If I remember correctly, Hisense has done this in one of their TVs. They achieved incredible contrast levels by using a second monochrome LCD layer to provide those inky blacks. Sadly, I don't think they've pursued it any further, synchronizing the LCD layers was quite difficult.
@@user-a6olopvmv9 Moire shouldn't be an issue if the tracing paper blurs the image enough to eliminate pixel boundaries. Effectively turning the back screen into a local dimming layer while the front screen creates a sharp image.
Something you didn't mention is that you have essentially doubled the display's gamma by multiplying all colors by themselves. Most displays have a built-in gamma of 2.2, but this raises that to around 4.4. You should be able to compensate by halving the gamma in software or in the monitor's settings, which will give you more accurate colors and fix over-saturation.
it would not be 4.4 since gamma is not on a linear function. More importantly, gamma does not describe the color saturation, it is instead a measurement of the luminosity of the pixels. We could assume that the gamma should actually be set higher than 2.2 because the overall luminosity has dropped with the combined LCD screens. As for color oversaturation, color saturation is also not on a linear function, and each color has a different curve. Meaning that each color acts differently on the same brightness, and colors even have different saturation changes with the same brightness change. And to top it off, since the contrast is being changed as well, the color saturations are changed even more. There is no way to calculate the required settings on your monitor. Which is why I'm sure he used a colorimeter like every youtuber out there to calibrate the monitor
@@KenTheSteak I’m no fun at parties because I attempted to correct someone who is wrong? Do you believe people who are wrong shouldn’t be called out? That’s what Karens believe, and Karens aren’t fun at life
@@bobsteven2363 I asked if you was fun, never said you wasn't, but thank you for answering, although I didn't ask "tell me you don't have friends, without telling me you you don't have friends"
I support this idea, I think he should go further with this! Maybe that will inspire some manufacturer to actually go to market, people don't care that much about thickness or consumption as companies may think.
Professional dual layer LCD monitors use a monochrome LCD as the rear panel, which is better for light to pass through, thus increasing the efficiency (though it is still significant less efficient than a single layer LCD). Also, with your method, you are significantly increasing the saturation, so you would need to get a calibration device to counter that. Still, this is a really cool project!
@@1337Shockwav3 I think that both are good for different purposes and budgets. The dual layered LCD monitors are designed for professional colour grading (especially for HDR) in high budget studios. QLED are great for other purposes, they are much cheaper, take less power, and are thinner. Mini-LED backlights are also gaining popularity, though they suffer from blooming. Personally, I don’t mind a bit of blooming as long as I am not in a dark room since I am not a professional filmmaker. I think this technology would be great in the hands of radiologists, but since medical diagnostic displays take a lot of time to be approved, this could be a while… though I think an OLED one specifically engineered and calibrated for medical imaging with an active cooling system would be the best. Edited: corrected.
Could he not use grayscale on the second screen? I know it technically still uses RGB components, but at least they would all get dimmer in proportion.
I think you could actually add a middle polarization layer between the other two. Weirdly, if the middle layer is at an angle between the other two, it allows light to pass through all three. I think this would actually allow more of the original light through than paper.
@@melty4204 Wasn't it that if you consider the polarized light coming through as a vector, the vector gets projected on the next layer so that if you have a 90 degree angle, the projection result is 0, but if you project it 2 times on a 45 degree angle, it isn't 0, which means that light still gets through.
If he took apart the panel stack and then stacked just the actual LCD's and then added the lens layers to that stack from one display in then he wouldn't need any of that nor the additional LED lights. The LENS layers serve multiple purposes and depending on the construction the polorizer often isn't affixed to the actual LCD and is just one of those layers in the LENS stack. On some panels it's glued directly to the LCD, but if it's not, then you can just stack the actual LCD's and place one stack of filter layers and that would eliminate all the Extra stuff he did with the paper and LED's. Note: When working with LCD panels films and lenses use gloves that leave no residue and do this in a dust free room while wearing something over your head or you will forever see dust, hair, and fingerprints on the image of the screen.
One thing I think you might have missed is the gamma. Monitors are non-linear with respect to their input, and there is typically a compensation curve (gamma) to fix that. However, with two screens, your nonlinearity will be doubled and you will need to adjust your gamma curve (easily accessible in desktop OSs).
Yes, although not so easily calibrated and you can run into precision issues with the required exponent of 4.8 that he now needs instead of the more typical sRGB exponent of around 2.2.
@@dorbie Really, this sort of display should be driven with different images sent to each monitor for this reason. It's got 16bpc in total at the higher gamma which could be remapped to the standard 8 (or even 12) with reasonable precision
@@jameshogge That would take a lot of highly refined calibration and systems integration that isn't really feasible in this context, it can buy you more bits but not 16 bits without independent backlights and a degree of control that doesn't exist here. You can do better than this though for sure. See my own response elsewhere and my referring to "esoteric reasons" for setting different display gamma values on each. Something that is feasible and easily within reach (if supported).
@@dorbie The calibration needn't be too hard: you can do well enough by assuming that the individual LCDs are correctly calibrated. The rest is just maths with the light intensities. One of the groups at my uni did a similar setup with a DLP projector as the backlight for an LCD. But yes, it is a pain to then drive the displays. Ideally, you use an FPGA. Alternatively you have to set up a virtual display in Windows and write a couple of shaders to produce the actual outputs.
i think removing the polarisation filter from the back lcd will help with brightness and eliminate the need of the tracing paper, you can then also bind both of the screens with uv glue used for screens, so there will be less blooming and blurryness in text
It won't work. The front layer of the back lcd and the back layer of the front lcd are 90 degrees apart in terms of polarization, so it would still completely block the light, unless you can find special LCD panel pairs which are oriented 90 degrees apart.
This seems like it would intuitively make sense, but I did some research and I don't think it would work. The liquid crystal film doesn't block much light. It's just red/green/blue translucent goop, organized into subpixels that can be polarized based on how much of that subpixel's color we want to shine through. Without the polarizing filter, the backlight shines through all the subpixels all the time, and it's almost impossible for our eyes to resolve an image. You can find video of what LCD screens look like without the filter, and they're basically just bright lights. Actually, I think this channel has done videos about upcycling broken LCD panels into lightboxes for exactly this reason. With only one polarization filter, it'd be the same problem as before -- not being able to block enough of the backlight to get the nice inky blacks and high contrast.
@@crystalsoulslayer How about getting a SLIGHTLY CLEARER polarizing filter that matches the second panel orientation?. I feel like he didn't explore this option enough and just brute forced the problem
While I usually find his DIYs trying the best & practical solution, this one has scope for a lot of improvement. 1. The large backlight led array: This is what makes this build thick, impractical and causes high power consumption. This can be avoided if tracing paper is not used. A better solution for the polarization problem is the way forward, which leads to #2 2. Tracing sheet: 2 stages of tracing sheet was used - one between the panels and other in front of the backlight. Both can be avoided. Fix / Recommendations: The polarizing element in the lcd panel is actually removable. He has to further separate layers of the lcd panel and remove the polarizing layer from the base panel. This removes the need for tracing paper and a more powerful backlight and solves blurry text issue. Second step is to change display settings in outer panel to high contrast black and white while inside panel displays full colour. This will solve the color gamma issue and the outer panel will only enhance blacks.
would be pretty interesting to test the optimised final version with an screen calibration tool, just to see actual numbers of things like nits, peak brigtness, color saturation, gamut, etc...and compare them with standard monitors
To be clear, this is not going to "compete" with mini-LED & OLED tech. Mini-LED displays get MUCH brighter and while this may match OLED brightness, it will fall short and in pixel response times and input lag. Youre also driving two displays off of a single output, so you're gonna need one hell of a GPU
@@erikhendrickson59 Nah it will be comparable to just driving a single screen in terms of GPU power if you use the "duplicate this display" option in windows.
@@erikhendrickson59 hey, watch the video before writing, you don't need a "hell of a GPU", duplicating the signal is enough, and even if you tried running both out of the GPU mirroring monitors is one of the simplest things GPUs do, it's the same thing twice, it doesn't have an overhead at all, people that make GPUs are some of the smartest people on the planet. Please stop thinking that just because you don't see a solution for something no one is capable of doing it
@@erikhendrickson59 GPU power is going to be the _exact_ same as with one monitor. But you are absolutely right about the brightness, and the power consumption is also absolutely crazy. It's not a viable product because of these
@@adamashworth6293 The HDMI splitter he's using means the PC only ever knows about a single monitor. The GPU is doing 0 extra work, because it doesn't even know the image is duplicated
You always come up with the most fascinating ideas. They aren't always practical, but the attention to detail you put in makes almost every build you do a piece of art.
If you fed both screens different HDMI signals, like in a "mirrored" dual monitor computer setup (using two outputs from your graphic card), you'd be able to adjust both layers individually. This could help making the picture's colours less "poppy" and maybe even help increase the white levels.
I don't know, but they might have slightly different vertical blanking intervals - meaning one monitor might start refreshing at a slightly different time than the other. I don't KNOW this to be true, but I suspect it might be true. That might be causing some problems, but really, I know about the technology of the LCD itself, but not the details of how it works.
@@fuzzywzheif running these from identical type outputs from the same quadro / Ax000 card or Radeon card (AMD is significantly less tight about locking up feature access on consumer GPUs, but low tier quadro cards are more bountiful harvested from old workstations), there's a ton of possibilities for adjusting stuff here to compensate for sync issues, gamma curves etc. Metrics and results should be pretty reliable, especially if you can get both displays sourced from the same production batch (almost guaranteed if bought new together). As that should avoid any internal revisions that wouldn't have noticeable impact on 2 individual monitors but could cause headaches combining 2 independent panels into a single display sandwich.
This is a project that I think deserves a version 2. You might get better results with different panel technologies (like TN VN and IPS), or you might be able to dig and find better solutions for diffusion, polarization, backlighting and driving the displays. Even if the final result is expensive and impractical I think it's worth it to see just how much is possible by going full DIY.
@Felipe Siqueira even if not removing one you can just get another polarizer sheet and rotate until it's corrected. Using paper for diffusion and polarizing is what kills this screen.
Yeah, absolutely, first of all because fine working monitors are filling the landscape, they are like everywhere, on every corner, so in terms of real economic behavior, not in terms of money, it definitely not just deserves, but needs further experimenting, there is a lot of tweaking that can be done that will make the result much better, like removing everything from the actual LCD panel would need much less light, what we need here are only the black pixels that are amplifying the darkening effect. Thank you for the comment, saved me an own one which much less people would see.
Nice project! The reason you need the paper between the two panels is that they have the same filter orientation (same brand/model). A polarizer filter lets light through with either horizontal or vertical polarization. The way LCD works is: there's a liquid crystal (hence LCD) layer between two polarizer filters of opposite orientation. So, for example: a horizontal filter in the back, a vertical filter in the front. This blocks all light (except some bleeds through) but when you apply voltage to the liquid crystal it rotates the polarization by some angle proportional to the voltage. So if it turns the horizontally polarized light by 90 degrees, then all of it goes through the vertical filter in front. Now when you put the two panels on top, the front vertical filter of the first panel lets through only vertically polarized light, but the back horizontal filter of the second panel blocks that. As someone suggested, you could remove one polarizer filter from between the two panels, but another option is to just flip one panel and have its front (horizontal) filter face the other panels front (horizontal) filter, so they would let light through without the need of the paper. Of course you would need to flip the image on that panel by software somehow - not sure how hard it would be. Another thing that someone already mentioned is that the two panels on top basically multiply the brightness, so you get your brightness squared - need to adjust your gamma.
