This idea has the 4 displays in a cuboid and 4 right angled prisms do the beam redirecting and merging all at once. To get the bad news out of the way, this cuboid is likely going to be larger than 100x100x60mm and I cant yet think of a solution that looks ok.

The benefits are:

• Uses the least amount of prisms so far. As well as price, this is important for fine tuning as I imagine it takes a bit of fine tuning to get a single prism to line up correctly in all 6 degrees of freedom.
• The cuboid could be detachable when not in use. 

I might as well try some ideas...

hmmm I feel like I'm going to hit doorways with that. And there goes my other idea of having it at the end of an antenna or something.

I mean... I like the top form:

But the front... Maybe if you were going for a more military style it would work

Or maybe double down and add a light?

Yeah like... it looks like it's equipment you'd wear at a job, but it's a lot more like a helmet and less like a hat, like the previous concept:

I guess I could merge the two ideas

Nah, the previous one was better.

Could I Dynamic Island my way out of this?

Uhhhh maybe? Filleted chamfers?

Probably not.

Also no.

I'll probably just go with this. For now, at these early stages of the project, I just need a solution that works and has a non-eyesore looking design.

It still looks like a futuristic tank turret though.

This might help...

Yeah like it looks more like a house cut into a hill with green windows. Combining the two ideas is also an option:

Yeah...? Perhaps...? Maybe...? It's... possible. Maybe fillets are what I need?

Nah the fillets mean that there are too many curves and not enough sharp edges. 

There's also this idea:

Yeah probably not.

This one looks the most like #Tetent Timespy [gd0136] and #Teti, so I'll go with this.

I downloaded a male head STL file and tweaked some dimensions so that it fit as I envisioned. 

I have a feeling that I'm going to encounter geometric intersections between the screens and decollimating optics. As you can see, the eye is quite far from the combiner, and that's why the image needs to be about 55x55mm. 

I plan to have some additional rear head support:

From the side, the concept actually looks suprisingly fine. I've noticed that 2022's releases of AR glasses look sleek right up until you're looking almost 90 degrees at them from the side, so I wanted to employ the same kind of strategy.

Unfortunately, it looks a lot boxier from the front. I only really mind the the top corners of the larger cuboid:

Also note that I'm expecing there to be a few knobs on those cuboids for manual focusing and adjustment.

The design also looks paletable from the back:

It looks a tad like a merge between a hard hat and a face shield. Due to the light from the fresnel being near parallel, I don't expect the image to be visible from this angle either from the fresnel lens or combiner. I do expect that the skin around the eye to be illuminated, so it could look pretty mystical having light seemingy appear from nowhere.

Speaking of mystical, the below image is how I expect the hat to look when mistified:

Same angle but transparency mode:

As expected, my first concept idea rarely looks as good when modelled in CAD. I'm not concerned about the part that goes around my head at the moment. I'm more trying to think of a good design for the top of the cap.

The cap thickness is 25mm and the tallest square is another 50mm ontop. At that point, I might as well mount the screens vertially. At least then I'd also have space for the decollimating optics too.

[05:00]

For a smaller design and lower unique part count, I've decided to use a convex and concave lens system to collimate the light from the display. For some reason, I thought the 55mm square fresnel lens I found on AliExpress had a focal length of 35mm, but it's twice that at 70mm. I also found a 21mm concave lens with a focal length of 17.835mm. The result:

Yeah this was quite a bit of a "snap back to reality" moment. The reduced beam is still much too large to use, and the overall system length of 54mm also needs to be reduced.

I found a 100x100mm fresnel lens with a focal length of only 35mm, so it's possible to obtain a smaller focal length. I'm going to assume I can get a 60x60mm fresnel lens with a 30mm focal length.

Next iteration:

I've changed the cross sectional axis so that the longest length is in view. Reading this webpage on combining lenses, I used the separated lenses equation. I've seen somewhere else that concave lenses have a negative focal length, so -17.835 is what I've used as the second lens.

