Active/Passive 3D Myth Revisited

About a month ago I wrote a post on this blog about what I like to refer to as the Passive 3D myth which states that

Passive 3D systems offer inferior resolution when compared to Active 3D because each eye only gets to see half of the image lines on the screen.

That post has received some good and insightful comments that made me rethink my position… and come to the exact same conclusion that the statement above is indeed a myth.

The comments did give a very good insight into where the source of the confusion lies when discussing this so I decided to revisit this subject with a more detailed explanation of why I think the statement is a myth and why the refresh rate of the television or the frame rate of the source material are largely irrelevant to the discussion of the merit of this statement.

Active glasses (back) shut out each eye in turn, passive glasses (front) filter the light

First, let us get some background on the subject and get our terminology straight as this turned out to be one of the major sources of confusion. Yours truly was not the least to blame as I myself succeeded in hopelessly mixing up terminology.

Active vs Passive
The challenge for any 3D system is delivering different pictures to each eye. Both active and passive 3D systems use glasses to achieve this but there is a fundamental difference. Active 3D glasses shut out each eye alternatingly whereas passive 3D glasses filter out the wrong picture, letting only the right one pass through to the eye.

Resolution
This is of course an important point to address as this term is used in the myth. Resolution actually means a lot of things and the definition of the term in Merriam-Webster’s free dictionary is a long page listing 6 different meanings some of which even have different variations. For this discussion we will be using this one:

6b. a measure of the sharpness of an image [..]

It is interesting to note that Merriam-Webster is not mentioning the use of the term resolution in discussions about audio though it obviously is used there a lot.

Spatial vs Temporal Resolution
In video, a whole sequence of images displayed in turn creates the illusion of movement. It is clear that both the spatial or angular resolution of the individual image frames as well as the temporal resolution of the image stream in time, the framerate, play a big role in perceived video quality. In 3D another factor is added, as the different images perceived by both eyes are blended in the brain to form a 3D video.

Progressive vs Interlaced
These terms originate from traditional CRT televisions and analog broadcasting systems. Basically when a picture is built progressively, the cathode ray in the television would start at one corner of the image and then beam the image, pixel by pixel, one line after the other until the full image was drawn, then start all over again for the next picture. In the beginning this could cause issues where the corner in the screen where the pixels were drawn first was already fading by the time the cathode ray gun would reach the last pixels of the image, giving a flickering, shimmering kind of effect. They came up with a way to counter this called interlacing. With interlacing the gun first scans all the even lines, reaching the other corner of the image twice as fast, and then would scan the odd lines of the image in the second pass. This was a very effective way to counter the shimmering effect back in the days of early CRT televisions and it has stayed with us ever since.

Framerate
This term refers to the rate at which different still video frames are shown to create the illusion of movement, usually measured in frames per second, or fps for short. Generally this term is applicable to the media source, e.g. a film or video game.

Film has traditionally been shot at 24 fps. This is also often called 24p. So in a traditional 2D film strip, one second of film consists of 24 video still frames, as shown below.


In a 24p 3D film, there are 2 times 24 images. For every moment in the film there is a still frame for the left eye (green) and one for the right eye (red).


Currently some film makers are experimenting with higher frame rates for their films. Both 48p and 60p frame rates are good candidates for future film standards. Below is that same second of 3D film, shot at 48p or 60p respectively.


Interesting as this subject may be, I will show that it is irrelevant to the discussion of whether the myth is true or not.

Refresh rate
Related to but distinct from the frame rate is the refresh rate. Where the frame rate is tied to the source material (i.e. the film you are watching), the refresh rate is related to the display technology (i.e the TV you are watching it on). We usually measure refresh rates in hz.

Traditionally televisions would have refresh rates of 60 hz (or 50 hz in Europe) but recently they more often than not have higher frequencies of 120 or even 240 hz. Again I will show that this is mostly irrelevant to this discussion.

Fields vs Frames
A television with a refresh rate of 60hz in interlaced mode renders only half frames, either the even or the odd scanlines. So we call what is rendered each pass a field. Two fields make up one frame.

Modern televisions don’t have a cathode ray gun anymore so the whole principle of interlacing is largely irrelevant for them, but when it comes to 3D it is making a reappearance in a new form in passive 3D television systems. So even though the term field from traditional interlacing isn’t directly applicable to modern televisions, it could be convenient to use it for the image that is rendered by the television each pass in it’s refresh cycle, to keep it distinct from the term frames which we will reserve for the source material.

Motion Interpolation
This is the technique of generating in-between frames to bridge the gap between the relatively low frame rate of the source material and the potentially much higher refresh frequency of the display technology. As this topic can make an already difficult discussion even more confusing and I think it is also irrelevant to this discussion I will largely ignore it.