I don’t believe liquid crystals will rotate the polarization if they aren’t aligned with the polarization on the side they come in. Another option is to flip the second lcd over.
Over ten years ago, students at RIT combined an Apple Cinema Display panel with a pair of projectors that, combined, matched the resolution of the display. The projectors projected the image in black and white behind the LCD panel displaying the same image in color. There were some off-axis alignment issues, but the first look at what HDR would eventually look like was stunning at the time.
@@ppsarrakis power is cheap in my country. It's ok, I would rather buy a good oled than this but it was interesting for me to see such thing can exist and actually work.
Yeah, Matt! Build an OLED display out of old electric teapot indicator lights and make a comparison! Also - the viewing angles on this double layer beast must be interesting.
@@CruelQuertos The only benefit, and why DL-LCD is available to studios is that it can reach 1000 cd/m2 which is the HDR mastering specification. It will never be available to consumers as regulatory agencies would never tolerate the near-squared power consumption.
The persistence blur would be pretty bad unless the panel can be refreshed very quickly. The new 500 frames per second panel thats in that new Alienware model has a pixel response time of 4 1/2 Ms at the slowest transition. The Hisense u9DG had ridiculously bad motion blur because of how hard it was to keep the panels synchronized
OLED will eventually replace LCD, in time. The problem with OLED is comparatively low yield in the manufacturing process, and they fade and die over time, blue in particular, will get less bright over time.
Awesome project! You could try use a half-waveplate instead of the tracing paper. This should more than double the brightness of the monitor. It's a clear film that rotates the polarisation axis by 90°. Right now, half the light is absorbed by the second monitor's polarising filter because the polarisation is random. Really nice work though. Looks amazing 👏
My first thought was to wonder whether the polarizing film could be peeled off the first LCD panel rather than polarize and then diffuse the light again. Couldn't you stack the panels THEN polarize?
@@VivekKumar-tn2ue yes normally without a polarizing film an LCD panel just looks white. But in this case there would still be a film in the correct orientation on the second panel, so I don't know if there's a particular reason it wouldn't work. It seems too dumb to be possible.
LCDs inherently twist their polarization - I can't remember if it is the dark or light pixels that are twisted. So simply flipping the front panel over MIGHT resolve the issue - but then requires mirroring its image when driving it. Also, the gamma is going to be really wrong here. Doing this properly would require something like taking 10-bit Rec. 2020 PQ in, and outputting two separate Rec.709/sRGB 8-bit signals calculated in such a way as to properly achieve the desired final transfer function.
@@K3Techs I got it just after posting the comment, the whole idea about this project was blocking more light using the polarising film so the blacks would be blacker, removing the film would defeat the entire purpose.
Awesome project! Adding a third polarized film that is perpendicular to the tv polarized film would allow you to see the image again and would probably be less fuzzy that trace paper. The other option is to remove the polarized film on the back panel.
My first thought was to use a third lcd screen but your idea would be a lot cheaper. It's not an intuitive outcome so I can see why he didn't think of it. Polarized light is basically witchcraft, after all.
You couldn't remove the diffuser between the screens, since the LCD rotates the polarization by 90 degrees. Adding another LCD would give you 180 degrees of rotation - back to the original.
To deal with the polarization problem, you can put a sheet of polarizing material between the 2 screens at a 45 degree angle to the polarization. This will "turn" the polarization without the need for diffusion. Another option: buy 2 SQUARE monitors. Assemble the screens a quarter-turn from one another. Run software to quarter-turn the images so they align.
That would probably not work because of syncronization issues. Even if the pc was able to produce the straight and 90° degrees-turned images perfectly at the same time the fact that one panel would be vertically syncronized while the other would be horizontaly would probably mess up everything (at least at higher refresh rates). It could work though if you designed your custom electronics to drive both panels from a single source and to account for rotation and syncronization.
I don't get why he didn't just flip the top LCD around so that the two identical top polarizers are touching in the middle of the sandwich and don't cancel eachother out, then flip the image on the top one.
Removing the polarizing layer would probably yield much better results, but I know it would be far harder to do without damaging the LCD. It would be cool to see you do it though, as I think you have the skills to do so.
Hot airgun and acetone, plus molybdenium wire used to separate touch on phones, that's easy if you ask any cellphone technician, just that when we do this by mistake it's a fking loss of a screen.
Removing the polarizing layer from one of the panels would likely raise the brightness floor and lower the contrast, though. The fact the two LCD panels resulted in black without a depolarizer in between indicates their output is polarized in perpendicular directions.
@@fss1704 I've partially pulled off the polarizing layer on a phone LCD by accident and managed the rehabilitate the screen by simply cleaning it off and reapplying some fresh self-adhesive polarizing film on top before. It's a royal pain in the ass, but it's not a total loss.
They could maybe just have flipped around one of the panels front to back. But then they would have to add signal processing to mirror one of the hdmi signals
Год назад+13
Front to back could work just by coincidence. What you want here is rotate the panel by 90 degrees. Better option though is to buy different panel with the same pixel layout that already runs on different polarization. Both vertical and horizontal are in use, just have to check which monitor does what. That happens even between minor versions of the same model. I had two Dell screens that were almost identical, sam layout, same size, resolution, look almost the same, one was just newer year model. They had different polarization , so when I was wearing driving sunglasses, I could see only one of them. Besides that, there are multiple types of LCD panels. They generally work by rotating polarized light with liquid crystals and blocking unwanted part. Liquid crystals are usually used to change light polarization, not filter it. So they either rotate and block dark pixels or rotate and let through the lit pixels. The one we can see here is the latter type as the output has rotated polarization. You can also buy the former type and those can be simply stacked with no additional effort. That's because the light that should go through is untouched and the light that should be blocked is rotated to be blocked by filter. This would be a perfect screen. IIRC it's more often used by VA or TN panels, I don't remember details.
This is quite similar to something that was done (very successfully) at the computer science lab at my university. Rather than using two LCD panels, they used a DLP projector as the backlight for one panel.
@@MsHojat Oh definitely! It wasn't practical as a replacement monitor but, by using a DLP projector, you have a very powerful backlight & absolutely zero light in the black areas
Unlike any other channels, this channel project literally take time, effort and innovation. But that makes it even more special. We're always waiting for the next video to release. Great job as always.
If only you'd put a similar time, effort and innovation into your comments instead of generic "praising others at the cost of putting everyone else down" statements
Guys I'm not trying to say its the only channel, or putting everyone else down. Thanks for reminding me. I'm not an english native speaker and I'm still learning english. 🙏
would love for you to have another crack at this! Youre the king of diy solutions and ive never seen you really make something where it wasnt fully complete. I think you should get back in the ring with this one !
One thing you could try to improve brightness and sharpness is removing the polarizers instead of unpolarizing the light. Just doing that should double the brightness, as 50% of unpolarized light is absorbed by a polarizer. The only problem is that it’s difficult to do without breaking the panel, but it’s doable.
@@danielxmiller hes not talking about the lcd panel, hes talking about the polariser on top of the panels, instead of removing the back one he's added a "depolariser" in the middle.
As a multimedia computing graduate, this project is simply priceless. Not just the project itself, but rather also understanding of the inner workings of a monitor. Throughout my four years of uni’ I’ve not once been explained that the LCD is taking advantage of polarisation to produce the images, as you’ve explained in very simple terms. I could not thank you enough.
Education these days is basically worthless -- only the most shallow details are covered, and the average person will learn far more if they simply do their own research.
@@hxhdfjifzirstc894 Nope 🤦♂. It's that Education often is severely underfunded and understaffed. Not to mention, if results are not "produced" funding gets cut so the material for general degrees is simplified to ensure a higher pass rate. Unless you choose very specialised subjects - that knowledge is actually needed as a foundational layer.
I find your comment about not being taught how an LCD display actually works as amazing as it is concerning. I am amazed there are people who can some how rationalize that gaffe. Of course my concern is obvious with my only real wonder being, "what the hell else is missing??"
@@DougDingus Sadly university is more about learning yourself than being taught everything. I'm certain that appropriate courses that cover display technologies can be found, as there are more complex things taught there as well. However, they mostly rely on people actively studying stuff on their own, and the main point is on how to find and process new information from various sources such as books, journals, and other publications both in paper and online. OF course, it's not perfect either. That's not what I'm saying. I was thoroughly disappointed with uni education at first during my undergrad studies, but I've found things here and there that are worth mentioning and appreciating for.
this result was unexpectedly interesting! I don't see much potential in this application in question, but rather its use in entry-level projectors, as the gain in contrast would allow the use of stronger light sources, bringing more luminosity and quality to these projectors! Send greetings to your Brazilian fans
An interesting project, thanks for the video. A couple of notes from a graphics engineer perspective: - By combining two LCD panels together and sending the same signal to both, you're getting color distortion. Each panel acts as a filter, multiplying the backlight radiance by the desired color: L * c. Two panels, therefore, square the color: L * c * c. That results in dark tones becoming darker, and colors shifting as if gamma correction with a factor of 2.0 was applied. To compensate for that, apply gamma = 0.5 to the output image. - The rainbow effect that you get with a camera is a Moire pattern, essentially a form of aliasing. When the screen is projected onto the camera sensor, its subpixels have a density comparable to the density of sensor elements, or probably somewhat higher. So, some sensor elements get more of the red subpixels, and some get more of the blue ones, for example. That issue is amplified by the Bayer pattern used on most camera sensors. To get rid of the rainbows, you can try to defocus the camera slightly, which will achieve optical anti-aliasing by removing the high frequencies from the input signal.
Very cool project Matt, even though this one isn't very practical as pulling 250W constantly for a monitor will easily offset just buying a more expensive monitor in the long term it's always fun seeing your ingenuity!
Not just the watts but your GPU would be effectively driving 2 monitors with only getting the resolution of one so WAY more GPU power is needed to. So yeah the cost of an Oled would be way more effective but yes still a interesting video.
This guy is great…every time I have that question lingering in my mind as he’s showing it, he immediately give the explanation or reason for how something works for his projects. Like he reads our thoughts. Great informative video. Wish I had the resources and time to make one myself. I’d want to try this on some older cheap lcd tvs
Matt, I absolutely love your content. you're easily my favourite content criator on youtube. thank you for giving us the most professional, entertaining, educational stellar content every single time
This just makes me want to find a high refresh rate monitor with a busted backlight and just drop it on top of an overhead projector from the thrift store so I can finally have my high refresh rate projector I've wanted for years now. Seeing you do the disassembly and backlight design was inspiring and empowering! Excellent as always
THIS project? Not the projector one? ruclips.net/video/YfvTjQ9MCwY/видео.html Admittedly that is a lot more effort than "slap monitor on overhead projector". But I can GUARANTEE you the overhead projector route is not great. Noise, light leakage. Aspect ratio. Awkward placement of ohp. Not saying it can't be done, as it has been done to high degree of success since ages ago on various forums. But it is also high effort for a "good solution". And the "low effort" method is very much, "you get what you pay for", so to speak. I suppose you are looking for a specific functionality that is missing in commercial products for a reasonable price, so it might be worth the effort. But if so, I imagine DIY Perk's projector project is also worth the effort. Just a thought.