If the first concave lens is 23.8mm from the fresnel, the focal point for the system is almost 46mm. Then, by placing the second identical concave 17.835mm away from this new focal point, I should get a parallel beam.

As you might be able to notice, a small change in the distance between fresnel1 and concave1 causes a larger change in the distance between concave1 and concave2. It doesn't seem that the resultant image size is affected as much though.

The target was 7mm, but a sub 9mm image should suffice too, requiring a beamsplitter of >=18mm.

It also seems that a assembly size under 30mm is possible. However, a problem arises when it comes time to expand the beam again, since the size would be twice as big and the 21mm lenses are barely enough already.

[some searching later...]

Oh wow:

This is one nice looking lens, and the company even offers a customisation service. Why AliExpress search didn't show this to me sooner, I don't know. 

There's also concave lenses here, and the listing says that the one below has a edge thickness of 5.2mm. The 21mm lens I found earlier only lists the centre thickness, so it's very likely that the above light path wouldn't work when the edge thickness is taken into account.

[About 1 hour later]

These lenses exist. Along with the F30mm fresnel, this is the next iteration of the optical path.

The dual 18mm diameter lenses reduce the 52mm square image down to 3.1mm. Merging 3.1mm beams is also going to be... interesting. After all the mergers, the beam size is doubled to 6.2mm. It then goes through a 3rd 18mm lens and then a larger lens before hitting another 30mm focal length fresnel lens. For the D43mm lens, an image of infinite distance is almost identical in size to the input. For the D40.5mm lens, an image of 5.97m distance is 13% larger.

I think I'm going to go with the D43 lens. Abberations are still a question mark, but hopefully they're not too bad considering the input image is approximately equal to the output image in terms of size and distance. 

It looks like it's about £5 per lens, so that's 5 * (4 * 3 + 2 * 3) = 5 * 6 * 3 = £90 in optics. Not great, but looking how much some singular sets of glasses cost, not terrible. If I use the 120Hz display driver board on AliExpress, the combined cost would lie somewhere around £400, depending on if there's a sale or not. This makes me think that the full BOM could be £599.

[5:20] There's a cheaper solution than using the D43 lens:

Image size: 52.75mm

[13:45] Turns out a fresnel lens focal point of 30mm was a good starting value, as 28 or 32 gives me an image way too large or small. Looking at IPD charts on the internet, I think targetting a 55mm IPD instead of 60 would be more inclusive. I'll see if I can get something like a 55x60mm or 55x65mm F30 fresnel lens so that I can hold the lens on the 2 shorter faces.

The following doodle is going to look really rough, but I finally figured out a solution that had an even number of reflections and the merging prisms were doing the merge on either the horizontal or vertical axis.

Red arrows and surfaces mean it's an inverted image or takes in an inverted image. The 180 degree on the first blue arrow is what happened to the image after going through the collimating system. Most right angle prisms have been drawn as planes, but the first blue mirror is actually a merging prism.

In this case, the bottom green panel (green beam) reflects on a standard right angle prism before going through the merger prism. The top green panel (the one with the orange bar) (blue beam) reflects on the merger prism.

This merger prism is different from the one last log to keep the amount of custom parts to a minimum. It should now be the only bespoke component.

I'm planning to use a 15mm beamsplitter cube, so the actual merger prism would likely be something like:

It doesn't really matter how long that 5mm dimension is. It just has to be thick enough to not break. It could be increased to 7.5mm so that the dimensions are that of a 15mm cube with a trapezoid cut out. The real question is if I can use 15mm or if inaccuracies would mean it would be better to go with 18 or 20mm prisms.

All right. i've done some more research, and I just got the 2 way mirror in the mail today. Let me talk about that first.

2 Way Mirror

This thing is actually really cool. I've only ever seen it in the rare supermarket store until now.

