Back to share more of our display measurement results from the new iPad. Side note before we jump in: this is a somewhat technical post, if you aren’t familiar with the general workings of an LCD, this great live teardown by Bill Hammack is worth watching: http://youtu.be/jiejNAUwcQ8
There are two ways to improve the color gamut performance of an LCD display: you can either make the backlight better or the color filters better. In both approaches, the goal is the same: to make red, green and blue light as pure as possible. The LCD display mixes these three primary colors to make all the other colors you see on screen, thus, the more pure the individual pimary colors are, the better all colors on screen are. Based on our measurements, it looks like Apple focused on the color filters for this new display, let’s take a closer look.
In the color spectrum chart below, you can see the result of some of the color filter changes that Apple made. Notice how the red peak (on the right, in the 600 nm range) has moved to a longer wavelength. This change in wavelength means reds on the new iPad will have a deeper hue, will be less orange and more distinctly red.
Another interesting thing to look at here is the blue peak at about 450 nanometers. In our last post, we noted that blue got the biggest boost with the new display. However, the blue peak did not change in wavelength or in shape, only amplitude (or brightness), which does not affect color. So what explains the dramatic improvement in blue seen on the new display?
The above spectrum isn’t telling the whole story. It was measured from a white screen, in other words a screen with all three primary colors turned on. We see very different results when looking at a screen with a blue image, where only the blue sub pixel filters are open.
This chart shows us only the light that is allowed to pass through the blue color filters. We can see the same blue peaks that we know from the white spectrum, but there’s also some extra light getting through – notice the two small tails to the right of the blue peak? That’s green light from the backlight leaking through the blue filter.
This means that when the iPad display needs blue light to make an image, some of that green comes along with the blue whether you want it or not. You will notice that the green blip is smaller on the new iPad, meaning less green is leaking through and a purer blue is displayed. Take a look at the comparison shot here and you can see how just a hint of that green leakage is making the iPad 2’s blue (on left) appear slightly aqua by comparison.
Leakage like this happens because its very difficult to make a truly perfect color filter and even harder to make one that is efficient enough for a mobile display. The reason is basic physics – a better color filter is narrower, allowing only the desired color through. However, the narrower you make the filter, the less light it lets through, and less light through means the display has to be driven harder to maintain brightness. This directly affects battery life, partially explaining the new iPad’s need for a larger battery. Based on our experience, we estimate that the color improvements alone in the new display probably cause it to consume about 20-30% more power than the iPad 2’s screen.
Perfecting the color performance of a display is a critical engineering challenge and worth highlighting because its one of those tiny details that Apple is so great at. Just making this small improvement in light leakage from iPad 2 to the new iPad accounts for a stunning amount of improvement in color performance and, most importantly, it makes for a richer user experience.
Pingback: Blog post: Apple's new iPad boasts better colors - how did they do it? | Nanosys
Thanks for posting this, very interesting to see. Here are some questions/comments:
1. Just to clarify, does the first chart come from measuring the spectral output of a “100% on” white for both devices? And the second for “100% blue”?
2. The graphs show an overall increase in brightness – is this due to the backlight, the filters, or both? Do you happen to know if the backlighting is the same? My guess is that the new iPad has a slightly brighter backlight.
3. “Blue got the biggest boost” – not really; green has about the same percentage increase at peak, and probably a larger absolute increase (curve integral).
4. “That’s green light from the backlight leaking through the blue filter.” – The backlight is white, so what do you mean by this? I think it’s more likely to be just part of the blue filter response.
Thanks for your questions
1. That’s correct, “100% on” for each color as you described it.
2. Seems like the new iPad does have a slightly brighter backlight. It’s worth noting that the new iPad filters also move the green peak a bit closer to a wavelength of green that our eyes are more sensitive to. This makes the display seem even brighter as well.
3. Probably was not clear enough on my part, I was referring to saturation not brightness there. If you take a look at a new iPad, the blue is the most striking difference in terms of saturation of the three colors.
