughing so you are stuffed (unless you can learn to fake sincerity)!

That's why need an S&M department.

The CAPS PC's are specified by Computer Audiophile's Chris Connaker.  He's local and we know him pretty well.  He's a very smart guy.

[jmone]

Then to me, that is your answer for the "Audiophile" Build.  Team up with him on what the build should include (his list sounds good and the presentation is great).  Get him to do a review/benchmark etc and you are off the the races!

I do like the overall idea of the JRiver prebuilt solutions for each of these areas and I think they will do well.  The only feedback I have on the "ID" (which a few of us bought) is that looked and felt "cheap", presentation is everything (just ask Apple).

The only "bits" you are really missing dev wise the the Gizmo changes at the low end, and video support if you want to use the Linux build for any of the scenarios. 

Support: Of course you will be up for all the HW support/wty requirements as a bunch of Beta Teamers won't cut it when the "In A Box" concept no longer "just works" as this is aimed at users without the technical knowledge that those who have built their own already have.  Again I'm thinking Apple Genius that take over the device and "fix" it for the users.  You may need want to build in some form of Remote Support Sessions for this.

[astromo]

Definitely a sensible next step that mirrors successful strategy that historically has worked well, when executed well.

I've said this out on Interact (somewhere?) in various spots but here's the most recent:
http://yabb.jriver.com/interact/index.php?topic=87784.msg601346#msg601346
In short, when I started on this whole computer based media journey my objective was to digitise my CD collection. Video was a secondary consideration.

I had a look for a one stop box that would rip, store and playback at a quality audio standard. It all got too hard and the price tags (especially out here in Australia where we get gouged) for kit was nuts, so I was prompted to go down the do-it-yourself path. I also took a view that I'd limit the advice I'd work with and took a view that the info on http://www.computeraudiophile.com, overall, made sense to me.

Then, as my understanding of MC grew, I made a concerted effort to standardise to one media platform, so movies and TV followed along. This development path echoes some of Jim's comments earlier. But if there was a sensible piece of hardware on offer that ticked the right boxes, I would have been strongly swayed to buy a boxed package.

I don't have golden ears and I don't see myself as a quirky audiophile (I'm in the bits are bits category and my engineering heritage grounds me with the scepticism that comes from applying science, which is why I really love the work mitchco does) but my take from what I've osmosed from places like CA are that a semi- or full audiophile would be looking for a system that includes:

• Quiet, preferably fanless
• Bit-perfect audio
• HiFi standard audio connection options (both digital and analogue). The approach that NAD has taken is worth bearing in mind http://nadelectronics.com/products/hifi-mdc-modules that gives the user a build menu approach with upgrade options for improvements that may come along in the future.
• Optical drive for ripping
• Preference for a VFD display or similar (though I'd agree with Jim that there is a mind shift going on being carried by smart phones and I'd push that along purely from the same point of view as Henry Ford, i.e. "What you don't fit costs nothing and needs no maintenance")

Carrying on from that base list of functional requirements, the ability to offer expansion of the platform to other media areas obviously makes sense along with distribution of media both locally and remotely. I'd also say that family useability is a priority too. I steered away from options like Vortexbox because I didn't want to learn a new bag of tricks and probably more importantly, didn't want to deal with aggravation from my significant other half.

I think jmone has put together an excellent graphical vision. The trick, is to pick the right cornerstone to work from.

If the audio user is the customer base that you want to build from (sounds sensible), unless I've missed it, a shortcoming of the NUC platform is that its audio out is limited to a 3.5mm stereo jack or HDMI/Display Port based on the latest offering here:
http://www.intel.com.au/content/www/au/en/nuc/overview.html
Refer my list above, an entry level audiophile will be looking for HiFi standard connections. I've read a lot of negative press around HDMI (that I don't believe by the way) and from what I know, a lot of HiFi gear doesn't have HDMI in, so a fair proportion of the prospective customer base will be looking at a USB DAC or a USB/SPDIF converter. Many will self cater but I really think you'd get nods of approval out in the crowd if you provide connection options that sound "quality".

Also, jmone's comment that gear that looks pro will work at a customer perception level is very sensible. The audiophile standards are good old black and brushed aluminium. The standard NUC plastic case won't necessarily win over an audio crowd.

[JimH]

I've used HDMI a lot.  It has some problems but has generally worked for me.  It's a solid digital connection to a receiver. 

Audio out on the NUC may also be digital.  It can be done on 1/8 stereo connectors.

A USB DAC would also work.

[jmone]

Re-reading Jim's original post.....  Sonos really leads with Speakers that have all the tech mounted inside the enclosure.  So.... if the plan is to compete with Sonos do you plan to team up with a Speaker MFR?  If so you really just need stuff like what was in the old "ID" and none of this is really "high quality" (but it does look good!).

Edit: What we have been discussion is a different set of products to how Sonos goes to market.  It is more like "I've graduated from Bose Cubes to something decent"

[wombat66]

...

What Sonos does extremely well is user-friendly zone control and synchronisation between zones. At the moment, there's no easy way to doing that with DLNA, USB DACs or even onboard audio. You need to crack that to make a road into an all-house solution....

Agreed!

To me, synchronization between zones is the one thing that makes Sonos appealing to me (oh, plus their Spotify integration).

[fitbrit]

Agreed!

To me, synchronization between zones is the one thing that makes Sonos appealing to me (oh, plus their Spotify integration).

Could that be achieved by using a mic built into the hardware and actually 'listening' to what's playing in the zones - introducing delays or advances until it matches? With different delays introduced by various DACs, receivers networking cards, etc. the only solution might be to monitor the end product - the actual sound made by the speakers. Any cross talk from neighbouring zones would have to be ignored of course, or even used as a reference.

[csimon]

What Sonos does extremely well is user-friendly zone control and synchronisation between zones. At the moment, there's no easy way to doing that with DLNA, USB DACs or even onboard audio. You need to crack that to make a road into an all-house solution.

See http://yabb.jriver.com/interact/index.php?topic=87986.0!

A DLNA renderer that does synchronised playback could actually be a huge selling point. However, until other DLNA controllers and servers get up to speed, it would be have to be a JRiver ecosystem anyway. But you could be a trailblazer! Presumably, this would all be done in the core MC product so existing owners could also take advantage.

[mwillems]

Some kind of robust syncing algorithm (especially one that uses a mic for a timing reference) would potentially add value even outside of the DLNA/streaming context, and could serve as a building block for other functionality.  

I see about a question a month on the forums about syncing several stereo DACs using zone link to create a virtual multi-channel soundcard (for surround sound, etc.).  At the moment that's not really a feasible option for most of the USB DACs out there. Some DACs natively support syncing, but most of them (especially the lower cost ones) don't currently natively support close syncing. If JRiver had a method of syncing multiple zones closely (it would need to be a pretty tight tolerance, talking about probably well less than a millisecond), then you'd be offering functionality that doesn't (to my knowledge) exist anywhere else, and would allow people doing multichannel a lot more flexibility.  It would also be particularly helpful for the folks with high channel counts (11.1 systems, etc.) who currently have a hard time finding interfaces with enough channels.  If you could get the sync tolerance very, very tight using a microphone (now probably talking about a tolerance of a couple of micro seconds), it would also potentially allow for bi-amping with multiple USB DACs that couldn't be synced otherwise.  

Very tight tolerance syncing like that would also potentially provide one of the building blocks necessary to allow JRiver to do certain kinds of automated room correction (i.e. hear the delay from the different channels in real-time and adjust timing so that all channels are perfectly synced).  In effect, it could contribute to functionality that would allow JRiver to do some or all of what Audyssey does, but do it better because it could be adjusted in real time (I think Matt had previously expressed some interest in doing something like that in JRiver). Obviously the development investment required to get something like that going would be considerable, but having a mic-based zone-syncing algorithm in place would reduce the unique investment required to get something like that off the ground.

This is speculative, obviously, and some of it would require significant additional development (beyond the already significant development required to get close syncing), but I wanted to point out that a tight tolerance syncing algorithm based on mic feedback could have potentially marketable applications outside of the DLNA context.

[jmone]

Re: The Quiet Fan thread.  You are not going to win expose the truth about the "Audiophile" argument that says any moving part is bad.  As I suggested in the builds above, I think the NUC is a good choice for the "HTPC in a box" but not for the "Audiophile ID" product.  Any the fan is just part of it.  For these guys they want to "know" it is excludes compromises and they will pay extra for it.

[Hendrik]

Audiophiles probably don't need the performance the NUC offers either, so they could get away with less horsepower, but passive design instead.
If you assume they just don't get video features, that is, or only basics at least.

[JimH]

I'm not trying to win an argument.  I'm trying to expose the truth.

There are audiophiles and there are audiophiles.  We're not going after the $200 USB cable buyer.

[JimH]

Audiophiles probably don't need the performance the NUC offers either, so they could get away with less horsepower, but passive design instead.
If you assume they just don't get video features, that is, or only basics at least.