While the projects and videos are amazing, This particular one is hardly new, Hisense was experimenting with it some time ago before moving on to a different way of doing things
@@Yuki2204 Hisense? Dude, dual layer LCD's been a thing since the 90s, LOL hardly new but what is new is how easy it is to accomplish at home with little effort.
@@Yuki2204 1990 Toshiba had dual layer SuperTwisted Nematic LCD displays which were two layers of TN panels together to increase contrast. Both LG and Samsung had dual layer prototype LCD panels in early 2000s, Samsung also produced dual layer LCD panels with ability for 3D which were mostly used in slot machines and arcades. Sony has a long history of dual layer LCD and later dual layer LED panels for their broadcasting monitors which they still produce to this day. So yes, they did exist, they did cost a lot of money, and no they were not popular, the first prototypes of LCD technology back in 1984 also had prototypes of dual and triple layer panels sandwiched together to create a two colour output, this was before they figured out they could sandwich a colour filter to get colour output from one panel.
OK, firstly, this is amazing, honestly, many kudos. And please make this a series, this can be improved vastly. On the top of my head a low res monochrome panel as the 2nd display or using controlled mini led for backlight, etc there's just too many possibilities. I cannot wait for more versions of this, Thank you.
Bad idea. The second panel must be identical to avoid image degradation with the different characteristics (transition timings, latency, gamma offset, etc.) For the controlled mini led for backlight, this is basically how the high end HDR LCD display with local dimming zones works. But its very hard to DIY as it needs extremely tight timings to analyse the signal and switch light with the same latency as the LCD panel. And the absence of blooming will vary with the number of zones and the size of the lighten elements on screen.
We have panels like this on a lot of our casino floors! IGT made a unit called the G23MLD (Multi-Layer Display) that has 2 LCDs stuck together to achieve a 3D effect on the slot theme. Most of the time the reels are on the back monitor and the line indicators are the front monitor
As soon as I saw this video I thought the exact same thing. We have them as well but most of ours have developed black spots. So I question the longevity of the monitor, but that one still looks much better then the IGT version ever did.
@@brandonf2562 Have you been making sure the fans are working? One of our older tech was saying that half the screen wouldn't be working soon after the fans died
great idea. I am stunned with the result of your experiment. if the problem is the dropshadow-like on letters especially during Ms task only, I suggest to make the lcds additional switch on and off so that if and whenever watching videos or gaming you can enable both lcds to take advantage the viewing performance and just disable the other lcds during light task only like Ms word.
9:09 there's a lot of contrast, sure, but did you adjust gamma to 0.5 to compensate for the multiplying effect? With two displays both displaying 50% gray you actually get 25% gray (0.5 * 0.5 = 0.25). A gamma of 0.5 (effectively sqrt) brings that back up to the desired value.
I cringed a bit at 9:49 when he said that it's "not at all" like increasing the contrast setting, because I'm pretty sure the contrast settings also increase contrast by messing with the gamma (or rather by changing completely the EOTF).
this gave me a more intricate look at a monitor than I ever thought I'd be seeing. So I thank you for that alone. Gave me a much better idea of how these work.
I suspect that you could get a slightly crisper image by using more uniform commercially available diffusion films and optically transparent epoxy to bond all of the panels and the intermediary diffusion film together. I do appreciate how you went straight to replacing the entire backlight system with something substantially brighter, though doing so without needing to utilize several layers of commercial optical dispersion films with different dispersion patterns before using a standard frosted dispersion film to make it uniform does necessitate a deeper enclosure. I do think this is a topic worth revisiting at some point in the future to further refine the idea. Maybe with a couple of cheap 43" 4K televisions. If you have a decently deep desk, then a 43" television mounted to the wall behind it is a great monitor size for most applications. If possible, it would also be a good idea to just utilize an inexpensive HDMI splitter to duplicate the HDMI signal without relying on your GPU to keep the images perfectly syncronized. Ideally, you would want to find a way to drive both panels from the same driver board so that the signal is sent at the same time and there isn't a risk of clock drift between the driver boards. This would reduce ghosting. Unfortunately, it's unlikely that the driver board can provide enough current on its output pins to handle two panels in parallel.
I've been wanting to build something like this ever since I saw coverage of a panel maker that was looking into something like this with a 4K main panel and a 1080p black and white panel to help control the backlight. I'm quite sad that it didn't work out, it definitely does seem worth it compared to the cost of higher end alternatives and the superiour amount of "local dimming zones". I would like to see you test the colour accuracy, as I prefer the image to be more accurate to being overly vibrant.
Hey man, just want to leave a random comment while it's on my mind. You're an inspiration! Using this insane technology around us and giving it new life, using some skill and know how to make things that people wish they could just buy. Or the exact crazy device that does what you personally want, how you want it. Amazing. You're an inventor's muse to be sure.
Super cool project, I always wondered if it was possible to do this. How about removing the polarizing filter of the first panel instead of using tracing paper? I know that people have made those "privacy screens" from old LCD's that are only visible when wearing polarized sunglasses. Or is the film impossible to remove in these modern panels?
Pretty neat, and very clever! The thickness bothers me 0%, it's still far lighter and thinner than many computers I've used in my life. But the increase in energy cost over time and the slight fuzziness are issues that i think would steer me away from making one
i wouldn't be surprised monitors are often lighter than computers if we're talking about desktops, but this isn't an AIO if that was the persuasion you caught
I mean the thickness could be fixed with a good diffuser that isn't baking paper would help the power consumption as well given the LEDs dont need to be as bright as before.
Have you tried to remove the polarization layer on the first screen, like a lot of projects do to make "secret" screens that only show up when wearing polarizing glasses?
@@Elemental-IT That's the point of the polarising glasses. You take the front polariser off the LCD, so the LCD is plain white and unreadable, yes. Unless you're wearing a pair of polarisers in front of your eyes. It's a silly hack but I suppose fun, I wouldn't do it to a monitor I actually wanted to use.
I came up with this idea myself a couple of years ago when I found out about how LED TVs with array backlights work, glad to see you tried it and that I was onto something!
I wonder if you could look into multi-layer 3D light field monitors. They use multiple LCD panels at a slight distance. Basically you can calculate how a certain light ray would move through the panels, and color their pixels accordingly. For example in this decade old paper: Tensor Displays: Compressive Light Field Synthesis using Multilayer Displays with Directional Backlighting
Ooooo amazing. I just said in another comment I'd love to see spatiotemporal superresolution as well as using old Sharp Quattro panels (the *panels* WEREN'T snake oil, but the TVs were, because some idiot at Sharp 'cut costs' & they used narrowband tuned RGB backlights instead of RGBY backlighting, effectively making the amazing panel useless), combining that with light field monitors o boot... *chefs kiss*
That is definitely one of the more most compelling uses for doubled-up LCDs. The tried and tested method for angular modulation has been to place a large array of lenslets in front of an LCD to modulate the angles of the pixels. Once you have computed the mapping from pixels in 2D to (hopefully tight & collimated) rays, you can trace rays backwards from the pixels to query their intersection with some simulated object, thus determining which pixels to switch on. This is great, but even with high resolution LCDs this ray optics model ignores the wave properties of light. Ideally, a second LCD in front of the lenslet array would have high transmission rate, and rather than modulate the brightness, modulate the phase of the transmitted light in order to simulate optical wavefronts. That would, in theory, make simulated objects in front of the panel focusable. A company called Light Field Labs appears to have done something like this using a static phase-only optical layer, but as far as I know nobody has demonstrated switchable phase modulation / dynamic focus using such a large panel.
This is why I usually prefer VA panels, they're much better these days than they used to be, viewing angles aren't really an issue on a good panel, mine has the same viewing angles as the same MSI IPS monitor. I got the curved version and I love the deep blacks and contrast. It's a high refresh rate too, 165hz. I have had bad VA monitors though a while ago but this one I have now is great.
@@heartbroken3344 actually not all, I recently tried the flat AOC AG275QXN and it's great in both response times and input lag, the only issues are viewing angles were very bad and the washed out colors probably because it is flat, also bad gamma shift (most noticeable on black color) so I returned it.
The right way to measure the convenience of "dual layerlcd" is to present an image like a chess board, 100% white squares along side 100% black squares. Then you measure the luminance with a luminance meter. A typical white luminance is about 80cd/m2, and a typical black luminance is about 2cd/m2. The contrast ratio is the white luminance/black luminance. A dual layer lcd is better than a regular lcd ONLY if the contrast ratio is greater in tje dual layer than in the regular lcd. You show us a darker black in the dual layer but also a brighter white in the regular lcd. We can say nothing without measuring contrast ratio.
In a typical display the contrast ratio is 40. So the white transmitance is 40 times the black transmitance. Lets say that the white transmitance is 0.4, so the black transmitance is 0.01. With 2 layers the CR=(0.4*0.4)/(0.01*0.01). The CR=40^2=1600. Wow, I was way way way wrong. Sorry by that.
10:35 Dual LCD is not super energy consuming when you do it right! There are some LCD monitors out there that use only 35 Watts for a 4K 144Hz screen. For sure a manufacturer could make a HDR 1000 dual LCD for slight under 100 Watts. LCD always was superior than OLED, but for some reason the industry don't want to innovate on dual LCD monitors. Your example shows how close you can come with just some wet finger work, now put that same work to a nano meter precision. There is the answer, you can scale your model with an easy 100x in better picture quality while doing the same thing on a nano scale precision.
@DIYPERKS Hey man I'm a big fan. Seeing you do this is awesome. You should try to get this thing as close to HDR as possible and add the ability to strobe the backlight for blur reduction. Because then that would be the ultimate monitor that you could have. Strobing the backlight would be able to improve motion resolution, and if you tune it right you can get it close to CRT like motion. One thing you might consider is using lasers and phosphor rather than LEDs for your backlight, because that would use substantially less power. If you wanted to go really ham with it using an mems system with lasers and phosphor would give you a built-in strobing backlight that would be very efficient and bright. You may even be able to remove the polarizers if you used lasers in a MEMS system because such a system would mimic an 7emissive display more closely.
Peak brightness is already an issue, so strobing the backlight might not be that feasible. I don't have much experience with displays that support that, do you know what's the typical duty cycle on strobed backlights at 60hz?
@@carnivorebear6582 it would depend on what Motion Picture response time you were looking for, as well as the native refresh rate of your panels and what refresh rate you want to strobe them at. A 240 hz LCD running at 240 FPS and pulsing the backlight with a 50% duty cycle (flash length) will let way more light through then if you just used 60 frames per second with a 50% duty cycle. A 50% duty cycle usually I believe so it cuts brightness in half, but you could also overvolt the hell out of the LEDs since you are pulsing them as opposed to leaving them on constantly, just like how fald back lights work. You can also adjust the duty cycle for lower MPRT and higher brightness or vice versa. You can pulse an LEDt at much higher brightness levels then they are rated for provided that you are only doing it momentarily. You can pump about five times to 6 times the brightness out of a typical LED for short duration without damaging them. Since he already has the heat sinks, he could get more. Other people have also mentioned that having one layer be monochrome would let more light through. This is how HDR works on Modern qled displays. Or how lightboost used to work in old 3D monitors. Adjustable pulse length and precise voltage control. Power and Heat considerations is why I think a laser phosphor based backlight would work way better then using LEDs. A micro electric mirror that steers a high watt laserbeam at a phosphor layer could get you very high brightness at low power usage, and strobing also improves the native contrast of the panel.
