July 29, 2014

Wikitude Makes a Major Leap Forward with the SDK 4.0

The next great advancement in vision-based augmented reality has arrived! As we mentioned a few weeks back, we’ve been busy putting the final touches on our new SDK. As leaders in the field, we’re committed to pushing the boundaries of what’s possible with AR. With the Wikitude SDK 4.0, we believe we’ve struck the perfect balance, and are delivering a library that’ll meet and exceed your expectations.

Entire new tracking engine working smoother and more robust

There’s a lot of computational science that’s happening in the image recognition component of vision based augmented reality, and we went back to the drawing board to see if there wasn’t a better way of looking at the entire process. Building off the previous advancements to our tracking engine, we reordered a few processes, retooled a few others, and a fiddled around with a few more until we’d found a solution that resulted in dramatically improved performance. 140718_WT_SDK4.0_KPI_all True to the classic phrase, “Time is money,” we know that the end consumer of your AR creations is someone on the go, and clearly in touch with cutting edge technology. To this end, one of our goals was to improve the quality and speed of tracking images. After optimizing our tracking algorithms and adapting them to the actual performance of each device we are now looking at a doubled tracking frame rate. This substantial improvement of our core algorithms leads to numerous effects you and your customers will benefit from. An aspect we looked at was how far away from a target image a user can be for a trigger to function. While the existing engine already could track an image a few meters away, the new Wikitude SDK 4.0 can track a target image (paper A4-size) up to 20 meters away. In other words, there’s no need to be right on top of an image target anymore, take a few steps back, and the scene will still work just as it should. Building on the tracking frame rate, we’ve also accounted for stronger movements in any direction. Users of your augmented reality scene can pan, tilt rotate or just move their phone faster or view the target image at an extreme angle – the Wikitude SDK 4.0 will still track the image. And the icing on the cake is the improved tracking recognition of even the lowest quality images. Those of you already using our SDK know about the 1-star rating, often forcing you to retool the intended graphic, or replace it with a higher rated image. You can forget about this now, as the Wikitude SDK 4.0 dramatically improves the tracking quality of these images. Having said that, it’s not 100% perfect, so those hopes of using a plain white sheet of paper are going to have to wait a tiny bit longer. Oh, and just in case these massive improvements weren’t enough to get you developing right away, we decided to slim down the library of both the iOS and Android SDKs, thus decreasing the memory footprint of your app. Additionally we packed support for x86 platform (Intel) into the SDK.

Fluid tracking on devices with low computational power

A design goal of the re-implementation of the tracking engine was to make image tracking more fluid on devices that are not equipped with the latest dual or quad-core processor. Naturally devices like Smart Glasses are based on low-power consuming architectures, which result in computing power that is not comparable with current smartphones. The optimized Wikitude SDKs for supported Smart Glasses benefit greatly from the new tracking engine and deliver a solid performance when it comes to recognize and track images. But not only Smart Glasses benefit from this improvement – also smartphones like the iPhone 4 show a respectable performance improvement when it comes to image recognition and tracking. This lead us also to the point, that we lowered our minimum hardware recommendation for image recognition and tracking to include devices like the iPhone 4.

Snap your augmentations – Snap to screen

Losing a target image while watching a video on top of it can be quite annoying. With the Wikitude SDK 4.0 you now have the ability to snap your content to the screen regardless of whether the target image is sill visible. You can customize the snap-to-screen feature as you need it (react on click or automatically) and place the augmentation everywhere on the screen. The video below gives you a first idea of the feature. We also added two new samples to our Sample App showing off the feature. All in all, we’re pretty darn excited about this revamping of how we approach vision based augmented reality. It’s been quite an undertaking, and a warm and heartfelt thanks goes out to all the developers we’ve spoken to that have provided tremendous amounts of feedback. We hope you enjoy developing with the Wikitude SDK 4.0 as much as we have in the building of it. The all-new Wikitude SDK 4.0 is now available via the download page. As an existing Wikitude SDK customer you can upgrade to the new version (SDK LITE: 399€, SDK PRO: 999€). Customers who purchased a Wikitude SDK license after 1st of July 2014 qualify for a free upgrade (the existing license key will work with SDK 4.0).

Questions, comments? Let us know!

If you have questions, suggestions or feedback on this, please don’t hesitate to reach out to us via Facebook, Twitter or send us an email at hello@wikitude.com!

Contact us

More Blog Posts

April 6, 2021

SDK 9.7: Explore the improved Expert Edition for Unity

February 23, 2021

Powerful cylinder targets - Download SDK 9.6

Subscribe to our Newsletter

Stay up to date with the latest Wikitude news.

Subscribe now

December 2, 2020

Object Tracking 30% faster: Download SDK 9.5

August 25, 2020

New in: Explore Wikitude SDK 9.3

July 9, 2020

Now available: 3D Model Object Tracking

May 19, 2020

Introducing the next generation of CLOUD and SDK 9.1