Busting the myth
To bust the myth I would like to present an example. A single frame of a 60p 3D movie displayed on a 240hz 3D television, comparing active and passive display systems side by side. As we are talking about 3D, a single frame of film actually consists of two images; one for the left eye and one for the right. I’ll color these red and green as we have done in the pictures shown before. Now when displaying 60 fps source material on a 240hz screen, every field would repeat the same frame four times. The way these are repeated differs in active and passive systems. Have a look at the diagram below:

The colored pixels are showing the image and the black pixels are ‘blocked’ or ‘blacked out’. On the left we see active 3D. Here alternatingly the left and right eyes are shown a picture and a black field. On the right we see two ways of doing passive 3D, either alternating horizontal lines or a checkerboard pattern. Columns are not a good choice as in 3D the two camera’s are only at different positions on the horizontal axis (they are next to each other) but not on the vertical axis (they are at the same height) so the images are vertically much more alike than horizontally.

As you can clearly see with both the passive technologies your left and right eyes combined get exactly the same amount of displayed pixels (as opposed to blocked or blacked out ones) as with active 3D. But the patterns on the right are much more regular than the alternating pattern on the left, which should result in a much more stable, flicker-free picture. Remember, this is a single frame of 3D video, consisting of two images, one for each eye.

In passive 3D, because only half the scanlines of the screen are designated for each eye it is hard to maintain that each eye would get the full spatial resolution compared to 2D. In active 3D each eye gets a full image, however only half of the time. This time it’s hard to maintain each eye gets the full temporal resolution compared to 2D. In both cases, resolution is halved for each individual eye, just over a different dimension.

Now even though resolution is halved in both systems, one could maintain that passive systems discard half of the information in each frame because it only displays half of the scanlines of each image. That is where 3D interlacing comes in. In 3D interlacing a similar technique is used as in traditional interlacing where first all even scanlines are shown and then all odd scanlines, but in the case of 3D interlacing the scanlines shown in the second pass are drawn in the same position as the ones in the first pass. If we were to apply 3D interlacing to the left image of a single frame from a 60p movie with line based passive 3D like we showed in the example above, the four fields rendered for that single frame would look like this:


In the full frame, on the left of the arrow, we colored the even and odd lines in a different shade of green to be able to distinguish between them. We can see in the four fields on the right that alternatively the even and odd lines are shown, albeit always at the position of the even lines. The brain does receive the extra detail from the odd lines but has to do some work to reconstitute the original frame.

LG is one of the companies pioneering this 3D interlacing technique and they recently received certification that their technique delivers Full HD in 3D by VDE, a german certification company.

Active vs passive 3D in practice
So far for the theory, but how does this work in practice? Which technology really does give most detail and which is most comfortable to look at? For a good comparison of both technologies Displaymate did some hands-on tests with four different televisions, two active and two passive systems. They are strong in their opinion that the myth is busted:

[..] small text (6 to 10 pixels in height) was readable on the FPR Passive Glasses, which definitively establishes that there is excellent 3D Image Fusion and the Passive Glasses deliver full 1080p resolution in 3D. Again, if the Passive Glasses only delivered half the resolution, as some claim, then it would have been impossible to read the small text on the FPR TVs. So those half resolution claims are manifestly wrong – no, ands ifs or buts!

So there you have it. Theory, practice and certification offices all agree: Passive 3D does not halve resolution compared to active 3D. Please share your thoughts on this subject in the comments below.

33 responses to “Active/Passive 3D Myth Revisited

  1. Pingback: Active vs Passive 3D Myth | Stijn de Witt's Blog

  2. Great explanation! In fact, I think you have cleared up a problem I was having with an LG 3DTV (the 47LK950). I was interlacing two images on my laptop and outputting the result to the screen via HDMI cable, but I couldn’t get the 3D effect to work. From my understanding of your explanation of LG’s 3D interlacing, I now think the problem is that the TV is actually looking for two full images to interlace, instead of the already-interlaced 1080p image I was passing in. The issue was frustrating me, as my approach works perfectly on a different 3D monitor, but not this particular TV. Do you by any chance know what form the input to LG’s TVs needs to take in order to use 3D interlacing technology?

  3. Thanks, glad you liked it!

    I don’t own an LG myself, but as far as I understand it, the 3D interlacing is done by the TV, so you just send it regular uninterlaced frames and it interlaces it itself.

  4. Now I understand better why cross-view 3D images give the best quality. Both eyes see the same time full brightness and resolution. Result, you obtain 2 times the resolution and brightness of a single image. In cross-view mode your eyes adding the resolution. This results in razor sharp 3D pictures. But cross-view is not practical because few people see this and you need both images side by side on the same monitor.

    I make a lot of 3D macro shots (high-speed insects in-flight) and I’m looking for the right monitor. I felt that active display was the best solution but now I doubt whether it is correct. It’s about a monitor, but the problems will just be the same. Thanks for the article, it gives a different vision and I will have to look further for the best monitor 3D vision system.

    fotoopa.

  5. But if you put two images side by side on the same monitor, still per eye you wouldn’t get the full resolution of the monitor. Each eye looks at one image which uses only half the monitor resolution. So in fact there is no difference there.