• As expected, 
  • I can use the 2 way mirror to also dim the outside world.
    • It's actually an ideall amount of dimming and it means that I can just get the standard, clear PLDC film. From the outside, it'll look like the visor went from a mirror finish to a block of snow or something I'd imagine. This fits in with the white main colour scheme seen in the majority of my projects.
  • Because of the reflective layer, I can't see any ghosting (double image) at all
• A nice suprise
  • I'd say that the percieved brightness from the reflection is like 85% or 90% of the source light.

The 1440x1440 Quad

The reflectance leads well into my next point, which is the computational issue I'm having with image stitching and enough light reaching the eye. 

My google search generally suggests that a screen in a cinema is 50 nits of brightness. Now, that's in a completely dark room and you can't see behind the screen, so it's not like the screen light needs to battle against ambient. The PLDC film, 2 way mirror film and the head cap design should help the displays power over the ambient light from outside, but I'm thinking that another 90x requirement should be "90 nits to eye brightness".

Now that I've seen the mirror, it sounds like [90 at eye] <- [100 after beam splitter] <-[200 before beamsplitter] is the minimum without significant light losses.

Non polarised beam splitters have significant light losses. The final beam splitter doesn't count, as the light is being used by the other eye, but beamsplitters where the light gets reflected into the abyss is a waste. The reason why I'm talking about beamsplitters is because that's one of a few ways I can stitch the 4 panels together. I'll get to that in a moment.

Team CuttingEdge

It took quite a while to find out what this is called, but what I'm desiring is a knife-edged mirror prism:

I'm wondering if there's some kind of mirrored cube that does this for cheaper, but this prism is £107 each from a site called ThorLabs. There's also slightly cheaper "pickoff" mirrors from Edmund (£62) that could also be used for a solution.

Team Prism

I assume that a Total Internal Reflection (TIR) 90 degree prism from AliExpress doesn't have ultra precision edges, because I could do this:

It might be more applicable to offset them so that I only have to get 1 beam at the edge:

I believe the prisms like the ones seen below would work nicely, but I haven't seen them around AliExpress yet. The bottom one is just 2 seperate prisms of the top one.

Team Beamsplitter

Now, I forgot up until writing this log that the 1440p screens will be focused to half the XY angle as the other, single screen solutions, which could make this viable. Nits = candela per sq m, and so if I apply a scale factor of 0.5 to the screen and the lumens of the screen stays constant, the candela per sq metre increases 4 fold. In other words, a 300nit screen should become at 1200nit screen if I understand things correctly. 

There's 2 versions of the 1440p 120Hz screen: a 150nit, 11% duty panel and a 300nit 100% (assumed) always on panel. Looking at the UFO test with black frame insertion to simulate lower duty displays, the lower the duty % the better. However, the 300nit has 1100:1 contrast instead of 600:1, and it's the one that comes with the 120Hz board so maybe there isnt even a choice in the matter.

I need 2 beamsplitters to merge all 4 screens:

The issue is that with a non polarisation beamsplitter, I lose 50% of the light. Half the blue light goes straight through, and half the green light reflects.

Team PolarisationSpinner

Now if I really want to get fancy, I could use a polarising beamsplitter. The question then becomes "How do I rotate the polerisation of the beams 90 degrees?". I haven't yet found anything that I can buy that can do the job, so the solution may have to come from software and mounting 2 of the screens at the required rotation:

Actually, now that I think about it, Teti would still only see a 2880x1440 monitor for the green section, so I might have to also split the display in software:

This would only work once though, now that I think about it. A standard beamsplitter would be needed to merge each colour.

Team Edgecase

I can simply just not strive for an edgeless seam and maybe have some kind of fly-eye-esque, 4 square panel thing. Or it would be the intended idea, but it would look like a standard video wall with virtual + bezel.

The other screens

Testing has shown that anything under 2560px isn't going to be enough pixels to feel like an enjoyable experience. Then, when the FOV is taken into account, it really does seem like 2880 and up is desirable, as 80ppd is as low as I'm willing to go.

