Or perhaps I'm wrong. Perhaps there's something genuinely unique in Apple's position as a cash rich company that controls its entire product and seeks competitive advantage through pushing key technologies. It may be that a ton of money up front is actually all you need to leap ahead, but no one was really positioned, till now, to make that a sound investment. Until the iPad there wasn't really any good reason to even want super high res 9.7" screen, unless it was for extremely expensive specialized applications. For all I know you can pay someone to fire up the foundry, work out the kinks, and go into full production at a price point that's sustainable for mass market. But then you'd still have to wonder at how you make a jillion of them, since Apple appears to have had a hard time just getting enough of their plain jane iPad 1 panels, and that's with several sources.
I think this sums up why we have not really seen panels at much higher resolutions.
I also think that a significant portion of the $500M deal with LG done a while back was for the R&D necessary to get to such a mass-produced panel with much higher resolution.
I've already linked to this, but it didn't seem to make any impression, so I'm going to link to it again.
Again, it's a look at the likly time frame of an iPad sized "Retina" display within the context of Moore's Law.
Quote:
Originally Posted by Flaneur
I did read this article last night, but thought the application of Moore's Law was maybe a bit of a reach, but then what do I know about LCD screen making, so I took the easy route back to wishful thinking. But thanks for the link. Grist for the old mill. Otherwise the stones get worn, grinding on themselves . . .
Edit: Just re-read your post above. You argue it better than the article.
Quote:
Originally Posted by asdasd
I am not aware of lcds following Moores law. If so, extrapolating backwards, not that long ago they would have had one or two dpi. Clearly the number if transistors on a screen is lower than a normal wafer of that same size.
Assuming an 18 month "generation" Moores law applied retrospectically would imply about 128 times less per inch a decade ago (2^7).
The "generation" therefore can be jumped, since the LCD generations are tardy. Just buy better equipment and more if it. If it works for one machine, then $3.9B will buy you lots of machines.
OK! I read the linked article!
I may be way off base here, but I also have some experience in set semiconductor industry. In 1963, I was Data Processing Manager for Hoffman Electronics Semiconductor division. They actually grew silicon crystals, cut them into wafers, deposited chemical coatings on the wafers to make transistors, cut the individual parts from the wafers, tested the parts and packaged the parts into semi-finished goods. Later, the semi-finished parts were tested to more restrictive parameters based on products desired to qualify for military or commercial use. Products that passed the tests were marked (labeled) and packaged as finished goods and put into inventory and sold.
Still later, some finished parts could be retested for different or more restrictive parameters (maintain consistent performance over a broader temperature range, etc.). Those that passed these tests were unlabeled and/or unpackaged, then repackaged and/or relabeled as the new part and put into inventory and sold.
There is one other example of how things worked. Sometimes, several parts that did not pass tests could be used in combination to form another product with the desired characteristics -- if one part was off by -1, it could be combined with another part that was off by +1 (or 2 parts off by +.5) to yield the desired characteristics.
The testing, packaging, marking, retesting, repackaging and remarking often was many times more expensive than the manufacturing process.
Then, sometimes, they could go through the the whole manufacturing process, on the same machines -- the exact same way that they had done many times before... and everything would turn out differently.
The key to it all was the % yield from manufacturing and testing.
OK! That was before microcomputers -- with thousands of transistors on a chip.
In the early 1980's we installed some Apple ][ computer networks (LANS) for a company named Applied Materials. This was pioneering at the time and Applied Materials even had several LAN connections in their Board Room -- so it was visible at the highest level.
During the sales, installation and support of the Applied Materials installations, I was given several tours and had the opportunity to discuss state-of-the-art semiconductor manufacturing -- I was brought up to date.
Applied Materials was/is the leading manufacturer of machines and processes that deposit chemicals on silicon to give it the ability to conduct electricity.
Quote:
Applied Materials is the global leader in providing innovative equipment, services and software to the semiconductor, flat panel display and solar photovoltaic industries.
So, my first challenge: Does Moore's law apply to LCD manufacturing?
The article says (emphasis mine)s:
Quote:
If you watched the Android vs. iPhone debate over the last couple of years, you saw a series of Android phones with successively higher DPI in the same physical area, followed by the iPhone 4, which had a just slightly higher DPI than the most recent Android phone at the time, but which happened to be double the resolution of the original iPhone, because the iPhone hadn't been trying to keep up. (Now that the pixels are so tiny that they're invisible, are Android phones still coming out with even higher resolutions? I haven't heard.)
The so-called "retina display" wasn't a revolution, of course - it was just the next step in displays. (Holding off until you could get an integer 2x multiplier, however, was frickin' genius. Sometimes the best ideas sound obvious in retrospect.)
Each of those successively higher resolution screens was a "new process" in semiconductor speak. With each new process came a higher density of imperfections per unit area, which is why smaller screens were the first to hit these crazy-high densities. We can assume that the biggest screens at "retina" density as of the iPhone 4's release - 960x640 in June 2010 - was the state of the art at the time. We can also assume that the iPad was the biggest you could go with the best process for that size at the time, 1024x768 in March 2010.
Here the author says that the retina display of 326 ppi on a 3.5-4" display had been evolving for several years -- and was just the next step.
Then he says, without citation, that 1024x768 or 132 ppi on a 9.7" display was state-of-the-art for March 2010.
So, my second challenge: What was state-of-the-art ppi for a 9.7" display in 2010?
The author gets a little sidetracked claiming that the iPad needs to be a multiple of the iPhone 4 retina display.
The iPhone to iPad question is already resolved -- all that needs to happen is that a new iPad rez needs to be able to handle 1x and 2x multiples of 480x320.
The need is for the iPad2 to handle 2x rez of the iPad1.
My final questions are:
If Moore's law doesn't apply and/or 1024x768 (132 ppi) on a 9.7" display is not state-of-the-art:
What is state-of-the art for 9.7" displays?
What is the expected yield/cost of these displays?
When will sufficient production capacity be in place to ramp up to 40 million units per year?
Or, if we assume a 2 tier product ratio 2 iPad1 :: 1 iPad2 -- when will sufficient production capacity be in place to ramp up to 13 million units per year?
My guess is that Phil emcees. He brings Cook out to talk the numbers. The new iPad is revealed by Phil then Mansfield comes on to talk about the hardware changes. Forstall then finishes up with any software developments. Finally, they roll a short clip with Steve extolling the virtues of the new iPad.
I think this provides the perfect opportunity to showcase the talent that Apple has in the executive suite.
iPad 2 is going to be great fir everyone who has been patiently waiting for it to land. That much is certain based on the obvious new features (A5, RAM, Cameras).
Whether or not the iPad 2 compells iPad 1 owners to trade up (like the iPhone) immediately upon launch, or at least within 6 months, depends entirely on other factors. With iPad prices in the $500-800 range, its not the easy upgrade you have with iPhone, which is nearly free if you pay $200, and sell your last gen for the same. No, I think (and hope) this means there are some very strong features of iPad 2 that are irresistible to current owners.
And I'm thinking it's this display.
I agree for current owners that the display is a must.
