5" 1080p screens becoming a reality just in time for P2?

[Exophase]

Soft rendering or mostly soft rendering is typically what you do in desktop stuff (file management, office stuff, browsers etc), and this is also the kind of stuff you really want to do at the native resolution because it involves lots of text in relatively small fonts that will degrade with scaling.

That stuff is mostly static and therefore not a lot of CPU time is needed rendering it. What you need to look out for is heavy animation like in games/emulators or window compositing/effects. Hopefully both of which will typically not involve software rendering at high resolutions.


Of course, even the things you refer to generally involve a bunch of solid color fills or repetitive texture blits that can be hardware accelerated; font glyphs can be cached to a small amount of texture space (you only tend to use a few dozen!) and efficiently hardware accelerated as well. I'm looking at my browser on this page and I see a lot of stuff that can be done faster if not with pure brute force software rendering.

I agree that the increase in RAM use and bus bandwidth is not gonna cause problems, but then again, I wouldn't count on the P2 to have 8GB of RAM, imo a more realistic number would be 1 or 2 GB. A doublebuffer framebuffer is still only 16MB at full HD, so it's certainly not problematic, but it is more than the 3MB it currently takes on the Pandora 1.

I wouldn't count on 8GB either, or at least I hope OPT has the sense not to push that design. But 4GB is likely, with 2GB being something of a minimum; would be genuinely surprised if they settled on only 1GB.

I agree that it would be nice to get 1080p output over HDMI, and that it should not be a huge challenge to find hardware that can do that (esp. since those phones will be able to do it). But I don't think the small screen should be at that resolution, because it's pointless to make the resolution that much better than what the human eye can see, and because we want a battery life that is better than that of those phones. Less pixels will mean better battery life, mostly because the screen itself will need less power.

You already know I'm on your side on this

Yes, and of course for emulation it would be better to have a 4:3 aspect ratio, say 1280x960. But that's probably impossible to find. I prefer 16:10 (1280x800) to 16:9 (HD or Full HD) though, because that's at least a bit closer to 16:12.

Okay, let's assume for a second that the maximum height of the display is fixed regardless of its aspect ratio; ie, that using a narrower screen won't let you use a taller screen. Then the question is, does it let you use a more narrow device? In Pandora's case I definitely don't think so - a relatively wide width is dictated by having the gaming controls and keyboard. So while emulators might waste space on the screen, switching to a different aspect ratio doesn't buy you anything if it doesn't allow you to either make the screen bigger or make the device smaller. You'd just be trading dead space on the screen for dead space on the unit.

 

[Grench]

So I don't think driving 1080p content will necessarily be a huge challenge, which is good, because regardless of whether or not it has a 1080p screen it IS nice to be able to put 1080p content out over HDMI.

Agreed.


Most SoC already have hardware scaling, right? So that would fairly well moot the point of magic multiples - but would potentially introduce issues with horizontal stretching if the aspect ratio isn't locked.


Having pixels 'too small to be seen' yields a scenario where some pixels from a low resolution image get 3, 4 or even 5 or more pixels worth on the result - but should still resolve pretty much as well or better than they would on a native low resolution screen.


Also - I like the idea of having what is on the small screen be pixel to pixel the same as what gets sent to a big screen - silly as it may be.


If we're going to have the 'ultimate palmtop computer' and if these aren't obnoxiously expensive and if we have the hardware to drive them and if they're, "any good," in the Top Gear / British use of that phrase, then I see lots of reasons to use them and not many reasons not to.


Right now we're entering a 'magical time' on the P2 specs. Hardware components being released inside the next 3+/- months should be -possible- for a Pandora 2 being released sometime 9-15 months from now.

 

[vcoleiro1]

I can only see one real point people have raised about these 1080p screens , and that is the possibility of it draining batteries a lot quicker. The only thing I can say to that, is we really need to wait and see the reviews that will shortly come out on that. Then we can decide - no point jumping the gate.


