Bing as a service: Bringing real world into applications

Microsoft is pushing Bing as a background platform for many of its products, including Windows 8.1. But after Build, Bing is also a component for your own applications.

In reality, Bing has long been available for your applications. The most common integration and demo scenario has always been maps: embedding Bing-driven maps in various scenarios, ranging from web sites, phones and Silverlight to Excel apps.

But now Microsoft is positioning Bing more and more as a high-abstraction-level cloud service that connects your applications with the world outside the computer. What they really mean is adding sight (OCR), speech (text-to-speech, voice control), linguistic understanding (translation) and external databases (search results, maps, synonyms etc.) to your applications.

Or as one of their slides puts it:


This makes sense, but isn’t that practical. The new Bing Developer Center puts more practical perspective into the equation.

The following services exist, but not all of them are new.


What’s new for Windows 8.1?

OCR – making text out of imagesimage

The keynote (at about 1h 25 minutes in) contained an elaborate demonstration of OCR. In practice, they took a photo of a business card for a restaurant and passed that over to Bing OCR service.

OCR is a new Bing service available through the Windows Azure marketplace. The idea is simple: you send in an image, and it spits out the parsed text along with their locations in the image (see the attached image for a demonstration). Excessively large or tilted images degrade result quality, but Bing works out minor blurs and other issues by itself.

Additionally, the team is shipping a Bing Optical Character Recognition Control for WinRT applications to make the UI building easy. The OCR service supports the following languages: English, Italian, German, French, Spanish, Dutch, Portuguese, and Simplified Chinese.

While in preview, 5,000 OCR requests per month are free. In fact, you even can’t pay for more at this time. The service is supposed to be generally available at the time of Windows 8.1 release.

3D Maps

imageMaps in Windows 8.1 will have 3D support, although this feature is not yet available in the preview version. Judging from the preview in the keynote, the quality of the 3D imagery is superior compared to e.g. Apple’s offering.

Microsoft promises that you will be able to embed the 3D maps into your application, but there is no word on which of the various Maps APIs will support which features. We’ll see.

image Note: In the keynote there was a demo (see the video at about 1:24:00) when the application was asked – using voice – “Who is the architect?” (of the building in view). The app could answer it by showing an information card of Santiago Calatrava.

Some of the bloggers report this as “Bing maps supporting voice queries”. I don’t think this is what was said in the keynote; rather, the voice understanding feature was a part of the “Road trip companion” demo app. But technically, the feature is powered by Bing services, as will be discussed further below.

In addition to these map changes, the Bing Maps API for Windows 8 applications has been bolstered with new classes and features such as drawing traffic data.

Voice recognition and text-to-speech

Both directions of voice interface are now available for Windows 8.1 applications, bringing Microsoft one step closer to reaching its Natural UI (NUI) nirvana.

imageThere is the Speech Synthesis API, which turns text into an audio stream. For this purpose, Bing technology is apparently integrated right into Windows – the synthesis process is entirely offline-based. The supported languages are English, French, German, Spanish, Chinese, Japanese, Korean, Italian, Russian, Polish and Portuguese.

The Voice Recognition part isn’t that well documented yet, but we’ll probably see more soon. Given that the Bing team’s goal is to make all platforms equal, it is probably reasonable to assume that the developer experience will be rather similar to what’s available on Windows Phone 8.

For more information on WP8 voice, check out the following links:

A word on the other services

Of the other APIs that didn’t get refreshed at this stage, particularly worth mentioning are the following.

MapsSearch is the classic Bing API, which comes in two varieties: web results only or full. First 5,000 queries a month are free, and further queries are priced at ~0,001 € apiece for the web-only version. The full version, which contains results from categories such as Images, News and Videos, is about 50 % more costly.

Translator ControlTranslation enables you to send a corpus of text and get it back, translated. This service is priced by characters translated per month. The first 2 million characters a month are free, and more can be bought at about 7,5 euros per a million characters. Supported translations are listed on the Bing Translator help page, but the translation quality between languages may vary.

Synonyms APISynonyms database is not a thesaurus, but rather intended to help parse user input by matching product, location and person synonyms. Their cited examples include “jlo” to “Jennifer Lopez”, “canon rebel t3i” to “Canon 600D” and “seatac” to “Seattle Tacoma International Airport”. Synonyms is in CTP like the OCR, and provides 5000 transactions / month for free. It is planned to reach production quality during August 2013.

Is this hot for you?

Most applications do not need these services, but for some, they may be indispensible. Maps is the first of the services to be almost universally useful – quite many applications benefit from geographic visualization or other map features.

But as time passes, natural UI technologies also improve and become more useful. Examples of Voice control scenarios are plentiful. For some applications, OCR and Translation may be killer features. Can you imagine a phone app that lets you to snap images of traffic signs and guideposts in foreign lands, providing immediate translations?

As the new Windows 8.1 search experience demonstrates, Microsoft is quite serious about bringing Bing into its products. Given all this information and past experience, it seems that leveraging Bing is a safe and well-supported approach to adding NUI features and real-world connectivity to your application.

Jouni works as a consultant focusing on Microsoft technologies and technology strategy. He is also a Microsoft Regional Director. Prior to his current job, he has an extensive background in development, IT administration and business management. He's been doing this for a living since 1995.

No comments.

Leave a Reply