4. The backlight is white as you say but the key thing in this case is the backlight’s spectrum. The iPad’s white LED backlight is not evenly composed of all wavelengths like sunlight for example. It actually has essentially two colors- a very narrow blue peak and a broad yellow green peak. What you are seeing in the chart is that narrow blue peak and the beginning of the green peak passing through a fairly wide filter. Take a look at this post from October where we discussed backlight spectrums in more detail.
Thanks for the clarifications. So it seems there are two significant differences in the new iPad – brighter backlighting and a narrower blue filter (which reduces the “leaking green” in question). I don’t suppose you have a spectral graph of both (iPad2 and new iPad) backlighting? (Not sure if your October post shows the iPad2 backlight in particular, or just a typical LED backlight.) I’m curious to see if it’s only a change in magnitude.
We haven’t had a chance to measure the backlight of either iPad yet. I will look into doing another post on that topic.
Pingback: Dot color explains why colors look so nice on the new iPad | Apple Stocks
Since this is a display panel made FOR not BY Apple, it seems that we will see the same improvements in other companies products, soon, yes?
That depends on if the engineering improvements were made by Apple, vs the manufacturer of the product. Made FOR doesn’t necessarily mean that the manufacturer designed it and has rights to offer that improvement to other buyers.
Case in point. The big honking sensor in the new Nikon D800 is actually made by Sony, which has its own line of SLRs. Yet Nikon paid them a hefty amount of money to embargo the chip for some amount of time, which means that Sony can’t even use their own sensor for something like a year.
Knowing about how Apple dominates their suppliers, it wouldn’t surprise me if they drive a similar bargain with Samsung/Sharp or whoever else is making these displays for them.
Yes probably. This display was manufactured by samsung (actually almost half of the iPads parts come from samsung) who is probably the largest display manufacturer in the world.
Manufactured doesn’t mean Samsung invented or thought about the Retina display. The display has been available for two years now but have you seen Samsung or other manufacturers offer the same retina display?
Possibly. Keep in mind that while it’s true the display was made FOR Apple, it almost certainly was designed to Apple’s specifications which they paid the display manufacturer to do. As such, Apple would own any and all rights to the design. Unless I’m mistaken. Perhaps somebody with more knowledge about this process could better elaborate?
No, that does not mean that anyone else is going to get this technology. Think about it for a moment, the iPhone’s screen is not manufactured directly by Apple, so why haven’t retina displays for other phones come out yet? It’s because Apple designs the components, they are just manufactured by other companies to Apple’s specifications. In fact LG, and Sharp are also going to be making this display panel for Apple. This isn’t a Samsung display, it’s a display manufactured by Samsung, there’s a big difference.
Apple sourced several manufacturers, of whom only Samsung reached Apples standards.
So many manufacturers know what Apple wanted, but not necessarily how to do it. As to whether they will go on to use it, I can only presume so – if Apple will license that specific technology.
Similarly specced displays will certainly become common over the next year or so, because the tech market just works that way.
Apple thinks of, engineers, designs, and patents this stuff!!! NOT Samsung. Sorry anti-apple zealots. Well I’m sure Samsung will copy this soon enough.
Pingback: Dot color explains why colors look so nice on the new iPad | TUAW - The Unofficial Apple Weblog
Pingback: Dot color explains why colors look so nice on the new iPad - Jay's Man Cave Magazine | Jay's Man Cave Magazine
Pingback: Dot color explains why colors look so nice on the new iPad - Veletech
Pingback: Dot color explains why colors look so nice on the new iPad | iMicroid
Pingback: Dot color explains why colors look so nice on the new iPad | Design City
Pingback: Dot color explains why colors look so nice on the new iPad | My Tablet News
Pingback: Dot color explains why colors look so nice on the new iPad | Brian's Blog Site
Pingback: New iPad color quality consumes 30% more power « Mobile Technology News
Pingback: Dot color explains why colors look so nice on the new iPad | Garret's Geeks
I heard that Apple sourced displays from both Samsung and LG. Is there any truth to that story? (It seems unlikely they would get the same quality and colour standards from two different suppliers).
Supposedly screens rejected by apple were sold to HP for their tablets – passing their spec requirements and standards.
