I had a chance friday to speak with the CEO of Real D, Joshua Greer, whose company acquired StereoGraphics back in February, hoping to get a good idea of how comitted the new company was to maintaining its support of the CAD and scientific markets.

Given the hardcore entertainment background of Joshua Greer and the other co-founder of Real D (Greer worked with James Cameron on the filmmaker’s Ghosts of the Abyss 3D film), I was concerned that the new company’s focus would be switching mostly in that direction.

But during our conversation, I came to realize that Greer is as passionate about the vibrance and reality a stereoscopic view of both entertainment and technical content, and that he was particularly proud of the company’s recent announcement that SolidWorks would be including support for the StereoGraphics products in the next version of its software.

If you thought stereoscopic 3D was dead or stagnant—for both CAD and the movies—guess again. This is just one of several recent stereo product announcements that I think portend a resurgence of this 150-year-old technology. I’ll tell you about the latest stereoscopic display system I saw at National Manufacturing Week later this week.

I’ve accomplished another step in installing the eDiemnsional stereo shutter glasses, unintentionally. I installed the latest NVIDIA display driver, and once that was done, an eDimensional installation tool that was still waiting in the background (after many reboots) noticed the new driver, and proceeded to install the NVIDIA stereo driver.

But still, even with the NVIDIA stereo driver installed and adding stereo settings to the NVIDIA-specific control panel, I can not get the stereo mode to work. Turning stereo mode on displays a test image that shows two strange-looking patterns, one on the left and one on the right. This appears to be the type of image meant for display systems that split the stereo pair on the left and right side of the screen, and then redisplay them via LCD goggles or something else.

But that’s not the frame-sequential method that the eDimensional glasses use, so it’s going to take some more sleuthing to figure out if I’ve got all the correct pieces installed.

Playing a hunch that the order of the devices in between my PC and monitor might make a difference, I plugged the eDimensional stereoscopic shutter glasses adapter directly into my computer, and the KVM switch into the adapter, and was pleasantly greeted with the image I expected on the CRT.

So, with that problem solved, I proceeded to install the eDimensional software and driver, and was troubled by several steps in the process. First was a warning that this install was only for Windows 2000 and Windows XP, and that I should quit and find the Windows 98/Windows Me installer in a different directory on the CD-ROM. I’m running Windows Media Center Edition 2004, which is based on Windows XP Professional, so it should not have given me this warning.

The second problem was that since I told it I had an NVIDIA graphics card, it wanted to install NVIDIA drivers, overwriting my existing driver. The last thing an application should ever have to do is to overwrite a system driver like this. And since I have no idea which driver this is, or if it’s compatible with Media Center, I have refused to install it, which I presume will make using the glasses problematic.

So I’ll be calling eDimensional tech support, to find out why this driver replacement is necessary.

The pair of eDimensional stereoscopic glasses I ordered yesterday arrived this morning, even though I had only paid for the 3-day FedEx shipping.

Unfortunately, when I plugged the interface box inbetween my KVM switch and my CRT display, the computer’s image failed to reach the monitor. I’m guessing that the use of two power-stealing devices is just too much for the poor-old VGA standard, and that I’ll have to ditch my KVM switch and find another way to get access to the other computers under my desk…

I’ve long advocated the stereoscopic viewing of CAD designs, as it greatly enhances the understanding of a design by non-technical reviewers, as well as the productivity of designers because it allows one to instantly see the real geometric relationships in 3D without having to spin the model around to orient one’s view of the model.

To test the consumer-level technology, I’ve just ordered a pair of E-D Wireless Glasses from eDimensional. This company appears to be the last company with a consumer gaming/entertainment focus selling stereo shutter glasses, so even though I’m not convinced their hardware uses the VESA standard 3-pin Mini DIN connector for stereo sync, I’ve decided that it’s the appropriate device to expect CAD visualization software to support. They can be had in a wired model for $70, little enough that governments and design clients alike could afford them for reviewing your designs.