This is what I've been wanting to do for weeks now. It actually started with the tv at home breaking down. I learnt all I could about tv, then I wanted to build one at home.
I have a idea of removing the need of the tracing paper for depolarization. You can flip the inner panel, so the inner panel is facing outside with its original inside side. The back light passing through the inner panel gets converted to the same polarization the outer panel's inner polarizer. Its kind of hard to explain it. The polarizers' configuration of yours is H-V-tracing_paper-H-V-backlight. If the second panel is flipped it becomes H-V-V-H-backlight. Connect both panel to the same graphics card, and in the graphics card driver set the display of second panel to be flipped.
Unfortunately, that wouldn't work. You'd end up with the RGB subpixels being swapped between the front and back lcd panel. While you could potentially get around this by using panels with different subpixel arrangements, text may still look weird due to the subpixel antialiasing used by most text renderers.
You should try this again but using SMD LED diffusers. They're a little plastic dome that just clips on over each LED to get the diffusion effect you are looking for instead of the tracing paper. Also, polarizing film isn't *that* expensive, changing the in-between layer for a sheet of that will likely give you even better contrast.
Matt. Thank you for making some of the most wholesome, calming content on the internet. mere minutes ago i was having a panic attack, and your channel was recommended to me. i clicked on this because it looked awesome. the calming music, your soothing voice and the interesting video idea have brought me back down to earth. Thank you so much.
There was a monitor that did this a while back and used a second black and white display to increase contrast without affecting colour balance. I think ltt did a video on it.
Why not flip or rotate the second LCD to match the polarization? Also if you don't need all the contrast reduce the black-level on the second LCD to get some more light transmission
To pass light you'd have to rotate it 90 degrees which has certain obvious geometric problems... Flipping it or rotating it 180 degrees just brings the polarization back to the same orientation it already is at.
@@tylisirn Flipping should switch polarization. Let's say the first is vertical in, horizontal out. Then the second unflipped is also vertical in, horizontal out. Which mean in the middle you have vertical + horizontal= no light gets through. With a flipped you have vertical, horizontal, horizontal, vertical. Which works
Presumeably, the first display panel polarizes light linearly, which requires a 90 degree orientation for the second panel to allow light through. (Ever had polarizing sunglasses on and turned your head to read an LCD display?) Now, if only the panels could polarize ths light circularly...that's tricky because they would need to be a multiple of 1/4 the wavelength thick. Polarizing filters on auto-focus cameras have to do this.
@@jeffbaker8808 Well a panel has not just one but two polarizers which are shifted 90° too each other. The LCD in between "turns" the light. So it's linear horizontal filter - LCD - linear vertical filter. That's why flipping works
Wouldn't flipping one panel cause issues with the pixel layout? One panel would be R-G-B the other would be B-G-R. Also flipping a panel 180 or 90 degrees causes inherent input delay or extra processing time in windows, pretty sure LTT did a video about this ages ago.
That is pretty cool. I feel like this would be extra nice for text in 4k. I want to go to OLED for my monitor but worried about burn-in. Dual LCD might work, if it was actually available for consumers.
My solution for the polarization issue would've been to flip the image around for one of the panels, and literally flip the front panel over so its "incoming" polarizer and the "outgoing" one from the rear panel would match. But to be fair, I'm not sure how you'd make the monitor flip around... could fix it in software if you make both monitors show the same thing but flip one, most OSes allow that, right?
I think the best way to do this mod would be to remove the polarized filter and diffuser from the second display. Not sure how feasible that is without completely destroying it, though.
It would have been polarised at right angles, so you would have needed to rotate the overlaying panel in the vertical orientation. So, would not be practical
Super neat. I feel one aspect for improvement is calibration between the position of the two panels to narrow down some of that 'drop shadow' effect. I'm sure you could probably also get better LEDs for the backlight as well.
HDMI splitters often cause problems with HDCP, I think it is better to connect both panels directly to the graphics card and clone the output in the driver. Btw: You can avoid the Moire artifacts in 2 ways. Increase the distance to the LCD and zoom in or tilt the camera by 45° and correct for it when editing the video
probably your best concept so far. with a more efficient scattering layer than the tracing paper i think you could get thhe brightness up and power down and it would be a viable commercial option
no the used dual layer oled displays wich is a completely different technology, plus they did not increase contrast but the brightness so essentially they did the opposite of whad matt did
I'd suggest to just remove the polarizing layer on the second display. Also, instead of projecting the same image on both display, you could try to project a grayscale version on the second layer just to enance the blacks. That would resolve the grey looking black areas without altering the colours accuracy that much.
There might be a way to run the second screen in greyscale. If you run both screens in duplicate mode from your GPU and set one of them to greyscale then you might be able to recover some of that brightness at the cost of color vividness.
even if the screen runs in grayscale, the LCD itself still has color filters that will block out 2/3rd of the light, he'd have to somehow remove the color filter from the LCD to let in more light. (the colors will probably be more accurate eitherway though)
The problem is light polarity (horizontal or vertical). The LCD twists the light 90 degrees, and polar filters (horizontal and vertical) block the light that doesn't match. Grayscale does nothing to twist the light to align with the polar filters.
MSI has a 1440p QD-OLED that’s currently around $500 brand new. Some of the lesser known brands on Amazon have W-OLEDs that dipped below $500. AOC also has a mini-LED that’s under $300
I like the irony! People who watched the video on a 1080p non-hdr cheap display couldn't appreciate much of the improvement your dual layer LCD provided. And i personally was amazed... And then i remembered my TV is HDR10+ Dolby Vision...
You could probably get even better results by pairing a single lcd panel with a high quality black and white crt. Monochrome CRTs can be pretty bright and sharply contrasted so they would make the perfect "pixel adjustable backlight" for the LCD. You also wouldn't need the diffusion layer in between because the light coming out of the CRT isn't polarized, so the contrast would be even sharper and without blooming.
Really cool project! I use a eizo prominence for my work, and it’s my understanding that it uses dual LCD tech to get the black blacks. It would be really interesting to probe your monitor and see what the gamma curve looks like, from what I see in your video, It looks like it might be a bit steep (your mids look dark). Either way, excellent work!
I have been sat here amazed by the performance of your creation, then I remembered that I have been watching it on my own traditional LCD monitor all along anyway lol. I would love to see something like this in person.
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
🥦 Let's hear it for the random broccoli
Lets gooo
Hello I'm the broccoli and I wanted to say that this is really cool. Thanks, Matt.
Or we could burn it with fire, encase it in lead and then eject it into space. Options.
Does that count as one of 5 veg's for the day?
Next try the laser tv
If I remember correctly, Hisense has done this in one of their TVs. They achieved incredible contrast levels by using a second monochrome LCD layer to provide those inky blacks. Sadly, I don't think they've pursued it any further, synchronizing the LCD layers was quite difficult.
Apple -did it- didn't do it
Edit: sorry, the Pro Display XDR is not Dual LCD. I swear I heard it was when it came out, but that was a myth
LTT has made a video on it
@@zachb1706 nope
@@zachb1706 did it on where?
I have one of these Hisense TVs it’s amazing except one of my back panel pixels died. Twice as many to go wrong I guess.
I'm just amazed how accurately you were able to align the two LCDs
Pixels on a 1080p monitor are not actually all that small, you can see them with your eyes if you get close, so it's not as hard as you might think.
Why does this video only have 700 likes?
Edit: my RUclips glitched
just use some sort of end stop lmao
Valid question. Misalignment can lead to Moire easily.
@@user-a6olopvmv9 Moire shouldn't be an issue if the tracing paper blurs the image enough to eliminate pixel boundaries. Effectively turning the back screen into a local dimming layer while the front screen creates a sharp image.
Something you didn't mention is that you have essentially doubled the display's gamma by multiplying all colors by themselves. Most displays have a built-in gamma of 2.2, but this raises that to around 4.4. You should be able to compensate by halving the gamma in software or in the monitor's settings, which will give you more accurate colors and fix over-saturation.
same thought
it would not be 4.4 since gamma is not on a linear function. More importantly, gamma does not describe the color saturation, it is instead a measurement of the luminosity of the pixels. We could assume that the gamma should actually be set higher than 2.2 because the overall luminosity has dropped with the combined LCD screens. As for color oversaturation, color saturation is also not on a linear function, and each color has a different curve. Meaning that each color acts differently on the same brightness, and colors even have different saturation changes with the same brightness change. And to top it off, since the contrast is being changed as well, the color saturations are changed even more. There is no way to calculate the required settings on your monitor. Which is why I'm sure he used a colorimeter like every youtuber out there to calibrate the monitor
@@bobsteven2363 Your fun at parties huh?
@@KenTheSteak I’m no fun at parties because I attempted to correct someone who is wrong? Do you believe people who are wrong shouldn’t be called out? That’s what Karens believe, and Karens aren’t fun at life
@@bobsteven2363 I asked if you was fun, never said you wasn't, but thank you for answering, although I didn't ask "tell me you don't have friends, without telling me you you don't have friends"
I would love to see a V2 version of this where you dissemble the panels further and experiment with various filter layers.
I AGREE! 2nd version soon!?
what sort of filter layer?
I support this idea, I think he should go further with this! Maybe that will inspire some manufacturer to actually go to market, people don't care that much about thickness or consumption as companies may think.
Maybe seeking a pair of LCD panels, one with and one without a filter layer might be a better goal to begin with.
@@b00marrows Or maybe just peel off polarising layer.
Professional dual layer LCD monitors use a monochrome LCD as the rear panel, which is better for light to pass through, thus increasing the efficiency (though it is still significant less efficient than a single layer LCD).
Also, with your method, you are significantly increasing the saturation, so you would need to get a calibration device to counter that.
Still, this is a really cool project!
Pretty smart idea to separate luminance and chrominance layers, rather than going with the QLED/rough zone lighting approach.
too bad there's no actual market with monochrome LCD's and conversion of colored LCD into monochrome one is nearly impossible task
@@1337Shockwav3 I think that both are good for different purposes and budgets. The dual layered LCD monitors are designed for professional colour grading (especially for HDR) in high budget studios. QLED are great for other purposes, they are much cheaper, take less power, and are thinner. Mini-LED backlights are also gaining popularity, though they suffer from blooming. Personally, I don’t mind a bit of blooming as long as I am not in a dark room since I am not a professional filmmaker.
I think this technology would be great in the hands of radiologists, but since medical diagnostic displays take a lot of time to be approved, this could be a while… though I think an OLED one specifically engineered and calibrated for medical imaging with an active cooling system would be the best.
Edited: corrected.
oled will be cheap in a few years
Could he not use grayscale on the second screen? I know it technically still uses RGB components, but at least they would all get dimmer in proportion.
I think you could actually add a middle polarization layer between the other two. Weirdly, if the middle layer is at an angle between the other two, it allows light to pass through all three. I think this would actually allow more of the original light through than paper.
this!
"Weirdly" is an understatement. The fact that this works seems to defy all rational. like some kind of glitch in the matrix.