    The fact you don’t use any glasses to filter away light means it does get a better brightness. I give you that. 🙂

  6. Panasonic is claiming on their product page that passive 3D reduces vertical resolution. Are they correct?

    “When viewed through polarizing eyewear, different images are shown to the left and right eyes, resulting in the perception of 3D images. This method reduces the vertical resolution to one-half, but it is an easy way to view 3D images using light, economical eyewear. Great for watching with a group of people.”

    http://shop.panasonic.com/shop/model/TC-L55ET5

  7. Joe, yes they are correct, *but* they fail to mention that active 3D has similar issues.

    The argument has been very carefully explained in my post. I call it a myth because people (including the Panasonic product page) keep repeating one side of the story without mentioning the other side.

    Passive 3D technology reduces the *spatial* resolution to one half, *per eye*. When the images from both eyes are combined, a full resolution image is perceived. 3D interlacing makes sure each eye receives all pixels, albeit in turn.

    Active 3D reduces the *temporal* resolution to one half, *per eye*. When the images from both eyes are combined, a full resolution image is perceived. High device refresh rates make sure each eye always gets to see each image at least once, but in the end each individual eye will be looking at a black screen half of the time.

    Both technologies have their pro’s and cons. People should just make sure to tell the full story. As it stands many sites and people are telling (often unknowingly) a half truth.

  8. Hi Stijn, nice article to read. I own an LG tv set and chose for this set because I did not want to have flickering glasses that give me a severe headache. I have never seen such vibrant 3d picture as with my 3d TV.
    And then I have only been watching SBS stereo material, so there, the full resolution is split up in a left and right picture halving the resolution anyway. I still have to buy a real 3d DVD to see the full 3d resolution, that must be even better. The SBS format is also the format that is widely used for 3d broadcasts anyway. So even if the passive glasses would half de resolution (which is not true), the source material is not full HD anyway so even an active tv would never be able to show that material in full HD.

    But I do have a comment on your blog. You state that the interlaced technology was invented because of a more stable picture? That is not true. Interlaced tv was invented to save bandwidth. The analogue channel’s bandwidth is too limited to send 50 full frames per second. To overcome this problem they decided to alter the even lines with the odd lines. This way you only have to send half the information, saving bandwidth. The fact you don’t notice this on a good old CRT is because of the afterglow of the phosphor layer in the tube. The problems start when you are using digital media to capture interlaced source material into non-interlaced media. Special algorithms need to be applied to eliminate the ugly effect interlacing has, especially when there is a lot of fast action where there is a huge difference between two frames.

    Today’s tv’s still cope with interlaced material via the analogue channels because still the bandwidth on cable tv is limited. Even digital HD tv is often broadcast interlaced 1080i to limit the required bandwidth but now being a smaller digital information stream. So even when you are watching HD tv via cable, most of the time even that is not a 1080p resolution. Even some Bluray discs that contain documentaries that were produced for TV, contain 1080i content.

    I hope this clarifies the interlaced vs progressive thing.

  9. Hi Stijn,

    After having seen two 3D movies in the cinema I can say that my main problem with passive 3D currently is the fact that the brightness is just about halved. Especially in the older cinema there were some scenes that became almost too dark to see.

  10. This article still doesn’t make much sense to me. I own a passive 3D monitor and an active 3D TV. The same content always looks sharper on the active 3D TV.

    I understand that it isn’t half the resolution when using 3D interlacing technique, however the picture quality must be of a lower standard. Why? With active 3D only 1 eye is seeing an image at a point in time, but a 1080p image is still a 1080p image whether you view it through 1 eye, or 2 eyes. With passive 3D, at a singular point in time, each eye is only seeing half the resolution (540 lines each). With interlacing, in the next pass/refresh I will now see the additional 540 lines of resolution, but the display has actually overwritten the previous 540 lines and I’m relying on my brain to remember what they were and piece together a full 1080p image for me. The brain does this extremely well, but it’s not as good as seeing the full 1080p image in one go (as per active 3D).

    You can see this as well. Edges to non-straight lines for example have a very distorted look to them, and overall the image looks significantly softer. I’m not saying it’s the same as playing the movie in 2D @ 1920×540, but it’s certainly not the same as playing in 2D @ 1920×1080. However, my monitor is only 60Hz, whereas the TV is 400Hz so it’s likely that more modern passive 3D TV’s with a much higher refresh rate can help assist the brain in putting it all together properly.

    Anyway, that’s how I see it. Please feel free to correct me if I am wrong. I always believed I was seeing the full 1080p image with my passive 3D monitor after reading reports like this, but after getting an active 3D TV I quickly changed my mind. Also, I really don’t know how that report by Displaymate showed text to be most readable on passive. I have an extremely hard time reading small text in 3D on my passive monitor – to the point where I’d say it’s almost impossible. With active it should be the same as 2D (providing there is no crosstalk).

  11. You all must have perfect vision which puts you in the minority. I wear prescription glasses like most people. Active glasses are not available in clip-ons But they are for passive. I tried active glasses. I went to several stores before I could find a working pair. Shuttering between two unequal eyes did not look so good. Active technology assumes you have good vision which is simply not the case in the real world. I read 64% of Americans have or need corrective lenses.