But, the iPad2 is/should target whole new markets -- medical, AV production, etc.
Yes, but what are we basing the idea that this is an "abrupt jump" on? As you say, nobody is using ultra-high-resolution 9.7" displays, but does that mean we can't make them? Do we really know when the iPhone 4's retina display became cost effective?
The article on Moore's Law and the retina display makes the following assumptions:
"We can assume that the biggest screens at "retina" density as of the iPhone 4's release - 960x640 in June 2010 - was the state of the art at the time. We can also assume that the iPad was the biggest you could go with the best process for that size at the time, 1024x768 in March 2010."
I think both of these assumptions are false. Firstly, June 2010 is, at best, an upper-bound on when the iPhone 4's retina display became cost effective. Surely it became cost effective many months before June 2010. Production would have ramped up long before then and Apple would have done its cost analysis earlier still. It's also possible that Apple could have used an even higher resolution display at the time but didn't because it wouldn't be exactly 2x resolution. Secondly, if Apple is planning to deliver a "retina display" for the iPad 2, it would have likely been aware of its plans when it chose to use a 1024x764 display in the current iPad. That display could have been explicitly chosen because Apple intended to use a 2048x1528 display in the second generation model. The iPad, I believe, was intentionally very conservative hardware-wise because it was an entirely new product category.
The figure he comes up with based on these two assumptions is 2012 for a 2x resolution update. If you believe the logic of the article but disagree with the assumptions, 2011 becomes a reasonable estimate.
I don't think even these revised assumptions are correct because I don't see any evidence at all that the computer and phone industry has been particularly interested in pursuing DPI. Android handset manufacturers, like PC vendors, choose displays based primarily on cost. Apple has been pursuing higher resolution displays and other manufacturers have responded by either offering higher resolution displays as an option or, eventually, moving to them as standard. But I've never thought of the PC industry, or the phone industry, as anything but extremely conservative on display resolution. So I don't think you can do this kind of analysis based on shipping products.
Personally I've yet to see any convincing evidence that a 9.7" 2x resolution display isn't cost effective right now. This kind of thing wouldn't be available to a start-up or small manufacturer, because they use off-the-shelf components, but to a cash-rich company that's heavily invested in the manufacturing plant and is looking to ship 20 million devices could do it if it's possible. And, like I say, I don't see any convincing evidence it's not possible. All that's been shown so far is that nobody else has done it yet.
You saw the same inconsistencies, plus a few more.
Maybe another way to ask the question:
Are there any displays larger than 9.7" with the required 2x ppi currently available?
Good point, perhaps the article should have made less of Moore's Law and stuck to silicon processes in general. The history of LCD display resolutions (like CPU perfromance) isn't marked by abrupt jumps beyond what anyone had generally imagined plausible; even the iPhone 4 was had just somewhat higher resolution than the incremental improvements that had reached the market since its release.
A 2048x1536 9.7" panel would be such an abrupt jump, certainly for a relatively inexpensive mass market product. I think the best argument against it being plausible is that it has never happened. Again, Apple's not the only CE company in the history of the world with some money to throw around; if skipping a generation of anything were economically feasible but for some upfront money it would happen more often.
Or perhaps I'm wrong. Perhaps there's something genuinely unique in Apple's position as a cash rich company that controls its entire product and seeks competitive advantage through pushing key technologies. It may be that a ton of money up front is actually all you need to leap ahead, but no one was really positioned, till now, to make that a sound investment. Until the iPad there wasn't really any good reason to even want super high res 9.7" screen, unless it was for extremely expensive specialized applications. For all I know you can pay someone to fire up the foundry, work out the kinks, and go into full production at a price point that's sustainable for mass market. But then you'd still have to wonder at how you make a jillion of them, since Apple appears to have had a hard time just getting enough of their plain jane iPad 1 panels, and that's with several sources.
No it wouldn't be an abrupt jump. The abrupt jump happened in 1999 when Lawrence Livermore Labs wanted 300dpi 30" LCD desktop displays for visualizing simulation results. They were delivered in 2000 at about $100K apiece to cover the custom R&D. Guess where the tech for lower res LCD televisions came from??? [also notice things went from ~150dpi SGI displays at 17" to 300dpi at 30" in less than 2 years. It wasn't conceptually that hard, they just had to scale up the fab machinery size. And it worked, pretty much first shot.]
The problems were in manufacturing the large panel without cracking, not in making the transistor density work. iPad screens are trivially sized for current glass panel sizes, yields will be very favorable because glass cracking in those sizes is minimal. The only issue is where is the producer? Somebody needs to set up a fab to do that.
The technical issues for a potential 2x iPad LCD are well in hand, and could be produceable in quantity if Apple was out in front of this for the past 6-9 months. Hmmmm. Who said they spent $3.9B on a major component purchase arrangement? That's the kind of cash that can reconfigure an LCD fab with around $3B left over.
I've already linked to this, but it didn't seem to make any impression, so I'm going to link to it again.
Again, it's a look at the likly time frame of an iPad sized "Retina" display within the context of Moore's Law.
....
Anyway, read the article, it makes a pretty good case.
It is an interesting thought experiment but really is no more or less convincing than any other speculation. It's possible LCDs follow Moore's law but I'm not at all convinced of that. I think LCD advances are driven more by economies of scale which comes from popular applications of the technology. In semiconductors, economy of scale is a given (Intel makes hundreds of millions of any one part).
Using his own arguments however, you can reach the opposite conclusion:
His basic assumption is that density doubles FOR THE SAME COST every 18 months.
Given that Apple wants 4x the density of the iPad, that should mean that it will take 3 years to get an 4X LCD for the same price.
The iPad was released in April 2010. There was very little discussion of the screen resolution as being particularly cutting edge and I wouldn't expect that Apple would use something particularly cutting edge in a version 1 product.
If we assume that the panel was state of the art some months before (let's say 6 months but we have no idea), then April 2011 is 18 months from then so 1 Moore's law generation.
So Apple could get a 2x panel now for the same cost. But the linked article supports the notion that you can get another 2x jump at double the cost.
So: the link would support the conclusion that Apple could use a 4x panel in the iPad2 as long as it was willing to pay double the cost for the screen.
Speculation is fun but pretty fruitless as none of us know the real constraints.
No it wouldn't be an abrupt jump. The abrupt jump happened in 1999 when Lawrence Livermore Labs wanted 300dpi 30" LCD desktop displays for visualizing simulation results. They were delivered in 2000 at about $100K apiece to cover the custom R&D. Guess where the tech for lower res LCD televisions came from??? [also notice things went from ~150dpi SGI displays at 17" to 300dpi at 30" in less than 2 years. It wasn't conceptually that hard, they just had to scale up the fab machinery size. And it worked, pretty much first shot.]
The problems were in manufacturing the large panel without cracking, not in making the transistor density work. iPad screens are trivially sized for current glass panel sizes, yields will be very favorable because glass cracking in those sizes is minimal. The only issue is where is the producer? Somebody needs to set up a fab to do that.
The technical issues for a potential 2x iPad LCD are well in hand, and could be produceable in quantity if Apple was out in front of this for the past 6-9 months. Hmmmm. Who said they spent $3.9B on a major component purchase arrangement? That's the kind of cash that can reconfigure an LCD fab with around $3B left over.
I'd like to see some links supporting the idea that it's trivially easy to scale up LCD densities as long as you're willing to set up the fab.
I guess we could reckon that there hasn't been a strong strategic advantage to releasing higher res displays than the competition, but that seems a bit counter-intuitive and I'm not sure what the explanation for not seeing these displays come to market would be otherwise.
Basically, you're arguing that there's this low hanging fruit of a leg up on the your competitors, but no one is picking it because.... they just haven't. But Apple will, because they're clever that way. I'm not quite seeing it.
Also, at some point we completely dropped the issues of power consumption and performance, that seem to me to be pretty significant issues for pushing a great many pixels on a power and CPU constrained device. A different issue, to be sure, but what I'm hearing is a kind of brusque assurance that Apple can and will make this happen, that it's actually just a matter of wanting to do it and getting it done. I think there are bigger hurdles than that.
I'd like to see some links supporting the idea that it's trivially easy to scale up LCD densities as long as you're willing to set up the fab.
The LLM 300ppi LCDs are ancient history. Pre-Google days for the contract, Google in beta during the delivery. There are no links unless you want to make blind searches in the Wayback Machine, be my guest. The production development contract to delivery was less than 2 years. You won't find that anywhere, you have to know someone who was asking for them. There were only minor press releases about the tech when it was delivered. It's not like doing nuc simulations was something that was going to get a lot of intarweb mileage.
It's the fact that the production ppi ratio went from ~150 on the best SGI displays to a delivered run of 300ppi by 30" in less than than 2 years despite the problems with > 22" glass cracking that is the basis of caling the scaling solved. Only because it was. 10 years ago.
Its not an open problem that needs R&D. Its a previously solved problem that needs a market willing to pay enough that reconfiguring the fab to do a new non-standard screen size.
Quote:
I guess we could reckon that there hasn't been a strong strategic advantage to releasing higher res displays than the competition, but that seems a bit counter-intuitive and I'm not sure what the explanation for not seeing these displays come to market would be otherwise.
Basically, you're arguing that there's this low hanging fruit of a leg up on the your competitors, but no one is picking it because.... they just haven't. But Apple will, because they're clever that way. I'm not quite seeing it.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
Quote:
Also, at some point we completely dropped the issues of power consumption and performance, that seem to me to be pretty significant issues for pushing a great many pixels on a power and CPU constrained device. A different issue, to be sure, but what I'm hearing is a kind of brusque assurance that Apple can and will make this happen, that it's actually just a matter of wanting to do it and getting it done. I think there are bigger hurdles than that.
The backlight is the same, the transistor draw in the LCD itself is quite small, but not trivial. The only issue would be the necessary GPU horsepower. The new rumored GPU is several years engineering, and a couple process steps smaller -- meaning it is probably close to on par power draw at the higher fill rates compared to the current GPU. And with the OpenCL support a whole bunch of CPU draining heavy calcs can be funneled to the GPU where those transistors are drawing non-idle power already anyways. That means less CPU power draw is traded for more work for about the same power on the GPU -- a power win.
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
I'm guessing that picking this "oddball size" must have ultimately been a visual and tactile choice, maybe based on whatever sense of harmonious proportion that led to such standards as 4:3 and the 8x10, to both of which it is close. Now that we have it, I think it feels absolutely right. But I think it was designed for the photograph, the page, and old cinema, and it flies in the face of the oppressively stretched 16:9, an alien form when it migrates from the big screen. 16:9 doesn't belong in the hands. All of which is to say that Apple has had this size all to itself. They're clever that way, as addabox says.
Quote:
Originally Posted by Hiro
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
So there are two completely compelling reasons to double the resolution.
One is display of text, and especially photographs and film, their main 'content' reasons for picking the size to begin with. We will be bowled over by how pictures will look, mostly because we will have never seen such clear and beautiful large images, outside of medium-format slide viewers with Kodachrome inside, and how many of us have seen those? "And yet it moves!" It will be irresistible.
The other reason is that they can't allow others to get the jump on them again, if what you say is true about Android tablets to come. If it doesn't happen in April, there will have been technical problems or cost considerations. They must do it eventually. I hope it's this time around, given what you say about the competition.
Oh, and thanks for the technical background. Bracing, if true . . .
Apple's current price line is $499, $599 and $699. Is it unreasonable to think they may now introduce a brand new product at the $799 and $899 price? The Mac Book air starts at $999. Apple will effectively fill all the price points they now have missing while introducing a much better iPad.
If price and price margins are the ONLY reason it can't exist, then why not simply make something way nicer than the current models for a $100 or $200 more.
Quote:
Originally Posted by uplate
I agree from the cost side alone. I'd love a 2x screen but I can't imagine what that will add to the cost of the unit.
I read an estimate of approx. $150 in extra cost (not necessarily including other increased costs of supporting it). 40% gross margin brings that to $240, so if that's accurate and there's to be such a Pad this year, $799 is a likely base (no 3g) price.
Quote:
Originally Posted by smithg
As I commented on another board, the unveiling presentation practically writes itself:
4 times the CPU (dual core A9)
4 times the graphics (dual core SGX 543)
4 times the RAM (1GB)
4 times the screen (2048x1536)
Still just $499.
Nope. Too much component cost increase - not enough "production experience" amortized for such new glass to bring a mass price this year. See above for estimated retail.
Quote:
Originally Posted by kube
Out of curiosity, does anyone know the resolution of a high-quality magazine? I'd imagine that this would roughly be a resolution end-point.
Quote:
Originally Posted by TalkingNewMedia
Not really a good comparison. The reason is this: the standard for magazine publishing photos is 300dpi. Photographers are instructed to supply art at this size and no smaller than the size the photo will be eventually laid out at. In other words, if the photo takes up a full page, the supplied photo must be 8x10 at 300dpi. But usually photos are downsized by the art director to fit the layout.
But printing and displays are so very different: for one thing ink spreads -- which is actually a good thing when you are trying to avoid the reader seeing individual drops of ink.
Quote:
Originally Posted by mstone
2400 dpi for line work such as black text. 175 line screen for images. The CYMK each have a possible dot size of 1 to 256 scale (256 being solid 100% overlap). It isn't exactly the same relationship to a monitor. The reason people say 300 dpi is because that used to be twice the line screen. Now we use 350 because we print at 175 line screen. The reason you want to be twice is because of antialiasing. In a worst case scenario the placement of your image would be exactly half way between the dot grid. That way you have enough resolution to accurately average the color values. You can kind of think of 175 dpi as the viewing resolution on a printed page although it is not entirely accurate since it is a completely different imaging science.
Quote:
Originally Posted by rain
The reason for 300 dpi is because of the substrate and dot gain. 300 dpi is what a coated stock can handle without killing the Rosetta.
Newsprint is 150 lpi or up to 220 dpi depending on the quality of the paper.
Flexo printing on a burlap sack can be as low as 14 dpi.
