Forum Moderators: bakedjake
A Scottish computer speech technology company has launched what it claims to be the world's first talking search engine.
[theherald.co.uk...]
Search results will be in an American voice, but newsfeeds will be in a Scottish accent.
Hmmm ... is this the future of speech synthesis?
Ummm ... no ... but if you were wondering whether speech synthesis could possibly be applied to dynamic content, well then it's not a bad demonstration.
Now, we have a mechanism to parse a dynamic web page and break the output down into syllabic phrases which could then be interpreted.
What about Farsi? Or Hebrew? Or Spanish?
How about adding an avatar to mouth the "words"?
Maybe a local search implementation with a custom avatar (an animation of YOU) housed on your hard drive that also accesses your local files and emails and allows you to rummage for a beer in the fridge while shouting out "cheap sailboats" and "next page" and "THAT one!" and "BUY IT!" ... without ever once looking at your screen?
I embrace and welcome this attempt. More before and after, I'm sure.
I decided to respond to some of the comments you have been making.
First of all you should note the big 'Technology Preview' stamp, this is equivilent to Google's BETA label.
The technology is still in development and is pretty good most of the time.
Doesn't accessibility software for the blind do this sort of thing already?
I really hope they refine the voices to be smooth.
For some reason, it kept skipping the 1st result when reading out the SERPS.
At present they are using Google for results, however their backend supports several other search engines.
Another excellent use for this technology which you have all missed is for people who do not speak English as a first language, people who can understand spoken English but can not read or write it. Speegle currently gets a very large number of hits from Japan and several other countrys.
For another example of why it is useful try searching for:
site:howstuffworks.com engine In regard to the comments about what do people do after they've found a result...
Speegle can read out (most) arbitrary webpages, they just don't let you do it yet.
Speegle is getting a huge number of hits and searches at the moment, which puts a lot of strain on the server.
They are currently working on sourcing additional hardware and bandwidth to provide a smoother and less laggy service.
Went back for a second look and test, and have to say, I like the keyboard shortcutting on the page. In fact, that alone impressed me more than anything. Such a simple thing to add, and yet you're the first to do it.
I've gone back and whacked a couple of searches out, and it still seems to be missing the first link or more when it reads out the results, this seems to be in direct correlation to overall response time, so I'm thinking it's somehow related to server load. The higher the server load, the longer it takes for the speech synthesis to kick in, and the more of the top few results you miss.
So far, worst response was the "read" starting at response #4.
BTW: This is not to be taken as a complaint, just an indicator of what the end user might be seeing/hearing, so you can work out bug fixes.
What caught my attention was that I have other speech software (with different voices) on my machine. And the search engine used it's own voices (voices that I don't have installed on my machine). So, I wondered if it was because I already had speech software installed or was the activex component it downloads the only thing needed. But you already answered in your post.
Thanks for the preview!
BTW, I'm not blind or anything that requires acessability. I just liked the concept that I could just listen instead of read. I guess maybe I'm just lazy.
To be useful for blind/visually impaired people shouldn't it have the ability to read aloud pages it returns in its search results? E.g. read aloud a cached version of the page? The best option would be a toolbar that could read any page it comes across.
The best results would be a merger of this type of technology with voice recognition/data input technology. No more using that annoying tiny keypad on my mobile/pda! Instead I could speak commands to my mobile ("search for hotmail.com"), listen to the results it finds, then dictate an email to it (software inputs/understands html forms) and send it. Role on the future! :)