@@melty4204 Wasn't it that if you consider the polarized light coming through as a vector, the vector gets projected on the next layer so that if you have a 90 degree angle, the projection result is 0, but if you project it 2 times on a 45 degree angle, it isn't 0, which means that light still gets through.
He could also flip over the second lcd, that should be the correct polarizations.
If he took apart the panel stack and then stacked just the actual LCD's and then added the lens layers to that stack from one display in then he wouldn't need any of that nor the additional LED lights. The LENS layers serve multiple purposes and depending on the construction the polorizer often isn't affixed to the actual LCD and is just one of those layers in the LENS stack. On some panels it's glued directly to the LCD, but if it's not, then you can just stack the actual LCD's and place one stack of filter layers and that would eliminate all the Extra stuff he did with the paper and LED's.
Note: When working with LCD panels films and lenses use gloves that leave no residue and do this in a dust free room while wearing something over your head or you will forever see dust, hair, and fingerprints on the image of the screen.
One thing I think you might have missed is the gamma. Monitors are non-linear with respect to their input, and there is typically a compensation curve (gamma) to fix that. However, with two screens, your nonlinearity will be doubled and you will need to adjust your gamma curve (easily accessible in desktop OSs).
Yes, although not so easily calibrated and you can run into precision issues with the required exponent of 4.8 that he now needs instead of the more typical sRGB exponent of around 2.2.
You will just have to hardware calibrate the end result. Might be easier if you calibrate both monitors before the sandwiching
@@dorbie Really, this sort of display should be driven with different images sent to each monitor for this reason. It's got 16bpc in total at the higher gamma which could be remapped to the standard 8 (or even 12) with reasonable precision
@@jameshogge That would take a lot of highly refined calibration and systems integration that isn't really feasible in this context, it can buy you more bits but not 16 bits without independent backlights and a degree of control that doesn't exist here. You can do better than this though for sure. See my own response elsewhere and my referring to "esoteric reasons" for setting different display gamma values on each. Something that is feasible and easily within reach (if supported).
@@dorbie The calibration needn't be too hard: you can do well enough by assuming that the individual LCDs are correctly calibrated. The rest is just maths with the light intensities. One of the groups at my uni did a similar setup with a DLP projector as the backlight for an LCD. But yes, it is a pain to then drive the displays. Ideally, you use an FPGA. Alternatively you have to set up a virtual display in Windows and write a couple of shaders to produce the actual outputs.
i think removing the polarisation filter from the back lcd will help with brightness and eliminate the need of the tracing paper, you can then also bind both of the screens with uv glue used for screens, so there will be less blooming and blurryness in text
I was thinking the same thing....removing one of the filters would do the trick.
Me too, but what do I know. It would make the job easier and better if it worked
It won't work. The front layer of the back lcd and the back layer of the front lcd are 90 degrees apart in terms of polarization, so it would still completely block the light, unless you can find special LCD panel pairs which are oriented 90 degrees apart.
This seems like it would intuitively make sense, but I did some research and I don't think it would work.
The liquid crystal film doesn't block much light. It's just red/green/blue translucent goop, organized into subpixels that can be polarized based on how much of that subpixel's color we want to shine through. Without the polarizing filter, the backlight shines through all the subpixels all the time, and it's almost impossible for our eyes to resolve an image. You can find video of what LCD screens look like without the filter, and they're basically just bright lights. Actually, I think this channel has done videos about upcycling broken LCD panels into lightboxes for exactly this reason. With only one polarization filter, it'd be the same problem as before -- not being able to block enough of the backlight to get the nice inky blacks and high contrast.
@@crystalsoulslayer How about getting a SLIGHTLY CLEARER polarizing filter that matches the second panel orientation?. I feel like he didn't explore this option enough and just brute forced the problem
While I usually find his DIYs trying the best & practical solution, this one has scope for a lot of improvement.
1. The large backlight led array: This is what makes this build thick, impractical and causes high power consumption. This can be avoided if tracing paper is not used. A better solution for the polarization problem is the way forward, which leads to #2
2. Tracing sheet: 2 stages of tracing sheet was used - one between the panels and other in front of the backlight. Both can be avoided.
Fix / Recommendations:
The polarizing element in the lcd panel is actually removable. He has to further separate layers of the lcd panel and remove the polarizing layer from the base panel. This removes the need for tracing paper and a more powerful backlight and solves blurry text issue.
Second step is to change display settings in outer panel to high contrast black and white while inside panel displays full colour. This will solve the color gamma issue and the outer panel will only enhance blacks.
would be pretty interesting to test the optimised final version with an screen calibration tool, just to see actual numbers of things like nits, peak brigtness, color saturation, gamut, etc...and compare them with standard monitors
srgb and adobe rgb too!
@@firdaus5093DCI-P3 too
I need very high brightness for outdoors
Love seeing these diy projects that compete with hardware 10x the price
To be clear, this is not going to "compete" with mini-LED & OLED tech. Mini-LED displays get MUCH brighter and while this may match OLED brightness, it will fall short and in pixel response times and input lag.
Youre also driving two displays off of a single output, so you're gonna need one hell of a GPU
@@erikhendrickson59 Nah it will be comparable to just driving a single screen in terms of GPU power if you use the "duplicate this display" option in windows.
@@erikhendrickson59 hey, watch the video before writing, you don't need a "hell of a GPU", duplicating the signal is enough, and even if you tried running both out of the GPU mirroring monitors is one of the simplest things GPUs do, it's the same thing twice, it doesn't have an overhead at all, people that make GPUs are some of the smartest people on the planet. Please stop thinking that just because you don't see a solution for something no one is capable of doing it
@@erikhendrickson59 GPU power is going to be the _exact_ same as with one monitor.
But you are absolutely right about the brightness, and the power consumption is also absolutely crazy. It's not a viable product because of these
@@adamashworth6293 The HDMI splitter he's using means the PC only ever knows about a single monitor. The GPU is doing 0 extra work, because it doesn't even know the image is duplicated
You always come up with the most fascinating ideas. They aren't always practical, but the attention to detail you put in makes almost every build you do a piece of art.
If you fed both screens different HDMI signals, like in a "mirrored" dual monitor computer setup (using two outputs from your graphic card), you'd be able to adjust both layers individually. This could help making the picture's colours less "poppy" and maybe even help increase the white levels.
Not in this case, as while you're squaring the contrast ratio, you're also squaring the total light blocking, which is already about 97% on LCD.
I don't know, but they might have slightly different vertical blanking intervals - meaning one monitor might start refreshing at a slightly different time than the other.
I don't KNOW this to be true, but I suspect it might be true. That might be causing some problems, but really, I know about the technology of the LCD itself, but not the details of how it works.
Does it then akt like a holo display (possebly if you increse the distance between them by a bit) but id like to see a DIY holo-skreen
or be able to watch 2 full screen RUclips videos simultaneously)))
@@fuzzywzheif running these from identical type outputs from the same quadro / Ax000 card or Radeon card (AMD is significantly less tight about locking up feature access on consumer GPUs, but low tier quadro cards are more bountiful harvested from old workstations), there's a ton of possibilities for adjusting stuff here to compensate for sync issues, gamma curves etc. Metrics and results should be pretty reliable, especially if you can get both displays sourced from the same production batch (almost guaranteed if bought new together). As that should avoid any internal revisions that wouldn't have noticeable impact on 2 individual monitors but could cause headaches combining 2 independent panels into a single display sandwich.
This is a project that I think deserves a version 2. You might get better results with different panel technologies (like TN VN and IPS), or you might be able to dig and find better solutions for diffusion, polarization, backlighting and driving the displays. Even if the final result is expensive and impractical I think it's worth it to see just how much is possible by going full DIY.
He should just remove one polarizer off, i'm not sure wich but it should improve the image if done right.
VA panels ftw
@Felipe Siqueira even if not removing one you can just get another polarizer sheet and rotate until it's corrected. Using paper for diffusion and polarizing is what kills this screen.
Yeah, absolutely, first of all because fine working monitors are filling the landscape, they are like everywhere, on every corner, so in terms of real economic behavior, not in terms of money, it definitely not just deserves, but needs further experimenting, there is a lot of tweaking that can be done that will make the result much better, like removing everything from the actual LCD panel would need much less light, what we need here are only the black pixels that are amplifying the darkening effect. Thank you for the comment, saved me an own one which much less people would see.
@@djayjp worst panel
Nice project!
The reason you need the paper between the two panels is that they have the same filter orientation (same brand/model). A polarizer filter lets light through with either horizontal or vertical polarization.
The way LCD works is: there's a liquid crystal (hence LCD) layer between two polarizer filters of opposite orientation. So, for example: a horizontal filter in the back, a vertical filter in the front. This blocks all light (except some bleeds through) but when you apply voltage to the liquid crystal it rotates the polarization by some angle proportional to the voltage. So if it turns the horizontally polarized light by 90 degrees, then all of it goes through the vertical filter in front.
Now when you put the two panels on top, the front vertical filter of the first panel lets through only vertically polarized light, but the back horizontal filter of the second panel blocks that.
As someone suggested, you could remove one polarizer filter from between the two panels, but another option is to just flip one panel and have its front (horizontal) filter face the other panels front (horizontal) filter, so they would let light through without the need of the paper. Of course you would need to flip the image on that panel by software somehow - not sure how hard it would be.
Another thing that someone already mentioned is that the two panels on top basically multiply the brightness, so you get your brightness squared - need to adjust your gamma.
I don’t believe liquid crystals will rotate the polarization if they aren’t aligned with the polarization on the side they come in. Another option is to flip the second lcd over.
Over ten years ago, students at RIT combined an Apple Cinema Display panel with a pair of projectors that, combined, matched the resolution of the display. The projectors projected the image in black and white behind the LCD panel displaying the same image in color. There were some off-axis alignment issues, but the first look at what HDR would eventually look like was stunning at the time.
Its funny that since the coming of lcd , we had to wait from crt to oled to get decent images again
Rochester Institute of Technology?
@@murray821
Decent image but you still lack the perfect vertical and horizontal motion resolution of a CRT.
@SMGJohn what's wrong with the movement in OLED ?
@@bongcloud_ken pixels
It is literally kind of a dimmer OLED but without any risk of burn-in 🔥
Really well made ! Great job man, your creations are incredible.
also this uses like 300++watts thats 4-5 times my 34inch+24 inch monitor combo
@@ppsarrakis power is cheap in my country. It's ok, I would rather buy a good oled than this but it was interesting for me to see such thing can exist and actually work.
@@ppsarrakis I don't care.
I would just buy a mini-led monitor
I would’ve loved to see a comparison between this and OLED. Would be cool to see the difference in black levels and persistence blur.
Yeah, Matt! Build an OLED display out of old electric teapot indicator lights and make a comparison! Also - the viewing angles on this double layer beast must be interesting.
Oled will have a better emd result. The pixels are off when black cant get better than that
@@CruelQuertos The only benefit, and why DL-LCD is available to studios is that it can reach 1000 cd/m2 which is the HDR mastering specification. It will never be available to consumers as regulatory agencies would never tolerate the near-squared power consumption.