As far as the extra pixels making a diff, again some reviews which will shortly come, will be good. The preliminary feedback from people who saw the LG 1080p 5" screens being demoed all say that it is noticeably better. Pixel density and and perception theories aside, what it is like in real life is what matters.


As for processing power, I don't think that will be an issue, we are getting to the point, especially now that Intel have entered the race, where mobile SOC's will handle 1080p in there strides - if we are not already at that point (which I believe we are)

 

[_wb_]

Of course people will say it looks noticeably better. Even if you give them the same thing twice, and tell them the second one is the new model, they will say it's noticeably better. That's called the placebo effect.


We don't have to wait for reviews to tell that 1080p screens will eat more battery than 800p or 720p screens. If it's the same technology, it will eat more power - probably not linearly, the efficiency per pixel will increase, but it will eat more power.

 

[vcoleiro1]

Of course people will say it looks noticeably better. Even if you give them the same thing twice, and tell them the second one is the new model, they will say it's noticeably better. That's called the placebo effect.


We don't have to wait for reviews to tell that 1080p screens will eat more battery than 800p or 720p screens. If it's the same technology, it will eat more power - probably not linearly, the efficiency per pixel will increase, but it will eat more power.

I think the point here is that tech reviewers are saying it is noticeably better. It's a bit disrespectfull to them, to say that this is just a placebo effect. Let's just wait until the flood of reviews come in. We will see then .


As for the Battery, you could assume it's going to eat more battery life. The point is you don't know how much it will reduce battery life ie whether it's an acceptable 10 minutes or unacceptable 3 hours or somewhere in between. Again , we will just need to wait and see.


No point discounting the technology without knowing any data on it.

 

[diligentcircle]

I think the point here is that tech reviewers are saying it is noticeably better. It's a bit disrespectfull to them, to say that this is just a placebo effect.

How is that disrespectful? Are you suggesting that reviewers are some kind of divine beings? If not, they can just as easily suffer from the placebo effect as anyone else.

 

[vcoleiro1]

^ Im suggesting that they do know a little of what they talk about and are less likely to just say somehting is better.


The point still remains, let's not dish the technology until we know some facts . Right now we know nothing about the actual impact on battery life or how much better the screen is (if any) .

 

[sswam]

of course for emulation it would be better to have a 4:3 aspect ratio, say 1280x960. But that's probably impossible to find. I prefer 16:10 (1280x800) to 16:9 (HD or Full HD) though, because that's at least a bit closer to 16:12.

4:3 might be better for retro gaming, but even with the current Pandora the dimensions of the screen are 5:3 which is closer to 16:9 than 4:3. The device is 16:10.2, which is also closer to 16:9 than 4:3. Pocket width is a constraint, depth not so much! We already have black bars down the sides of our 4:3 games, both on the screen, and the speakers area.


I think we don't really lose anything much with a hi-res screen. It will surely be able to cope well enough with the desktop, if the software is well written; and games can run at any resolution. The hardware scaler can scale them up to 1080p free of charge. There will be some slight blurring, I don't think it will be very obvious or annoying. Even now scaling a 640x512 res to fit in 800x480, there is some blurring but it is not really annoying. We can also use exact pixel doubling, and just accept the 6.25% borders on top and bottom.


For new games such as your microbes, and even the desktop, we could optionally run at 980x540, and zoom it up with the hardware scaler. That resolution is still higher than the current Pandora's 800x480. It might be useful to do this for apps where the fonts appear small at the higher resolution, or to boost performance (but that will not likely be necessary).


I'm working on emulator for the BBC Micro, which has 5:4 aspect, with unsual resolutions such as 640x256. The hardware scaler with Notaz SDL can scale it up to the Pandora screen quite nicely. It can even scale down higher resolutions such as 640x512 to fit in 640x480, and can still see distinct pixels.


The iCP2 is more elongated than the Pandora, I suppose the Pandora 2 will be the same shape - closer to 16:9, and less thick than the Pandora 1. This can fit nicely in our pockets!