  12. @Gary: Some good points. Both technologies have their drawbacks. I’m not siding on any of them (I need a lot more experience with both before I can draw a conclusion which is best for me and this may differ per person), but I just wanted to give a bit more nuance to the discussion, which I think *is* often very one-sided.

    @Hans: Good point! I never even thought about this not wearing any glasses myself, but that is indeed a very important consideration!

  13. I wear glasses and have no problem with either technology – though I admit that clipons are the ideal method. Both 3D technologies require you to see a different image in each eye so I’m not sure why passive is better for you in terms of 3D ability. It’s not like you can use one eye to compensate for the other…

    I had no problem putting active glasses over my prescription lenses – it’s not the most comfortable thing in the world but it’s not that bad.

  14. Hi Stijn, thanks for your article. It is good to see more discussion around this topic because there is a lot of misinformation. You did quite a good job describing the technology and terminology. Unfortunately though your arguments and reasons are adding to the misinformation. The two main problems with your article are these statements:

    “In both cases, resolution is halved for each individual eye, just over a different dimension.”
    That is a HUGE difference. With active 3D a full field of view is shown half the time to each eye (TEMPORALLY). The eyes and brain do an amazing job to blend this into a coherent moving scene and is the reason why video (which is comprised of displaying successive static images) works at all. With passive 3D half the content is shown to each eye (PHYSICALLY). The brain does a poor job of filling the physical gaps regardless of whether or not the opposite eye sees the other half of the image.

    “For a good comparison of both technologies Displaymate did some hands-on tests with four different televisions, two active and two passive systems. They are strong in their opinion that the myth is busted”
    Several things in that article are plain wrong! I’d be ashamed as the author to stake my reputation on that article. Even he says “On the other hand, there are instances when 3D Image Fusion [combining alternate lines from each eye with passive 3D] may not work well with FPR [passive 3D]. They arise when the brain is unable to properly match up the right and left image content or when there is fine computer pixel and line graphics”.
    And in the section “Instances When FPR 3D Image Fusion May Not Work” he says all the reasons EXACTLY WHY passive 3D does not have the resolution of active 3D – it’s the same with ALL content, not just the specific examples he mentions.
    Although he doesn’t state it explicitly the main reason active 3D in his tests did not perform as well was because of crosstalk (ghosting – causing blurred images). But he chose to test only two active 3DTV’s and one of them is the worst available for crosstalk! He should not have drawn general passive vs. active 3D conclusions from such a limited sample – he could have easily found discussion online to tell him that the Sony HX729 is very poor for 3D. And he’s right that crosstalk is detrimental to picture quality – but passive 3D suffers from crosstalk too – it’s generally fairly good but some of the best passive 3DTV’s are actually WORSE for crosstalk than some of the best active 3DTV’s.

    If you’re still prepared to call passive 3D “full HD” or 1080p then you’d better start calling active 3D “twice full HD” or 2160p because both eyes get to fuse two 1080p images!

  15. It wasn’t me who called it Full HD, that was LG, who has been certified for this term by an external company as I stated in the article.

    I agree with most of what you say and thank you for your valuable input. My article is written in a biased way, not necessarily for active or passive 3D, but against a statement that I have found time and time again, that passive halves the resolution when compared to active, without telling the whole story about the temporal halving of resolution. Whether the gain in temporal resolution passive 3D has can compensate for the loss in spatial resolution I honestly don’t know, but I truly believe that it *could* do so.

    When looking at the future, where we see resolutions of 3840×2160 (‘4K’) coming, I believe there is a lot of potential for the loss in spatial resolution to be made insignificant. Just as with interlaced full HD video as is sent by a lot of broadcasting companies, where you really get only half the scanlines per frame, in the future there will probably be a lot of full HD material being upscaled to 4K. And in such situations active 3D cannot profit from it’s potential advantage of more spatial resolution, whereas passive 3D can *always* profit from it’s advantage of more spatial resolution. No matter how high the refresh rate becomes, each eye will *always* see a black frame 50% of the time.

    About the crosstalk, I have to admit that when I got the chance to have a viewing on an LG passive 3D television set, I was hugely disappointed by the *huge* amount of crosstalk. Basically, when I closed one eye, I still saw 2 images superimposed over each other. One was just a lot more vague than the other, but it was clearly there. I am sooo hoping that that was due to some cheap (bad) polarizing glasses, because otherwise I will never buy such a set. And that is probably the wisest to do for everyone interested in 3D: Just compare some sets. Take the time to look at different aspects of the experience and then decide for yourself which is best for you.

  16. To complicate things, the currently popular passive 3D implementation isn’t the only option. iZ3D for example used to display a full screen (not an interlaced one) for each eye at a time, alternating between both eyes. This effectively halved the framerate per eye, instead of the resolution per eye per frame.

  17. I just purchased an LG 55LM5800 (LG 55″ 1080p passive 3D) and I am able to verify that looking at small print in the background of a movie, is indeed full resolution when in 3D mode. In fact, switching between 1080 2D and 3D, the small text actually became more distinct and easier to read. I switched back and forth several times and each time the 3D version made the text clearer.

