Tech: New Text to Speech (T2S) use

Subject: Tech: New Text to Speech (T2S) use
From: JohnInKansas
Date: 26 Apr 07 - 11:20 PM

One of my regular magazines, Technology Review, has just announced a new policy that will make all articles posted at their website accessible as "machine synthesized audio" that you can listen to online or can download for listening on your "music machine."

Some of our people have mentioned having Text to Speech (T2S?) "voice" readers that can be used to "read a text" to them; but this is an apparently "new and different" feature – for a web site to provide the audio version using T2S technology. Of course, they claim to have a new and improved T2S technology that may also be of some interest.

The article on this new feature (which I first saw in the print version) is featured prominently at the current home page linked above, or you can go directly to Podcasts of Every Article for information. [I note that there is no audio of the information page about the new audio feature.] If you go to any article from the magazine, there should be a header at the top of the article where you can click to listen or download.

The audio versions when you click at the website appear to be provided as mp3 recordings. My dial-up can't support the "real-time" playback, but I can download the mp3 and play it back. The first article I looked at was a 6.9 MB download, which with my connection is about an hour to save. Playback (mp3) was at 128 KB/sec. It's obvious that it's a "synthesized" speech, but is understandable and apparently pretty accurate on a word-by-word basis.

A "subscription" and a software download are required to get the magazine's RSS feed which is separately available, but gives you automatic updates when an article you select is edited or updated. RSS generally isn't compatible with my connection speed, but others may be interested. For the RSS version, you can get a Linux compatible (Juice) version, or you can choose iTunes, Zune, or Juice versions for your Windows/OS-X players.

It must also be noted that some of the magazine's "best" articles are "limited access" and require some special qualifications or paid subscriptions, although quite a bit of good stuff is available for free public access.

Whether the articles at this magazine are of interest or not, those with an interest in the general technology of "audio versions of text" may be interested in taking a look at the AudioDizer web page to see whether it really is something new, and whether more widespread use might be something to look forward to, something to add to a website, or something to ask for at one of your favorites(?).

John

Subject: RE: Tech: New Text to Speech (T2S) use
From: GUEST,Mike
Date: 28 Apr 07 - 01:25 PM

John,

Nice article - I agree that they could have put some audio file to describe the information on Technology Review. I think overall idea is great. You can subscribe for free to the magazine through mp3 audio on iTunes or on their website. So, I have a choice to listen or read the articles.

Maybe you should ask AudioDizer/Technology Review to create a lower bandwidth file so that you can listen to the content?

Subject: RE: Tech: New Text to Speech (T2S) use
From: JohnInKansas
Date: 28 Apr 07 - 02:27 PM

G Mike -

Technology Review is a long-established magazine and has been on the web for quite a while. A free signup/login or just a visit to the site gets you some of the articles, but at least in the past a few articles may require a paid subscription, or individual article purchase (or a degree from MIT and participation in the Alumni Association).

The AudioDizer is a "student generated startup company" presumedly based on or developed out of some class/project activity at MIT where Technology Review originates.

I was a bit disappointed at the sparse information at the AudioDizer web page, but it appears their main targeted customer base is websites that may want to provide the audio conversion.

If you click any of the "real articles" at the magazine site, you can try out the voice version - as you may have done. You can click for audio on any individual article, or you can sign up for an RSS feed for related stuff, I believe. You can visit the site any time.

I haven't looked at any back issues to see if the audio feature has been automatically applied there.

John

Subject: RE: Tech: New Text to Speech (T2S) use
From: GUEST,.gargoyle
Date: 28 Apr 07 - 02:55 PM

MACs have had the feature for ten years - you can even select voices for the reading like Duke of York or Burl Ives.

Sincerely
Gargoyle

Subject: RE: Tech: New Text to Speech (T2S) use
From: GUEST
Date: 28 Apr 07 - 03:23 PM

FROM http://tips.vlaurie.com/

Google 411

April 26th, 2007

As noted before, Google is exploring voice-powered search. The company has a page that explains one of its current efforts:

Welcome to Google Voice Local Search Google Voice Local Search is Google's experimental service to make local-business search accessible over the phone.

To try this service, just dial 1-800-GOOG-411 (1-800-466-4411) from any phone.

Using this service, you can:

search for a local business by name or category. You can say Giovanni's Pizzeria or just pizza.

get connected to the business, free of charge.

get the details by SMS if you're using a mobile phone. Just say text message.

Subject: RE: Tech: New Text to Speech (T2S) use
From: JohnInKansas
Date: 28 Apr 07 - 07:14 PM

Gargoyle -

Windows XP has had text-to-speech capabilities built in since it's introduction (2001), but like the one from Mac the results people got from it were variable. Some people could get useful results and for others it was just another pain-in-the-ass "feature." Similar programs were available with Win2K and possibly with some earlier Windows versions, but were third-party or separate download packages.

Some aftermarket (Third Party) programs have been able to give quit consistent and useful results for some time, althugh I don't know the detailed history of the better ones.

The difference with the program here is that the website can incorporate the spoken version as something available to all who visit the site, and it's not necessary for the "user" to make any arrangements, do any special setup, or have any program other than mp3 playback capability. (And a fast enough connection to be able to download rather large mp3 files at useful speeds.)

John

Subject: RE: Tech: New Text to Speech (T2S) use
From: GUEST,.gargoyle
Date: 28 Apr 07 - 11:22 PM

Three blind friends (the are probably my friends because they are blind) ALL use MACS and its built-in features for audio web-browsing (a teacher, a keyboardist, a piano tuner)

IBM's Dragon Naturally Speaking (crap for S2T) is wonderful for text 2 speaking and it has the accents and any voice you want to mimic has been composed.

A Finance Friend for at six years (before PodCast) would pick up the front page of the WSJ online edition....process and burn a cd while he showered/shaved and listen to it on the 45 minute commute to work.

This is OLD Tech

Sincerely,
Gargoyle

Subject: RE: Tech: New Text to Speech (T2S) use
From: GUEST,Mike
Date: 30 Apr 07 - 08:30 PM

Gargoyle,

I agree that the technology is not new - i think the thing that is new is the automation behind publishing an article and getting the mp3 file so that it has music, a commercial, and multiple voices. I don't think your friend was doing all of that.

This production makes the files easier to listen to and allows magazines like Technology Review the ability to podcast all their articles.

I agree with John - this process eliminates all the work required for a user. now i just subscribe via iTunes and no work is required on my part.