SPI-V 1.3 UPDATE, ONE YEAR LATER by Michelle Bienias It was just over a year ago that Aldo Hoeben released the SPi-V Engine with the goal of creating the best viewer for panoramas. It was quite the draw at the IQTVRA Summit in Sedona that year and had panographers waxing prophetic, to wit: “[SPi-V] will revolutionize the field of VR Photography”; and “it is the most exciting thing to happen in panoramic delivery since QuickTime 5”. Indeed, high praise, and expectations.Given such fanfare, we thought it’d be interesting to check back in on Hoeben and his SPi-V Engine, particularly since he recently released the 1.3 version. We were particularly interested to learn what additions were made to SPi-V 1.3, how SPi-V was being used, and by whom. Hoeben answers all these questions below, along with revealing the website that blew him away, and the special gift he bought for SPi-V’s one year anniversary. Hoeben also extends a generous discount (good for one week only!) to VRMag readers on all license types. When VRMag featured the SPi-V engine a year ago, you had just released the initial version. What has the market response been over the past year?' It's been quite a ride so far. Both rewarding, when seeing people use SPi-V to display their content, and at times frustrating, when I can't get fixes, features and/or ideas out as quick as I want to. I can tell you it hasn't been a year where I had a lot of free time though, working only part-time on SPi-V and panoramic imaging...One of my reasons for developing SPi-V was to give an impulse to developments in the panoramic software industry. I will not claim that I have personally revived the panorama viewer industry, but I pride myself on having inspired at least some of the new (hardware accelerated) panorama viewers that have appeared over the past year. Fortunately these developments in turn help adoption of SPi-V, as more and more panorama photographers realise there are more exciting panorama delivery methods now. You’ve made many fixes and additions to the new SPi-V release, which ones are the most important and/or will have the biggest impact on users? The first major update after the launch in November 2004 brought compatibility with older versions of Shockwave, down to version 8.5. This meant a larger audience could view SPi-V content without having to install or update Shockwave. A suitable version of Shockwave comes preinstalled with OS X. Next up was version 1.2, adding support for XML 'snippets'; The features that set SPi-V apart are unlocked by writing XML files to describe tours, scenes and effects. And though I'll tell everyone who wants to hear it that writing an XML document isn't complicated at all, I'll be the first to admit that it has very little to do with making (panoramic) photos. Snippets are bits of reusable XML 'code' that can be included in an XML file to easily add certain features such as a navigation toolbar, or things like auto panning. They are comparable to the ready-made wired sprites that you can add to QuickTime VR files. Late last year, just weeks after SPi-V's first birthday, I finally released the 1.3 release of SPi-V. The initial focus of version 1.3 was to finally solve a long standing issue with certain video cards showing 'seams' or 'black lines' in certain panoramas. The cause of these artefacts lies in the underlying hardware accelerated '3d engine' in Shockwave, which was co-developed by Intel and the Director team at what used to be Macromedia (now Adobe). While it is this engine that enables SPi-V to display panoramas as characteristically smooth as it does, the engine heavily relies on video hardware and video drivers. Subtle differences between manufacturers or even between driver versions influence how the engine displays 3d scenes. To make a long, complicated story short, SPi-V 1.3 should finally remove these graphic glitches on all hardware/driver combinations. But there's a lot more to SPi-V 1.3... Since the first version of SPi-V it has been possible to just give SPi-V a panoramic image, without an accompanying XML file. In 'simple mode', the engine will have a look at the image dimensions, and make a quick guess about the projection and pan/tilt limits. This makes it a lot easier to use SPi-V to 'just display panoramas', but you're missing out on all the special features that SPi-V has to offer, such as adding branding, a dynamic cap image, etc. However, looking at how current licensees are using the SPi-V engine, I saw that this 'simple mode' is the most popular use of SPi-V. Adding further support for this way of using the engine, SPi-V now accepts a template for simple mode. This way you can 'spice up' the way panoramic images are displayed. The template can include the snippets that were introduced in SPi-V 1.2, so it is now possible to add a toolbar, branding and/or automatic panning functionality to panoramas displayed with SPi-V, without having to write an XML file for each panorama. Other changes include requests for cursor-key panning and right mouse button zooming, as well as controlling the viewer from a Flash interface or browser-side Javascript. The deveolper's website has a comprehensive list of SPi-V changes . Many of us have seen your splitscreen Euromast panorama. Are these splitscreen panoramas also a new feature of SPi-V? In the Euromast demo, you are looking at two panoramas taken from roughly the same point atop the Euromast in Rotterdam, six months apart. Instead of blending from one of the 'states' to another, like I have done before, the display is split through the middle, showing one half of the view in one state, and the other half in the other state, six months later. While you look around, the division moves with you, allowing you to interactively compare the two panoramas.The splitscreen viewing is actually not a standard feature of the SPi-V engine. It's a specifically authored extension to SPi-V. The goal of SPi-V is, and has always been, to create the best viewer for panoramas. The viewer is basically a layer between panoramic images and the full-fledged hardware accelerated 3d engine in Shockwave. It extends the underlying 3d engine in some parts and limits it in other parts to protect panoramic photographers from the real complexities of hardware accelerated 3d. With special knowledge of the Shockwave 3d engine and the SPi-V engine innards, it is possible to extend the basic functionality of SPi-V. This way I can add special functionality to the SPi-V engine, which is only used in special cases, without having to make changes to the SPi-V engine itself, which would make it bigger and more complex. Special one-off features, such as the splitscreen viewer don't bog down the general-purpose viewer itself. Current extensions include the fun-but-wild unfolded cubic and cylindrical reprojections of panoramic scenes, as well as a remake of my original Space Needle demo. Anticipated features such as the inclusion of sound and video will probably be implemented as extensions before they are included in the core SPi-V engine. A number of these extensions are now available from the SPi-V developer's website, so users of the SPi-V engine can play around with these special effects on their own images. When SPi-V turned one, you said on your weblog you bought a special gift for SPi-V?