Shockingly, iOS is not color managed, so if you want to take best advantage of that glorious display, you must send photos to it already converted to the display profile. I used a Spyder3 to profile my iPad1 and when I used the resulting profile to render photos in Lightroom for the iPad, the results were noticeably better than sending sRGB images. So I’ll use that profile to update my Lightroom-to-iPad workflow article [ http://regex.info/blog/lightroom-goodies/ipad-howto2 ].
However, I’m sure you could create more accurate profiles for the various iDevices than I could, so I’d like to ask if you wouldn’t mind doing so? I’d love to be able to link to a source of quality profiles.
Pingback: Improved colour on new iPad | SRPS Blog
Pingback: Dot color explains why colors look so nice on the new iPad | iPhone 4 and 4S Tutorials
Pingback: Dot color explains why colors look so nice on the new iPad | Forum on Cameras
Pingback: Dot color explains why colors look so nice on the new iPad | Top PC News from the latest Tech News/Blogs
Pingback: Uno de los secretos de la Retina Display pueden ser los filtros de color | TechyMag
Pingback: New iPad’s Color Filters Alone Eat 30% More Battery
Pingback: New iPad’s Color Filters Alone Eat 30% More Battery | Apple Stocks
Pingback: I colori del Nuovo iPad, consumano circa 30% in più di batteria rispetto all’iPad 2 - AppleZein – Il blog Italiano Apple News per gli utenti Apple
Pingback: New iPad’s Color Filters Alone Eat 30% More Battery « The Wall Street Geek
Pingback: Dot color explains why colors look so nice on the new iPad
Where do I get one of those spectroradiometers to produce the nice spectral power distribution graphs? Do you recommend any budget versions?
evolvability, we use a Photo Research PR-655 SpectraScan Spectroradiometer. Definitely not a budget option. You may be able to get some decent data out of something like X-Rite’s i1BasicPro (http://www.xrite.com/product_overview.aspx?ID=1912) for a lot less.
Pingback: Shopping for a tablet this holiday season? Don’t forget to look at color performance | dot color
Great post. I am new to this field. Does the height of the blue peak mean that more light of 450nm emanates from the screen than higher wavelengths? If so, why is this? Is it possible to get blue color without 450nm wavelength say by mixing yellow and green? Thanks!
Thanks! You have it right- the “color filter spectrum” chart shows the relative amounts of blue, green and red for a pure white screen on the iPads. The balance of the three colors affects the white point of the display. For example, more red would make for a warmer-looking white and more blue would make the white look cooler. The iPad has a very good, standard white point of about 6700K which makes whites appear slightly cool and crisp. More on color temperature: http://bit.ly/106JSLt
Displays make color by mixing light, which behaves very differently from mixing ink or paint. This kind of color mixing is called “additive mixing”. In additive mixing red, green and blue are primary colors. More on additive color mixing: http://bit.ly/14muO1N
I’d like to add some details about color mixing. The green that is “leaking” through is most likely not a “leak”, but rather by design.
The main “blue” peak is at 450 nm. While light at 450 nm is part of the “blue” spectrum, it appears violet / purple in color, rather than what most people would consider “blue.” When a pixel has a color of Red=0, Green =0, Blue=255, the display is actually mixing the violet-colored 450 nm light with a smaller amount of green light to get the blue color. If they were to filter out all green, the 450 nm peak would appear a violet / purple type color.
Interestingly, the older CCFL displays had a spectral peak at 438 nm, again appearing violet / purple in color. They also had a peak at 490 nm, appearing blue/ green in color. If an entire screen of pixels was colored at Red =0, Green =0, Blue = 255, the blue color was created by mixing significant amounts of 438 nm light with 490 nm light. The 438 nm violet-colored peak is observed on virtually all fluorescent lights, not only monitors, but also household light. I’ve used a spectroscope to observe how several LED and CCFL devices produce blue.
I’ve noticed that a few newer monitors are moving the blue peak to 460 nm in response to concerns that light in the blue-violet range may be causing eyestrain and retina damage. Philips refers to their devices using 460 nm as “soft blue.”
Perhaps in the future a company will move the blue peak to 470 nm and no longer use any green light when displaying pixels set to Red =0, green =0, blue=255.
Thank you for your interesting posts.