Coated inkjet paper has special (expensive) chemicals on it to sustain a smaller dot... But doesn't use the same method as commercial offset printing.
Imagine a mag printed on an inkjet or dye-sub... The thing would cost $200.
I could get into 'art' printing, like for expensive books, but that again is a different process.
In terms of the original question no one's specifically mentioned the frequent confusion between "dpi" (a measurement of printed dots of ink per inch) and "ppi" (pixels/per inch on the screen). See: http://www.rideau-info.com/photos/mythdpi.html. You can make a 300 dpi print of the most pixelated source material, which will quite faithfully translate its unglory to a page as easily as one with deep pixel density that looks great at smoooth retina resolution
Quote:
Originally Posted by digitalclips
I hate 'mobile' web sites you cannot avoid with a passion. Some simply refuse to let you get around it. There should always be an option (and a savable one) to select the real site if you wish. Now having said that an option to access a true iPad / iPhone style version (as opposed to something designed for a BB user) is nice at times as an option (like when on slow connections) but only as an option. I think I said option too many times here
The LLM 300ppi LCDs are ancient history. Pre-Google days for the contract, Google in beta during the delivery. There are no links unless you want to make blind searches in the Wayback Machine, be my guest. The production development contract to delivery was less than 2 years. You won't find that anywhere, you have to know someone who was asking for them. There were only minor press releases about the tech when it was delivered. It's not like doing nuc simulations was something that was going to get a lot of intarweb mileage.
It's the fact that the production ppi ratio went from ~150 on the best SGI displays to a delivered run of 300ppi by 30" in less than than 2 years despite the problems with > 22" glass cracking that is the basis of caling the scaling solved. Only because it was. 10 years ago.
Its not an open problem that needs R&D. Its a previously solved problem that needs a market willing to pay enough that reconfiguring the fab to do a new non-standard screen size.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
The backlight is the same, the transistor draw in the LCD itself is quite small, but not trivial. The only issue would be the necessary GPU horsepower. The new rumored GPU is several years engineering, and a couple process steps smaller -- meaning it is probably close to on par power draw at the higher fill rates compared to the current GPU. And with the OpenCL support a whole bunch of CPU draining heavy calcs can be funneled to the GPU where those transistors are drawing non-idle power already anyways. That means less CPU power draw is traded for more work for about the same power on the GPU -- a power win.
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
Well I guess you're right either way then....
If Apple fails to deliver the fabled high res iPad, they just decided it wasn't the "correct economic time", even though, apparently, there's absolutely no technical or financial hurdle between them and doing so.
I think you're conflating "is technically possible", which I've never disputed, and "is financially viable", which you're hedging on and which is the basis of my skepticism.
You seem to want to argue that it's actually nothing at all, aside some startup investment, to start cranking out the high res iPad screens, while leaving yourself an out with wondering after Apple's commitment to the technology, but I don't think you can have it both ways. Your insisting on a scenario wherein it would make no sense whatsoever for Apple not to take the iPad 2 to 2048x1536, so we'll see.
As far as Android devices with somewhat higher resolution screens and dual core processors, sure, but what of it? "Somewhat higher resolution" being broadly marketed doesn't make "massively higher resolution" more likely.
I'm guessing that picking this "oddball size" must have ultimately been a visual and tactile choice, maybe based on whatever sense of harmonious proportion that led to such standards as 4:3 and the 8x10, to both of which it is close. Now that we have it, I think it feels absolutely right. But I think it was designed for the photograph, the page, and old cinema, and it flies in the face of the oppressively stretched 16:9, an alien form when it migrates from the big screen. 16:9 doesn't belong in the hands. All of which is to say that Apple has had this size all to itself. They're clever that way, as addabox says.
Yes, and all I meant by oddball size of the rumored 2x iPad display was it's uniqueness resolution-dimension pairing which absolutely nothing else uses.
Yes, and all I meant by oddball size of the rumored 2x iPad display was it's uniqueness resolution-dimension pairing which absolutely nothing else uses.
I think QXGA is quite common for certain applications and has been for awhile.
If Apple fails to deliver the fabled high res iPad, they just decided it wasn't the "correct economic time", even though, apparently, there's absolutely no technical or financial hurdle between them and doing so.
I think you're conflating "is technically possible", which I've never disputed, and "is financially viable", which you're hedging on and which is the basis of my skepticism.
You seem to want to argue that it's actually nothing at all, aside some startup investment, to start cranking out the high res iPad screens...
Mostly agreed.
Quote:
while leaving yourself an out with wondering after Apple's commitment to the technology, but I don't think you can have it both ways. Your insisting on a scenario wherein it would make no sense whatsoever for Apple not to take the iPad 2 to 2048x1536, so we'll see.
It's not a matter of me having it both ways. The decision tree is simple, either Apple has an agreement with a producer to crank out unique screens, or they don't. Technically, there's nothing special about producing screens in that size, but there are zero known fabs doing it. Meaning either Apple contracted a quiet set-up or Apple didn't.
That's not having it both ways, it's just saying all the poo-poohing over !ZOMG! 2x display density isn't possible! Is just plain bunk.
That simply leaves us with the flat-out obvious, Apple only does something tech-wise when they think it is economically advantageous.
Quote:
As far as Android devices with somewhat higher resolution screens and dual core processors, sure, but what of it? "Somewhat higher resolution" being broadly marketed doesn't make "massively higher resolution" more likely.
There I'll just plain disagree. I'm convinced Steve and company aren't happy about all the screen size grief they took over the first half of last year. If there was ever a second-order financial motivation of removing a potential marketing weakness while simultaneously allowing a huge pronouncement of massive technical progress (even though it really wouldn't be that massive) and using that to make it EXTREMELY hard to be technically passed in the next 9 months or so, this would fit the bill. Using it to close out any significant threats in the market space until Apple is as long term strong in tabs as they were in year two of the iPhone.
What the real $20 question is is this: Was Apple thinking this way in first quarter last year? If so the 2x screen has a real chance of showing up sooner rather than later. If they didn't start thinking this way until sometime in the summer I don't think it could happen yet.
Either way I'm not predicting, I've merely stated why the "it is not possible arguments" are incorrect. And then added my own 2 cents on why it might make sense on an aggressive timescale. Very different from saying it will happen. Hedging, yes a bit. I've been wrong on both sides of predictions enough before, so now I just dabble in the technical discussion part where I actually have a clue and leave the actual business to Apple.
Comments
Or perhaps I'm wrong. Perhaps there's something genuinely unique in Apple's position as a cash rich company that controls its entire product and seeks competitive advantage through pushing key technologies. It may be that a ton of money up front is actually all you need to leap ahead, but no one was really positioned, till now, to make that a sound investment. Until the iPad there wasn't really any good reason to even want super high res 9.7" screen, unless it was for extremely expensive specialized applications. For all I know you can pay someone to fire up the foundry, work out the kinks, and go into full production at a price point that's sustainable for mass market. But then you'd still have to wonder at how you make a jillion of them, since Apple appears to have had a hard time just getting enough of their plain jane iPad 1 panels, and that's with several sources.
I think this sums up why we have not really seen panels at much higher resolutions.
