multi-galvo system for 'virtual' increase in point rate

So recently I have been working a bit more on my 3DS to ilda converter software and have noticed in this and in quite a few other softwares (such as LazyMame) that the ability to project 60,90, or 120kpps would be quite beneficial to the graphical presentation.

I thought some today about ways to improve laser graphics projection speed but as everyone who has written optimization software knows, there are limits to the galvo speed based on physics principles that cannot be overcome without exponentially rising costs. If the scan speed could be increased with linear rising costs and high extensibility it would be a great benefit for detailed graphics projection.

A relatively straightforward solution to the scan rate limitation problem is to use two or more closely positioned galvo pairs and to feed each galvo pair a subset of the total point 'work' of the frame. The galvos would have a parallax offset that could be set in software on a computer or on a parallel galvo driving control board based on the distance between the galvos and the distance to the projection screen. Parallels of this concept can be found in many industries such as the computing industry and graphics card industry.

What would be required is software to choose the points to split up and create 'subset frames' for each galvo, plus the parallax offset and other geometrical correction as necessary. This software could be run on a computer or on a dedicated multi-galvo driving card that could be fed whole frames and it would split them out as necessary. In the case that the software runs on a computer, it could be in the form of a driver such that the show software does not need to know that it is there, the show software just thinks it is spitting out frames at 60 or 120kpps and the driver performs calculations to split the frame up, perform geometric corrections, and feed it to multiple DACs. In this way the concept would be compatible with much existing software such as HE-laserscan, LDS, zoofgames, Lazymame etc, only the software must have the capability to produce such high pps frames. It is possible that the show playing software could have this functionality built into itself and could make galvo splitting decisions based on the additional frame information that it has access to. An additional benefit is that the projectr would be able to produce frames twice, 3x, 4x or more than what is capable given any existing or future glavo technology.

In high power laser projectors, often times multiple lasers of one color are combined before being passed through a dichroic mirror to add to the main beam before being scanned. In a setup such as the one described here, the lasers would not be combined, but fed to two or more galvos. In the case of one laser per color, a beamsplitter could be used to feed the separate galvos but a method of independantly modulating the beams such as PCAOM must be used for each galvo. With these methods many combinations of sources and scanners could be devised.

I've talked enough you get the idea. It wouldn't be hard to write up a dual projector driver for EzAudDac cards as a proof of concept.

What do you think? Old news that I just haven't heard of yet?

thats how they scanned the images on the wall in Real Genius...

its a great idea and i dont think that a hardware interface is required.. i think it could be handled pre-dac in software..
you could basically "render" the show to a format that works for the quad setup and then play it.. the time coding is handled by the cpu/dac so it should work fine..

It sounds very plausible and it would surprise me if it isn't already done for some larger production show.

I see your point for needing pcoams if you want to use the same lasers. However, the whole setup can be simplified tremendously by using two similar projectors side by side.
The software would need some calibration wizard to align the two projectors in terms of position and orientation and some software correction provisions for the exact match.
The software would decide which projector draws which part of the frame.
Cost would be exactly twice as high given that the software supports this setup.

I would also leave speed out of the equation but define the gain in frame detail.

Alternatively the same setup can be used to create a wide-screen projection, allowing the same increase of detail but spread out over twice the surface.

To cut cost, maybe tricks can be done with color separation. One projector full RGB and the other is just green/RGY. The software sends the green parts of the frames to the green projector. Of course to enjoy the same increase in level of frame detail half of the frame would have to be green/GRY.

I've seen galvo pairs in series, I've seen multiple head shows, but I've never seen NxN pairs for the same image. With the decrease in costs of foreign galvos it starts to look like a possibility. Cost, has always been the dealbreaker. It is not hard to align two pairs to overlap.

Steve

Quote:

---

Originally Posted by mixedgas

I've seen galvo pairs in series

Steve

---

I would like to know more about that - especially the application. The idea has crossed my mind a few times. They idea is frequency separation right? One pair (the second) would have bigger mirrors anyway so that pair would take the lower frequencies.
Driving pairs of galvos in parallel seems less complicated and more COTS / low cost.

Didn't lobo do that?

Quote:

---

Originally Posted by Zoof

I see your point for needing pcoams if you want to use the same lasers. However, the whole setup can be simplified tremendously by using two similar projectors side by side.

---

Yep, but for stability purposes and like Steve said the galvos are getting cheaper so it might make a lot of sense to build a single high speed projector. Another benefit is that you could switch the projector 'mode' from high speed graphics to wide scan angle beams by not overlapping the projected area of the galvos.

Quote:

---

Originally Posted by FourDee

Didn't lobo do that?

---

Yes, LOBO has done this as part of their "Tri Dome" product. I am not sure if their LACON takes it any further for the more general case.

In fact, this could be done with our software, albiet using what I might call a "semi-automatic" approach rather than fully automatic. This is in fact, one of the key reasons why Projection Zones were developed. The semi-automatic approach places the division of imagery among available scanners in the hands of the artist rather than the computer, so that consistent results could always be counted on. The fully-automatic approach will lead to inconsistent results, not to mention a stronger requirement for perfect geometric correction.

Bill

drlava: Scary- that was your 666th post I think!
I think the splitting of a single frame data destined for multiple scanners was mentioned a while back in order to accomodate the heavy point count in the game Star Wars for the original LaserMame. I currently only have one EasyLase, but I will have a sound card DAC soon with a little luck.