    I also observed that on almost horizontal lines there is some slight disturbance along the edge. However, I was able to clear this up by adjusting the 3D parameter associated with viewing angle. The same adjustment also cleared up some slight ghosting.

  18. If after wearing 3-d glasses, the picture appears darker, then is it active 3D or passive 3D? In India, we get 3D movies in both formats in cinema theatres. Initially, we did not know the difference between the technologies or the glasses they gave us in the theatres, Now that we know, we can’t seem to remember which technology was it that caused darker pictures when glasses were worn. We would like to go for that 3D technology that maintains the brightness. Pls advise.

  19. If after wearing 3-d glasses, the picture appears darker, then is it active 3D or passive 3D? In India, we get 3D movies in both formats in cinema theatres. Initially, we did not know the difference between the technologies or the glasses they gave us in the theatres, Now that we know, we can’t seem to remember which technology was it that caused darker pictures when glasses were worn. We would like to go for that 3D technology that maintains the brightness. Pls advise.

  20. Excellent explanation but I think the final conclusion that “half resolution” is a false myth is wrong. Traditional passive 3D definitely reduces resolution to 1920×540. With the emerging passive interlacing technology the resolution can be recovered. However, as this is an emerging technology that is mostly not there yet, I think it wise to clarify it’s only a false myth with the newer technology. I also agree that with 4K, passive 3D will give 1080p resolution which will surely conquer the world!

  21. I have 3 3D TVs, a Sony and Samsung active, and LG passive.
    All active have slight crosstalk which ruins the quality, and the LG passive has zero crosstalk, making it perfect.
    So the whole argument of resolution is a wasted argument, as the brain does not perceive the 3D as well when there is any crosstalk.
    Passive is way better. Not to mention the cheap lightweight glasses with no batteries to replace, and you can also keep the glasses from the cinema and use those.

    I use my Samsung as a computer monitor. I notice that when I switch it to 3D mode, the resolution clearly drops, as all the text goes blocky like the resolution is halved, even though the resolution of the image sent to it is 1080P/24Hz. It must be discarding half the image.
    So even though technically it is possible to display full resolution to each eye, I wonder how many TVs actually do.

  22. Please note that you can’t just look at numbers but also have to consider what the real world effect is and how the human system interperates input.

    That said there is one major issue to contend with and that is that the human eye and brain handle lost temporal informatio far better than lost physical information (I believe those are the terms you used).

    Your eye and brain have natural image retention that helps alleviate lost temporal information however it is far less able to resolve lost physical information.

    I own a passive 3D LG and while you do get used to the lost info, you have only to view a hard contrasty edge at any point to see the result of lost resolution on passive. Aliasing is strongly visible in many cases.

    I believe the reasons is that in an attempt to meld both eyes images together your brain chooses techniques that are appropriate for the information it gets – during a scene with no regular and complex detail, the overal best fit is to overlap the images offset so as to effective deinterlace (or deinterlave as it were) the lines on a passive 3D display.

    However when you hit strong contrasty subject your brain sees stripes and decides the black stipe from the left eye must match up against the black stripe from the right eye… thus instead of staggering the images and using the data form one eye to cover the black line from the other eye, it aligs the black stripes and bright stripes from each eye showing a strong scanline effect and lots of aliasing because the left and right eye info tend not to to match horizontally, so when forced to match vertically you end up with an incongruity.

    Sadly this means both 3D methods have issues – active with brightness, cross talk, motion tearing and shutter headaches and passive with resolution loss and aliasing.

    BTW if you saw lots of ghosting while watching passive you were probably above or below midline on the TV. Passive has a small vertical sweet spot as far as ghosting goes – active has no such vertical limitation but usually color accuracy falls off with active significantly as you move vertically….

    It’s as if the two technologies are truly opposites in that the benefit of one is the con of the other in almost every respect.

  23. Nice explanation but you fall into the same trap so many others have – thinking the math tells the whole story…

    You mentioned you would need a neurologist to figure out how your brain handles either form of missing resolution.. but you don’t… you only need your eyes and maybe a quick look at existing video technology.

    Here are a few points you missed:

    Spatial vs temporal loss of resolution – does your brain (and or eyes) have a method to cope with either one? Yes indeed it does… it’s what makes motion imaging possible with current technology. It’s called image retention.

    It’s what makes a CRT Tv look like it’s drawing a whole picture when indeed it’s drawing one dot at a time… it’s what makes it so waving a stick quickly in front of a flashlight or projector lets you see the beam/picture even though the stick never completely bears the entire beam/picture.

    You humans have a built in tool for overcoming temporal issues in imaging.

    The have no such ability for static interferance… this is why if you hold a comb in front of your eye while you watch TV, you never stop seeing the comb. Unlike rapid black frames between image frames, your eye and rain have no method to resolve the static interference that is the teeth of the comb… and this is very much like what happens with the scanlines on a passive TV. There is no equivalent of image retention to help get rid of them.

    So myth #1 – spatial vs temporal interference are similar issues. False. Spatial is far more an issue than temporal on the scales we are talking about.