The miniLED display driver is still being developed by the Alibaba company, and considering that even with the standard beamsplitter option there should be about 120 nits of preceived brightness, so it seems that I've eliminated all but 1 option.

Next steps

Conviniently enough, I've also finally computed an aesthetic solution for the quad 1440p in the event that I went ahead with the idea. It's loosely inspired by these hexagonal stones, but my design uses squares.

I'll look a bit more to see if there are any prisms like the bottom ones in Team Prism, but I'm also concerned about minimum brightness. I only use the max 600 nits of my smartphone outside under the sun, and there's no plug sockets for #Teti [gd0022] out there. I've never even seen 100% brightness on my 400nit 15.6" display. It might be a waste of optical power unless I can also get some DC dimming on the LED backlights, but the standard beamsplitter sounds like a way to get a certified CanDoTheTaskApp. I can think about CanDoTheTaskBetter some other time. 

Then again, I have heard gaming in HDR looks amazing. I don't think I'd have the contrast ratio for it though.

I've been doing some more reading on aspherical lenses (such as here, here and here), and I'd still like to give standard ones a go first, at least for the large lenses. The reason is because of the high quality result DIY Perks achieved in one of his camera builds. He explains the use of fresnel lenses here, and I'd assume due to the size, they're the standard kind.

In the above animated diagram, there's the lens at the left, the (lime green) fresnel lenses and a diffusing sheet in the middle. That sheet takes the parallel rays from the first fresnel lens and rescatters them. In other words, an image is formed on it. In my application, this would be the display panel. 

So that failing computation was being mined in the back of my mind for a few days. Yesterday, I found this video:

(Favourite lines: What do you do if you wanted to view a tree on land? It can't be inverted! So there's a problem!)

The notable slide I want to talk about is this:

It was an eye opening moment, realising that after the focus point, the rays once again diverge and can be focused again as if the actual object was at that point. The image is just rotated 180 degrees.

So the actual light is likely doing something like this:

• Pretend that the dotted lines aren't there. 
• The arrows is to show if the image is 0 or 180degrees. 
• The yellow lines are the light rays from the lens. 
• They hit the fresnel and become parallel. 
• Those then hit the difuser and light goes in all directions, originating from the difuser. 
• The orange line from the diffuser are the rays that are still parallel to the optical axis. 
• The second fresnel converges those rays and the image flips again before hitting the camera lens, or in my case, my eye.

Now, a diverging lens were the first shown in the binocular video. I think the reason for a converging lens setup is because those seem to be more common. I haven't come across a diverging fresnel yet.

Optical Diagram

I found out from the OpenAR project that "first surface mirrors" are the kind of mirrors that don't have glass infront of it.

With the combiner, there are 2 reflections and a 180 degree rotation. Since the reflection axes are different, the reflection component is also 180 degrees, cancelling the rotation out. Reflection multiples of 2 also cancel out.

The nice thing about this is that the screen area can be a square prism, like half of DIY Perks' camera, which could help solve an aesthetic solution for Quad 1440px. The image can be compressed to a smaller area, meaning that smaller lenses, mirrors and beamsplitter cube can be used.

[04:45]

Some more quick updates from research. This was going to just be a comment in the previous log, but I thougt that I could use the better formatting options of a full log:

• I've got an idea for a design that would 
  • 1) fit with the new cuboid design language, 
  • 2) will fit the 2.9 or 3.2 inch screens whilst hopefully not looking unreasonably bulky (a.k.a like it's part of the design and not just a hilarious looking block ontop of the hat).
  • 3) That an even number of reflections are taken from the screen to the eyes. I looked into it, and flipping the images sent to the screen (e.g. for a teloprompter) isn't a standard GPU feature.
  • The design is inspired from the phone seen below as a texture to a square prizm going across the top (and slightly angled). I'll make a render eventually, but first I need to have some idea of optical requirements before I start.