I'd like to find a device that would also support video when we're ready.

[jmone]

I'm not trying to win an argument.  I'm trying to expose the truth.

Fixed my post 

There are audiophiles and there are audiophiles.  We're not going after the $200 USB cable buyer.

Sure but I'm not suggesting anything "stupid" either.  Even the advice from Chris Connaker advocates a fanless build and it is just about the right case.  Here is an example of one for NUC from Tranquil PC (who I used to get stuff from) and I'm sure there are others.

& Yes.... it should have enough power that you can "turn" on Video ... so the 2820 would be the minimum.

[jmone]

Here is the pic of the chassis for NUC made by them (you could offer it as a fanless upgrade option).

[jmone]

They even offer an OEM Service - http://www.tranquilpcshop.co.uk/nuc-h2-oem/ for customisation....

[Hendrik]

That would be neat, shiny metal case, a logo laser edged into the unit.
One can dream...

[csimon]

<wipes screen>

[jmone]

It does look good and is not badly priced at under "100 Pounds" and ticks most boxs the audiophiles want.  This example is made for other NUC formats (not the 2820) but I was thinking the concept would be:
- The MC-HTPC-NUC package US$500: Uses the Intel NU2820, preconfigured etc etc
- Audiophile Chassis Upgrade Option US$250: Something like the above (eg just replaces the Intel plastic box with the Aluminium one - sans fan).

Even with the above - what about the DAC?  Do you
1) test the Intel inbuilt DAC
2) try to bundle one into the chassis if the stock is not that great
3) or just let them pick their own external DAC

I'm guessing Point 1 or 3 makes the most sense.

[Hendrik]

The Onboard DACs are definitely not audiophile quality, its your standard low quality Intel HD Audio chip.
I couldn't find an actual spec on the one used in the NUC, but I doubt you can rely on it for high quality, not to mention that its stereo only (although for Video use, most people will probably use HDMI instead)

[jmone]

Yup the HDMI works for most, so I guess it would be #3 for the audiophiles.

[cncb]

Maybe you should consider not using the term "audiophile" at all since it may attract somewhat unreasonable people like the guy in the "quiet fans" thread?

[JimH]

A DLNA Renderer for the Complete Slacker?

[cncb]

Someone will post that they thought it came with Slacker Radio.

[cncb]

How about "A DLNA Renderer for the Reasonable Audiophile (No Offense)"?

[jmone]

FYI - I contacted Tranquil-PC to see if they make a chassis for the N2820 NUC but they don't (at this stage), the Pic I posted is for the i3 / i5 NUC kits so it starts to add up...

[jmone]

Q for Jim - what are you thinking for the JRiver HW platform at this stage?

[AndrewFG]

The Onboard DACs are definitely not audiophile quality, its your standard low quality Intel HD Audio chip.

Far worse than the "quality" of the chip, is the matter of the environment where it is sitting: If you have a DAC sitting on a PC MoBo, it is surrounded by CPU, video, Ethernet and RAM chips that are pumping out high frequency noise; it gets its power from a chopped DC converter; and it is sitting next to a fan (dare I mention fans again) that throws out a rotating chopped E/M field. In such an environment, even the best DAC chip in the world would be dead in the water...

[jmone]

I'm so sorry!  I can not believe with all this talk of HW and Whole of House media I dropped the ball and did not mention rubberduckies once!

[JimH]

Split multi-room hardware

[rpalmer68]

I'm so sorry!  I can not believe with all this talk of HW and Whole of House media I dropped the ball and did not mention rubberduckies once!

Or Intercoms, gee Nathan you must be getting slow in your old age.

[jmone]

Mmmm me thinks I smell some editing from a admin with dry wit 

[BryanC]

How about something based off of the Intel MinnowBoard?

[JimH]

Can't some of those SOCs run Linux such that you could put MC on there and it would be much cheaper?  What about the Raspberry Pi?

Old post.  I know, but I'm re-reading this thread.

Raspberry Pi is ARM, but Intel may have one we could use.  I was looking at this today:
http://www.ecs.com.tw/LIVA/index.html

At the low end of hardware, the priece difference between ARM and Intel is probably not going to make or break the Id.  If we can prove the concept on Intel, we can move to ARM later.

[JimH]

I've always loved this vision and for me it ticks all the boxs:
- It is enduring
- It can not be done overnight
- It allows me to bang on about all the stuff I want (and have for years)!

Back to the HW.  I see three different pieces of JRiver HW:

Sub $100 "Sticks" - Market is for the Cheap, easy and just works over "quality"

Based on Android & Gizmo, and needs to support:
- HDMI
- Wireless N
- Analogue Out
- USB Ports
- Micro SD Cards
- Ability to optionally support a mouse/keyboard
- Ability to optionally support wired Ethernet (through a USB dongle)

Gizmo Development Required:
- Local File Playback
- Gizmo can be "pushed to" (eg DLNA Renderer) so you can see the Gizmo instance from an MC / Other Gizmos and push content to it

http://static.businessinsider.com/image/4ff3041c69bedd5941000004-400-300/equiso-smart-tv.jpg

approx $500 "MC HTPC in a Box" - Market is for those that want a preconfigured full MC environment who don't want to build their own but offer full "quality"

Based on MC and Windows / Linux (when it does Video) support:
- Single SSD / msata
- HDMI
- Analogue Out (quality DAC?)
- USB Ports
- IR Port
- Optionally comes with a RC/mouse/keyboard
- Optionally comes with a BD Drive
- Wireless N and Ethernet

MC Development Required:
- Windows base : would need to get and supply OEM Windows Licences
- Linux base : would need to support Video

The NUC is a good candidate as you can get it with various CPU specs and the the rest of the requirements are built in.  It would be worth testing to see how good/bad/indifferent the DAC in these boxs are.  They may be pretty good???
http://yabb.jriver.com/interact/index.php?action=dlattach;topic=87901.0;attach=12481;image

approx $500+ - pending config "MC Server in a Box" - Market is for those that want a preconfigured full MC environment to serve media who don't want to build their own but offer full "quality"

Based on MC and Windows / Linux (when it does Video) support:
- One Drive std but with support for adding multiple drives
- HDMI
- BD Drive
- Analogue Out
- USB Ports
- Optionally comes with a mouse/keyboard
- Wireless N and Ethernet

MC Development Required:
- Windows base : would need to get and supply OEM Windows Licences
- Linux base : I don't know if the current version supports the streaming/transcoding of Video

http://www.lian-li.com/en/files/2013/01/PC-Q02-color.png

#2 and #3 are probably going to be the same Id.  Keeping the functionality all in one package is important.  

Even shucky duckies have feelings.

[glynor]

http://techreport.com/news/25344/move-over-atom-intel-quark-is-one-fifth-the-size
http://www.intel.com/content/www/us/en/do-it-yourself/galileo-maker-quark-board.html

Intel has been... Slow, certainly, on their move to truly developing chips appropriate for product classes where ARM now competes.  Their attempts in this space have been largely flops, and since the death of the "netbook era" they've had essentially no real, competitive answers in the Tablet space.  And, it was really no surprise.  After all, those low-cost devices stood in basically polar opposition to their profit motives (to push customers, and the market as a whole, towards larger, and more expensive, chips).

But... No more.  They've found the faith, it seems to me.  (And, I wouldn't be surprised at all to see the A8 or A9 fabbed at otherwise idle Intel fabs).

And so... Where does that leave us?

I don't know.  Processor manufacturing is not easy.  It requires billions of dollars of investment, and there is a huge benefit in cost structure and performance from better, smaller process (and the ability to more quickly move to even smaller process nodes).

And Intel has the best fabs.
And now they've decided to do it.

They've been behind before, with the Athlon 64 and Northwood.  And they decided to do it, and AMD can't do anything to compete, has sold off its fabs, and is a shell of their former selves (if it wasn't for their ATI business, they'd be gone and folded and in someone's IP portfolio by now).

We'll see.  It is still a challenging market, and there is an entrenched competitor.  I don't see Samsung or other manufacturers that have their own Fabs moving, unless forced.  But... They have the best fabs in the world.  And not just by a hair, they're years ahead of their nearest competitors (and their nearest competitors are generally not in the ARM space, at least as far as process tech).

I would not bet against Intel, just because they've been slow thus far.  They may be a lumbering giant, but they're still a giant, and they have the biggest club of them all.

[astromo]

Old post.  I know, but I'm re-reading this thread.

Raspberry Pi is ARM, but Intel may have one we could use.  I was looking at this today:
http://www.ecs.com.tw/LIVA/index.html

At the low end of hardware, the priece difference between ARM and Intel is probably not going to make or break the Id.  If we can prove the concept on Intel, we can move to ARM later.

Interesting:
www.ecs.com.tw/LIVA/feature.html
With this concept it's a lot easier to claim portability given that it can be run off a battery pack without plugging into A.C. Wonder how long it will run on a "reasonable" sized battery?

Good to see more options in this area. If the hardware is so cheap, looks like the right place to be is in software.