    Next up – 540p+540p = 1080p.

    This makes sense… if you are a computer deinterlacer. You take each set of image lines and slide them between the other set to get 1080p.

    Too bad your eyes and brain are not a computer deinterlacer… what your brain and eyes are used to doing is resolving two views of the same item in each eye. They do this by matching up similar points as best they can. Bright shiny spots or dark spots, matching patters etc.

    What happens when you feed both eyes seperate interleaved images? Your brain matches up all the black lines on top of each other and all the image lines on top of each other because that is a better match than the alternative (actually deinterlacing/deinterleaving them).

    Think about it… your brian sees a totally black line in each eye and says “Well that’s obviously one thing that each eye is seeing, it looks exaclty the same in each ey!” It sees the image data and while it’s probably a little different due to parallax shift, your brain matches it up becuase it’s probably about 90-100% the same as the other eye sees.

    The result? You don’t get a deinterlaced/deinterleaved image you get black scanlines around each set of incorrectly overlapped image data.

    Test this yourself: Get a 3D passive TV, get up close where you can easily see the scanlines with 3D glasses on. What do you see? You will see definite solid horizontal black lines alternating with solid image lines. This is your brain matching each black line up with it’s (incorrectly paired) black line in the other eye.

    This happens until such a point as you get so far from the display that you cannot resolve the black lines anymore. However at this distance you cannot (by definition) resolve 1 full pixel height of resolution meaning you might as well be watching a half resolution or worse image to begin with.

    Myth #2 – Your brain combines both images to make 1 solid 1080p image. It doesn’t, it overlaps them in a very incorrect manner which is why you see scanlines and aliasing.

    Test #2 on this same myth: Load up a 2D scene with some high contrast imaging (credits on a movie or large solid computer generated logo work well). Put on the 3D glasses. Now every line of data is a perfect set of 2D data, this means if your brian interlaces it right, you will resolve all the data there was as a 2D data set (unlike 3D) has no parllax shift and so you will indeed see every pixel there is to see, nothing thrown away.

    Now while doing this you will notice aliasing and scanlines. This would not happen if your brain deinterleaved the two fields like the myth would have you believe, you would see exaclty the same image as the 2D without glasses.

    Lots move on to the idea that it’s even 540p to begin with.

    It’s not.

    It’s 1080p. But every other line is black.

    Remember the days of CRT TVs and the scanlines that appeared on them? When we moved to progressive displays the picture got much smoother despite there not being any more pixels worth of data.. what happened?

    You removed an image artifact – now your only gap between image eliments was the pixel spacing and with exception of some early and poorly made panels, pixel spacingon LCD TVs is small enough that it’s more or less unresolvable at any decent viewing distance.

    FPR does not show you a 540p image, it shows you a 1080p image to each eye with a huge image artifact equivalent to 1 pixel height for a full line every other line.

    To test this take an LG TV (I assume other passives can do this too but I happen to own an LG so know it’s possible). Load up a 1080p Top/Bottom 3D encoded video.

    These videos are two 1920×540 images stacked on top of each other, the TV anamophically stretches them each to 1920×1080 and then displays them. If you have the glasses on, the FPR removes every other line essentially showing the original 540p for each image.

    So take one of these videos, hit the 3D button and turn off 3D mode. What this will do is show you only 1 field on the TV without the other. Essentially you are now showing 1920×540.

    See how it looks?

    Now put on the 3D glasses and close one eye…

    Now see how that looks?

    Right… it looks much worse… suddenly you see massive scanlines and any hard edges have significant aliasing… this would be massively more pronounced if the image is moving.

    That’s the difference between 540p and 1080p with every other line black.

    Myth #3 – Passive shows you two 540p images. It doesn’t – it shows you 2 1080p images each with black lines every other line.

    You cannot discount the black lines just because they aren’t image data – to your eyes they are image data, they are black line images.

    This can all easily be verified without a neurologist and with a little common sense and preferrably access to test one of these TVs.

    Lastly LG has recently introduced a function in its 120hz and higher 3D sets where the missing frames of image data are shown after the original ones at 120th of a second each so every 60th of a second you get the full 1080 lines of data for each eye.

    This sounds like it means you get full 1080 resolution now!

    It doesnt.

    For 1 they filter the highest frequency data out to prevent aliasing (so you are throwing out some amount of each line of data so not full 1080 lines as something is missing in most cases) and the lines appear right where the previuos lines were… so sure you are seeing all the lines but half of them are out of place directly on top of the previuos line.

    Let’s say I read you a book on tape – but I left out every other word… but wait! To fix it I add back in every other word, but it gets said at the same time as the previous word.

    Does this mean you now have 100% of the words read to you? Yes it does…

    Does it mean that every word is understandable and not interfearing with the other words due to being out of place?

    No it does not.

    Myth #4 – You can overcome the missing image lines by jamming them in every half frame. You can’t… you can however pollute the existing image lines with filtered versinos of the missing lines… whether that makes things better or not… probably not as all it really ends up doing is averaging every two lines of resolution and then showing them as one line… remember our friend image retention? Yup.. it’s working here… working against you.

