Surface Pro 2017 Backlight Bleed

The average new Surface Pro user may find this backlight bleed issue difficult to understand, and you may be asking, "Is this significant or not?  The info to follow may help in gaining perspective on the validity of the issue. 

Most know the feeling when others, who seem to be experts, assert there is no problem, but you timidly (or not) assert that you think thereis a problem; It feels not so good.  But remember, bloggers and pundits may see themselves as experts, deservedly or not.  However, real experts who study this stuff diligently are refreshing to listen to because of the clarity they bring to a matter.

Somewhere in this forum’s thread, I previously mentioned a research scientist who had written an excellent review on the display of the new Surface Pro's predecessor, the Surface Pro 4.  This scientist has a company that does LCD testing.  And, in fact, the Surface Pro 4 technical review performed by Dr. Raymond M. Soneira was prompted by Microsoft; they provided him with a Surface Pro 4 (to test).  His company, DisplayMate Technologies, tested that Surface Pro 4 with “well-known objective and comprehensive display Lab tests, measurements, and analysis.” The purpose of those tests, the review said, “is to explain in-depth display performance results for consumers, reviewers, and journalists.” The link to that review is:

• http://www.displaymate.com/Surface_Pro4_ShootOut_1.htm

I found the article to be excellent, but please read on about his comments on this current backlight bleed issue.

Almost a week ago, I wrote Dr. Raymond M. Soneira about the "New Surface Pro" backlight bleed issue, and I provided him with a link to this forum's thread.  I did not expect a response from him -- I don't know him -- but he kindly replied with the following comments and opinion about this issue.  He said (emphasis mine ):

Almost all edge lit LCDs show some degree of backlight edge leaks and hot spots (sometimes from individual LEDs in the backlight assembly). We check for leaks and hot spots using an all black full screen test pattern viewed in the dark, which is extra sensitive for this.

Normally most leaks and hot spots are not visible when viewing normal content. When looking at screen content that has a white background, like in many text applications, the light bleed is unlikely to be noticeable. But for photo and video content, particularly for letterboxed widescreen movies that use black borders, the light bleed can really standout and be rather objectionable.

The screen shots in the link you sent [in this forum thread] show rather considerable visible leakage on a blue background screen.This is either a design or manufacturing defect. Microsoft should replace such defective units.

Raymond Soneira

President of DisplayMate Technologies Corporation

http://www.displaymate.com

I appreciate these comments of Dr. Soneira's because he is an expert.  His comments provide perspective and help to validate what can be considered normal backlight edge leaks, and also what can be considered defective.

Thank you, Andrew! I had a feeling the pattern was the same. Let's hope more people like you and Brain&Force will try it out.

As a next step, beyond solid grays, I'm thinking of implementing a way where we can take any snapshot (say one of those notorious OneNote pages with the side bleed), and apply the compensation to it. Then if you put that snapshot on your device in full screen, you should see the problem mostly fixed. If it works with snapshots, then it's even closer to becoming a solution.

The final step, which I'm not sure of, is how the eventual compensation could be implemented and whether it can be efficient enough (in computation and power) to apply on-the-fly. I'm not sure if the Surface firmware can manipulate hardware look-up tables (LUTs) or if a simple pixel shader has to be implemented into the graphics/video drivers. If anyone here has the technical knowledge about such things, please let me know.

Superficialis, I tried your little test and the bleeding reduced significantly. It was a noticeable improvement at all brightness levels. This suggests (as you say explicitly) that this is a systemic problem since the pattern is the same across many devices.

It would be wonderful if a patch or update or calibration could "fix" the issue. That would make me perfectly happy with an otherwise stellar device (and otherwise stellar screen).

B&F, see here, this is your same photo with only the bright part cut-out.

You don't see any dark band anymore, do you? Not any more than what you have at the top, just rotate the picture around or flip it upside down. Because the "dark band" part of the image is actually the same shade of gray as the part above it. Your eyes tell you it's darker in the original, because there's the bright area just below. You can try the same thing in MS Paint (bless its soul, it lives!). Or go ahead and try in real life to cover the bright area of your screen with a notepad or a book or something (just half an inch or so, only the bright area), and you won't see any dark band. I think the reddit thread is very misguided in this regard.

There are two reasons why people are confused about why the problem mostly appears at mid-brightness:

1. There's almost no light bleed when the screen is black because this Surface seems to have rather smart dynamic LED dimming, where it practically turns off the backlight for dark content. So the issue disappears.
2. When the brightness is turned all the way up, we are in high-exposure mode and the contrast is less perceptible to the human eye. You can emulate this with the camera (at any brightness) by increasing the exposure, and the camera won't "see" / your pictures won't have the non-uniformity. Conversely, even at 100% brightness, if you set your camera at very low exposure, it will see the same non-uniformity that your eye fails to see. It's just how we perceive things in bright contexts versus dark, it's complicated.

Bottom line: There is no dark band. It's non-uniform backlight in the form of excess brightness at the edge, aka bleed.

Whomever reads this: please download and test the following compensation patterns in full screen in Paint and tell us, does it "fix" your non-uniformity? (See my earlier posts for explanations). It's not a cure yet, but may signal a possible cure. Thanks.

(I know Brain&Force believes the problem is "dark bands", but that's mostly an optical illusion combining the effect of excessive brightness on the edge, and the dimming due to viewing angles.)

I don't think the dark banding is an illusion. If anything, I think backlight bleed is not the real issue here, but the dark band is. My Surface Pro 2017 has less light bleed than my Surface Pro 3. The light bleed on either device does not affect regular usage either - this is with the screen on max brightness and a completely black background, which isn't normal usage.

The problem, at least for me, is certainly the brightness variation of the screen. At normal viewing angles it's very visible, and it's visible on backgrounds which aren't black and at full brightness.

Yes, the dark bands look worse at the edges, but even still, it's something I consider unacceptable. Without the dark band, the screen would look more uniform.