Hound is now available on iOS, a voice search app with finer word segmentation than Siri

Hound was born out of the music radar application SoundHound. The iOS version was officially launched recently. The application is free and supports iPhone and iPad.

Music recognition was Hound's first step. The team conducted in-depth research on speech recognition and natural language processing technologies, hoping to make Hound a voice assistant similar to Siri, or in other words, a faster, more accurate, and better voice assistant than Siri.

Hound now provides a variety of search results, including weather, hotel reservations, navigation, local search, stocks, time zones, geographic information, calculations, and even food nutrition facts.

When you open Hound, a card-like text similar to Google Now will prompt users on how to express their voice needs on different topics. There is a prominent voice button at the bottom of the interface. In addition, Hound also has a voice activation function, which can be awakened by calling "OK Hound".

But in fact, what really makes Hound popular is its finer granularity in understanding natural speech. For example, when I ask Siri, "Show me Asian restaurants nearby, but not Korean or Japanese, that are open after 8PM", Siri will list some Japanese, Korean and other Asian restaurants, while Hound will recommend nearby Asian restaurants other than Japanese, Korean, etc. to me.

You see, this subtle difference is enough to make people feel that Siri is lacking in the role of a personal assistant. Moreover, on 9to5Mac, there is a video comparing the functions of Hound and Google Now. When sending a voice request at the same time, Hound's processing speed is obviously faster than Google Now, and the result interface presented is more user-friendly.

In fact, in June last year, Hound was launched on the Android internal version. At that time, it was mentioned that Hound has strong performance in some aspects, such as being able to understand complex and long sentences, understand continuous questions about different subjects, and have low latency. The latter is because Hound combines cloud recognition and natural language processing in one engine.

In addition to Hound, there are actually quite a few new players in voice search, all of which are trying to "beat Siri."

For example, at the end of last year, speech recognition technology company Semantic Machines announced a new round of financing of $12.3 million. Its chief technology officer, Larry Gillick, used to work for Apple Siri. Semantic Machines said that it hopes to do something beyond Siri and will also be ahead of other digital assistants on the market. Moreover, Semantic Machines can not only understand commands, but even conversations.