The persistence blur would be pretty bad unless the panel can be refreshed very quickly. The new 500 frames per second panel thats in that new Alienware model has a pixel response time of 4 1/2 Ms at the slowest transition. The Hisense u9DG had ridiculously bad motion blur because of how hard it was to keep the panels synchronized
OLED will eventually replace LCD, in time. The problem with OLED is comparatively low yield in the manufacturing process, and they fade and die over time, blue in particular, will get less bright over time.
Awesome project! You could try use a half-waveplate instead of the tracing paper. This should more than double the brightness of the monitor. It's a clear film that rotates the polarisation axis by 90°. Right now, half the light is absorbed by the second monitor's polarising filter because the polarisation is random.
Really nice work though. Looks amazing 👏
My first thought was to wonder whether the polarizing film could be peeled off the first LCD panel rather than polarize and then diffuse the light again. Couldn't you stack the panels THEN polarize?
@@K3Techs I was thinking the same thing, i saw some videos doing so and the user not able to see anything without polarising sunglasses.
@@VivekKumar-tn2ue yes normally without a polarizing film an LCD panel just looks white. But in this case there would still be a film in the correct orientation on the second panel, so I don't know if there's a particular reason it wouldn't work. It seems too dumb to be possible.
LCDs inherently twist their polarization - I can't remember if it is the dark or light pixels that are twisted. So simply flipping the front panel over MIGHT resolve the issue - but then requires mirroring its image when driving it.
Also, the gamma is going to be really wrong here. Doing this properly would require something like taking 10-bit Rec. 2020 PQ in, and outputting two separate Rec.709/sRGB 8-bit signals calculated in such a way as to properly achieve the desired final transfer function.
@@K3Techs I got it just after posting the comment, the whole idea about this project was blocking more light using the polarising film so the blacks would be blacker, removing the film would defeat the entire purpose.
Awesome project! Adding a third polarized film that is perpendicular to the tv polarized film would allow you to see the image again and would probably be less fuzzy that trace paper. The other option is to remove the polarized film on the back panel.
My first thought was to use a third lcd screen but your idea would be a lot cheaper. It's not an intuitive outcome so I can see why he didn't think of it. Polarized light is basically witchcraft, after all.
This is a clever idea.
You couldn't remove the diffuser between the screens, since the LCD rotates the polarization by 90 degrees. Adding another LCD would give you 180 degrees of rotation - back to the original.
I was thinking that to, if he removed the rear panels filter he would also get more light and a clearer picture.
To deal with the polarization problem, you can put a sheet of polarizing material between the 2 screens at a 45 degree angle to the polarization. This will "turn" the polarization without the need for diffusion.
Another option: buy 2 SQUARE monitors. Assemble the screens a quarter-turn from one another. Run software to quarter-turn the images so they align.
That would probably not work because of syncronization issues.
Even if the pc was able to produce the straight and 90° degrees-turned images perfectly at the same time the fact that one panel would be vertically syncronized while the other would be horizontaly would probably mess up everything (at least at higher refresh rates).
It could work though if you designed your custom electronics to drive both panels from a single source and to account for rotation and syncronization.
He could also just have removed the bottom screens polarization layer and anti glare film which would have left him with what he needed
I don't get why he didn't just flip the top LCD around so that the two identical top polarizers are touching in the middle of the sandwich and don't cancel eachother out, then flip the image on the top one.
@@peshozmiata I'll be the same, it has to be perpendicular pass the polarisation, that's why the comment above mentions squared screens
Removing the polarizing layer would probably yield much better results, but I know it would be far harder to do without damaging the LCD. It would be cool to see you do it though, as I think you have the skills to do so.
Hot airgun and acetone, plus molybdenium wire used to separate touch on phones, that's easy if you ask any cellphone technician, just that when we do this by mistake it's a fking loss of a screen.
Removing the polarizing layer from one of the panels would likely raise the brightness floor and lower the contrast, though. The fact the two LCD panels resulted in black without a depolarizer in between indicates their output is polarized in perpendicular directions.
@@fss1704 I've partially pulled off the polarizing layer on a phone LCD by accident and managed the rehabilitate the screen by simply cleaning it off and reapplying some fresh self-adhesive polarizing film on top before. It's a royal pain in the ass, but it's not a total loss.
They could maybe just have flipped around one of the panels front to back. But then they would have to add signal processing to mirror one of the hdmi signals
Front to back could work just by coincidence. What you want here is rotate the panel by 90 degrees. Better option though is to buy different panel with the same pixel layout that already runs on different polarization. Both vertical and horizontal are in use, just have to check which monitor does what. That happens even between minor versions of the same model. I had two Dell screens that were almost identical, sam layout, same size, resolution, look almost the same, one was just newer year model. They had different polarization , so when I was wearing driving sunglasses, I could see only one of them.
Besides that, there are multiple types of LCD panels. They generally work by rotating polarized light with liquid crystals and blocking unwanted part. Liquid crystals are usually used to change light polarization, not filter it. So they either rotate and block dark pixels or rotate and let through the lit pixels. The one we can see here is the latter type as the output has rotated polarization. You can also buy the former type and those can be simply stacked with no additional effort. That's because the light that should go through is untouched and the light that should be blocked is rotated to be blocked by filter. This would be a perfect screen. IIRC it's more often used by VA or TN panels, I don't remember details.
That ad segment was like watching interdimensional cable.
This is quite similar to something that was done (very successfully) at the computer science lab at my university. Rather than using two LCD panels, they used a DLP projector as the backlight for one panel.
Ahhh, interesting. That seems like it would take up even more space though. Plus probably more work as well.
@@MsHojat Oh definitely! It wasn't practical as a replacement monitor but, by using a DLP projector, you have a very powerful backlight & absolutely zero light in the black areas
Projector stacking. We did that to increase projector lumens. Fancy seeing this used to enhanced darker areas. Cool.
Unlike any other channels, this channel project literally take time, effort and innovation. But that makes it even more special. We're always waiting for the next video to release. Great job as always.
This isn't the only channel on this website whose projects take time, effort, and innovation
If only you'd put a similar time, effort and innovation into your comments instead of generic "praising others at the cost of putting everyone else down" statements
Guys I'm not trying to say its the only channel, or putting everyone else down. Thanks for reminding me. I'm not an english native speaker and I'm still learning english. 🙏
would love for you to have another crack at this! Youre the king of diy solutions and ive never seen you really make something where it wasnt fully complete. I think you should get back in the ring with this one !
One thing you could try to improve brightness and sharpness is removing the polarizers instead of unpolarizing the light. Just doing that should double the brightness, as 50% of unpolarized light is absorbed by a polarizer. The only problem is that it’s difficult to do without breaking the panel, but it’s doable.
isn't the purpose to use the two polarizers to remove the "grey" and turn it black though?
@@danielxmiller hes not talking about the lcd panel, hes talking about the polariser on top of the panels, instead of removing the back one he's added a "depolariser" in the middle.
Does brightness really matter that much ? If you spend
The polarizers will get out fairly easily with just iso alcohol or acetone, the glue isn't strong.
Bro isn't the point of the polarizer to block unwanted lights so you can actually see the image?
As a multimedia computing graduate, this project is simply priceless. Not just the project itself, but rather also understanding of the inner workings of a monitor. Throughout my four years of uni’ I’ve not once been explained that the LCD is taking advantage of polarisation to produce the images, as you’ve explained in very simple terms. I could not thank you enough.
Education these days is basically worthless -- only the most shallow details are covered, and the average person will learn far more if they simply do their own research.
@@hxhdfjifzirstc894 Nope 🤦♂. It's that Education often is severely underfunded and understaffed. Not to mention, if results are not "produced" funding gets cut so the material for general degrees is simplified to ensure a higher pass rate. Unless you choose very specialised subjects - that knowledge is actually needed as a foundational layer.
I find your comment about not being taught how an LCD display actually works as amazing as it is concerning. I am amazed there are people who can some how rationalize that gaffe. Of course my concern is obvious with my only real wonder being, "what the hell else is missing??"
@@hxhdfjifzirstc894 A pretty wild exaggeration.
@@DougDingus Sadly university is more about learning yourself than being taught everything. I'm certain that appropriate courses that cover display technologies can be found, as there are more complex things taught there as well. However, they mostly rely on people actively studying stuff on their own, and the main point is on how to find and process new information from various sources such as books, journals, and other publications both in paper and online.
OF course, it's not perfect either. That's not what I'm saying. I was thoroughly disappointed with uni education at first during my undergrad studies, but I've found things here and there that are worth mentioning and appreciating for.
this result was unexpectedly interesting! I don't see much potential in this application in question, but rather its use in entry-level projectors, as the gain in contrast would allow the use of stronger light sources, bringing more luminosity and quality to these projectors! Send greetings to your Brazilian fans
Congratulations, you are getting more and more creative with every single video you upload. I am satisfied to live in the same decade with you😁
Poor guy's only gonna live for a decade? :(
@@theyruinedyoutubeagain RIP 😥
@@theyruinedyoutubeagain damn
RIP in pieces
RIP, you will be missed🗿
An interesting project, thanks for the video. A couple of notes from a graphics engineer perspective:
- By combining two LCD panels together and sending the same signal to both, you're getting color distortion. Each panel acts as a filter, multiplying the backlight radiance by the desired color: L * c. Two panels, therefore, square the color: L * c * c. That results in dark tones becoming darker, and colors shifting as if gamma correction with a factor of 2.0 was applied. To compensate for that, apply gamma = 0.5 to the output image.
- The rainbow effect that you get with a camera is a Moire pattern, essentially a form of aliasing. When the screen is projected onto the camera sensor, its subpixels have a density comparable to the density of sensor elements, or probably somewhat higher. So, some sensor elements get more of the red subpixels, and some get more of the blue ones, for example. That issue is amplified by the Bayer pattern used on most camera sensors. To get rid of the rainbows, you can try to defocus the camera slightly, which will achieve optical anti-aliasing by removing the high frequencies from the input signal.
This really needs a part two.
Very cool project Matt, even though this one isn't very practical as pulling 250W constantly for a monitor will easily offset just buying a more expensive monitor in the long term it's always fun seeing your ingenuity!
It's obviously winter monitor. During summer, he uses his DIY outdoors waterproof screen 😄
But it cuts your heating bill..
Not just the watts but your GPU would be effectively driving 2 monitors with only getting the resolution of one so WAY more GPU power is needed to. So yeah the cost of an Oled would be way more effective but yes still a interesting video.
@@WDMurphy no the HDMI signal was just split. GPU not doing anything more
@@WDMurphy but gpu is free heater. XD
You just built one of my long standing tech-dream.
Crazy guy you are!
Thanks for tinkering around!
I thoroughly enjoy your channel buddy and have for years. The things you do and make are phenomenal. You're extremely talented. Great work
This guy is great…every time I have that question lingering in my mind as he’s showing it, he immediately give the explanation or reason for how something works for his projects. Like he reads our thoughts. Great informative video. Wish I had the resources and time to make one myself. I’d want to try this on some older cheap lcd tvs
Matt, I absolutely love your content. you're easily my favourite content criator on youtube. thank you for giving us the most professional, entertaining, educational stellar content every single time
It's nice to see a typical diy perk again. I like these relatively simple diy ideas
This just makes me want to find a high refresh rate monitor with a busted backlight and just drop it on top of an overhead projector from the thrift store so I can finally have my high refresh rate projector I've wanted for years now. Seeing you do the disassembly and backlight design was inspiring and empowering! Excellent as always
maybe even mod the overhead with a high CRI COB LED
@@Angel_Bob_ Yeah, definitely go LED for major heat savings ---> longevity.