To celebrate SPi-V's first birthday and then-imminent release of SPi-V 1.3, I indulged myself in an eMagin 3DVisor. Head mounted displays have always been 'cool' but at the same time hugely disappointing; way too expensive to have, and too cumbersome (heavy, calibration-prone) to use. eMagin is a company that makes OLED microdisplays for lightweight HMDs geared towards the military, and the Z800 was their first consumer product, at an affordable price. For a first consumer product, I am rather impressed. The effect of wearing the display is not all-immersive; it is more like looking at about a 21" screen from a comfortable distance (about 2ft). What makes these visors particularly cool however is the integrated 3DOF motion sensor; a number of accelerometers and compasses measure pan, tilt and roll, or '3 degrees of freedom'. With some custom 'glue' for the 3DVisor drivers (thanks Maarten!) and a SPi-V extension like I mentioned before, I can feed these values directly into the SPi-V engine, allowing someone wearing the display to look around in the current scene by... looking around. The effect is surprisingly convincing; this product seems to be made for panoramas. Of course the current product is just a first step; these displays will only become lighter and more immersive, and quite possibly a lot cheaper if volumes pick up. And hopefully one day you won't look like a complete and utter geek when wearing these things... In the meantime I will be improving the experience, looking into ways of navigating multinode tours, etc. But the promise is there, and if I'm all stressed out I'll just pop on my VR glasses, select one of my favorites Sedona desert panoramas and relax for a while.
 Now that it’s been out there for a year, do you have a good feel for who your user base are? Any surprises there? Out of what I find online, most users seem to be using SPi-V to just display panoramic images. They tend to use 'simple mode' instead of writing intricate XML files, which is of course hardly surprising. While XML files will unlock the full potential of SPi-V, writing them is an acquired taste, and the fact that the XML file documentation is still incomplete does not help (though with the release of SPi-V 1.3 I have also made a big step towards completing the XML specification). On the other hand, among the group of panoramic photographers are enthusiasts who have mastered Panotools (before the GUIs became any good ;-) ) and do some pretty intricate stuff scripting PTViewer. So every now and then I bump into a piece of SPi-V content that surprises me and that's very stimulating.One time I was looking through my site referrers (I do that sometimes when I get bored), and I found a personal site of a Finnish guy by the name of Mika Tanninen. That was the first time I was really blown away by what I saw, wondering how on earth he made the floor plan in his apartment tour. That's when I decided to add a 3rd party gallery to the SPi-V dev website. The most intricate SPi-V work so far was done by Bernhard Vogl, combining Adaptive Dynamic Range scene, a before and after scene, and multiple other scenes into a well-rounded tour of a fire station. Anything else you want to mention or point out? Well, now that you mention it, the release of SPi-V 1.3 came rather close after SPi-V's one-year anniversary celebration, and I have not had a real chance to celebrate the 1.3 release. So as a special gesture to the VRMag readership, I'd like to announce a special one-week-only discount of 25% on all license types from February 13 - 20, 2006. When ordering a license, use this coupon code: VRMAG06- For information about the SPi-V engine, visit: http://fieldofview.nl/spv - For more content developer information, visit the SPi-V development website: http://fieldofview.nl/spv-dev Email Aldo Hoeben: aldo[at]fieldofview[dot]com |  | | | The purpose of this banner is to raise funds for a new VR community project VRMag will launch in a few months. | |