 

[_wb_]

Even if they can somehow manage to make a 1080p screen that uses less power than a 720/800p screen, that should mean that the next generation of 720/800p screens, updated with the same technology, will use even less power. That's something we know already, no need to wait for test results to arrive. Of course it's possible that for some reason the industry will not incorporate that 1080p low-power technology in the 720/800p screens, and in that case, we should consider the 1080p screen anyway. But I don't think that scenario is very likely.


Consider the following analogy: say you want to have a light in your living room. You currently have a 60W old-fashioned light bulb, and you're considering to replace it with a more efficient LED light. You could install a 40W LED light, and it will consume less power, but it will be so bright that you get blinded, so you end up dimming it a lot, which (to make the analogy correct) does not really reduce its power use and certainly does not reduce its purchase price, but at least it doesn't blind you anymore. All I'm saying is: maybe it's better to install a 15W LED light instead, which will consume even less power, be cheaper, and still provide more light than the original bulb you had, without blinding you. And you can say: wait, maybe there's a new kind of LED, let's call it SuperLED, which is even better: a 30W SuperLED produces even more light than the original 40W LED light, but we have to wait for the test results to be sure. And then I say: OK, but even if that's true, then I would still prefer to install, say, a 10W SuperLED rather than a 30W one.

 

[Grench]

Even if they can somehow manage to make a 1080p screen that uses less power than a 720/800p screen, that should mean that the next generation of 720/800p screens, updated with the same technology, will use even less power.

Not necessarily. It is the back light that consumes a big chunk of the juice on the screen. Using different masks to allow more of the back light through, etc - those take some too.


I could see a scenario where a 1080p screen only uses >0% to 5% more juice than an equivalent technology 720p screen which would only use >0% to 5% more juice than an 800x480 on equivalent tech - IF anyone bothered to fit that tech into something so deprecated.


Yes, there is probably a point of diminishing returns. But consider this - these new 1080p screens are being designed specifically for a phone-based market that is VERY sensitive to thickness, power consumption AND display quality. They aren't going to produce a screen that uses dramatically more power than the ones on the market today - it simply wouldn't sell.


We'll have to see - but the reviewers seem to be a bit gaga over these 1080p screens. Whether or not the eye can resolve the difference between a 5 pixel light blob or a 6 pixel light blob isn't as relevant as the end product's look, feel and real performance.


I'm not opposed to a 720p screen, but I see a lot of advantages to a 1080p IF they're any good.

 

[Exophase]

So how do you explain iPad 3 using so much more power than iPad 2, regardless of whether or not the GPU (only major addition in the SoC) is strongly exercised?

 

[Grench]

So how do you explain iPad 3 using so much more power than iPad 2, regardless of whether or not the GPU (only major addition in the SoC) is strongly exercised?

Sorry, not relevant. The power drain from a 10% power increase on an already hungry 10" screen is probably at least four times the net drain from an equal % increase on a 5" screen.


Please stick to 5" screen examples comparing 1080p, 720p and 800x480.


You can't - because the numbers aren't in yet.


So you're rejecting the latest technology based on assumptions that you can't prove - I'm saying, wouldn't it be neat if it is 'any good'. Can you give it a chance so we can find out if it is or is not, 'any good' rather than dismissing it out of hand?


It is being presented as a possibility that is worthy of evaluation. That is what the P2 SPAM forum is for. Chill out a bit.

 

[rohezal]

One of my professor worked with the DVBT standard and told us a bit about screen size, distance screen / user and resolution. It was suprising, at a given relativ small distance (and a given screen size), you can't see single pixels. On a small display it doesn't matter if its full hd 1920*1080 or just 1024*768. Just make sure a full website fits on the screen without scroll bars.


But it is true that the most produced components are the cheapest ones. Try to get old ram these days, you pay the double amount of money for it (often the double amount of money for half the amount of Megabytes).

 

[Grench]

The iCP2 is more elongated than the Pandora, I suppose the Pandora 2 will be the same shape - closer to 16:9, and less thick than the Pandora 1. This can fit nicely in our pockets!

This is probably the most overlooked of highly important factors - pocketability.


I'm a big guy. Big pants. Capacious pockets. I carry my Pandora in the front left pocket.


