Scannermax 506s

[mixedgas]

hmn going to give it a go at the test pattern too. Got some ideas
Will post when I got a graphic done.
I have the lasermedia issue on one angle too.
However I have it when scanned without rotation. with rotation it's gone, I don't have this on another scanner so this is an opertunity to make a test pattern which forces it at all angles if occuring.
Will include a circkle with low point count and square and a color delay test (rising and falling and color shift) and feedback test. along with a small graphic that only shows up properly when a system is perfectly tuned (it will also show any occuring resonance).

Any other things you want to see from the scanner please PM me and I'll add it to the graphic.

The graphic will not flicker so test results can be shared more easly.

PJ, ask Dr. Lava if he'll let you have his color delay test pattern. It works real well and could be incorporated in a large pattern.

Steve

[dkumpula]

Look at the markings on your DT30 amps, you'll find they are DT40 amps

My amps that I'm running my 506's on are actually from a DT30 set.

Essentially, all Dragon Tiger did (afaik) is take the DT25 galvo and paired it with the existing DT40 amp to make the DT30 scan set.

The only difference between DT30 and DT40 is the galvo

Thanks, Norty!

I should have known that. Well that's problem "solved" in my case. Now just to decide how many pjs to convert to 506s and how many get to wait on Saturns.

David

[The_Doctor]

I had thought about rotating the grid in my suggestion but the small line sizes plus the fact that it would make rounding harder to spot put me off the idea. The reason I didn't go for the laser media patter is because it left no room for the elements.

I accept that a new pattern will likely help more, but wouldn't the ability to halt rotation at an arbitrary fixed angle, or change rotation speed help? That's what I was doing to explore the WideMove problem... It meant I could use far less data in my tests, if I could get fine rotational control on the data I used. My small video was actually based on 180 stills advanced by 1° each time. It showed the same thing a full video did, with less data by far and with large detailed image frames. Similarly, a small pattern that can be set to any angle at will may do this too. My instincts say that the smaller the data set in any test pattern, the better. The aim being the neatest, smallest expression that covers the most useful bases. If we end up with a 'kitchen sink' approach to test patterns it might go wrong in all manner of ways. The more important the tool, the more situations in which it must work easily and reliably.

[White-Light]

I accept that a new pattern will likely help more, but wouldn't the ability to halt rotation at an arbitrary fixed angle, or change rotation speed help? That's what I was doing to explore the WideMove problem... It meant I could use far less data in my tests, if I could get fine rotational control on the data I used. My small video was actually based on 180 stills advanced by 1° each time. It showed the same thing a full video did, with less data by far and with large detailed image frames. Similarly, a small pattern that can be set to any angle at will may do this too. My instincts say that the smaller the data set in any test pattern, the better. The aim being the neatest, smallest expression that covers the most useful bases. If we end up with a 'kitchen sink' approach to test patterns it might go wrong in all manner of ways. The more important the tool, the more situations in which it must work easily and reliably.

I agree totally.

There's no reason why however why a rotating line couldn't actually vary in speed automatically if when assembling the test pattern a keyed rotation approach was used. Whether this would satisfy what you're test was aimed to achieve, I don't know, but obviously there are limited controls when you automate something but equally it's not possible to selectively rotate part of a pattern manually.

However, it's also important to note that what I posted above was a photoshopped suggestion not a cue based solution. No test pattern is ever going to be definitive in it's own right as the number of tests vs the number of points the scanners can reasonably draw is always going to dictate that a number of test patterns are going to be necessary to test every aspect. I believe what Bill is trying to currently achieve is a test pattern that can be used to test different scanners at different speeds and show up their failings (if any) obviously to the casual observer so you don't need a PhD in Test Pattern interpretation to see that something is wrong. Something for the ebay laser generation as well as the enthusiast / professional that shows at a glance whether a scanner is achieving the claimed performance so it's easier for end users of limited knowledge to see the difference between something that simply performs and something that performs better. That's why I included the grid and your rotating line in my original suggestion because apart from putting the circle in the box, those seem to be the two patterns that most vividly show errors in scanning, when either the speed or angle are too great for the scanners to perform at their optimum.

But like I say it was a suggestion not a solution. Bill is far better qualified than I am to assemble a test pattern that actually works and achieves what he had in mind. What I was doing was merely playing to see what might fit and filling a little spare time for fun.

[mixedgas]

Again, what is wrong about reading the directions for the original test pattern? They are a sticky here now.

