The Hardware
NVIDIA’s SHIELD mobile Android gaming device is here!
Dear NVIDIA,
It has come to my attention that you are planning on producing and selling a device to be called “NVIDIA SHIELD.” It should be noted that even though it shares the same name, this device has no matching attributes of the super-hero comic-based security agency. Please adjust.
When SHIELD was previewed to the world at CES in January of this year, there were a hundred questions about the device. What would it cost? Would the build quality stand up to expectations? Would the Android operating system hold up as a dedicated gaming platform? After months of waiting a SHIELD unit finally arrived in our offices in early July, giving us plenty of time (I thought) to really get a feel for the device and its strengths and weakness. As it turned out though, it still seemed like an inadequate amount of time to really gauge this product. But I am going to take a stab at it, feature by feature.
NVIDIA SHIELD aims to be a mobile gaming platform based on Android with a flip out touch-screen interface, high quality console design integrated controller, and added features like PC game streaming and Miracast support.
Initial Unboxing and Overview of Product Video
The Hardware
At the heart of NVIDIA SHIELD is the brand new Tegra 4 SoC, NVIDIA’s latest entry into the world of mobile processors. Tegra 4 is a quad-core, ARM Cortex-A15 based SoC that includes a 5th A15 core built on lower power optimized process technology to run background and idle tasks using less power. This is very similar to what NVIDIA did with Tegra 3’s 4+1 technology, and how ARM is tackling the problem with big.LITTLE philosophy.
Speaking of Tegra 3, the new T4 processors claim to have 6x the graphical performance of T3 thanks to 72 GPU cores and a clock speed as high as 1.9 GHz. We’ll have some graphics and processor performance tests later in the article that will showcase the performance jump from T3 to T4.
It should be noted that even though I don’t have power consumption numbers on Tegra 4, the NVIDIA SHIELD does utilize an active cooling solution. A fan spins up during HD video decode and gaming scenarios to exhaust heat out the rear of the unit. Though that might sound startling to someone that is used to passively cooled ARM-based phones and tablets, the noise of the fan was never a distraction in my use even while resting SHIELD on my chest and watching a movie without headphones.
SHIELD ships with 2GB of low power DDR3 system memory and 16GB of internal flash memory. A microSDXC slot resides on the rear of the unit as well for storage expansion up to 64GB with current card availability.
The display is a 5-inch 1280×720 multi-touch screen that has a surprisingly high DPI: 294 ppi. The quality of the screen is damn good and even up close I didn’t have any issues with pixel pop. The orientation of the display is fixed in landscape mode which is best for gaming but not for a lot of other Android applications. NVIDIA addressed this with an OTA update that allows portrait apps from the Play Store to install but holding the controller in that way is rather frustrating.
Another aspect of the design that NVIDIA is very proud of is the audio implementation. Because SHIELD has a deep cavity design it allowed NVIDIA to use a port audio system for more bass than you would typically be able to get out of a tablet or phone design. The forward facing stereo speakers are actually quite good and I found it was decent as a “portable” music device around the house or the office. Obviously these are no substitute for a good set of headphones or external speakers but for a mobile device, they are top notch.
Network connectivity on SHIELD is provided by a solid 2×2 MIMO 802.11n WiFi controller that works at either 2.4 GHz or 5.0 GHz. This allows SHIELD to connect to the much less crowded 5 GHz space and is actually recommended by NVIDIA for PC streaming use. SHIELD also includes Bluetooth 3.0 support, and a GPS just incase you happen to be lost in the woods with your SHIELD. If you have to, a micro-USB to Ethernet adapter can be used when performance is critical.
On the back of SHIELD you’ll find a handful of connectivity options that allow for expansion. A mini-HDMI connection allows you to directly connect it to a TV or entertainment system for full 1080p video support (though NVIDIA claimed 4K support at the CES press event). A micro-USB 2.0 port can be used for accessories or data transfer from a PC, but is most likely going to be utilized for charging with the included 10 watt wall adapter. A 3.5mm headphone jack can be used for headphones and also supports a microphone so you can use a headset for Skype calls, etc.
