ILDA Format BS

[The_Doctor]

Actually, there's been a good bit of speculation around exactly that idea, Gary: Moving the digital-to-analog converstion process *off* the laser show controller and into the scanner amp itself. Several people were talking about it on the ILDA cruise. Rumor has it that great minds are investigating the concept, and at least one person has claimed to see a prototype (that evidently didn't work very well, but it was an early attempt.)

This is about the only way that it's going to overcome the need for an analog signal to be sent, and often stored (as this is cheap and easy and fast). But to do it is equivalent to making a closed loop servo active loudspeaker with a DAC on board, with all tailoring of responses meticulously designed for that hardware, as a single and EXPENSIVE unit. So far as I know, this has not yet been invented and sold commercially, even to top studios. While there are esoteric projects that can be found via Google, perhaps, they're rare, last one I saw was huge and cost the kind of money people buy large houses with.

People might still be discussing how to make a digital signal go to a specialised scan renderer decades from now but what's the point? More speed to let you fill in blocks of colour within a frame? To scan more lines per second? Right now you can do this with a video projector, or a few laser projectors. So long as you move a mass that scans a single beam, that can only point at one spot at a time, it makes no sense to prevent the analog signal being open to all to access it their own way. If top bods of the ILDA are seriously suggesting hardware that prevents this, they're more likely to fuel continued speculation that they're trying to create further exclusion than anythign else.


It doesn't matter how you work out what the beam will draw, it's still got to draw it. The only thing the scan amp should be doing is to do this faithfully, except where such instruction endangers the scanners, in which case all it needs to do is impose a voltage and slew rate restriction so the scanners don't take the flak.


EDIT: keeperx, precisely so.

[The_Doctor]

...it's hard to send analog signals over long distances with the current ILDA cable standard. A digital link would be far superior.

Agreed. But this isn't a reason to argue for a complex control language to extend to the scan amp! I accept that it's an excellent argument for a digital conversion of a signal though. For now, people already have to have laser controllers generate waveforms. So just send them digitally via whatever cables are cheap and appropriate and use cheap audio DAC's as a simple add-on to most scan amps. I guess the question of what format of digital audio stream might be used will arise, but why not? Someone said that using audio stuff was 'forcing' use of inappropriate tech, but that;s a silly argument. Audio actually appropriates tech from other fields more than they do from audio. The one thing that audio has done is make it plentiful, cheap, and reliable. I think that failing to take direct advantage of that is foolish.

EDIT: I recommend S/PDIF. If you need more channels for speed and flexibility, use them. It's been invented to solve the problem you mentioned, so why not use it? It even has a SYNC feature built in!

[James Lehman]

The whole reason for using modified sound cards to control laser projectors is that we get technology that is perfectly suited for the job, but has NOTHING to do with the laser display biz.

It's always a good idea to look for a generic solution that already exists.

Haven't you noticed that anything made specifically for laser display is not very common and usually quite expensive?

Use your brain before you use your wallet.

James.

[keeperx]

S/PDIF is optical so thats cool...

and your right.. an audio dac is a does nothing more than convert digital signals to analog signals so you can get analog sound out of your computer.. its not actually audio technology at all.

[carmangary]

Maybe you are missing the whole point. The point is that current scanner amps do NOT faithfully reproduce what you send to them. The ILDA test pattern is a perfect example of this. It is precisely an exercise to make your scanners suck the exact same amount as everyone elses. The reason the circle is inside the square is because the scanners can't handle what is actually being sent to them. This means that if you display an image at 30K that was intended to be displayed at 15K it might look wrong. With an advanced control system, it wouldn't matter what speed you used as long as it was not too fast for your galvos. Then you could double your framerate and the image would look excactly the same. Can you not see the advantage of that? The only way to do that presently is to recompute the frame in software by pulling or adding points or whatever. And even then, you aren't guaranteed that it will look right unless the speed and angle is exactly right.

Money is of no matter. Technology that costs thousands today will costs 10s in the future.

Am I the only one that sees the benefits of what I am talking about? Or does everyone need to spend time writing software to control DACs before they get it?

[The_Doctor]

My point is that by placing more intelligence nearer to the galvos you can have much better control. And the likely way of interfacing to this "smarter" scanner is via a digital signal of some kind.

From what I've read, dwell point duration is dependent on the angle the scan draw is changing though, though it's actually specific to each galvo. This can be calculated by the rendering engine. I found that WideMoves actually were happier with a dwell time that changed with overall speed, but not with angle! Probably because that's what they're for, they have a fairly slow step response but it changes far less over a wide angle change than with most other types, probably. So, what's needed isn't a complex dedicated hardware scanner, you just need a model for each scanner, that is plugged into the software end, it modifies the slew rate, curve, whatever it takes to optimise the drawn path for speed and precision. This is entirely feasible, because it's easily added, and easily switched off, and there aren't that many scanner types in widespread use.