    Lastly something you didn’t specifically mention – lost vertical data. LG has argued that it’s ok to toss out every other line of data as really all you are losing is vertical data and the left right parallax data is there in the other field just one line off.

    First off this is not true… vertical data is important to image quality. Light reflects differently off every part of a surface and the angle at which you see it changes that… that means that no two vertical points have the same light/color characteristics in the vast majority of cirucmstances.

    This means you are losing image detail and even worse each eye is trying to make one set of image information with a set that is entirely missing from the other eye.

    Myth #5 – lost vertical data is not important. It’s LESS important thatn left right data, but it’s definitely important.

    These can all be tested on a myriad of sources however a handy one is Call of Duty Black Ops II. (probalby Black Ops 1 also but I don’t own that one).

    This game was made in partnership with LG to promote passive 3D TVs.

    Load the game up and on any lobby screen look at the players names in 3D mode with no glasses on. You will notice they are smoothon all sides – no aliasing. These are 2D no parralax parts of the images. Put on the glasses – aliasing.

    Where did that come from? That’s right… its’ your brain shifting the image to match black lines onto black lines.

    Now watch the bottom right corner of the screen – this game uses a scrolling oscilloscope looking line as a sort of loading icon sometimes… with the 3D glasses on there is massive crawling along this line as it scrolls… without them on no such image artifact.

    BTW here are is one big myth the other way that no one seems to realize:

    540p is half the resolution of 1080p…

    It’s not.

    It’s half the VERTICAL resolution… but it’s the same horizontal resolution.

    In the world of resolution it’s assumed that when you say half, you are halving it both ways… thus half of 1920x1080p is really 960x540p.

    What each eye sees on a passive 3D TV is 1920×540.

    This is largely why a pasive image does not look worse than PAL resolution (576p).

    It may have less vertical resolution per field, but it has more horizontal resolution per field.

    Hope I’m not too late to this party to make my point!

    BTW in closing – remember resolution is only a small part of image quality… let your eyes be the judge. If you have bad eyesite, sit far back from your TV or somehow just aren’t effected by the image artifacts of passive 3D, you are lucky! You get 3D for a much lower price, with often much punchier colors and lighter, cheaper glasses!

  24. I just realized the last comment was also by me… just forgot I ever made it… weird reading your own post and not even realizing till the end!

  25. @New Deep:
    In both cases the picture will get darker. That’s because filtering is going on. With active 3D, the picture for the wrong eye is filtered out so only the right eye sees it. So each eye is looking at a black frame half of the time. ==> Darker picture. With passive 3D, the parts of the picture which are for the wrong eye are filtered out. So all odd scanlines for exampe. So each eye is looking at an image which is half black all the time. ==> Again, darker picture.

    @Frank:
    Yes, both forms of 3D halve the resolution when compared to 2D (though they add an extra dimension which may compensate). However, I am stating that it is a false myth to state that passive halves resolution when compared to active. They both halve resolution, just over a different dimension. Spatial versus temporal.

    @John:
    Thanks for chiming in, you learnt me something. What you say makes sense. We will just have to wait for 4k passive probably before we really get to enjoy 3D in full resolution without any flickering.

    @Devedander:
    “You humans have a built in tool for overcoming temporal issues in imaging.”
    Lol this totally makes you sound like a super natural being :p
    Jokes aside, I think you posted one of the most informative comments I ever received (though I have to admit there aren’t that many :). Do you work in (3D) imaging?
    I don’t and to be honest, I don’t know that much about it at all. I just as you said, ‘did the math’. However, once you are thinking about imaging in the nuanced way you obviously are, you probably don’t need my post anymore. 🙂

    @Everyone:
    Wow! Just wow!!
    I posted some articles before and even had my own website in ‘the early days’. Feedback from other people is always very cool to get, but never before have I received so much feedback on a post of mine as on this one. What a great warm feeling! Thank you so much and I wish you all the best for 2013!

  26. Haha yes I didn’t re read what I typed and so lots of silly stuff ended up in there 🙂

    I do not work in 3D and in fact only recently started caring about 3D when I bought a 3D LG Passive on sale. Of course AFTER I bought it I started to do real research which is not the best way to do things of course 🙂

    During my research I found a lot of info that seemed questionable and after a little while figured out how to test it and refute it myself and was pretty annoyed how accepted these half truths are… I actually thought about putting a blog together to lay out my points (and may still) but found it convenient to tag onto the existing works for now.

    I don’t mean this to be insulting to anyone but it honestly surprises me that these things are not more commonly understood as none of this was hard for me to grasp and debunking it was all also very straightforward.

    I see industry experts doing things with fancy test patters and tiny font and wonder to myself why they never just looked at the picture and see that the brain does not indeed just mesh the two things together.

    I actually just found out one more thing:

    On my LG the left and right eye both show the same line! ie Right1 and Left1, then Right3 and Left3 etc… they do not show alternating lines so it really IS 540p data as it’s as all even (or odd not sure which set) is totally thrown out for BOTH eyes!