I also think that a significant portion of the $500M deal with LG done a while back was for the R&D necessary to get to such a mass-produced panel with much higher resolution.
Or, maybe I'm just dreaming.
1) Moores law does not apply to DPI ( or LCDs in general).
2) He applies this law - an 18 month time to double - at the wrong time assuming that the law applies on the very release day of the retina screen.
For that reason I'm oot.
I've already linked to this, but it didn't seem to make any impression, so I'm going to link to it again.
Again, it's a look at the likly time frame of an iPad sized "Retina" display within the context of Moore's Law.
I did read this article last night, but thought the application of Moore's Law was maybe a bit of a reach, but then what do I know about LCD screen making, so I took the easy route back to wishful thinking. But thanks for the link. Grist for the old mill. Otherwise the stones get worn, grinding on themselves . . .
Edit: Just re-read your post above. You argue it better than the article.
I am not aware of lcds following Moores law. If so, extrapolating backwards, not that long ago they would have had one or two dpi. Clearly the number if transistors on a screen is lower than a normal wafer of that same size.
Assuming an 18 month "generation" Moores law applied retrospectically would imply about 128 times less per inch a decade ago (2^7).
The "generation" therefore can be jumped, since the LCD generations are tardy. Just buy better equipment and more if it. If it works for one machine, then $3.9B will buy you lots of machines.
OK! I read the linked article!
I may be way off base here, but I also have some experience in set semiconductor industry. In 1963, I was Data Processing Manager for Hoffman Electronics Semiconductor division. They actually grew silicon crystals, cut them into wafers, deposited chemical coatings on the wafers to make transistors, cut the individual parts from the wafers, tested the parts and packaged the parts into semi-finished goods. Later, the semi-finished parts were tested to more restrictive parameters based on products desired to qualify for military or commercial use. Products that passed the tests were marked (labeled) and packaged as finished goods and put into inventory and sold.
Still later, some finished parts could be retested for different or more restrictive parameters (maintain consistent performance over a broader temperature range, etc.). Those that passed these tests were unlabeled and/or unpackaged, then repackaged and/or relabeled as the new part and put into inventory and sold.
There is one other example of how things worked. Sometimes, several parts that did not pass tests could be used in combination to form another product with the desired characteristics -- if one part was off by -1, it could be combined with another part that was off by +1 (or 2 parts off by +.5) to yield the desired characteristics.
The testing, packaging, marking, retesting, repackaging and remarking often was many times more expensive than the manufacturing process.
Then, sometimes, they could go through the the whole manufacturing process, on the same machines -- the exact same way that they had done many times before... and everything would turn out differently.
The key to it all was the % yield from manufacturing and testing.
OK! That was before microcomputers -- with thousands of transistors on a chip.
In the early 1980's we installed some Apple ][ computer networks (LANS) for a company named Applied Materials. This was pioneering at the time and Applied Materials even had several LAN connections in their Board Room -- so it was visible at the highest level.
During the sales, installation and support of the Applied Materials installations, I was given several tours and had the opportunity to discuss state-of-the-art semiconductor manufacturing -- I was brought up to date.
Applied Materials was/is the leading manufacturer of machines and processes that deposit chemicals on silicon to give it the ability to conduct electricity.
Applied Materials is the global leader in providing innovative equipment, services and software to the semiconductor, flat panel display and solar photovoltaic industries.
http://www.appliedmaterials.com/
It's been a long time, but I do have some knowledge of the industry -- if not actual experience and bona fides.
I went through all that because I want to challenge parts of the "linked to article".
Like the writer of the linked article, I don't have experience with LCD manufacturing.
I did some surfing, and the LCD manufacturing appears quite different than transistors or microcomputers:
http://www.crystec.com/crylcde.htm
So, my first challenge: Does Moore's law apply to LCD manufacturing?
The article says (emphasis mine)s:
If you watched the Android vs. iPhone debate over the last couple of years, you saw a series of Android phones with successively higher DPI in the same physical area, followed by the iPhone 4, which had a just slightly higher DPI than the most recent Android phone at the time, but which happened to be double the resolution of the original iPhone, because the iPhone hadn't been trying to keep up. (Now that the pixels are so tiny that they're invisible, are Android phones still coming out with even higher resolutions? I haven't heard.)
The so-called "retina display" wasn't a revolution, of course - it was just the next step in displays. (Holding off until you could get an integer 2x multiplier, however, was frickin' genius. Sometimes the best ideas sound obvious in retrospect.)
Each of those successively higher resolution screens was a "new process" in semiconductor speak. With each new process came a higher density of imperfections per unit area, which is why smaller screens were the first to hit these crazy-high densities. We can assume that the biggest screens at "retina" density as of the iPhone 4's release - 960x640 in June 2010 - was the state of the art at the time. We can also assume that the iPad was the biggest you could go with the best process for that size at the time, 1024x768 in March 2010.
Here the author says that the retina display of 326 ppi on a 3.5-4" display had been evolving for several years -- and was just the next step.
Then he says, without citation, that 1024x768 or 132 ppi on a 9.7" display was state-of-the-art for March 2010.
So, my second challenge: What was state-of-the-art ppi for a 9.7" display in 2010?
The author gets a little sidetracked claiming that the iPad needs to be a multiple of the iPhone 4 retina display.
The iPhone to iPad question is already resolved -- all that needs to happen is that a new iPad rez needs to be able to handle 1x and 2x multiples of 480x320.
The need is for the iPad2 to handle 2x rez of the iPad1.
My final questions are:
If Moore's law doesn't apply and/or 1024x768 (132 ppi) on a 9.7" display is not state-of-the-art:
What is state-of-the art for 9.7" displays?
What is the expected yield/cost of these displays?
When will sufficient production capacity be in place to ramp up to 40 million units per year?
Or, if we assume a 2 tier product ratio 2 iPad1 :: 1 iPad2 -- when will sufficient production capacity be in place to ramp up to 13 million units per year?
My guess is that Phil emcees. He brings Cook out to talk the numbers. The new iPad is revealed by Phil then Mansfield comes on to talk about the hardware changes. Forstall then finishes up with any software developments. Finally, they roll a short clip with Steve extolling the virtues of the new iPad.
I think this provides the perfect opportunity to showcase the talent that Apple has in the executive suite.
I like that scenario!
iPad 2 is going to be great fir everyone who has been patiently waiting for it to land. That much is certain based on the obvious new features (A5, RAM, Cameras).
Whether or not the iPad 2 compells iPad 1 owners to trade up (like the iPhone) immediately upon launch, or at least within 6 months, depends entirely on other factors. With iPad prices in the $500-800 range, its not the easy upgrade you have with iPhone, which is nearly free if you pay $200, and sell your last gen for the same. No, I think (and hope) this means there are some very strong features of iPad 2 that are irresistible to current owners.
And I'm thinking it's this display.
I agree for current owners that the display is a must.
But, the iPad2 is/should target whole new markets -- medical, AV production, etc.
Yes, but what are we basing the idea that this is an "abrupt jump" on? As you say, nobody is using ultra-high-resolution 9.7" displays, but does that mean we can't make them? Do we really know when the iPhone 4's retina display became cost effective?
The article on Moore's Law and the retina display makes the following assumptions:
"We can assume that the biggest screens at "retina" density as of the iPhone 4's release - 960x640 in June 2010 - was the state of the art at the time. We can also assume that the iPad was the biggest you could go with the best process for that size at the time, 1024x768 in March 2010."
I think both of these assumptions are false. Firstly, June 2010 is, at best, an upper-bound on when the iPhone 4's retina display became cost effective. Surely it became cost effective many months before June 2010. Production would have ramped up long before then and Apple would have done its cost analysis earlier still. It's also possible that Apple could have used an even higher resolution display at the time but didn't because it wouldn't be exactly 2x resolution. Secondly, if Apple is planning to deliver a "retina display" for the iPad 2, it would have likely been aware of its plans when it chose to use a 1024x764 display in the current iPad. That display could have been explicitly chosen because Apple intended to use a 2048x1528 display in the second generation model. The iPad, I believe, was intentionally very conservative hardware-wise because it was an entirely new product category.
The figure he comes up with based on these two assumptions is 2012 for a 2x resolution update. If you believe the logic of the article but disagree with the assumptions, 2011 becomes a reasonable estimate.
I don't think even these revised assumptions are correct because I don't see any evidence at all that the computer and phone industry has been particularly interested in pursuing DPI. Android handset manufacturers, like PC vendors, choose displays based primarily on cost. Apple has been pursuing higher resolution displays and other manufacturers have responded by either offering higher resolution displays as an option or, eventually, moving to them as standard. But I've never thought of the PC industry, or the phone industry, as anything but extremely conservative on display resolution. So I don't think you can do this kind of analysis based on shipping products.
Personally I've yet to see any convincing evidence that a 9.7" 2x resolution display isn't cost effective right now. This kind of thing wouldn't be available to a start-up or small manufacturer, because they use off-the-shelf components, but to a cash-rich company that's heavily invested in the manufacturing plant and is looking to ship 20 million devices could do it if it's possible. And, like I say, I don't see any convincing evidence it's not possible. All that's been shown so far is that nobody else has done it yet.
You saw the same inconsistencies, plus a few more.
Maybe another way to ask the question:
Are there any displays larger than 9.7" with the required 2x ppi currently available?
If yes, in what quantities, and at what costs?
Has more and better already been done?
The Internet is an echo chamber.
Good point, perhaps the article should have made less of Moore's Law and stuck to silicon processes in general. The history of LCD display resolutions (like CPU perfromance) isn't marked by abrupt jumps beyond what anyone had generally imagined plausible; even the iPhone 4 was had just somewhat higher resolution than the incremental improvements that had reached the market since its release.
A 2048x1536 9.7" panel would be such an abrupt jump, certainly for a relatively inexpensive mass market product. I think the best argument against it being plausible is that it has never happened. Again, Apple's not the only CE company in the history of the world with some money to throw around; if skipping a generation of anything were economically feasible but for some upfront money it would happen more often.
Or perhaps I'm wrong. Perhaps there's something genuinely unique in Apple's position as a cash rich company that controls its entire product and seeks competitive advantage through pushing key technologies. It may be that a ton of money up front is actually all you need to leap ahead, but no one was really positioned, till now, to make that a sound investment. Until the iPad there wasn't really any good reason to even want super high res 9.7" screen, unless it was for extremely expensive specialized applications. For all I know you can pay someone to fire up the foundry, work out the kinks, and go into full production at a price point that's sustainable for mass market. But then you'd still have to wonder at how you make a jillion of them, since Apple appears to have had a hard time just getting enough of their plain jane iPad 1 panels, and that's with several sources.
No it wouldn't be an abrupt jump. The abrupt jump happened in 1999 when Lawrence Livermore Labs wanted 300dpi 30" LCD desktop displays for visualizing simulation results. They were delivered in 2000 at about $100K apiece to cover the custom R&D. Guess where the tech for lower res LCD televisions came from??? [also notice things went from ~150dpi SGI displays at 17" to 300dpi at 30" in less than 2 years. It wasn't conceptually that hard, they just had to scale up the fab machinery size. And it worked, pretty much first shot.]
The problems were in manufacturing the large panel without cracking, not in making the transistor density work. iPad screens are trivially sized for current glass panel sizes, yields will be very favorable because glass cracking in those sizes is minimal. The only issue is where is the producer? Somebody needs to set up a fab to do that.
The technical issues for a potential 2x iPad LCD are well in hand, and could be produceable in quantity if Apple was out in front of this for the past 6-9 months. Hmmmm. Who said they spent $3.9B on a major component purchase arrangement? That's the kind of cash that can reconfigure an LCD fab with around $3B left over.
I've already linked to this, but it didn't seem to make any impression, so I'm going to link to it again.
Again, it's a look at the likly time frame of an iPad sized "Retina" display within the context of Moore's Law.
....
Anyway, read the article, it makes a pretty good case.
It is an interesting thought experiment but really is no more or less convincing than any other speculation. It's possible LCDs follow Moore's law but I'm not at all convinced of that. I think LCD advances are driven more by economies of scale which comes from popular applications of the technology. In semiconductors, economy of scale is a given (Intel makes hundreds of millions of any one part).
Using his own arguments however, you can reach the opposite conclusion:
His basic assumption is that density doubles FOR THE SAME COST every 18 months.
Given that Apple wants 4x the density of the iPad, that should mean that it will take 3 years to get an 4X LCD for the same price.
The iPad was released in April 2010. There was very little discussion of the screen resolution as being particularly cutting edge and I wouldn't expect that Apple would use something particularly cutting edge in a version 1 product.
If we assume that the panel was state of the art some months before (let's say 6 months but we have no idea), then April 2011 is 18 months from then so 1 Moore's law generation.
So Apple could get a 2x panel now for the same cost. But the linked article supports the notion that you can get another 2x jump at double the cost.
So: the link would support the conclusion that Apple could use a 4x panel in the iPad2 as long as it was willing to pay double the cost for the screen.
Speculation is fun but pretty fruitless as none of us know the real constraints.
No it wouldn't be an abrupt jump. The abrupt jump happened in 1999 when Lawrence Livermore Labs wanted 300dpi 30" LCD desktop displays for visualizing simulation results. They were delivered in 2000 at about $100K apiece to cover the custom R&D. Guess where the tech for lower res LCD televisions came from??? [also notice things went from ~150dpi SGI displays at 17" to 300dpi at 30" in less than 2 years. It wasn't conceptually that hard, they just had to scale up the fab machinery size. And it worked, pretty much first shot.]
The problems were in manufacturing the large panel without cracking, not in making the transistor density work. iPad screens are trivially sized for current glass panel sizes, yields will be very favorable because glass cracking in those sizes is minimal. The only issue is where is the producer? Somebody needs to set up a fab to do that.
The technical issues for a potential 2x iPad LCD are well in hand, and could be produceable in quantity if Apple was out in front of this for the past 6-9 months. Hmmmm. Who said they spent $3.9B on a major component purchase arrangement? That's the kind of cash that can reconfigure an LCD fab with around $3B left over.
I'd like to see some links supporting the idea that it's trivially easy to scale up LCD densities as long as you're willing to set up the fab.
I guess we could reckon that there hasn't been a strong strategic advantage to releasing higher res displays than the competition, but that seems a bit counter-intuitive and I'm not sure what the explanation for not seeing these displays come to market would be otherwise.
Basically, you're arguing that there's this low hanging fruit of a leg up on the your competitors, but no one is picking it because.... they just haven't. But Apple will, because they're clever that way. I'm not quite seeing it.
Also, at some point we completely dropped the issues of power consumption and performance, that seem to me to be pretty significant issues for pushing a great many pixels on a power and CPU constrained device. A different issue, to be sure, but what I'm hearing is a kind of brusque assurance that Apple can and will make this happen, that it's actually just a matter of wanting to do it and getting it done. I think there are bigger hurdles than that.
Digitimes has ZERO credibility:
http://www.ecommercetimes.com/story/71552.html
" No change in resolution"
I'd like to see some links supporting the idea that it's trivially easy to scale up LCD densities as long as you're willing to set up the fab.
The LLM 300ppi LCDs are ancient history. Pre-Google days for the contract, Google in beta during the delivery. There are no links unless you want to make blind searches in the Wayback Machine, be my guest. The production development contract to delivery was less than 2 years. You won't find that anywhere, you have to know someone who was asking for them. There were only minor press releases about the tech when it was delivered. It's not like doing nuc simulations was something that was going to get a lot of intarweb mileage.
It's the fact that the production ppi ratio went from ~150 on the best SGI displays to a delivered run of 300ppi by 30" in less than than 2 years despite the problems with > 22" glass cracking that is the basis of caling the scaling solved. Only because it was. 10 years ago.
Its not an open problem that needs R&D. Its a previously solved problem that needs a market willing to pay enough that reconfiguring the fab to do a new non-standard screen size.
I guess we could reckon that there hasn't been a strong strategic advantage to releasing higher res displays than the competition, but that seems a bit counter-intuitive and I'm not sure what the explanation for not seeing these displays come to market would be otherwise.
Basically, you're arguing that there's this low hanging fruit of a leg up on the your competitors, but no one is picking it because.... they just haven't. But Apple will, because they're clever that way. I'm not quite seeing it.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
Also, at some point we completely dropped the issues of power consumption and performance, that seem to me to be pretty significant issues for pushing a great many pixels on a power and CPU constrained device. A different issue, to be sure, but what I'm hearing is a kind of brusque assurance that Apple can and will make this happen, that it's actually just a matter of wanting to do it and getting it done. I think there are bigger hurdles than that.
The backlight is the same, the transistor draw in the LCD itself is quite small, but not trivial. The only issue would be the necessary GPU horsepower. The new rumored GPU is several years engineering, and a couple process steps smaller -- meaning it is probably close to on par power draw at the higher fill rates compared to the current GPU. And with the OpenCL support a whole bunch of CPU draining heavy calcs can be funneled to the GPU where those transistors are drawing non-idle power already anyways. That means less CPU power draw is traded for more work for about the same power on the GPU -- a power win.
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
I'm guessing that picking this "oddball size" must have ultimately been a visual and tactile choice, maybe based on whatever sense of harmonious proportion that led to such standards as 4:3 and the 8x10, to both of which it is close. Now that we have it, I think it feels absolutely right. But I think it was designed for the photograph, the page, and old cinema, and it flies in the face of the oppressively stretched 16:9, an alien form when it migrates from the big screen. 16:9 doesn't belong in the hands. All of which is to say that Apple has had this size all to itself. They're clever that way, as addabox says.
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
So there are two completely compelling reasons to double the resolution.
One is display of text, and especially photographs and film, their main 'content' reasons for picking the size to begin with. We will be bowled over by how pictures will look, mostly because we will have never seen such clear and beautiful large images, outside of medium-format slide viewers with Kodachrome inside, and how many of us have seen those? "And yet it moves!" It will be irresistible.
The other reason is that they can't allow others to get the jump on them again, if what you say is true about Android tablets to come. If it doesn't happen in April, there will have been technical problems or cost considerations. They must do it eventually. I hope it's this time around, given what you say about the competition.
Oh, and thanks for the technical background. Bracing, if true . . .
Apple's current price line is $499, $599 and $699. Is it unreasonable to think they may now introduce a brand new product at the $799 and $899 price? The Mac Book air starts at $999. Apple will effectively fill all the price points they now have missing while introducing a much better iPad.
If price and price margins are the ONLY reason it can't exist, then why not simply make something way nicer than the current models for a $100 or $200 more.
I agree from the cost side alone. I'd love a 2x screen but I can't imagine what that will add to the cost of the unit.
I read an estimate of approx. $150 in extra cost (not necessarily including other increased costs of supporting it). 40% gross margin brings that to $240, so if that's accurate and there's to be such a Pad this year, $799 is a likely base (no 3g) price.
As I commented on another board, the unveiling presentation practically writes itself:
4 times the CPU (dual core A9)
4 times the graphics (dual core SGX 543)
4 times the RAM (1GB)
4 times the screen (2048x1536)
Still just $499.
Nope. Too much component cost increase - not enough "production experience" amortized for such new glass to bring a mass price this year. See above for estimated retail.
Out of curiosity, does anyone know the resolution of a high-quality magazine? I'd imagine that this would roughly be a resolution end-point.
Not really a good comparison. The reason is this: the standard for magazine publishing photos is 300dpi. Photographers are instructed to supply art at this size and no smaller than the size the photo will be eventually laid out at. In other words, if the photo takes up a full page, the supplied photo must be 8x10 at 300dpi. But usually photos are downsized by the art director to fit the layout.
But printing and displays are so very different: for one thing ink spreads -- which is actually a good thing when you are trying to avoid the reader seeing individual drops of ink.
2400 dpi for line work such as black text. 175 line screen for images. The CYMK each have a possible dot size of 1 to 256 scale (256 being solid 100% overlap). It isn't exactly the same relationship to a monitor. The reason people say 300 dpi is because that used to be twice the line screen. Now we use 350 because we print at 175 line screen. The reason you want to be twice is because of antialiasing. In a worst case scenario the placement of your image would be exactly half way between the dot grid. That way you have enough resolution to accurately average the color values. You can kind of think of 175 dpi as the viewing resolution on a printed page although it is not entirely accurate since it is a completely different imaging science.
The reason for 300 dpi is because of the substrate and dot gain. 300 dpi is what a coated stock can handle without killing the Rosetta.
Newsprint is 150 lpi or up to 220 dpi depending on the quality of the paper.
Flexo printing on a burlap sack can be as low as 14 dpi.
Coated inkjet paper has special (expensive) chemicals on it to sustain a smaller dot... But doesn't use the same method as commercial offset printing.
Imagine a mag printed on an inkjet or dye-sub... The thing would cost $200.
I could get into 'art' printing, like for expensive books, but that again is a different process.
In terms of the original question no one's specifically mentioned the frequent confusion between "dpi" (a measurement of printed dots of ink per inch) and "ppi" (pixels/per inch on the screen). See: http://www.rideau-info.com/photos/mythdpi.html. You can make a 300 dpi print of the most pixelated source material, which will quite faithfully translate its unglory to a page as easily as one with deep pixel density that looks great at smoooth retina resolution
I hate 'mobile' web sites you cannot avoid with a passion. Some simply refuse to let you get around it. There should always be an option (and a savable one) to select the real site if you wish. Now having said that an option to access a true iPad / iPhone style version (as opposed to something designed for a BB user) is nice at times as an option (like when on slow connections) but only as an option. I think I said option too many times here
It was your option....
But I agree!
The LLM 300ppi LCDs are ancient history. Pre-Google days for the contract, Google in beta during the delivery. There are no links unless you want to make blind searches in the Wayback Machine, be my guest. The production development contract to delivery was less than 2 years. You won't find that anywhere, you have to know someone who was asking for them. There were only minor press releases about the tech when it was delivered. It's not like doing nuc simulations was something that was going to get a lot of intarweb mileage.
It's the fact that the production ppi ratio went from ~150 on the best SGI displays to a delivered run of 300ppi by 30" in less than than 2 years despite the problems with > 22" glass cracking that is the basis of caling the scaling solved. Only because it was. 10 years ago.
Its not an open problem that needs R&D. Its a previously solved problem that needs a market willing to pay enough that reconfiguring the fab to do a new non-standard screen size.
No. What other form factor needs a screen like a 10" iPad screen? None, they don't exist.
The iPad itself is the same argument. <sarcasm> Why would anyone make a 10" tablet. They just haven't. It couldn't be successful or it would already be produced.</sarcasm> But Apple did, because they're clever that way. I'm not quite seeing your argument anymore because it's the exact same argument. But now that a form factor exists that would use this admittedly oddball sized screen your lack of seeing the potential is just your lack of seeing the potential.
The backlight is the same, the transistor draw in the LCD itself is quite small, but not trivial. The only issue would be the necessary GPU horsepower. The new rumored GPU is several years engineering, and a couple process steps smaller -- meaning it is probably close to on par power draw at the higher fill rates compared to the current GPU. And with the OpenCL support a whole bunch of CPU draining heavy calcs can be funneled to the GPU where those transistors are drawing non-idle power already anyways. That means less CPU power draw is traded for more work for about the same power on the GPU -- a power win.
These aren't Apple special engineering tricks, these are just the established pattern in mobile device (including laptop) evolution. All quite plausible. Leaving the question -- does Apple see it as the correct economic time to pop the tech into the market, rather than is the tech generically ready. I can guarantee the Android tablets will have a couple that ship with greater than 1024x768 screen resolutions by the end of the year and use similar multi-core GPU technology and dual core CortexA9 derives SoCs. Most of those specs are already in play for spring summer tabs, the collective community won't miss an opportunity to attack the iPad exactly the same way Android phones with higher resolutions as a main selling point ate into iPhone growth before iPhone4 last summer.
Well I guess you're right either way then....
If Apple fails to deliver the fabled high res iPad, they just decided it wasn't the "correct economic time", even though, apparently, there's absolutely no technical or financial hurdle between them and doing so.
I think you're conflating "is technically possible", which I've never disputed, and "is financially viable", which you're hedging on and which is the basis of my skepticism.
You seem to want to argue that it's actually nothing at all, aside some startup investment, to start cranking out the high res iPad screens, while leaving yourself an out with wondering after Apple's commitment to the technology, but I don't think you can have it both ways. Your insisting on a scenario wherein it would make no sense whatsoever for Apple not to take the iPad 2 to 2048x1536, so we'll see.
As far as Android devices with somewhat higher resolution screens and dual core processors, sure, but what of it? "Somewhat higher resolution" being broadly marketed doesn't make "massively higher resolution" more likely.
I'm guessing that picking this "oddball size" must have ultimately been a visual and tactile choice, maybe based on whatever sense of harmonious proportion that led to such standards as 4:3 and the 8x10, to both of which it is close. Now that we have it, I think it feels absolutely right. But I think it was designed for the photograph, the page, and old cinema, and it flies in the face of the oppressively stretched 16:9, an alien form when it migrates from the big screen. 16:9 doesn't belong in the hands. All of which is to say that Apple has had this size all to itself. They're clever that way, as addabox says.
Yes, and all I meant by oddball size of the rumored 2x iPad display was it's uniqueness resolution-dimension pairing which absolutely nothing else uses.
Yes, and all I meant by oddball size of the rumored 2x iPad display was it's uniqueness resolution-dimension pairing which absolutely nothing else uses.
I think QXGA is quite common for certain applications and has been for awhile.
Well I guess you're right either way then....
If Apple fails to deliver the fabled high res iPad, they just decided it wasn't the "correct economic time", even though, apparently, there's absolutely no technical or financial hurdle between them and doing so.
I think you're conflating "is technically possible", which I've never disputed, and "is financially viable", which you're hedging on and which is the basis of my skepticism.
You seem to want to argue that it's actually nothing at all, aside some startup investment, to start cranking out the high res iPad screens...
Mostly agreed.
while leaving yourself an out with wondering after Apple's commitment to the technology, but I don't think you can have it both ways. Your insisting on a scenario wherein it would make no sense whatsoever for Apple not to take the iPad 2 to 2048x1536, so we'll see.
It's not a matter of me having it both ways. The decision tree is simple, either Apple has an agreement with a producer to crank out unique screens, or they don't. Technically, there's nothing special about producing screens in that size, but there are zero known fabs doing it. Meaning either Apple contracted a quiet set-up or Apple didn't.
That's not having it both ways, it's just saying all the poo-poohing over !ZOMG! 2x display density isn't possible! Is just plain bunk.
That simply leaves us with the flat-out obvious, Apple only does something tech-wise when they think it is economically advantageous.
As far as Android devices with somewhat higher resolution screens and dual core processors, sure, but what of it? "Somewhat higher resolution" being broadly marketed doesn't make "massively higher resolution" more likely.
There I'll just plain disagree. I'm convinced Steve and company aren't happy about all the screen size grief they took over the first half of last year. If there was ever a second-order financial motivation of removing a potential marketing weakness while simultaneously allowing a huge pronouncement of massive technical progress (even though it really wouldn't be that massive) and using that to make it EXTREMELY hard to be technically passed in the next 9 months or so, this would fit the bill. Using it to close out any significant threats in the market space until Apple is as long term strong in tabs as they were in year two of the iPhone.
What the real $20 question is is this: Was Apple thinking this way in first quarter last year? If so the 2x screen has a real chance of showing up sooner rather than later. If they didn't start thinking this way until sometime in the summer I don't think it could happen yet.
Either way I'm not predicting, I've merely stated why the "it is not possible arguments" are incorrect. And then added my own 2 cents on why it might make sense on an aggressive timescale. Very different from saying it will happen. Hedging, yes a bit. I've been wrong on both sides of predictions enough before, so now I just dabble in the technical discussion part where I actually have a clue and leave the actual business to Apple.