One of my first posts after joining this forum was a thought about a missile command game that incorporated a second high-power beam scanner to emulate the shooter's phasers. I bet there would be all kinds of fun stuff you could do with this.

Also, if you could gang up banks of scanners, think what you could do with raster scanning. It would be like at the electronics store where all the TV's are ganged together to form a giant image.

Lastly, aren't a lot of DAC's out there limited to the number of PPS they support anyway? If that is the case, the DACs will need to catch up, or something like this is bound to happen.

Cool topic!

-Mike

Quote:

---

Originally Posted by Zoof

I would like to know more about that - especially the application. The idea has crossed my mind a few times. They idea is frequency separation right? One pair (the second) would have bigger mirrors anyway so that pair would take the lower frequencies.
Driving pairs of galvos in parallel seems less complicated and more COTS / low cost.

---

END quote,

Zoof, thank you for my new patent......... Just kidding. No, the idea was more angle.
More frequency was not a permutation I was expecting. If you want to do that, put a AO deflector in series with the galvo.

Steve

Quote:

---

Originally Posted by mikkojay

drlava: Scary- that was your 666th post I think!

---

Hmm I almost don't want to post again.. oh well :)

Quote:

---

I think the splitting of a single frame data destined for multiple scanners was mentioned a while back in order to accomodate the heavy point count in the game Star Wars for the original LaserMame. I currently only have one EasyLase, but I will have a sound card DAC soon with a little luck.

---

yeah it would be great for games and 3d-generated graphics!

Quote:

---

Lastly, aren't a lot of DAC's out there limited to the number of PPS they support anyway? If that is the case, the DACs will need to catch up, or something like this is bound to happen.
Cool topic!
-Mike

---

YES a lot of DACs are limited to 30 or 35 kpps, but that's not a big issue to overcome. Audio DACs natively do 48kpps and some 96kpps. But this would work best if computed before the DAC, IF there is a DAC. It would work great with an ILDA digital projector interface standard, too.

Quote:

---

Originally Posted by Pangolin

Yes, LOBO has done this as part of their "Tri Dome" product. I am not sure if their LACON takes it any further for the more general case.

In fact, this could be done with our software, albiet using what I might call a "semi-automatic" approach rather than fully automatic. This is in fact, one of the key reasons why Projection Zones were developed. The semi-automatic approach places the division of imagery among available scanners in the hands of the artist rather than the computer, so that consistent results could always be counted on. The fully-automatic approach will lead to inconsistent results, not to mention a stronger requirement for perfect geometric correction.

Bill

---

If the projection zones are static in position then a complex figure would have to span two zones to have any gain in scan speed. If the complex figure moved mostly into one of the projection zones during some effect then the scan speed would be slowed. If the projectors are tied to sprites or segment groups manually or as they appear then that should work well.

Why not use a cheap webcam and a little image processing to automatically compute the parallax for different parts of the target zone?

I love to see people adding to this.

Interesting idea, cheap webcams are typically 640x480 native resolution but to hit 0.2mrad alignment you would need 3,000 pixels for a 35 degree field. The resolution requirement could be halved by sub-pixel scanning and detecting peaks but it's still pretty high.

In the y direction a 640x480 webcam with sub-pixel scanning technique could get you about 0.62mrad if the scanning field filled the frame.

a more expensive 1024x768 native camera could get you 0.3x0.4mrad which is probably close to acceptable.

EDIT: the above calculations assume black and white mode on the camera if the bayer pixel arrangement of the camera starts to play a role then things get much worse.

Quote:

---

Originally Posted by drlava

Yep, but for stability purposes and like Steve said the galvos are getting cheaper so it might make a lot of sense to build a single high speed projector.

---

Cost, I figured that is why you should avoid adding the cost of a pair(!) of PCAOMs. For overlapping projection, both projectors need half the power minus pcoam losses compared to a single high power one.

Quote:

---

Originally Posted by drlava

Another benefit is that you could switch the projector 'mode' from high speed graphics to wide scan angle beams by not overlapping the projected area of the galvos.

---

Didn;t I just say that??? ;)

Quote:

---

Originally Posted by Zoof

Cost, I figured that is why you should avoid adding the cost of a pair(!) of PCAOMs. For overlapping projection, both projectors need half the power minus pcoam losses compared to a single high power one.

---

well you were suggesting separate projectors which would need double lasers anyway so PCAOMs aren't needed. At this point it just depends on usage. If someone expects to use the high speed or wide angle projector most of the time it makes sense to have it all built into one. But if not, then the flexibility of two makes sense.

Quote:

---

Didn;t I just say that??? ;)

---

Hmm re-reading it yep!

Quote:

---

Originally Posted by drlava

If the projection zones are static in position then a complex figure would have to span two zones to have any gain in scan speed. If the complex figure moved mostly into one of the projection zones during some effect then the scan speed would be slowed. If the projectors are tied to sprites or segment groups manually or as they appear then that should work well.

---

You misunderstand the concept of zones, as implemented by our system. Understandable since you don't have one of our systems... The latter part of your statement is closer to how zones work in our system. But anyway, you will figure out all of this, plus the limitations soon enough...

Bill

Borrow from the eye and put the camera on a piezo or two, shift it 1,2 , or 1/2 pixel.

Digital saccadic motion :-)

You have a precision position sensor in the galvos, and gratings create known images. Think about it...

Steve