But despite the benefits, the use of stereoscopic visualization in the CAD market has been little enough that the longest stalwart—StereoGraphics—has been acquired by REAL D—a startup with its sights directly on the entertainment market.

I can’t imagine REAL D will leave behind StereoGraphics’ existing customers in the automotive, aerospace, and scientific fields, but whether the new company expends significant efforts in these markets remains to be seen. In any case, if they do, any harmonization between the entertainment and design markets could increase the likelihood of stereoscopic content becoming successful.

I spent several hours yesterday researching what consumer stereo shutter glasses are still available—I even went down to Fry’s in a fruitless attempt to find them in a retail store. What amazed me was the number of different glasses that have been sold over the years, as well as the number of different methods used to connect the glasses to the computer.

From reading the Stereo3D charts—and just from common sense—it becomes clear that the only sensible interface standard is what this site calls “VESA-3″—the three-pin mini-DIN connector initiated by StereoGraphics and adopted as a standard by VESA.

After finishing that fishing expedition, I decided to visit StereoGraphics’ Web site, to see what they were up to, and discovered a pleasant surprise. The company announced at SolidWorks World that the next version of SolidWorks will be supporting its shutter glasses. This is great news, and tells me that that stereoscopic display is far from dead in the CAD industry.

So with this encouraging news, I’ll be adding stereoscopic support to my list of must-have features that I test with all CAD software.

Somewhere last week, I found a link to a cool NASA software project—World Wind—which appears to be a revival of the Blue Marble Viewer (for the Blue Marble project) with a different code base.

The ability to view and zoom in on a 3D globe of the Earth is wonderful, but the program only works on Windows, and since its code base is C# and DirectX, instead of C++ and OpenGL, it will be very difficult to port to other platforms. I applaud NASA for creating this, but since our tax dollars are being spent, they should be spent on tools that work on Windows, Mac, and Linux, to ensure equal access.

And the UI is also as frustrating and confusing to use as The National Map. Even though the whole thing is rendered in 3D, there is, of course, no support for either 3D mice such as the SpaceBall or stereoscopic viewing.

I’ll be testing this tool this week, comparing it with the commercial competition—Keyhole, GlobeXplorer, and Earth Explorer. (I might also take a NOAA thingie for a spin.) In software like this, I would expect to be able to extract geo-referenced images easily, view the content in stereo, and use my SpaceBall to manipulate the view.

This might keep me busy long enough to get an answer from Autodesk and Bentley about including their mapping software in the Field Day mapping project/software review.

Since many of the documents I view and create on my computer claim to be based on real-world sizes (drawings in inches or mm, text in points, etc.), I’ve decided to finally calibrate my computer’s monitor, using the built-in Windows display control panel features.

For my CRT, with its current settings, the correct value turns out to be 85% of normal size—82dpi. I determined the value by holding a tape measure up to the screen, and adjusting the value until the on-screen scale matched the real one.

This feature has existed in Windows as far back as I can recall (at least back to Win98, and probably Win95), but its existance is little known, and rarely used correctly by applications. My use of it will allow me to test any application that has a “show actual size” feature. I’m not expecting it to be pretty, but it is the only way you should ever expect such a feature to display your work in its “actual size”.

There are many problems with its implementation, all the way through Windows XP:

• The on-screen ruler is too short to be entirely accurate. It should extend the entire length of the screen, to ensure the correct value can be entered.
• Windows only allows you to set this value globally, even if you need a different value for each monitor.
• The Custom DPI Setting is also burried where few people will ever find it. You must open the Display Properties control pannel, select the Settings tab, press the Advanced button, choose the Custom Setting… from the DPI Setting drop-down listbox, and then have a ruler handy with which to compare the on-screen ruler agains.
• Although you can adjust the calibration using either the percentage drop-down listbox or by dragging the ruler, you can do so only in whole percentage steps—which is usually more than one pixel—so the result might not be as accurate as it should be.
• Many of Windows’ built-in display items, including system fonts and icons, are pixel-based and don’t scale well when you choose a DPI setting other than the built-in ones, which are rarely accurate.