THIS project? Not the projector one? ruclips.net/video/YfvTjQ9MCwY/видео.html Admittedly that is a lot more effort than "slap monitor on overhead projector". But I can GUARANTEE you the overhead projector route is not great. Noise, light leakage. Aspect ratio. Awkward placement of ohp. Not saying it can't be done, as it has been done to high degree of success since ages ago on various forums. But it is also high effort for a "good solution". And the "low effort" method is very much, "you get what you pay for", so to speak. I suppose you are looking for a specific functionality that is missing in commercial products for a reasonable price, so it might be worth the effort. But if so, I imagine DIY Perk's projector project is also worth the effort. Just a thought.
Congratulations, you created Tandem LCD.
You amazingly and regularly come up with original projects. Keep up the good work!
While the projects and videos are amazing, This particular one is hardly new, Hisense was experimenting with it some time ago before moving on to a different way of doing things
@@Yuki2204
Hisense? Dude, dual layer LCD's been a thing since the 90s, LOL hardly new but what is new is how easy it is to accomplish at home with little effort.
@@SMGJohn fair enough, the Hisense panels are the only time I encountered such a thing so
@@Yuki2204 I've never heard of this before... and I was around in the 90s!
@@Yuki2204
1990 Toshiba had dual layer SuperTwisted Nematic LCD displays which were two layers of TN panels together to increase contrast.
Both LG and Samsung had dual layer prototype LCD panels in early 2000s, Samsung also produced dual layer LCD panels with ability for 3D which were mostly used in slot machines and arcades.
Sony has a long history of dual layer LCD and later dual layer LED panels for their broadcasting monitors which they still produce to this day.
So yes, they did exist, they did cost a lot of money, and no they were not popular, the first prototypes of LCD technology back in 1984 also had prototypes of dual and triple layer panels sandwiched together to create a two colour output, this was before they figured out they could sandwich a colour filter to get colour output from one panel.
You are the modern inventor, DIY Perks. I'm struck by your creative brilliance in every video I watch.
Crazy project! No other channel consistently blows my expectations away with every video lol
OK, firstly, this is amazing, honestly, many kudos. And please make this a series, this can be improved vastly. On the top of my head a low res monochrome panel as the 2nd display or using controlled mini led for backlight, etc there's just too many possibilities. I cannot wait for more versions of this, Thank you.
Bad idea. The second panel must be identical to avoid image degradation with the different characteristics (transition timings, latency, gamma offset, etc.)
For the controlled mini led for backlight, this is basically how the high end HDR LCD display with local dimming zones works. But its very hard to DIY as it needs extremely tight timings to analyse the signal and switch light with the same latency as the LCD panel. And the absence of blooming will vary with the number of zones and the size of the lighten elements on screen.
We have panels like this on a lot of our casino floors! IGT made a unit called the G23MLD (Multi-Layer Display) that has 2 LCDs stuck together to achieve a 3D effect on the slot theme. Most of the time the reels are on the back monitor and the line indicators are the front monitor
As soon as I saw this video I thought the exact same thing. We have them as well but most of ours have developed black spots. So I question the longevity of the monitor, but that one still looks much better then the IGT version ever did.
@@brandonf2562 Have you been making sure the fans are working? One of our older tech was saying that half the screen wouldn't be working soon after the fans died
This was one of my favorite videos from you!
great idea. I am stunned with the result of your experiment. if the problem is the dropshadow-like on letters especially during Ms task only, I suggest to make the lcds additional switch on and off so that if and whenever watching videos or gaming you can enable both lcds to take advantage the viewing performance and just disable the other lcds during light task only like Ms word.
I would love to know the results of a checkerboard contrast test before and after the DIY.
I really LOVE the way you change ORGINAL things into awesome stuff 👍🏻👍🏻👍🏻👍🏻👍🏻
9:09 there's a lot of contrast, sure, but did you adjust gamma to 0.5 to compensate for the multiplying effect? With two displays both displaying 50% gray you actually get 25% gray (0.5 * 0.5 = 0.25). A gamma of 0.5 (effectively sqrt) brings that back up to the desired value.
I cringed a bit at 9:49 when he said that it's "not at all" like increasing the contrast setting, because I'm pretty sure the contrast settings also increase contrast by messing with the gamma (or rather by changing completely the EOTF).
You inspired me to turn two old laptops into a router and the other one is now a media PC thanks!
this gave me a more intricate look at a monitor than I ever thought I'd be seeing. So I thank you for that alone. Gave me a much better idea of how these work.
I suspect that you could get a slightly crisper image by using more uniform commercially available diffusion films and optically transparent epoxy to bond all of the panels and the intermediary diffusion film together. I do appreciate how you went straight to replacing the entire backlight system with something substantially brighter, though doing so without needing to utilize several layers of commercial optical dispersion films with different dispersion patterns before using a standard frosted dispersion film to make it uniform does necessitate a deeper enclosure. I do think this is a topic worth revisiting at some point in the future to further refine the idea. Maybe with a couple of cheap 43" 4K televisions. If you have a decently deep desk, then a 43" television mounted to the wall behind it is a great monitor size for most applications. If possible, it would also be a good idea to just utilize an inexpensive HDMI splitter to duplicate the HDMI signal without relying on your GPU to keep the images perfectly syncronized. Ideally, you would want to find a way to drive both panels from the same driver board so that the signal is sent at the same time and there isn't a risk of clock drift between the driver boards. This would reduce ghosting. Unfortunately, it's unlikely that the driver board can provide enough current on its output pins to handle two panels in parallel.
I've been wanting to build something like this ever since I saw coverage of a panel maker that was looking into something like this with a 4K main panel and a 1080p black and white panel to help control the backlight.
I'm quite sad that it didn't work out, it definitely does seem worth it compared to the cost of higher end alternatives and the superiour amount of "local dimming zones".
I would like to see you test the colour accuracy, as I prefer the image to be more accurate to being overly vibrant.
Hisense had a TV using this tech
Hey man, just want to leave a random comment while it's on my mind.
You're an inspiration! Using this insane technology around us and giving it new life, using some skill and know how to make things that people wish they could just buy. Or the exact crazy device that does what you personally want, how you want it.
Amazing. You're an inventor's muse to be sure.
Super cool project, I always wondered if it was possible to do this. How about removing the polarizing filter of the first panel instead of using tracing paper? I know that people have made those "privacy screens" from old LCD's that are only visible when wearing polarized sunglasses. Or is the film impossible to remove in these modern panels?
Yes. Removing one polarizing should help a lot with the brightness.
@@tristanwegner Problem with removing the polarizing layer is you loose the image entirely. it was something he thought about though.
I think that, for this kind of project, you actually need the polarization filters - due to how LCD displays work.
@@LiteLotus Wouldn't the polarizer on the front screen polarize both screens?
@@newnuu6448 I have the exact same question, leaving a comment in case someone knows more
Pretty neat, and very clever! The thickness bothers me 0%, it's still far lighter and thinner than many computers I've used in my life. But the increase in energy cost over time and the slight fuzziness are issues that i think would steer me away from making one
i wouldn't be surprised monitors are often lighter than computers if we're talking about desktops, but this isn't an AIO if that was the persuasion you caught
I mean the thickness could be fixed with a good diffuser that isn't baking paper would help the power consumption as well given the LEDs dont need to be as bright as before.
I just thinking about that, because I have seen 2-layer OLED screen. and you do this.
GJ!
Have you tried to remove the polarization layer on the first screen, like a lot of projects do to make "secret" screens that only show up when wearing polarizing glasses?
I had the exact same thought!
when you remove the polarizer, the lcd becomes plain white - negating the effect.
@@Elemental-IT Well, one of the displays still has it's polarizer.
@@Elemental-IT You'd have to remove the polarizer of the back layer from the top panel and the front polarizer from the bottom one.
@@Elemental-IT That's the point of the polarising glasses. You take the front polariser off the LCD, so the LCD is plain white and unreadable, yes. Unless you're wearing a pair of polarisers in front of your eyes. It's a silly hack but I suppose fun, I wouldn't do it to a monitor I actually wanted to use.
I came up with this idea myself a couple of years ago when I found out about how LED TVs with array backlights work, glad to see you tried it and that I was onto something!
Lol This is one of those thing I didn't expect to work. Very interesting to see that it did and then all pros and cons with it!
Theon grey finally moved on and is doing some of the most creative stuff...Respect for you Man
his hands are looking better than ever #blessed
I wonder if you could look into multi-layer 3D light field monitors. They use multiple LCD panels at a slight distance. Basically you can calculate how a certain light ray would move through the panels, and color their pixels accordingly. For example in this decade old paper: Tensor Displays: Compressive Light Field Synthesis using Multilayer Displays with Directional Backlighting
Ooooo amazing. I just said in another comment I'd love to see spatiotemporal superresolution as well as using old Sharp Quattro panels (the *panels* WEREN'T snake oil, but the TVs were, because some idiot at Sharp 'cut costs' & they used narrowband tuned RGB backlights instead of RGBY backlighting, effectively making the amazing panel useless), combining that with light field monitors o boot... *chefs kiss*
That is definitely one of the more most compelling uses for doubled-up LCDs. The tried and tested method for angular modulation has been to place a large array of lenslets in front of an LCD to modulate the angles of the pixels. Once you have computed the mapping from pixels in 2D to (hopefully tight & collimated) rays, you can trace rays backwards from the pixels to query their intersection with some simulated object, thus determining which pixels to switch on. This is great, but even with high resolution LCDs this ray optics model ignores the wave properties of light. Ideally, a second LCD in front of the lenslet array would have high transmission rate, and rather than modulate the brightness, modulate the phase of the transmitted light in order to simulate optical wavefronts. That would, in theory, make simulated objects in front of the panel focusable.
A company called Light Field Labs appears to have done something like this using a static phase-only optical layer, but as far as I know nobody has demonstrated switchable phase modulation / dynamic focus using such a large panel.
@@stewartforshaw8558 Sounds like what Lytro was doing with cameras.
your channel really deserves its name!
This is why I usually prefer VA panels, they're much better these days than they used to be, viewing angles aren't really an issue on a good panel, mine has the same viewing angles as the same MSI IPS monitor. I got the curved version and I love the deep blacks and contrast. It's a high refresh rate too, 165hz. I have had bad VA monitors though a while ago but this one I have now is great.
But unfortunately they are still terrible in terms of response times and hence the ghosting, so no go for competitive gaming
@@heartbroken3344 actually not all, I recently tried the flat AOC AG275QXN and it's great in both response times and input lag, the only issues are viewing angles were very bad and the washed out colors probably because it is flat, also bad gamma shift (most noticeable on black color) so I returned it.
The right way to measure the convenience of "dual layerlcd" is to present an image like a chess board, 100% white squares along side 100% black squares. Then you measure the luminance with a luminance meter. A typical white luminance is about 80cd/m2, and a typical black luminance is about 2cd/m2. The contrast ratio is the white luminance/black luminance. A dual layer lcd is better than a regular lcd ONLY if the contrast ratio is greater in tje dual layer than in the regular lcd. You show us a darker black in the dual layer but also a brighter white in the regular lcd. We can say nothing without measuring contrast ratio.