Frankly, the Pandora prints like I'm smuggling a brick - it pushes the boundaries of acceptable 'thickness'. If it weren't for all of the cool things it can do, I wouldn't even consider carrying something that thick around with me.


Every mm that can be shaved off the Pandora's thickness when going to the Pandora II is worth the effort.

 

[Exophase]

Sorry, not relevant. The power drain from a 10% power increase on an already hungry 10" screen is probably at least four times the net drain from an equal % increase on a 5" screen.

Please stick to 5" screen examples comparing 1080p, 720p and 800x480.


You can't - because the numbers aren't in yet.


So you're rejecting the latest technology based on assumptions that you can't prove - I'm saying, wouldn't it be neat if it is 'any good'. Can you give it a chance so we can find out if it is or is not, 'any good' rather than dismissing it out of hand?


It is being presented as a possibility that is worthy of evaluation. That is what the P2 SPAM forum is for. Chill out a bit.

What you're saying makes no sense. Why would increasing the resolution 4x but keeping the size, luminosity, etc the same greatly increase the power consumption on a 10" screen but negligibly increase it on a 5" screen? It isn't a 10% increase, it's a 70% increase - for the entire unit. Meaning that if everything else takes the same power, and there's no reason to think it wouldn't if you aren't using the new features like the bigger GPU, then that means that the power consumption for the screen alone has increased by substantially more than 70%. Yes, the screen is 4 times larger than a 5" would be in a Pandora 2 - the battery is also over 10 times the capacity. On mobile devices like phones the display takes a LOT of the active power. Pandora happens to be using a very low power LCD compared to others in its class (which may also be reflected in its response time which is not exactly class leading; there's probably some correlation between power consumption and response time, ie wrt drive strength), but you could hardly call its impact on power consumption minor; from what I know of the numbers I estimate it's at least 33%.


And who exactly do you think you are telling me to chill out? I gave you a single sentence asking you to explain it and you go off on a poorly reasoned rant about how it doesn't need to be explained. Just because you're the (blind) optimist and I'm the realist doesn't mean I'm the one who needs to chill out. I'm not dismissing it out of hand, I'm just giving a reasonable data point that casts doubt on your 0-5% estimation. I asked you to explain it and you couldn't, and you're getting pissy over it again. Sure it's worthy for evaluation, and I'll be the first to admit that if it has no disadvantages whatsoever it should be used, but it's not helpful to get people's hopes up by telling them you expect this with no good basis.

 

[Grench]

Sorry, not relevant. The power drain from a 10% power increase on an already hungry 10" screen is probably at least four times the net drain from an equal % increase on a 5" screen.

Please stick to 5" screen examples comparing 1080p, 720p and 800x480.


You can't - because the numbers aren't in yet.


So you're rejecting the latest technology based on assumptions that you can't prove - I'm saying, wouldn't it be neat if it is 'any good'. Can you give it a chance so we can find out if it is or is not, 'any good' rather than dismissing it out of hand?


It is being presented as a possibility that is worthy of evaluation. That is what the P2 SPAM forum is for. Chill out a bit.

What you're saying makes no sense. Why would increasing the resolution 4x but keeping the size, luminosity, etc the same greatly increase the power consumption on a 10" screen but negligibly increase it on a 5" screen? It isn't a 10% increase, it's a 70% increase - for the entire unit. Meaning that if everything else takes the same power, and there's no reason to think it wouldn't if you aren't using the new features like the bigger GPU, then that means that the power consumption for the screen alone has increased by substantially more than 70%. Yes, the screen is 4 times larger than a 5" would be in a Pandora 2 - the battery is also over 10 times the capacity. On mobile devices like phones the display takes a LOT of the active power. Pandora happens to be using a very low power LCD compared to others in its class (which may also be reflected in its response time which is not exactly class leading; there's probably some correlation between power consumption and response time, ie wrt drive strength), but you could hardly call its impact on power consumption minor; from what I know of the numbers I estimate it's at least 33%.


And who exactly do you think you are telling me to chill out? I gave you a single sentence asking you to explain it and you go off on a poorly reasoned rant about how it doesn't need to be explained. Just because you're the (blind) optimist and I'm the realist doesn't mean I'm the one who needs to chill out. I'm not dismissing it out of hand, I'm just giving a reasonable data point that casts doubt on your 0-5% estimation. I asked you to explain it and you couldn't, and you're getting pissy over it again. Sure it's worthy for evaluation, and I'll be the first to admit that if it has no disadvantages whatsoever it should be used, but it's not helpful to get people's hopes up by telling them you expect this with no good basis.

OK - lets simplify this. There are very basic physical reasons why a % increase in power usage in a 5" screen is about 1/4 the net power requirement of the same % increase in a 10" screen.


-A 10" screen is roughly four times the surface area as a 5" screen - regardless of resolution.


-Ignoring the CPU/GPU drain for the moment, the biggest power draw in screen technology is the back light - regardless of resolution.


-For the sake of simplicity, lets say we have a 10" tablet that has a back light consuming 1W/hr to produce 300cd/cm^2.


-Under that same logic, a 5" tablet would consume ~0.25W/hr to produce 300cd/cm^2.


A 10% rise in power usage on the 10" screen above is 0.1W/hr.


A 10% rise in power usage on the 5" screen above is 0.025W/hr.


So, logically, smaller screen, less backlight, less drain on a quadratic function. The screen tech itself for 1080p vs 720p isn't likely to be a gigantic power hit. Assuming they use anything remotely modern, the GPU in the SoC will be already set and suited for driving 1080p on two channels. I'm betting the GPU/SoC power draw difference between on/running 720p and on/running 1080p is actually relatively small.


Is there likely to be a power usage difference between the two? Sure - could be. But the 1080p screen is being made with the most bleeding edge efficient tech that they can throw at it - and the 720p screen is based on tech a year old. It MAY be that the 1080p screen has thinner masking, better back light transmission and higher efficiency than other screens. They may not even back-run the technology to lower resolutions as the process change from good 720p efficiency to better would cost millions. I.e. for all we know the 1080p screen could take less power to operate than available 720p screens. We don't know - yet.


So, back to why I asked you to chill out. The numbers aren't in on these. We don't know if they're 'any good'. The timing is about right though, the tech looks cool, it's getting great reviews. I think it's worthy of a good look. If OpenPandora can get one and compare it side by side with whatever other screen tech they're considering - that's worth the time. Quit saying it can't/won't work and rejecting it. You don't have any more information than any of the rest of us yet. It may suck - it may not. It may be the most gorgeous screen ever put in a device and still have ghosting so bad it's useless for anything but a picture wallet. We don't know. It IS worthy of consideration though.


Me, I'd still love to see a PixelQi screen in it, but the last I knew they weren't in production of anything of appropriate size.

 

[Grench]

One of my professor worked with the DVBT standard and told us a bit about screen size, distance screen / user and resolution. It was suprising, at a given relativ small distance (and a given screen size), you can't see single pixels. On a small display it doesn't matter if its full hd 1920*1080 or just 1024*768. Just make sure a full website fits on the screen without scroll bars.


But it is true that the most produced components are the cheapest ones. Try to get old ram these days, you pay the double amount of money for it (often the double amount of money for half the amount of Megabytes).

You don't need to be able to see single pixels for there to be a visible difference in display quality at equal sizes. Is there a limit? Probably. Are we there yet? Getting close in the 380ppi and greater range. There probably won't ever be a need for displays beyond 1080p at 5". I can tell you, though, there is a very notable difference from 2.5' away between 1280x720 at 4.8" (GSIII) and 800x480 at 4.3" (Pandora). You'd be hard pressed to see any individual pixels in either, but the image quality is dramatic.


This gets back into the big question of, 'What is the primary purpose of the Pandora II?'.


If your answer is, 'just a game console', then lower resolutions in discrete multiples are potentially better.


If your answer is, 'as a general use computer' or 'as a movie player' or 'as a presentation device' or 'image editing' or 'as a way to take over my desktop/server/Remina/etc...' then there is a logical argument that higher resolutions are better - and let the hardware scale anything that doesn't fit.


For me, games is a secondary level function for my Pandora. Your use will vary - which is why I like this device so much.

 

[Exophase]

OK - lets simplify this. There are very basic physical reasons why a % increase in power usage in a 5" screen is about 1/4 the net power requirement of the same % increase in a 10" screen.

-A 10" screen is roughly four times the surface area as a 5" screen - regardless of resolution.


-Ignoring the CPU/GPU drain for the moment, the biggest power draw in screen technology is the back light - regardless of resolution.


-For the sake of simplicity, lets say we have a 10" tablet that has a back light consuming 1W/hr to produce 300cd/cm^2.


-Under that same logic, a 5" tablet would consume ~0.25W/hr to produce 300cd/cm^2.

None of this matters. NONE OF IT.


You made the statement that the power draw increase of the screen only could be only 0-5%. According to numbers I worked out, the iPad 3's screen, while keeping the same size and luminosity of the iPad 2's, uses around twice as much power. I don't have my work for that math here so I'll have to show it to tomorrow. But if you don't like that, it's at least easy to show it's over 70% - because the battery is 70% bigger and it lasts less time in Anand's web browsing test, and since that's using wifi and not stressing GPU NOTHING ELSE in the iPad 3 should be contributing to that power increase.


So please, explain to me why quadrupling the resolution in a 10" device would increase the power consumption by > 70%, while increasing it by 2.25x only nets a 0-5% increase. Give me one reason to think they wouldn't scale similarly. Your argument of "the 720p screen is year old tech!" is awfulunless you seriously think that when Pandora 2 comes out everyone - even the lowest end devices! - is going to be using 1080p screens. You don't believe something that stupid, right? Good, then you should realize that the 720p screen tech will move with the 1080p screen tech, assuming that screen tech really goes down in power very much over that time span.


Stop telling me to chill out because I'm applying common sense and you don't like that it interferes with your fantasies.

 

[_wb_]

This gets back into the big question of, 'What is the primary purpose of the Pandora II?'.


If your answer is, 'just a game console', then lower resolutions in discrete multiples are potentially better.


If your answer is, 'as a general use computer' or 'as a movie player' or 'as a presentation device' or 'image editing' or 'as a way to take over my desktop/server/Remina/etc...' then there is a logical argument that higher resolutions are better - and let the hardware scale anything that doesn't fit.


For me, games is a secondary level function for my Pandora. Your use will vary - which is why I like this device so much.

For me it could be any of those things, and for some of them (movie player, presentation device, something you connect to an external monitor to do things like image editing, web browsing, etc.), it makes sense to have hdmi output at full HD, because that's clearly going to be a widely supported standard for several years to come. For none of those things however, it makes sense to me to have a full HD 5" screen.


Car analogy: suppose we have to choose between a car that can drive 400 km/h and one that can drive 800 km/h. I would say: 400 km/h is enough, because you're only allowed to drive 120+-10 km/h anyway in most countries.


Piano analogy: a normal piano has 88 keys (A0 to C8). We want to build a new piano, and we have to choose between 100 keys (C0 to G8 or something) and 250 keys (C-4 to C20 or something). I would say: 100 keys is enough, because those extra 150 keys are in a range that the human ear cannot perceive anyway.


Lightbulb analogy: we can choose between a 30W LED light and a 80W LED light (both can be dimmed) to put in our living room. You typically need only 10W worth of light - much more than that and it starts to become too much, blinding you. I would say: the 30W one is enough.


etc.

 

[ZXDunny]

Grench -


I might be being a tiny bit stupid here, but wasn't the point of improved screen hardware that there would be no backlight? Ie, we'd likely be looking at an OLED or similar system. Hence your point about power consumption being mainly the backlight wouldn't hold water.


Or do you think we should going for LCD+backlight for our next display?


You mentioned that there was a marked difference in image quality between the S3 and the Pandora - I doubt that's entirely down to resolution, and more down to superior display lighting.


D.