So far, the problems I have found that appear to result from using the correct value include:

• Windows scaling of some icons and text (the icons in the quick launch bar and the task bar) to the point of being distorted, when they shouldn’t be scaled at all, since they are designed in pixels.
• Some text and hyperlinks in the Mozilla browser being cut-off incorrectly at the top, bottom, or in the middle.
• A similar problem with images, with multiple white horizontal lines tarnishing some images, some of the time. In either case, highlighting and un-highlighting the text or graphics displays them properly, until you scroll the page. I’ve not seen this problem in Internet Explorer yet, just Mozilla.

The SpaceBall 5000 USB that I received at AU impliments the USB HID (Human Interface Device) specification for a “multi-axis controller”, but because no software I have yet tried supports this cross-OS, cross-manufacturer standard for 3D input devices, I must install 3Dconnexion’s system driver and then a sub-driver for each program I’d like to use the controller with—AutoCAD, Photoshop, Office, and so-on.

Sometimes this is a good thing, because not all software vendors may not realize that their 2D application would benefit from a 3D mouse, but in most cases, it makes the configuration of the controller more difficult and less seamless than it would be if the app supported it natively. For any 3D design or viewing program—even the free ones—to not support this natively is as not supporting a 2D mouse natively.

When I first plugged the SpaceBall in, Windows XP informed me that it had found and installed a driver, and that my new hardware was now ready to be used. That should be all I needed to do, and the fact that it isn’t is the fault of CAD software vendors; not 3dconnexion.

With that preface out of the way, let me detail what I found installing the 3dconnexion driver. (Keep in mind that I’ll frequently disable it when testing software, to find out which applications support it natively.)

I was surprised when the installation program complained that Adobe Acrobat was running, and that I must stop it before continuing. At the time, I had no instances of Acrobat running, and the icon was not to be found in the system tray. So I was forced to hit Ctrl-Alt-Del to bring up the Task Manager, re-sort the list by application name, and end the Acrobat.exe process. Yuck.

After dealing with that, I chose the Custom install, and found that it had detected the following software installed and selected the appropriate SpaceBall plug-ins for them: Office, AutoCAD, Acrobat, and Photoshop. It then asked me if it should install and launch the system driver (3DxWare) on startup.

And then it proceeded to direct my browser to the 3Dconnexion Web site and a form asking me to register the product; magically selecting the applications it had found and installed plug-ins for. Unfortunately, it did so without launching a new browser window or tab, so it did so in my current window: this in-progress posting, just about wiping out all my input!

Note to the installation team: Opening new browser windows at the drop of the hat is annoying, but changing the current content of my browser window without permission or warning is far worse!

Walking out of the AUGI meeting a few minutes early, I was slightly delayed by a woman wearing an HP shirt, passing out flyers.

The flyer I got had the number 12 (or some other nearby number), and the gal from HP was saying “Would you like a free SpaceBall? Take this to the HP booth” to me and the next person behind me. If there had been any mention of this, I had totally missed it. Apparently, I was one of 20 or so other lucky individuals who just happened to pass through there at about the same time.

Surely all the others had the same goal I did—of beating the crowd into the line for the AUGI Beer Bash on the show floor, but alas, the line was already longer than any I had seen at AU.

Upon reaching the HP booth, I learned that the giveaway appeared to be a clever geurilla marketing effort by HP to encourage people to explore the benefits of the device, and quite likely to evangelize it back at their office. To receive the SpaceBall 5000 USB, each of us had to turn in our numbered flyer (no double-dipping!) and sit through an automated training session with the device.

Being quite familiar with the device already, I didn’t really need the training session at all, but agreed to try it anyway. I was impressed with the exercises the tool used to tutor new users, and offered the fellow from HP that gave me the demo (and the SpaceBall) a few tips on how it might be improved.

Not to be overly fateful, but I did see this as a sign that my desire to return to covering the CAD market after a year’s hiatus was clearly in the cards.