All posts by eolvera

Eduardo Olvera is a Senior VUI Designer with a passion for Information Architecture, Usability and Dialog Design, for touch-tone and speech recognition systems, in English, Spanish and French. For more information:

Vista Speech Recognition – The new voice of hacking

I’ve recently being readying about all the ongoing efforts to improve the security of Microsoft‘s new Vista Operating System, but I never thought I would find out that one of the first flaws being publicized since the public launch of Vista earlier this week would have to do with its Speech Recognition capabilities.

Apparently an attacker can run malicious programs by using prerecorded verbal commands, ultimately meaning you can “shout-hack” a system.

“Microsoft has also recommended that users who are concerned about having their computer shout-hacked should either disable the speaker or microphone, turn off the speech recognition feature, or shut down Windows Media Player if they encounter a file that tries to execute voice commands on their system.”

Full article

Top 10 Survival Strategies for VUI Design in Spanish Applications

By popular demand, here’s a list of the Top 10 Strategies VUI Designers can follow to design a Spanish application, as published in SpeechTEK’s magazine:

  1. Make Spanish as important as any other language
  2. Apply Spanish Marketing 101: understand your market
  3. Follow user-centric design methodologies: combine usage-case scenario analysis with actual observation
  4. Seek professional advice on localization: translations aren’t enough, consider all language, cultural and branding aspects
  5. There’s no such thing as “neutral Spanish”: carefully design your persona and coach your voice talent
  6. Anticipate recognition, tuning and text-to-speech challenges early: Spanish modules and tools aren’t as robust as English
  7. Adapt the transcription process: specific transcription conventions for multilingual utterances
  8. Eliminate duplicate documentation: keep all languages synchronized in a single document
  9. Context, context, contex: localize the entire dialog and check and check for coherent flow
  10. Identify divergence during development: adapt code to dialogue, not the other way around

You can also download the full article: Bueno – Are you listening to your Spanish Speakers

Voice-Recognition – Huge Opportunities

In this month’s article from Business2.0 titled “Now You’re Talking”, Jeanette Borzo performs an interesting analysis of the industry, it’s current status and its future.

In particular, there are a couple of aspects I found intriguing:

  1. Nuance getting more PR related to the competition against Ben cook, the 17-year-old Guinness Book of World Record holder for text messaging, where dictation proved to be 3 times as fast (you can watch the video below). Unfortunately they forgot to mention the fact that Nuance had to perform grammar customization and tuning to the system prior to the competition, which again causes the misconception on the market that speech recognition can flawlessly work out of the box.
  2. Tellme continues to appear as an established player, while BeVocal wasn’t mentioned at all.
  3. Most new players coming into the market are not related at all to the classical speech recognition implementations (a.k.a. Call Centers) but rather are targetting new markets such as video/audio search, personal communication devices, car navigation systems and real-time translations
  4. The rapid growth of the voic-recognition technology market, exceeding $1.2B in sales in 2006, estimated to reach $2.6B by 2009, yet most speech platform and service vendors still seem to struggle to find new customers and start new projects.

Welcome! Bienvenido! Bienvenue!

Welcome to I got tired of reading and listening to self-proclaimed “gurus” of Voice User Interface Design, in particular around the Usability area, and decided it was time for me to make a stand!

I can’t believe most of the “expert advice” given by them is nothing but things that the web community has learned, tested and improved over a good number of years to the point where they are well-known standards.

That’s right. Even though it may sound as if they are “visionaries” in the telephony field, most of the time it just requires looking around at what’s happening in other fields or over the last few years on the web world, and voilá, we can see it’s nothing more than history repeating itself on a new medium (something like the battle we’re currently seeing between Blu-Ray and HD-DVD… anyone remembers Beta and VHS? )

Anyway, my objective is to share with you those things that worked and that failed on the field, interesting development and articles from the VUI community, and any other design “similarities” that I run across everyday: things I read, website I see, places I go, products I use, etc., which in my opinion provide us with “hidden gems” that can be generalized and applied to any Design to make them better and easier to use… in particular, in the field of Voice User Interface Design.