• I've found out that aspherical lenses don't have strange optical issues like chromatic abberation (colour splitting) at the edges like the simple standard spherical ones do. Issue is, aspherical's are much less easy to find.
• Images focused at infinity do not produce an image. It will have to be refocused after the beam splitter.
• So far, the only aspherical (fresnel) lenses I've found (about 10 mins before starting this log) are these ones:
  • https://www.edmundoptics.co.uk/p/23quot-x-23quot-13quot-fl-aspheric-fresnel-lens/2429/
  • All the fresnels in the range are either £27 or £29, regardless of size.
  • When I wake up again, I'll start scanning their solid lens range, as those may offer a clearer image than fresnel.
• Due to the size of the 1440p panels, and their need for at least 1 lens to collimate (make parallel) the light for each, it's likely something that I'll skip on persuing. The result is some endgame specs, but it sure looks like I need to be pulling off some engame plays to get such a solution.
• 2160 is still likely to be a consideration because everything seems straightforward for it. It's also a similar size to the 2880, so it's very well possible that I could design for the 2880 in mind and then have 2 SKUs of TyMist. 
  • I might feel that 1620x2160 is a tablet being aircast to a projector, but that's still more pixel work area than the many 1080p laptops being sold today.
  • The seller of the 2880 said that the board is still under development though, so I'll have to see if that ever comes out if I ever get to the point where I could buy and install it.
• I wouldn't say that the 2560 is out, but I do like the sound of upgrade paths (see #SecSavr Suspense [gd0105]) 

[02:30]

I've been thinking over my options all day, and even scanned over Panelook to see if there were others I hadn't considered. I'd like to think of something else like the PCB of #Tetrinsic [gd0041], but it seems I sold the solution too well to myself and now my subconsious won't let it go.

Anyway, the screen options:

• Single 2560 Si OLED 
  • https://www.alibaba.com/product-detail/SeeYA-1-03-inch-2560-2560_1600425378462.html
  • https://mansky.co.uk/assets/uploads/pdfs/seeya_SY103WAM01_specification_v1-0_20210415.pdf
  • Leaves the option to upgrade to a second screen for AR or the next option below.
  • 2560 square could also fit dual FHD desktops with a bit of padding around them.
  • The visual area of content will not change for 16:9 content. I'm both limited by comfortable ppd, which I believe is between 80 and 90, closer to 90, and the area I can look at without subconsiously moving my head, since the entire view will move along with.
  • I imagine that the hat would look like it has a somewhat fat antenna on one side.
  • Framerate is less than 90Hz (1920x1920 is 90Hz) according to seller. I think it's 75Hz but would have to confirm.
  • I tried 48Hz -> 60Hz to get the same kind of expected feel as 60 -> 75Hz and it's still a welcome framerate bump.
  • It's small, it's OLED and it's very bright. I can use a beamsplitter cube and I'd still have a bright image to be sent to the rest of the optics.
  • So far, the compute believes that using this would have the lowest decimal odds of success. 
  • (decimal odds: 1 = 100% chance, 100 = 1% chance, lower is better)
  • Assuming that they'd actually courier it over, it'll cost $320 (£280).
• Dual 2560 for a 4550 * 2560 desktop
  • Highest resolution and most expensive option.
  • Only does 60Hz according to seller.
  • Most versatile. Since the ppd and viewing distance is kept constant, the only thing that can change with a lower resolution is a lower image size. It also means that there's enough pixels to potentially do some head movement corrections
  • The icons for other screens will be one the left / right sides of the destop area.
  • Requires dual antenna-looking things.
• Dual 1440 LCD for a 2560 * 1440 desktop: https://www.aliexpress.com/item/33030720934.html
  • Similar to the above but uses the 120Hz 1440px square LCD panels. It's seemingly the only 120Hz panel available at a usable size.
  • Unlike the 2160 edition, the 1440 ones sport a 300 nit brightness, so a beamsplitter will cut that into 150 nit images sent to the rest of the optics.
  • It's on AliExpress for under £180 (2 screens + 120Hz controller).
  • The icons on the left and right would be 128px, which should be servicable but can't be anything fancier like a preview of the window.
  • The desktop will be the same size as a 15.6" screen at a distance of 70cm. Whilst mentally simulating what it'll be like, I noticed that 1) even at distance, it's kind of small and 2) 16:9 is suprisingly wide.
  • To further talk about that second point, whilst it's fine for watching videos, I think I really want more vertical pixels and portrait mode. I think the main reason why I don't usually use portrait mode is because:
    • It's physically cumbersome to mount my current screen in portrait.
    • 9:16 looks team extreme for an aspect ratio. Now I'm starting to realise why I liked the look of Oppo's folding phone.
    • Websites don't agree with the low resolution caused by portrait mode.
  • There's so many situations though where the aspect ratio in landscape is nothing but wasted space. Exhibit A:

  • I did some scrolling on Aliexpress on my 60Hz monitor and 120Hz phone, and yeah that would be a very nice bump to go with 120. Considering that I may be able to use my PC much faster with #Tetent [gd0090], it might even be necessary. I haven't yet figured out a way to compare 90 and 120Hz on mobile yet though.
  • Unlike the OLEDs, the 2 "antennas" is likely to just look like 2 cubes on top. 
  • I'm not sure if viewing angles of +/- 80 XY means that there's less light pointing at 0 degrees than the OLEDs which are +/- 30 degrees, but I may also need to collect light over an even wider area than the 52 * 52mm active area. The image will then need to be shrunk to 30mm or less before going into the beamsplitters.
• 2160 LCD
  • This is the current proposed solution, but since there isn't a board that splits the MIPI output to 2 displays, the only solutions are to cough out my own board or buy 2 control boards. I'm not sure why, but the controller board looks the same as the ones used for other screens but seems to cost £50 more when the respective screen's price has been deducted. 
  • Thus, the price is something like £210. I'm going to have to ask about that controller price discrepancy.
  • The visual image would only be the height of the body of my 15.6" monitor, which is 223mm, if viewed from 70cm away.
    • (Mmmm 3 different units in the same sentence)
  • Thus, 16:9 videos would look even smaller than they do now. 
  • 100 nits makes this the dimmest option out of them all
• 2880 MiniLED: https://www.alibaba.com/product-detail/3-2inch-2880RGB-2880-LCD-Modul_1600539433491.html
  • Another Alibaba exclusive that runs at up to 90Hz.
  • Only has 190 nits of brightness. It might be possible to beamsplit it for 95 nits.
  • Currently seems to be £170, driver not included.
  • Would have the largest FOV. 
  • Has the exact same amount of pixels as UHD.
  • (added [17:45]) It's a Mini LED panel with a focused 20 degree viewing angle. I haven't seen a driver board for it though, so this panel might not even be an option.
• Quad 1440 for 2880 * 2880 (added [13:45])
  • The ultimate in "Could it be done??" technology.
  • Likely would require 2 cables.
  • As nits = candelas per square metre, preceived brightness should be higher, as the images of individual panels would be focused to a smaller area. 
  • Fitting so many screens (and related optics and electronics) might be an aesthetic challenge.
  • Blacks not as deep as OLED, and I wouldn't be suprised if colour accuracy is lower too. 
  • Obtains a larger image and framerate than the 2560 OLED.
  • Costs £340

My initial thought is to make the CAD design somewhat option agnostic, so that if I initially planned to use one and then decided to use something different instead, I haven't hard-coded myself hours of rework. I'm also thinking of eliminating the 2:1 ratio options in favour of being able to switch from landscape to portrait desktops.

I have also finally found a video that has the black PLDC film option against light:

I've been worried for a few days now that the film was going to be too dark, but this seems to be fine. I also looked around for electrochromic glass that (eventually) tints, but I can't find any purchase locations: