Avegant Glyph:KickStarter Success Triple funded on day2

[Dilip]

A good HMD is on horizon some time back it was posted with WEIRD thread title like "What the WHAT?"

All things seem promising apart from their idea of adding HEADPHONES no actually they are building their HMD inside HEADPHONES
THats weird then What the WHAT

I wish they just build HMD reduce those 100 USd from price and put simple universal stero socket to attach 3.5mm universal audio jack
as most of us now have their own chosen high fidelity headphones. even sony has learned this bitter truth i can't get why Avegant is not getting instead strip them from HMD cost.

They still require awesome Head Tracker and a nice PC suit + they got to reduce that 8Feet Projection distance to compete with OCULUS RIFT though.

Fresh Review here.
http://www.theverge.com/2013/12/18/5222 ... ty-headset

Own web page
http://www.avegant.com

Lets see how good they improve
I wish oculus won't make too late to come that people got attracted to elsewhere.

[Dilip]

Tight lip-lock here,all pretend they don't want to see...

May be this the one who has high potential threat to every1's favorite...not so near but really???

[unsilentwill]

The Rift crowd has moved on a bit. As far as Avegant Glyph (say it out loud slowly) I don't think it's a competitor for VR, a huge threat to Sony's personal viewer though. It's concerning how they talk about VR, and all the support for consoles and just about any game right out of the box, hopefully people will see for themselves that it's not even close to a full VR game.

[Dilip]

Of course not direct completion to oculus till they eliminate that 8 feet projection distance and small 45' FoV

Their strong points will be unique to them, as they don’t use display.

[Fredz]

Of course not direct completion to oculus till they eliminate that 8 feet projection distance and small 45' FoV

I wouldn't hold my breath for any meaningful FOV upgrade, NVIS could only achieve a 80° FOV when using microdisplays, and with a very complicated optical layout. And they aren't exactly newcomers to VR, contrary to Avegant.

NVIS relay starting design.png

Their strong points will be unique to them, as they don’t use display.

Doesn't make a difference whether the lighting comes from reflection (DLP) or is behind the display (LCD). LCoS is also a reflective technology like DLP and has already been using in HMDs as well (in the SMD ST1080 for example).

[geekmaster]

Their strong points will be unique to them, as they don’t use display.

Doesn't make a difference whether the lighting comes from reflection (DLP) or is behind the display (LCD). LCoS is also a reflective technology like DLP and has already been using in HMDs as well (in the SMD ST1080 for example).

The HeadPlay HMD also used LCoS. And Avegant uses lenses for focusing as well. With lenses, any transmitted or reflected light source can be projected onto a surface (such as your retina). Even the Rift DK1 does that...

Avegant is just trying to set themselves apart with terminology obfuscation, IMHO...

[Fredz]

Yes exactly my thought. For me its only interesting point is that it's cheaper than the SMD ST1080 and the Sony HMZ-T1/2/3, nothing more. The SMD ST1080 is even higher resolution (1920x1080 per eye).

[Dilip]

The HeadPlay HMD also used LCoS. And Avegant uses lenses for focusing as well. With lenses, any transmitted or reflected light source can be projected onto a surface (such as your retina). Even the Rift DK1 does that...

Avegant is just trying to set themselves apart with terminology obfuscation, IMHO...

What you say about HeadPlay HMD is True, but Avegant are first to implement DLP chip Which when compared to LCoS has..

1) Better brightness then LCoS.
2) Nil Motion Blur. (Limitation of LCD even LCoS as it also involves liquid crystal on silicon)
3) LCoS is more vulnerable to dead pixels due to its core LCD structure.

Besides, DLP can also support 120Hz output so no judder or motion blur.

They are bringing this package at lower price point and thus accessible to more people, as i see is also one achievement.

Limitations of LCD. (that DK1 Deploys)

1) Dead pixels are common, They tend to be there.
2) Contrast ratios are lower than those with DLP Chip.
3) Shadow detail and absolute black levels fall short then with DLP Chip.
4) LCD & OLED panels are organic and lose image quality over time.
5) It needs 8K resolution to eliminate screen door at current magnification, which is not here. if you even get processing power is not yet there.
6) LCD has pixel switching time higher then what DLP Chip can offer.

Though DK1 has larger FOV and Better Tracker, it also induces blurring, nausea and motion sickness and let’s not forget SCREEN DOOR EFFECT.

DLP is different and fresh approach, which may bring some better outcome if it improves in right direction.

[Dilip]

something gives me strong vibes, Avegant people are reading this blog
too as i seen them mentioning 120Hz which was not there, till now in any articles....ha ha

The current prototypes have proved successful in head-tracking testing, so there's potential for immersive gaming, and reported system benefits include high resolution 2D or 3D viewing, absence of a screen-door effect and low latency.

"The image is very clear and refreshes at up to 120 Hz with minimal motion blur, "Banks tells Gizmag.

see the full article here...
http://www.gizmag.com/avegant-glyph/30233/

[geekmaster]

something gives me strong vibes, Avegant people are reading this blog ...

They do read and post at the OculusVR forums:

Hey guys. I work for Avegant. ... The VRD that we've developed has a lower field of view because we want it to be useful for non-gaming applications (direct-streaming movies, desktop apps, etc) as well as games. ... Anyway, expanding the FOV is definitely on our radar. We've gotten a lot of good feedback from gamers on what their VR needs are. We just have to move carefully...

Notice in the photos and videos at the link that it has adjustable focus lenses.

ANY display with a lens "projects" an image on the retina. You can test this with the Rift in a dark room. Cover one lens, and hold it near a wall or sheet of paper. The uncovered lens will project a focused image at some point (but not all in focus at the same time due to a flat wall).

In fact, people used to build a "poor man's" projector with a CRT and a lens, mounted with a shroud to prevent stray light emission. It needed a dark room and a good projection surface to be useful.

Technically, your eyes already have lenses, so everything you see is "projected" onto your retina.

The whole "Virtual Retinal HMD" thing here is mostly hype, but it sure makes an impression on those who do not consider it relative to existing common technology (including the Oculus Rift, and any other HMD for that matter).

Realistically, the most impressive thing about Avegant's approach here is the hype traction it gets from its choice of buzzwords...

It's true that any display technology will "project" an image, but not like the micromirror array that we're using. The tech and the optics that we have are what produce the high quality image.

So it is likely they are following relevant posts here too.

@geekmaster: Ha! Yes! I remember building one of those poor-man projectors years ago. Had to tape brown-paper bags over the windows to block out all light. And it barely worked, but it was still tons of fun. Good times.

[Dilip]

developer.oculusvr.com, a more gem place for VR and in general HMD tech 'Know-How', Thanks for pointing out. Soon will join it too..

I have also seen brantlew's review of ST1080. so Micro Displays have limitation of FoV and Projection Size other then blurriness around corner issues. So how they are claiming researching to expand FoV?

Anyway I still think there are few merits of DLP chip(Which are of the Chip, Not necessarily their implementation) which are better to LCoS in some aspects. best i be waiting for CES Reviews of GLYPH

Again I wish I could see this in action, but sadly I am on other part of glob. I like to read your Glyph review too, if you could see this thing at CES(here I am assuming you will visit them at CES )

[Fredz]

1) Better brightness then LCoS.

Brightness in near-eye displays is very different from brightness in projectors. And concerning contrast ratio, DLP is worse than other projection technologies.

From : DLP vs LCD vs LCoS: Projector tech pros and cons

Contrast ratio :
Winner: LCoS
Loser: DLP
Runner-up: LCD

"LCDs have come a long way in recent years, and some now offer some excellent performance in their own right. DLP, however, has not. Its native contrast ratio has changed little in the past few years, and has fallen behind LCD and LCoS."

2) Nil Motion Blur. (Limitation of LCD even LCoS as it also involves liquid crystal on silicon)

Motion blur is present on DLP as well, and it's neither a limitation of LCD or LCoS. LightBoost is precisely a type of LCD monitor with no motion blur :
http://www.blurbusters.com/zero-motion- ... boost-faq/

"LightBoost is a strobe backlight invented by nVidia for 3D Vision to serve three purposes simultaneously: 3. Eliminate motion blur."

3) LCoS is more vulnerable to dead pixels due to its core LCD structure.

If dead pixels were a problem the overwhelming majority of monitors, TVs, laptops, tablets and smartphones wouldn't be using LCD technology.

Besides, DLP can also support 120Hz output so no judder or motion blur.

120Hz doesn't mean no judder or motion blur. Read this article from Michael Abrash about judder :
http://blogs.valvesoftware.com/abrash/d ... ng-judder/

"I can tell you, though, that 100 Hz is nowhere near enough. 200 Hz would be a significant improvement but still not enough"

[Dilip]

I Do not Say "Oculus Sucks Avegant Wins'
And definitely not am saying 'Oculus invest in DLP,instead of LCD/LED'

That said,now..
http://reviews.cnet.com/8301-33199_7-57 ... -and-cons/


Same Link says

Motion blur
Winner: DLP
Loser: LCD and LCoS

Light output (brightness)
Winner: LCD and DLP
Loser: LCoS

Dead pixels do matter when display is magnified 10/20 times. i have seen reveiws people mentioning dead pixel/always green/lit pixles for many HMD including VUZIX , Emagin Z800 and RIFT Too.

Anyway, i just wrote that to convey Geekmaster that DLP has few merits above LCoS.Specially when compared to HMD he was talking about "Head play" which was 2006 conceived and released in 2007/08.
http://www.thinkartificial.org/machine- ... d-display/
using single LCoS for both eyes and having rainbow ghosts. i don't know about Headplay but some articles and WIKI says rainbow ghost occur due to mechanical RGB wheel between light source and display which in end makes picture and that color ghost effect has been almost nullified with PHLAT LEDs having high speed and hi refresh rate

This thinking of mine grew stronger when i read brantlew's review of ST1080 which is 2012 product. in that review he talked about many things like blur at upper edge of picture, cylinder effect in screen and blurry corner edges. But nothing such as rainbow ghost trails so this thing might have been taken care of in evolution of technology. Even other article also did not mentioned it
http://eyegadgets.com/silicon-micro-dis ... 80-review/

Stragely though when Geekmaster pinched Avegant (with words ) with his question regarding colorghost they did not responded in exact answer. i too did asked them pointing out the GIZMAG review where a quote from Avegant insider talked about color ghosting

http://www.gizmag.com/avegant-glyph/30233/

but strangely, no reply,.......Of course it was expected.

ST1080 also uses LCoS and its appears from reviews to be free from rainbow ghosts or atlest must not be that prominent.

Considering all other benefits i think it has good value as HMD not necessarily VR HMD but can dent in to funds people kept reserved for RIFT due to fact that it does not require tailor-made games.

Avegant doesn't seem to be standing by their product and they do not focus on PC Gaming. using almost indecent hardware to display capability of their product like MC book air and PS3 will not put them PC Stereo 3D realms anyway.

Those of you got lucky to see GLYPH in action at CES.please Do post here about Colour ghosts you notice in demonstration at CES.

[Fredz]

Motion blur
Winner: DLP
Loser: LCD and LCoS

As I said motion blur can be eliminated with strobed backlights in LCD. I guess strobing can also be used to eliminate motion blur in DLP, so I don't see this as a selling point for this technology in HMDs.

Light output (brightness)
Winner: LCD and DLP
Loser: LCoS

As I also said, brightness in a HMD is very different from brightness in a TV/monitor/projector. The light has a very short travel path from the display to the eye in near-eye displays compared to current displays, so I don't think higher brightness will offer any advantage. Brightness in the Oculus Rift devkit is surely high enough although it's using LCD.

Dead pixels do matter when display is magnified 10/20 times. i have seen reveiws people mentioning dead pixel/always green/lit pixles for many HMD including VUZIX , Emagin Z800 and RIFT Too.

Sure, but that's an epiphenomenon and basically a moot point. DMD chips used in DLP projectors can also have stuck mirrors or other kind of failures.

some articles and WIKI says rainbow ghost occur due to mechanical RGB wheel between light source and display which in end makes picture and that color ghost effect has been almost nullified with PHLAT LEDs having high speed and hi refresh rate

I think this problem always happen when the RGB color components are not updated at the same time as is the case in the Avegant Glyph. The effect can be lowered in projectors, but I think it's a lot more prevalent in HMDs because of eye saccades over a wide field of view. The only way to get rid of that would be to use three DMD chip per eye, not really practicable.

This thinking of mine grew stronger when i read brantlew's review of ST1080 which is 2012 product. in that review he talked about many things like blur at upper edge of picture, cylinder effect in screen and blurry corner edges. But nothing such as rainbow ghost trails so this thing might have been taken care of in evolution of technology. Even other article also did not mentioned it

I think it's just because it's less apparent with a small FOV (~40° horizontal FOV for the ST1080) just like with projectors in home cinema conditions (~35° horizontal FOV). But a review of the Avegant Glyph showed that there was color fringing :
http://www.roadtovr.com/avegant-hmd-vir ... iew-video/

And Michael Abrash clearly explains why in his blog :
http://blogs.valvesoftware.com/abrash/l ... ar-and-vr/
http://blogs.valvesoftware.com/abrash/w ... our-brain/

Anyway I don't dismiss DLP as a technology, I even own a DLP projector. But for now the drawbacks (low FOV, bad contrast, rainbow effect, color fringing, cost) seem to clearly outweigh the benefits (better resolution, better brightness, reduced motion blur, small form factor, less screen door).

[Dilip]

I think it's just because it's less apparent with a small FOV (~40° horizontal FOV for the ST1080) just like with projectors in home cinema conditions (~35° horizontal FOV). But a review of the Avegant Glyph showed that there was color fringing :
http://www.roadtovr.com/avegant-hmd-vir ... iew-video/

And Michael Abrash clearly explains why in his blog :
http://blogs.valvesoftware.com/abrash/l ... ar-and-vr/
http://blogs.valvesoftware.com/abrash/w ... our-brain/

Anyway I don't dismiss DLP as a technology, I even own a DLP projector. But for now the drawbacks (low FOV, bad contrast, rainbow effect, color fringing, cost) seem to clearly outweigh the benefits (better resolution, better brightness, reduced motion blur, small form factor, less screen door).

Thanks for links they were quite an informative.

Won't increasing color cycling speed help? what if they employ 3 different light source on single board controlled by a chip rather then using single lamp cycle RGB colors.

something likeCBM-290-RGB
http://luminusdevice.weebly.com/mix-leds.html
they need a custom built though, as this one might get burning hot to burn the eyes

Honestly i think using single diode and cycle through RGB can be slower than using 3 diodes alternating R/G/B with chip controlled period that can be synchronised with DLP chip's refresh rate. using 3 source can be faster to minimize that fringing.

This is just an idea, don't know in-depth but yes here are many illuminaties to help that’s best thing.

[MSat]

You can indeed modulate all 3 colors at the same time, but you would probably need to implement custom driving logic, since I don't believe any commercial product does this with DLP.

The two biggest issues I see with DLP is cost and packaging. AFAIK, even the super high resolution cellphone displays are cheaper than the DMDs alone, and you still need a lot more components to make the DMD useful unlike LCDs which have a much higher level of integration. Having to use two DMDs would destroy any potential value (if it even had any in the first place). Now incorporating the numerous optical elements (which further erode value) have to be implemented in a way that makes the package a reasonable size too. There's really no way to beat the simplicity inherent in something like the Rift's design.

Until we see mass production of wafer-level LED displays, I don't think anything will be able to beat the price/performance ratio of standard LCDs/OLEDs.

[Fredz]

Won't increasing color cycling speed help?

Probably, but it'll still be sequential and I guess there must be a minimum duration for each flash of color to maintain persistence of vision.

Considering the human eye can move at 100°/s in smooth pursuit movement and has a 900°/s peak angular speed I think even 1ms per color wouldn't prevent color fringing.

That may not be a problem for smooth pursuit (0.1°/0.5 pixel error) but it would most probably be a problem with fast saccades (0.9°/5 pixels error), considering the 115° horizontal FOV and 640 pixels width of the Oculus Rift. For 1280 horizontal pixels like in the Avegant Glyph for the same FOV the error would be 1 pixel in smooth pursuit and 10 pixels in fast saccades.

For the current FOV of the Avegant Glyph it's 3.2 pixels and 28 pixels resp., which is IMO the reason why color fringing has been apparent when playing a game in the Road To VR review, since it's more likely to require fast eye saccades than a movie.

what if they employ 3 different light source on single board controlled by a chip rather then using single lamp cycle RGB colors.

That won't change anything I think, color will still be sequential. The only way to not have a color sequence is to use 3 DLP chips per eye and one colored LED for each.

[Fredz]

The two biggest issues I see with DLP is cost and packaging. AFAIK, even the super high resolution cellphone displays are cheaper than the DMDs alone

You can find 1080p chips for less than $200 on Amazon or ebay : http://www.ebay.com/itm/BRAND-NEW-TV-DM ... 1027186026

Driving them is another matter though...

[MSat]

The two biggest issues I see with DLP is cost and packaging. AFAIK, even the super high resolution cellphone displays are cheaper than the DMDs alone

You can find 1080p chips for less than $200 on Amazon or ebay : http://www.ebay.com/itm/BRAND-NEW-TV-DM ... 1027186026

Driving them is another matter though...

And yet 1080p cellphone/tablet displays are well under $100 if iSupply's cost analysis of various high-end phones and tablets is anything to go by.

[Fredz]

And yet 1080p cellphone/tablet displays are well under $100 if iSupply's cost analysis of various high-end phones and tablets is anything to go by.

Problem is there is no way to drive them from a DVI/HDMI signal though, or not at this price.

[MSat]

And yet 1080p cellphone/tablet displays are well under $100 if iSupply's cost analysis of various high-end phones and tablets is anything to go by.

Problem is there is no way to drive them from a DVI/HDMI signal though, or not at this price.

I didn't mean to imply that they are, but even the support circuitry for LCDs is still vastly cheaper than that required for DLP, which was part of the point I was trying to make. The LCD, driver chips, and all the other components will still be well below the price of just a DMD. That's significant.

So to recap, a complete HMD using DMDs cannot be cheaper than one using LCDs or OLEDs, at least assuming that the resolution is the same.

[Jobe Smith]

I think they hacked the prototype up using 2 off the shelf handheld LED DLP 720p Pico Projectors. Pico's already have driverboard, dmd, 3 color rgb leds and hdmi inputs built in. A totally from scratch TI version could have supported1080p right from the start.

[Fredz]

According to their FAQ they are not using off-the-shelf pico projectors :

We’re working with TI utilizing brand new un-released, cutting-edge micromirror technology.

And I tend to believe them since the PCB supporting the DMD chip is clearly designed for the form factor of their glasses, it couldn't be cut from a pico projector :

Avegant close up.png

TI doesn't offer any public documentation about how to drive their DMD chips, you have to either use their digital controllers (DLPCXXX) or work a partnership with them to design a custom controller.

Or reverse engineer the protocol of the DMD, but that's an entirely different matter...

[MSat]

IIRC, at least some of the DMDs utilize a 32-bit DDR interface, as the DMDs act much like a memory device (they store 1bpp I believe).

[Fredz]

According to this document each mirror is attached to a Dual CMOS memory element and there are 32 data input lines for XGA chips (1024x768) so this makes sense :
http://www.ti.com/lit/pdf/dlpa008

But loading a memory cell is not enough to change the mechanical state of the mirror, you need to issue a reset command to synchronize them. The protocol seems quite complicated (timing requirements, loading by row (32 bits), reset by block (16 bits), four reset modes), I guess that's why a digital controller is needed.

I guess OzOnE2k10 could shed more light on this since he did experiment with DLP already.

[MSat]

It doesn't seem too complicated. 16 to 32 bits are loaded in parallel on both the high and low transition of the clock (DDR), and then depending on how you want it to function you use one of 4 "reset" modes to tell the mirrors to update their positions. It's a fairly straight forward protocol - you just need to adhere to the timing requirements. Of course, this is only part of the DLP process, as you still need to incorporate the process of lighting.

So yeah, it's still more complex and expensive than LCD/OLED

[Dilip]

Avegant has launched KICKSTARTER and in just 2 Days its triple funded of required amount and is constantly growing.
http://www.kickstarter.com/projects/ave ... me_popular

I was indeed very excited about this product, the day i first saw the article on net. They have strong points like 120 Hz capable DLP chip, ability to support all the media be it smart phone/tablet/Shield/PS3/PS4/Xbox 360/Xbox One and not so tailor made output requirement all media supported out of box.

The limitation is small FOV ~45 Degree but same as SONY HMZ and its half of price for that so not RIFT competitor but will send SONY HMZ run for its money. Colour fringing may also be the deal breaker but as study shows not all the people actually get affected by it or notices it. Besides they can increase the flickering frequency of LED to cycle RGB thus can greatly reduce fringing. Dr.Evan also have claimed to have some trick under sleeve to tackle this.

I Still see lot of potential in Glyph as VR is amazing but still not backed by great titles to enjoy with and Virtual Reality itself can induce discomfort and disorientation for many when not handled content properly regradless of absolute hardware. While not all people are in VR but many do enjoy 3D.

While this device can be benefitted from fact that any game that works with DDD Drivers for SBS output can be played in 3D on this device. Which makes huge library of over 500+ games in stereo 3D and there also won’t be any requirement for SDK or warping at least for display, head tracking may require.Head traking is not strong point either due to colour fringing but they can offer great screen for mobile devices good headphone is still bonus they can carve their diffrent path offering personal cinema and big scren gaming while not messing in VR.

I always dreamed to have device which can be attached with standalone small device which can be used at any place on the go and have bare minimum cables and can offer great display which is in my buying capacity.

Most of the check box appeared to be ticked here.

They can improve

FOV
Reduce Price
Reduce overall size of Unit
Tackle colour fringing while employing Head Tracker
All in the plan as per their kick-starter narrative. Looking forward for their December delivery of KS units.

I Still wish to see more gaming reviews. At present only one or two those are also with wrong devices and and bad choices.

[Dilip]

Display - Is chromatic aberration noticeable?

There is no chromatic aberration visible, no screen door and no pixelation. The micromirror array creates a very sharp image. The image is beautiful, better than any other display I've ever seen, head-mounted or not.
Last updated: Mon, Jan 27 2014 2:34 AM +05:30

https://www.kickstarter.com/projects/av ... u?ref=live

They are saying it does not exists.
Means solved?
Only beta deliverables can say

Interesting though they are pretty close to hit 1 Million- Still 21 Full days to Go

May be its going to be another RIFT Kick starter Story
http://www.gizmag.com/eyes-on-avegant-glyph/30660/

[Dilip]

Nice solid Development Prototype with JUST SINGLE CABLE.

https://www.kickstarter.com/projects/av ... sts/998964

This product is shaping up nicely parallel to rift, Has more surety of reaching consumer version lot more faster. There are less hurdles for them though evidently, as
most of path is carved for them once they made their first prototype. Though they did mentioned single output for twin display DMD was challenging for them.

It will be interesting to see how they make their place in market. They are no competition to RIFT but even backers and users demanding larger FoV then their current diagonal 45 degree.

Also found an interesting video of DLP chip projector showing rainbows.

http://www.youtube.com/watch?v=OFRsPNEY6aw

between 5.36 to 5.39 Minutes i have noticed heavy color cycling if this is what called rainbow effect indeed its bugging one.
don't know how they are going to fight this. i think this projector employes color wheel which could be slower while single LED cycling RGB can be hypothetically lot more faster. Wish they were free from this.

[Haloar]

As strange as it is to comment a dead thread, I would just like to say, that a "virtual retinal display" is a real class of display and not just marketing hype.
This display type was built by Thomas Furness III (of Boeing (?) fame) and Eric Seibel (of Magic Leap and spiral endoscope fame) at HIT Lab U. Washington among others.

It uses laser sources and a raster mirror (Microvision) or spinning fibre optic or other method to create a display which is then seemingly shone so to form a single point at the surface of the eye.

The negligible width and divergence of each pixel ray bundle creates a "Maxwellian View Condition" where the display will always be in focus no matter the focus/ accommodation of the eye lens (presumably with a changing apparent image size?).

It would be amazing if the Glyph had created this using LED/ broadband sources.

[Dilip]

Your perception of VRD is absolute right..
implementation of Avegant is somewhat derivative of VRD not exactly tech in its pure form. (Vibrating fiber optic (piezoelectric to vibrate the tip of the fiber in a spiral pattern so that it "scan" a much higher resolution image) is may be what magic leap is doing i guess)

They are basically using dlp pico projectors (dmd + Light source for each eye) beaming on iris rather then beaming on wall or screen
its RGB led source sequencing to create full color image and its not a steered laser.

Still their approach and product both are some what first of kind. Though resulting image they kept at 40 Degrees, may be good for movies
but i more enjoy movies seating just a feet and half away from 24 inch monitor which i think is much much higher FoV.

To investing in such display for sake of private entertainment on the go is very small and specific market they are targeting so i am bit curious about its reception .