    I tested this when I noticed, again with COD BLOPS2 that when switching to 3D mode all the overaly text (which is zero parallax) got a very low res chunky look… on closer inspection without glasses on you can clearly see every row of data is duplicated directly above (or below depending on how you look at it) meaning the TV really is showing the same line for each eye and not even alternating!

    For instance in the letter O Line 1 should be just a small dot at the top and line 2 should be 2 slightly sperated points and line 3 2 slightly more seperated points etc…

    But when you examine the display, both eyes see the point at line one, then they both see exactly the same spaced points for the next line of each eye… if each line were being show then each eye would see slightly different seperation and only one eye would see the point at the top of the O.

    Now I believe LG has suggested in some models they “wobulate” for lack of a better term so all lines are shown, but the truth is still that there is a full line gap between each line and so any aliasing artifacts will continue to occur as you never at any time see every line at the same time!

    Anyhow, thanks for following up and thanks for putting this on your blog in the first place!

    Oh and BTW 4K passive… a few things to remember:

    I don’t think higher res passive can ever really be a good solution for the shortcomings of passive because of timing of technology.

    4K is going to be largely for larger sets for the near future so far as I can tell as the expense of a 4K panel means it makes little sense to stick one in a 42 inch set when it will be un noticeable on 2D content so only valuable to show 1080p in 3D content – the extra cost just can’t be worth it for quite some years.

    And remember 4K solves the 1080p 3D issue (mostly except still each eye is 1 pixel line off vertcally) but 4K sets open the door for 4K active which will again be twice the resolution of passive.

    Also remember that an 84 inch 4k set is literally twice the height of a 42 inch set and thus even with 4x the resolution, each pixel will be the same size as the 42 inch model. This means the black lines will be the same size as well.

    This means you will have to sit back just as far to not be able to see the black lines anymore… while certainly 10 feet from an 84 inch screen is more acceptable than 10 feet from a 42 inch screen I think a lot of us home theater enthusiasts prefer to sit closer than that for good immersion.

    And of course it still means you have to sit back far enough (by definition) to not be able to resolve that level of resolution.

    Sad… because I really like the passive technology idea, cheaper light glasses and no flicker, but I think my next display will have to be active as it really does bother me the lack of resolution and image artifact passive introduces 😦

  27. Oh BTW you mentioned that with active each eye sees black half the time… I think that’s not true… I think in order to prevent crosstalk they actually blank one duty cycle before and after each frame. So really each eye is only open about 25% of the time (which is why it’s so dark). So while you would think active goes Closed/Open/Closed/Open for a 50% open, it’s actually more like Closed/Closed/Open/Closed.

    On closed to actulaly block the wrong frame, and one before and after it to buffer any timing errors of the other eyes frame.

    With passive the glasses block about 35% of the light straight out. Of the remaining light, half of it is lost (by definition) due to the fact the FRP blocks every other line to each eye, so 50% of the light is lost).

    Ironically this gets you down to about 30% light to your eyes which is virtually the same as active…

    Yet passive looks brighter than active, why?

    Illusion – one most passive TV’s crank up the brightness when entering 3D mode (check with yoru glasses off you can easily tell) and two your eyes see the surrounding light around the TV as well – with active EVERYTHING is darkened by about 25%, with passive the non TV part isonly darkened by about 35%, the image is the only thing that loses 75% of it’s light.

    Interesting no? 😀

  28. Sorry to keep stringing these together, if you can merge them please feel free to.

    The last point about passive being brighter is related to the same issue of comparing temporal to spatial resolution –

    You would think that because both technologies block about 70% of the light, they would appear the same brightness right? But that’s not true, your eyes often determine brightness not by the total light output, but by the brightness of the brightest regions.

    So while passive results in an overall loss of about 70% image brightness, the parts that aren’t FPR filtered are only about 35% loss of brightness and so your brain registers the whole thing as being only slightly less bright.

  29. “So really each eye is only open about 25% of the time”

    I knew this, but left it out on purpose to keep things simple and it basically just strengthens my argument. 🙂

    Interesting points on the brightness illusion.

    I don’t own a 3D system as of yet. I bought my current TV just before the hype. But I am really interested in the subject and read a lot about it… Which makes me just the right expert to explain this topic no? 🙂

  30. Haha yes! It means you can be technically right… which is the best kind of right!

    But 3D is getting very cheap now… with 55 3D LCD dropping well below $1000 and the BenqW1070 in about the $1000 price range I would think someone intersted could get into the market pretty easily around now! Assuming prices are similar in your neck of the woods 🙂

  31. And yet passive 3D has more or less died out in favour of active. The main problem is that when looking at a passive screen, the gap between alternating scan lines is clearly visible and similarly gaps in the checkerboard pattern are too. In the future when retina-like displays become common for external display monitors and we are able to view images with 600+ pixel pitch then maybe passive can makes its way back into the market. This would imply 24 inch displays with more than 8000 horizontal pixels. As far as I know, cables won`t even carry that sort of band width currently and max out at around 4K, hence the newly popular 4K displays, but these are still less than 300ppi at monitor sizes.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s