Some thoughts:

I have a hard enough time to illustrate to non-technicians how the fixed pattern works. Now I'm supposed to explain a spinning line animation?
Here, randomly turn these two pots/move these sliders till you see the lines become one?

Remember any thing you design has to be global, it has to work as a playback on a SD card reader as well as high end software.
We would not want fighting in the war room over the test pattern gap, now would we?

I think you'll find any new pattern that can measure long jump performance will be ONE axis at a time. Just so the Humans can read it.
It will involve square waves.

Steve

[kecked]

since this is a function of the scanner design and you can't change it, does it matter? I mean once the test is done it isn't going to change much for that model. just post the results

[The_Doctor]

I'll risk making a clumsy answer to the last three posts.
If the bearing wear or lubrication changes, or a scan tuning pot goes dodgy, an interpolated rotating line test can reveal it later, even if all was well first time. On the other hand if the letters in an ILDA pattern start to look mushier than they used to, that will be warning too. Perhaps the original test patterns should not be changed, but I see no problem with rotation as a way to extend the ability of a test, partly because we've seen it do exactly that, usefully. As Whitelight said, we can't selectively rotate part of one pattern, so keeping them separate may be best. As it is, there is nothing in my rotating line test that cannot be done by rotating the LaserMedia pattern. That looks like the best bang per buck to me, although there may be ways to add colour modulation info to it to test laser mod timing and intensity as well. I think the less change to existing patterns the better. The ILDA pattern is already packed with various mechanics tests. The LaserMedia pattern had been described to me a couple of times as 'less useful these days' or words to that effect, so this rotation plus a bit of colour testing might be a good way to update it to parallel the ILDA pattern in usefulness. As to the three tests I mentioned, it comes down to whether the tool showing the pattern can rotate it and switch interpolation on or off. These two details do add complexity, but not a lot, besides they will help teach newcomers the power of their tools.

[White-Light]

As Whitelight said, we can't selectively rotate part of one pattern, so keeping them separate may be best.

Just one quick note, you can selectively rotate one part of a frame in Pangolin software using an automated rotation effect (use a synth and only apply the rotation effect to a single element). Making a universal or manual test cue of that type though is impossible as different software's have differing effect methods for achieving rotation and manual rotation of a part of a frame cannot be done on any software I'm familiar with.

@ Steve, I did have a look at the ILDA documents and many thanks for getting these for everyone. However, the one thing that was missing for me to actually design a physical cue, was the relationship between the size of the square and the scanner speed. Incorporating a 30K test circle / square is easy enough as you adapt the existing 30K pattern through editing. However, adding the same for 40K and 60K is impossible if you don't know the relationship between square size and scanner speed.

[The_Doctor]

That might be why I had a hard time getting people to do this rotation test in the past, lack of means to set it up easily. Maybe it could be added to a static frame if it jumped in 5° increments clockwise though 180° total. It would add data, but solve two things: no need to add extra rotation, and it can be timed to give eyes and brains long enough to take in what it's showing at each step. Personally, I'd just set up one rotating line at full scan angle as I did, but for systems with no way to manage rotations the stepped form might work, so long as interpolation can be switched on or off depending on what's being tested for.

A stepped form can be square waves (is by default if no interpolation or slew rate limiting is used), and includes 0° and 90° as Steve suggests, but there are uses for the other angles too. (Laser Media already includes 45° and allows matching tuning for X and Y axes.) To make it easier to spot which angle is actually displayed at each step, colour can be used, for example make it white (or brightest) at 0, 45, 90, 135, some other colour (or dimmer) for 15, 30, 60, 75, 105, 120, 150, 165, and the other 5° steps in a third colour, (or dimmest). This might be ridiculous if it makes people wait for white verticals or horizontals or diagonals when they're tweaking stuff, but that's why I prefer to set up rotation specifically to order, then put it wherever I want.

Thanks Steve for those ILDA thingers. I'll try to navigate those as much as I can.

White-light, if you work that thought in reverse you might get that relation. Start with scan angle for 30K, widening until the diagonals in a rotating Lasermedia pattern begin to bow apart. (No interpolation to be used). Then drop scan angle low, step up scan rate to 40K, or 60K, and widen it again looking for the same conditions. The smaller angle, in relation to what it was for 30K, probably gives a useful relation, assuming resonances in scanners and mirrors don't mess it up (which in many cases I think they might, if I understand Bill correctly). As it will depend on the point where a particular scanner gives way from small step response to large step response, it may be the only way to do it, rather than try to embed such a test in one fixed pattern.

[White-Light]

Thanks for the suggestion but my scanners are only rated at 28K.