And the moment you tune scanners to modify the damping and gain much, all bets are off anyway. About the only advantage that a digital system in the scan amp has is storage and repeatability of settings, allowing easy derivations of copensation templates for scan renderers to use. So I agree with you that having some digital stuff at the scan amp end is worth more than just a DAC to overcome line noise, but there's no need for it to be 'intelligent', as it's cheaper and more versatile to just send fast data streams. Logically, this is the same as the analog methods we use now, but with the expected improvements that digital signals and storage can offer.

[carmangary]

Yes, I have proposed the scanner model idea on here before. For now, that is only what you can do and then hope for the best. But what happens when your bearings wear and your model is no longer correct?

When you drive down the road, do you see people flying off the road into fields? Each car/passenger is unique, going different speeds, etc. But, they all manage to follow the same path on the road without going into the ditch. The reason is because there is intelligence coupled directly to the car via your body. Your brain is the computer. Your eyes are the sensor. Your foot and hands are the controller. That is what I am proposing.

What we have now is basically a blind person driving down the road. We say, go exactly 55mph and turn when I saw NOW! As long as they do it just right they'll make the corner just fine. But, if the road is wet or they are going a bit too fast they fly off the road. I have a hard time finding this to be acceptible.

I don't mind spending a little extra money or dealing with the added complexity.

I'm not proposing digital for the sake of proposing digital. We already use digital. WAV files are digital. ILDA frames are digital. The whole point to it is that if there is such a thing as what I am proposing in the future, it makes no sense to send the data in WAV forms that represent voltages because you are no longer dealing with voltages, you are dealing with images. I keep hearing talk about it being about the laser art yet you want to control voltage. What I am proposing allows you to deal only with the art without knowing jack about the hardware.

[The_Doctor]

The reason the circle is inside the square is because the scanners can't handle what is actually being sent to them. This means that if you display an image at 30K that was intended to be displayed at 15K it might look wrong. With an advanced control system, it wouldn't matter what speed you used as long as it was not too fast for your galvos. Then you could double your framerate and the image would look excactly the same. Can you not see the advantage of that?

Yes, but how complex does it have to be? The rendering engine draws the scan waveform. It's designed for the current hardware. Given that a 30K scan needs a signal with a few tens of thousands of samples, you could drop alternate samples to halve the rate while preserving realtime display, and modify line optimisations to suit the different scanner system. As those differences are based in gain, damping, mass, it should be possible to get that info plugged into a rendering engine.

This ought to be cheaper and better for the same reason that many electronics tasks become easier if you know how to program PIC's and Atmel IC's instead of having to build dedicated models in hardwired systems. Once you've done this, you have an appropriate signal to send the easy way.


EDIT: Got to say I DON'T know how to do this... But I've been told it many times by people who do.


More edit, rather than a 6th post (I'm in danger of overdoing it here)..
Worn bearings will either result in a change the closed loop can adapt to, or it will be nonlinear enough to be beyond simple compensation, wherever it is attempted. Re the car analogy, I like it, BUT, tell me where the intelligence is in the crankshaft and steering column. (More importantly, the fact that people can manage this feat is due to forethought, observation, and experience. These are HARD to model! Besides which, they don't work by being close to the wheels. They need the long view. This problem won't be solved by confusing higher and lower order functions.) Ok, that's flawed, because scan amps already use the kind of analog logic of PID controllers. But if it can't be controlled there, perhaps with a simple adjustable correlation of those controls and the incoming frame and scan rates, then it's definitely a higher order thing, and best not done at the scanner end. I think you're best off doing the scanner models. If a galvo becomes so worn that the models can't cope with a small tweak, the scanner is likely unfit for safe shows.

[The_Doctor]

What I am proposing allows you to deal only with the art without knowing jack about the hardware.

Ok, I had to post a separate one for this.. It's a VERY attractive prospect, I agree, but at what cost, assuming you can do it? The history of art is full of people who assert that you cannot separate the art from the craft. Once you see that trying to do so can result in the equivalent of riding a bike with stabilisers on it, it's not even a an attractive choice any more.

[heroic]

More edit, rather than a 6th post (I'm in danger of overdoing it here)..
Worn bearings will either result in a change the closed loop can adapt to, or it will be nonlinear enough to be beyond simple compensation, wherever it is attempted. Re the car analogy, I like it, BUT, tell me where the intelligence is in the crankshaft and steering column.

Does your car have traction control? Antilock brakes?

Ok, that's flawed, because scan amps already use the kind of analog logic of PID controllers. But if it can't be controlled there, perhaps with a simple adjustable correlation of those controls and the incoming frame and scan rates, then it's definitely a higher order thing, and best not done at the scanner end. I think you're best off doing the scanner models. If a galvo becomes so worn that they can't cope, they the scanner is likely unfit for safe shows.

A hard drive is basically a magnetic head controlled by a galvo. They manage to be fast and accurate to within hundredths of thousandths of a millimetre. They do this every time. They very seldom fail. They are controlled not by the computer sending them a voltage, but by the computer telling them which block to seek to.

They also cost fifty bucks a pop. Do you think they'd be that cheap if they had to do all the modelling in the computer? Everything in the mechanical system would have to be more accurate and the electronics would have to be an order of magnitude more complex to compensate for the fact that the drive couldn't compensate for its own inadequacies.

-J.