Just like all phones and tablets, NVIDIA SHIELD integrates a 3-axis gyro and accelerometer.
The controller on SHIELD is probably the most important feature as it will be the most used interface to the consumer. Though it is a combination of the Xbox 360 and Playstation 3 designs, the button layout resembles the 360 controller quite closely – which is a great thing considering its dominance as the controller of choice for PC gamers. The D-pad is decent though a bit mushier – kind of like the Xbox 360 controller in that regard as well. Using the thumb sticks was a positive experience although the acceleration at the mid-point seems to pick up very quickly.
Other buttons on the SHIELD include volume, back/return, home, start, and the SHIELD button. The volume is controlled by tapping the button then using the touch screen context menu to raise or lower the sound levels. Back and home buttons perform just as they do with any other stock Android device. The start button is used for selective context usage in Tegra-based games and PC streaming emulation. Finally, the SHIELD button is used for powering on and off the device as well as entering the TegraZone sub-menu to access games and PC streaming.
The battery used in the SHIELD is rated at 28.8 Whr, and in my testing was able to get about 16 hours of web surfing time over WiFi. We are still doing more testing on game play times, but with the variance in usage from the titles we are testing with, we need more time to compare SHIELD to other Android gaming platforms.
A big plus for NVIDIA SHIELD is that it is a stock Android device in nearly every way, therefore you can expect to see updates from NVIDIA for newer iterations of the OS sooner than most other phones or tablets. Hopefully this means that we’ll see an upgrade to Android 4.3 soon, putting the SHIELD just behind Google’s own Nexus devices.
Setup and First Impressions Video
I am actually looking forward
I am actually looking forward to this, we need something to take over the portability area, the PS Vita “idea”, and the DS doesn’t have great graphics.
I think this would be a winner provided Android could make enough quality games….and the spectrum that was “gameboy/DS”
Mechanical or rubber membrane
Mechanical or rubber membrane buttons? Production shots make it seem like it is rubber membrane. http://blogs.nvidia.com/wp-content/uploads/2013/01/someassemblyrequiredweb.jpg
Membrane makes a lot of
Membrane makes a lot of sense. There are only a handful of mechanical gamepad switches that I am aware of… a knockoff ALPS switch for some Guitar Hero strum-bars and some arcade sticks (well, buttons on the arcade stick… not the stick itself).
Apparently Razer makes at least one mouse with mechanical buttons… which also surprises me… but meh.
All three variants of Razer’s
All three variants of Razer’s 360 controller (Onza, Onza Pro and Sabertooth) have all used mechanical switches for the face and shoulder buttons, but uses rubber membrane for the directional pad.
I have no idea what switch type the Vita uses, but it sure feels nice, with a definite audible and tactile “click” with each press. https://d3nevzfk7ii3be.cloudfront.net/igi/45JHbSwE51rIAPQd.huge
They appear to be the same type of switch that is used in the DSi, which also has a very tactile clicky feel. https://d3nevzfk7ii3be.cloudfront.net/igi/Ive2yIEWZn2AUhYj.huge
A little too early for this
A little too early for this device. Should have done this with Logan on board. Would have made much more sense allowing games to run on processors that don’t have the thermal constraints they might have on a thinner form-factor.
A waste of money in my opinion.
personally i don’t think
personally i don’t think there is too early or too late with this device. yes it meant for gaming (android and pc) but the very purpose of this device is to showcase the capabilities of nvidia tegra processors and other nvidia tech at the same time (for example the pc gaming stream feature is something related to their GRID technology, so actually it giving the glimpse to the public and developer what Geforce GRID should like given the internet network is not an issue). also this device will not compete with the likes of of PS Vita/3DS or even other android console like many like to believe. it has the capabilities in terms of performance but what nvidia want the most i believe is for those company to use their chip instead competing directly with them.
Support for gaming related
Support for gaming related stuff in Android means that better games can be built. It doesn’t have to compete with any product. If they were able to get the Logan GPU to support this, then it would make the Shield an interesting product. Right now, I still can’t see who is it going to be good for. Being a niche product is good and all, but the niche itself is too small. nVidia will have to eat the costs like Microsoft did with RT. This stinks of desperation. This will remain a proof of concept until the 20nm chips are released.
ROFL…It was only 10mil to
ROFL…It was only 10mil to dev. They are already sold out. No idea how many they made but I’m guessing it was at least 100K run which means they already broke even (assuming 100/device, based on BOM of Vita $159 and 3DS $100.71). This can’t cost more than $200 vs. those two so they are making $100 easily IMHO. If they sell 100K they are at 10mil which is the cost to dev this thing.
I wish Ryan would have told if there is any difference when say, setting up the HDMI out to TV (is quality any different than over miracast??). I’d like to know that as this is a feature I’d want to use (a portable roku/bluray replacement etc).
20nm will be better but it has no effect on it’s ability to play games now. Clearly it does the job already and the list of PC streaming games will only grow all the way to xmas and into next year when they surely put out a kepler 20nm rev.
I’m guessing by xmas this will add 100mil to their bottom line. I can’t see how they won’t sell 1mil+ of these considering 65% of us own NV gpus and they are already sold out. Did they already sell 1mil? Who knows. I suspect we’ll see a price drop by xmas to $250 or so, so they can nix some xbox1/ps4 sales and possibly sell even more discrete cards because of it. Then do it all again for every soc release…LOL. Each year making a console sale even harder to fathom.
I think AMD will rue the day they signed up for SINGLE digit margins, with only barely into double digits as the process matures. So if AMD is lucky they’ll make 100mil in the next year (of course that isn’t profit for quite a while, as they had to dev both sony/ms chips – what was that cost? I’ve guessing far more than 10mil for shield), while shield will probably hit that in a few months by xmas and only get better stealing from console sales. AMD made a mistake. If margins on those were say 20% from the get go and moving to 35% over the 8yrs or something, then it would have been a great deal. But single digits moving to double sucks. Realizing they have about 6 consoles coming to compete makes AMD’s move even worse and NV’s low risk 10mil go at it a NO RISK device. Google/Apple both have consoles coming, madcatz, ouya, shield, wikipad, gamepop etc. There is a lot of people after MS/Sony/Nin and wiiu is already off 50% after xmas. Sales of it and Vita are awful at best. I expect the same after xmas for MS/Sony. Only the 3DS escaped the same fate because of selling before phones/tablets really took off on gaming. Now Vita/3DS are useless and Vita is seeing this (which is the better device of the two BTW…Again showing they’re dead).
Finally I think we get to wave goodbye to stuck in stone for 8yrs consoles 🙂 YAY.
Nobody special, I could not
Nobody special, I could not agree more. Jen-Hsun Huang is my new product development hero. I’ve been a product developer (in other fields) for many years, and have nothing but admiration for how quickly, beautifully and profitably (I must guess at the last, but the sell out is a great sign) they conceived and executed project Shield. If, as I think, this is their first full-on consumer device, it’s amazing to me how glitch-free and smooth the experience seems to be, from the reviews I’ve read. We should also bear in mind that this is the kind of thing nVidia MUST do if it’s going to maintain high margins — because supplying processors for Nexus devices or gaming consoles, both of which are sold at break-even, can never pay the bills. I read a fascinating article on the nVidia b,of called “How Shield was Built” and am preparing an article of my own in which I speculate on the marketing reasons that drove the development process. It’s tempting to think of this as a reaction to the OUYA, but I kind f doubt even nVidia could have turned around a product so polished that quickly. I agree with the comment that Logan would have provided an even better experience, but you gotta start somewhere, I think nVidia wants to dominate mobile gaming — I agree that Android and iOs are where it will happen — and think nVidia is off to a roaring start. Making the screen detachable seems to me to be the next big hurdle, but I doubt we’ll see that until Shield 3.
Nobody special, I wish you’d
Nobody special, I wish you’d tell us who you are, because you sound like Somebody Special who knows a lot about marketing!
“Support for gaming related
“Support for gaming related stuff in Android”
This alone made me laugh.