In a typical display the contrast ratio is 40. So the white transmitance is 40 times the black transmitance. Lets say that the white transmitance is 0.4, so the black transmitance is 0.01. With 2 layers the CR=(0.4*0.4)/(0.01*0.01). The CR=40^2=1600. Wow, I was way way way wrong. Sorry by that.
@@guatagel2454 edit your original comment with this. I was about to comment that you were way off
videos like this are the reason i m subbed to you since 3 years now !!
Keep it up brother !!
Couldn’t you have removed the polarizer from one of them?, or am I just stupid? 😭
THATS WHAT I WAS THINKING
I guess he gotta scrap the monitor or split the layers. Manufacturer can do it way easier my making dual layer with a single polarizer only.
10:35 Dual LCD is not super energy consuming when you do it right!
There are some LCD monitors out there that use only 35 Watts for a 4K 144Hz screen.
For sure a manufacturer could make a HDR 1000 dual LCD for slight under 100 Watts.
LCD always was superior than OLED, but for some reason the industry don't want to innovate on dual LCD monitors.
Your example shows how close you can come with just some wet finger work, now put that same work to a nano meter precision.
There is the answer, you can scale your model with an easy 100x in better picture quality while doing the same thing on a nano scale precision.
that is incredibly impressive. well done that man.
After seeing the new iPads, I remembered I've seen it before. I did, from your video.
@DIYPERKS
Hey man I'm a big fan. Seeing you do this is awesome. You should try to get this thing as close to HDR as possible and add the ability to strobe the backlight for blur reduction. Because then that would be the ultimate monitor that you could have. Strobing the backlight would be able to improve motion resolution, and if you tune it right you can get it close to CRT like motion.
One thing you might consider is using lasers and phosphor rather than LEDs for your backlight, because that would use substantially less power.
If you wanted to go really ham with it using an mems system with lasers and phosphor would give you a built-in strobing backlight that would be very efficient and bright. You may even be able to remove the polarizers if you used lasers in a MEMS system because such a system would mimic an 7emissive display more closely.
Peak brightness is already an issue, so strobing the backlight might not be that feasible. I don't have much experience with displays that support that, do you know what's the typical duty cycle on strobed backlights at 60hz?
@@carnivorebear6582 it would depend on what Motion Picture response time you were looking for, as well as the native refresh rate of your panels and what refresh rate you want to strobe them at.
A 240 hz LCD running at 240 FPS and pulsing the backlight with a 50% duty cycle (flash length) will let way more light through then if you just used 60 frames per second with a 50% duty cycle.
A 50% duty cycle usually I believe so it cuts brightness in half, but you could also overvolt the hell out of the LEDs since you are pulsing them as opposed to leaving them on constantly, just like how fald back lights work.
You can also adjust the duty cycle for lower MPRT and higher brightness or vice versa.
You can pulse an LEDt at much higher brightness levels then they are rated for provided that you are only doing it momentarily.
You can pump about five times to 6 times the brightness out of a typical LED for short duration without damaging them. Since he already has the heat sinks, he could get more. Other people have also mentioned that having one layer be monochrome would let more light through.
This is how HDR works on Modern qled displays. Or how lightboost used to work in old 3D monitors. Adjustable pulse length and precise voltage control.
Power and Heat considerations is why I think a laser phosphor based backlight would work way better then using LEDs.
A micro electric mirror that steers a high watt laserbeam at a phosphor layer could get you very high brightness at low power usage, and strobing also improves the native contrast of the panel.
This is what I've been wanting to do for weeks now. It actually started with the tv at home breaking down. I learnt all I could about tv, then I wanted to build one at home.
I have a idea of removing the need of the tracing paper for depolarization. You can flip the inner panel, so the inner panel is facing outside with its original inside side. The back light passing through the inner panel gets converted to the same polarization the outer panel's inner polarizer. Its kind of hard to explain it. The polarizers' configuration of yours is H-V-tracing_paper-H-V-backlight. If the second panel is flipped it becomes H-V-V-H-backlight. Connect both panel to the same graphics card, and in the graphics card driver set the display of second panel to be flipped.
Unfortunately, that wouldn't work. You'd end up with the RGB subpixels being swapped between the front and back lcd panel. While you could potentially get around this by using panels with different subpixel arrangements, text may still look weird due to the subpixel antialiasing used by most text renderers.
@@Lucy-sf6zm You are absolutely right! Totally forgot the rgb sub pixels
You should try this again but using SMD LED diffusers. They're a little plastic dome that just clips on over each LED to get the diffusion effect you are looking for instead of the tracing paper.
Also, polarizing film isn't *that* expensive, changing the in-between layer for a sheet of that will likely give you even better contrast.
This is one of those channels where watching the whole video before the payoff at the end is always worth it.
Matt. Thank you for making some of the most wholesome, calming content on the internet. mere minutes ago i was having a panic attack, and your channel was recommended to me. i clicked on this because it looked awesome. the calming music, your soothing voice and the interesting video idea have brought me back down to earth. Thank you so much.
Happy to hear that it helped! I hope things improve for you.
8:15 You can use a Optical Low Pass Filter to mitigate the Moiré effect.
There was a monitor that did this a while back and used a second black and white display to increase contrast without affecting colour balance. I think ltt did a video on it.
This should improves black, but not the color purity, like with the full dual LCD version here.
Great project! One suggestion for the next project:
A dual lcd display based 4k projector!
Why not flip or rotate the second LCD to match the polarization?
Also if you don't need all the contrast reduce the black-level on the second LCD to get some more light transmission
To pass light you'd have to rotate it 90 degrees which has certain obvious geometric problems... Flipping it or rotating it 180 degrees just brings the polarization back to the same orientation it already is at.
@@tylisirn Flipping should switch polarization. Let's say the first is vertical in, horizontal out. Then the second unflipped is also vertical in, horizontal out. Which mean in the middle you have vertical + horizontal= no light gets through. With a flipped you have vertical, horizontal, horizontal, vertical. Which works
Presumeably, the first display panel polarizes light linearly, which requires a 90 degree orientation for the second panel to allow light through. (Ever had polarizing sunglasses on and turned your head to read an LCD display?) Now, if only the panels could polarize ths light circularly...that's tricky because they would need to be a multiple of 1/4 the wavelength thick. Polarizing filters on auto-focus cameras have to do this.
@@jeffbaker8808 Well a panel has not just one but two polarizers which are shifted 90° too each other. The LCD in between "turns" the light. So it's linear horizontal filter - LCD - linear vertical filter. That's why flipping works
Wouldn't flipping one panel cause issues with the pixel layout? One panel would be R-G-B the other would be B-G-R. Also flipping a panel 180 or 90 degrees causes inherent input delay or extra processing time in windows, pretty sure LTT did a video about this ages ago.
That is pretty cool. I feel like this would be extra nice for text in 4k. I want to go to OLED for my monitor but worried about burn-in. Dual LCD might work, if it was actually available for consumers.
I shouldn't worry about burn in, there are safety measures with any new oled monitor like pixel shifting which you won't even notice.
If you look after your display burn In isn't a problem
@Joao_M hmm yeah makes sense if you use it for office programs. The pixels shift but 99% will shift to the same color
I just found this channel a few days ago... i love it.
My solution for the polarization issue would've been to flip the image around for one of the panels, and literally flip the front panel over so its "incoming" polarizer and the "outgoing" one from the rear panel would match. But to be fair, I'm not sure how you'd make the monitor flip around... could fix it in software if you make both monitors show the same thing but flip one, most OSes allow that, right?
I think the best way to do this mod would be to remove the polarized filter and diffuser from the second display. Not sure how feasible that is without completely destroying it, though.
@@victorbarroscoch Still need to flip it depending on the LCD tech
for that he'd need to plug both hdmi ports to a pc...with this he is using an hdmi splitter that basically duplicates the displays.
It would have been polarised at right angles, so you would have needed to rotate the overlaying panel in the vertical orientation. So, would not be practical
@@Tossphate flip it, not rotate
Super neat. I feel one aspect for improvement is calibration between the position of the two panels to narrow down some of that 'drop shadow' effect.
I'm sure you could probably also get better LEDs for the backlight as well.
Your channel is easily one of the best here
Diy Perks: Tandem LCD
Apple: Tandem OLED
HDMI splitters often cause problems with HDCP, I think it is better to connect both panels directly to the graphics card and clone the output in the driver.
Btw: You can avoid the Moire artifacts in 2 ways. Increase the distance to the LCD and zoom in or tilt the camera by 45° and correct for it when editing the video
probably your best concept so far. with a more efficient scattering layer than the tracing paper i think you could get thhe brightness up and power down and it would be a viable commercial option
One year later, Apple did similar things on their iPad.
no the used dual layer oled displays wich is a completely different technology, plus they did not increase contrast but the brightness so essentially they did the opposite of whad matt did
@@eliasgotzfried1131🤓
Outsourcing at its finest
I'd suggest to just remove the polarizing layer on the second display. Also, instead of projecting the same image on both display, you could try to project a grayscale version on the second layer just to enance the blacks. That would resolve the grey looking black areas without altering the colours accuracy that much.
Exactly. You can just remove the polarization layer. Kinda annoying that that wasn’t even discussed
Just like the action lab
absolutely love your videos and the way you explain things, good job!
There might be a way to run the second screen in greyscale. If you run both screens in duplicate mode from your GPU and set one of them to greyscale then you might be able to recover some of that brightness at the cost of color vividness.
even if the screen runs in grayscale, the LCD itself still has color filters that will block out 2/3rd of the light, he'd have to somehow remove the color filter from the LCD to let in more light. (the colors will probably be more accurate eitherway though)
The problem is light polarity (horizontal or vertical). The LCD twists the light 90 degrees, and polar filters (horizontal and vertical) block the light that doesn't match. Grayscale does nothing to twist the light to align with the polar filters.
Quick update: OLED displays now cost less than 500 bucks on the used market
MSI has a 1440p QD-OLED that’s currently around $500 brand new. Some of the lesser known brands on Amazon have W-OLEDs that dipped below $500.
AOC also has a mini-LED that’s under $300
@NuggetsXInfinite I am seeing lots of OLED monitors for 600 new, but less than 500? that's gotta be a sale
I like the irony!
People who watched the video on a 1080p non-hdr cheap display couldn't appreciate much of the improvement your dual layer LCD provided.
And i personally was amazed... And then i remembered my TV is HDR10+ Dolby Vision...
You could probably get even better results by pairing a single lcd panel with a high quality black and white crt.
Monochrome CRTs can be pretty bright and sharply contrasted so they would make the perfect "pixel adjustable backlight" for the LCD.
You also wouldn't need the diffusion layer in between because the light coming out of the CRT isn't polarized, so the contrast would be even sharper and without blooming.
how would you account for the bulge of the crt? a flat piece of sheet glass in front of the crt?
Really cool project! I use a eizo prominence for my work, and it’s my understanding that it uses dual LCD tech to get the black blacks. It would be really interesting to probe your monitor and see what the gamma curve looks like, from what I see in your video, It looks like it might be a bit steep (your mids look dark). Either way, excellent work!
I have been sat here amazed by the performance of your creation, then I remembered that I have been watching it on my own traditional LCD monitor all along anyway lol. I would love to see something like this in person.
Do you want to send it over for testing?
Yeah