I have zero interest in this
I have zero interest in this as an Android device. My specific interest in this is only as a tethered device to stream games from my PC around the house.
Unfortunately, it lost my interest back when I saw that it was $350, the streaming games thing was only beta, and only about six games supported it.
at it’s current
at it’s current pricepoint…a guaranteed flop.
regardless of specs and price of parts, unless it’s impulse price of $99 or below, very few going to touch it.
flop or not it’s not that
flop or not it’s not that nvidia expect to make tons of profit with this device. and nvidia will not be stupid enough to price Shield so low just so they can sell the unit and making the mistake by competing with another android console maker directly
Well one day later it’s sold
Well one day later it’s sold out. That can’t be considered bad…LOL.
Microsoft did the same trick
Microsoft did the same trick the with the Surface RT, they severely limited initial supply to only 1-3 units per entire store for the first 2 months and claimed what a wild success RT was always being “sold out”.
Don’t be so gullible next time.
You have no proof NV did
You have no proof NV did that. Thus no proof I’m gullible 🙂
I never thought RT would sell with a 1.3ghz vs. what should have been in there 1.7gh T3+. I do believe shield will sell out of first production and probably second. Don’t forget it is far better than vita/3ds power wise, plays movies out to tv via hdmi on the go, plays all android stuff, etc etc…Titan still is pretty much selling every chip they make. The first run was 100K and sold out in hours. There are MILLIONS of 600 series owners and adding 700 series owners daily. All of these are candidates for this device. Considering the power (and the power of nearly all phones next year being about T4 level), devs will start aiming FAR above Vita/3DS on android. Android gaming will kill these 2 devices shortly, so their entire audience will be looking at android (they are already) instead of vita/3ds which devs are abandoning according to GDC2013 survey. When less than 3% of 2500 Devs are planning a game for your device and 60% are planning mobile games you’re dead.
NV hasn’t said they are sold out. It’s just an observation at newegg gamestop, microcenter (found one in stock in houston) etc. On NV’s site says 2-3weeks will ship, but on newegg says back in stock 8/5 supposedly. Until I see evidence of them NOT selling well, I think they’ll do ok. This device has way too many things over Vita/3ds to be a failure and consoles are 3 months off giving it plenty of time to sell before xmas. We will know in 3 months. I haven’t seen a bad review yet either. 8/10 of the buyers on newegg are verified owners (so they bought there). I’m surprised anyone took the time 1 day after it went on sale. I’ll wait for the Kepler model next year (have to purchase maxwell first anyway…LOL) but this thing will sell regardless, and in enough quantity to make NV some decent cash and further push us to mobile gaming on gamepads (which is the point, get us off consoles, thus using more of NV’s stuff).
Spoken like a foreigner to
Spoken like a foreigner to handheld gaming.
There are a lot of things we don’t even know about when it comes to the Shield. Very basic gamer things; how many times the controller reports inputs, what types of switches it uses, or even the commitment to long term software and hardware support.
Before the Shield, *all* controllers for Android have sported a loathsome report rate of once per frame, with an actual average of once every three frames. This is completely unacceptable for gaming purposes, and I’m not going to trust anyone’s review on how responsive it “feels.” I want hard numbers, and Nvidia is tight lipped as ever.
Throughout the entire process of the Shield, from project to final product, we have not been able to see anything about the inner workings of this device. We have no idea what sensors are being used for the analog sticks, or how long they’ll remain accurate. We have no confirmed information about what switch types are used for the buttons. We don’t even know how the directional pad compares to the timeless Sega Saturn directional pad. You know, things gamers care about.
And finally there’s the massive bugbear of how long Nvidia will offer support for the operating system. If it is anything like other Android devices, you’ll get three or four updates in the first six months for the operating system, and then it’ll be forgotten to the ages, while the 3DS and Vita will have support from their respective companies for years to come. I still get updates for the DSi, a five year old system. I’d wager that the Shield will never get 4.3 Jelly Bean update.
A “handheld” at nearly the
A “handheld” at nearly the price of upcoming new consoles themselves and within 3-4 months of their release?
What dufus at nvidia approved this price point?
This device is strictly for fanboys. That pricepoint is going to force a very steep price cut in the near future implying it’s a business failure, whether it is or not, and perception is everything.
A shame, as it doe s appear to be a very interesting concept but ridiculously overpriced.
@ nobody special:
if it’s
@ nobody special:
if it’s worth your yime defending it, how come you haven’t bought one, done a youtube unboxing and submitted your full review to PC Per?
(insert your lame excuse here)——–>
My Maxwell card comes first.
My Maxwell card comes first. At which point I think rev2 will be out. So no point in rev1 for me as I want to stream my PC’s maxwell. I currently have a radeon 5850. So a good chunk of this things use is not usable by me.
That is not a lame excuse. It’s reality. If I own an NV card I would have bitten already, then handed it to my nephew when rev2 comes along by june next year.
Likely developed as in in
Likely developed as in in house remote terminal for an Nvidia GPU/GRID equipped Steam Box, where it would actually make sense, it’s release into the wild indicates Nvidia ultimately lost out to AMD to provide that hardware, probably due to publisher/developer pressure to go with a more console compatible solution, an AMD HSA APU that they could port to easily and cheaply.
If it is AMD HSA APU based, and is supposed to be upgradable, then I would guess Kaveri + Sea Islands GPU. Though Steam Box is to be part of a larger Linux based ecosystem, whatever is in Steam Box will become the default hardware of choice for that ecosystem as Steam Box game ports will target that particular hardware. The beauty of HSA based APUs is future iterations will retain nearly complete hardware compatibility, so HSA based Steam Box game ports will be fully hardware compatible with future AMD HSA based APUs and GPUs. Not only makes the publishers and developers job very very easy, but moves them substantially toward their nirvana, a single future proof PC hardware standard to port to, so it would actually be in their interests to, over time, not only heavily optimize their ports for AMD HSA based hardware, but specifically NOT optimize their ports for Intel or Nvidia hardware. Death by (not so) benign neglect as it were … at the very least make Intel and Nvidia foot nearly the entirety of whatever optimizations do occur.
As for timing … if the above is the case it is unlikely Steam Box will arrive before Kaveri is available to meet that larger ecosystem demand, and as recent reports indicate Kaveri is delayed into early 2014, I would expect Steam Box to arrive sometime after … say E3, 2014 … they’d have the new console reveal center stage spotlight all to themselves and the time to get a healthy selection of their Linux based game and application ports ready to go. It would be hard to imagine a launch scenario better geared to ensure the success of Steam Box.
Several months ago a leaked (and quickly pulled) AMD developer PDF showed up on Semi-Accurate forums showing Kaveri with some 2014 roadmap ‘system integration’ features. Makes sense as Kaveri was on the same engineering/fabrication timeline as the custom console APUs, so it’s logical, and likely, knowing exactly what was going into both console processors, AMD simultaneously optimised Kaveri’s architecture based on that knowledge and pulled some 2014 GPU/context switching HSA elements into Kaveri and making sure Sea Islands was system integration ready. They would want all the HSA elements pertinent to porting to and from console games and future proofing fully in place with Kaveri. Kaveri’s delay might be related to such optimizations along with sharing console APU fabrication lines creating capacity issues. If Kaveri is delayed into 2014, I wouldn’t be surprised to see a ‘soft’ launch for Kaveri during the holiday season to put review site next gen game performance numbers on prospective buyers radar. If those If those gaming performance numbers on next gen games blow away the Intel/Nvidia competition on cost/performance like I think they will, it would be well worth AMD taking a ‘soft launch’ PR hit.
Such an optimised Kaveri would be utterly ideal for an upgradable, future proofed Steam Box and it’s associated third party ecosystem. It was pretty obvious at CES Gabe was intending an Intel/Nvidia solution for Steam Box, but he also stated developer support was ‘critical’ to the success of Steam Box. This was before confirmation of AMD APUs for both consoles. Valve said they were there mainly to liaise behind closed doors with software and hardware vendors. I’m guessing Gabe was in for quite a surprise when he saw the reaction when he then, and subsequently, approached the major game publishers and developers, already in deep liaison with AMD and months and eyebrows deep into optimizing their game engines and toolsets specifically for AMD hardware in general and HSA APUs/GPUs in particular, and who already knew Kaveri was going to be super compatible with the console APUs, with an Intel/Nvidia based Steam Box solution.
Note that these same factors will also make future AMD HSA APUs the overwhelming choice for future Sony and Microsoft console iterations as there will be built in software/hardware compatibility and almost no associated development costs. No reason a fully hardware backward compatible/4k capable Xbox One or PS4 ‘Turbo’ edition couldn’t be released in three or four years.
I think given his druthers, Gabe would still prefer an Intel/Nvidia solution, particularly in light of his stated desire for Steam Box to have a ‘multiple simultaneous game streaming’ capability (GRID), to remote in-house terminals (Shield). It’s worth noting that concurrent with the PS4 reveal an AMD spokesperson pointed out the PS4 APU was capable of multiple game streams and AMD was well along with their own GRID like game and application streaming solution.
While in his latest public appearance on YouTube Gabe is basically shilling for Intel CPUs, it’s likely it just isn’t economically/publisher/developer/Steam Box success feasible to go with anything but an AMD HSA APU solution.
I agree with a few of your
I agree with a few of your buried points: I do not think people appreciate what seems to be right around the corner. Virtualized GPUs allowing Linux users to run Windows applications at full speed and, probably more importantly, purchase one "beefy computer" and have all the rest of your tablets/handhelds/netbooks/laptops/HTPCs/whatever simply stream in video (and stream out keyboard/mouse/gamepad input) from that one PC.
Thanks for the great
Thanks for the great review.
This is a really nice surprise from nvidia and I hope it will inspire developers to port console games to it.
This looks really nice.
This looks really nice.
PS the first Shield may
PS the first Shield may indeed be strictly for fanboys, but even if so, said Fanboys are plentiful enough to fund the development cost. Bear in mind that, as Tim Stevens observed at CES, this is also AN INVESTMENT IN ANDROID GAMING. Logan is plainly built for powerful gaming, and nVidia wants to prime the pump for it, getting Devs excited about Android as a platform for serious gaming.
Looks like Tegra 4 sucks if
Looks like Tegra 4 sucks if Snapdragon 600 can rival it @ 1080 and Snapdragon 800 kill it.
http://www.androidauthority.com/snapdragon-800-overwhelms-tegra-4-gpu-249086/
Tegra 4 products will be limited to below 1080p. Guessing that why Shield stuck to 720p. An the Tegra 4 tablet is 768 resolution.
Will other tegra 4 tablets be
Will other tegra 4 tablets be able to do the pc streaming feature as well, or will this most likely only be available on the Shield.
XBMC is a great program as a
XBMC is a great program as a home theater program. I use the SkystreamX to run XBMC on my TV. It comes preloaded with all of the good add ons. I didn’t have to spend hours programming it. I just plugged it in, hooked it up to my internet and starting watching all of my favorite TV shows, Movies and live sports. It also turns your TV into a smart TV with over 800,000 Android Apps. I highly recommend this Android TV Box
http://www.skystreamx.com
They would want all the HSA
They would want all the HSA elements pertinent to porting to and from console games and future proofing fully in place with Kaveri. Kaveri’s delay might be related to such optimizations along with sharing console APU fabrication lines creating capacity issues. If Kaveri is delayed into 2014, I wouldn’t be surprised to see a ‘soft’ launch for Kaveri during the holiday season to put review site next gen game performance numbers on prospective buyers radar and more that No reason a fully hardware backward compatible/4k capable Xbox One or PS4 ‘Turbo’ edition couldn’t be released in three or four years.Android Box TV
Nothing really better than
Nothing really better than the G-Box MX2 in my opinion. Plays games and does XBMC like nothing else out there. I found it at this android xbmc website. I would steer clear of the fakes out there like the Skystream box and other chinese brands. They dont have any support and you can forget about them about as quickly as buy them. Ouya is also not too bad at XBMC but overall. I’d just grab the G-Box and call it a day.