twitter radio, another sneak peak

[Matt L. Hackett]

HI all-

As the new Moon approaches, I wanted to give you all another taste of
TweetRadio, now more audible, more in English, and with lots of
complicated nerdy behind-the-scenes improvements that I won't get
into.

http://bit.ly/twradio_beta [links to an MP3]

Feedback is appreciated. Also, if you have any more ideas for the
"stock phrases" that are used to make the feed more radio-call-in-
like, that would be helpful. (For example, right now these are things
like "___ issued a statement that ___" and "Here we have a shout out
from ___ in ____ to ____, ____.").

Cheers,
-- Matt

[Zack Sultan]

THIS IS SOOOO FUCKING COOL.

[Bradley Hope]

This is very cool. I think we should make it open source --- how
exactly does it work, anyway?

[Matt L. Hackett]

I'm definitely a supporter of open source, but at this point it's not enough of a well-polished piece of software that I'd feel comfortable having the source out there with my name on it. It may get there as I finish up, we'll have to see.

I'll be writing up both a technical and a prose blurb about the project for the site, the former explaining how it works. The very simple version is that there are three components (one of which is not yet done):
1. A service that collects tweets, filters, scrubs, reformats them, and saves to a database. Google's Language APIs are used to pick out only coherent English tweets, amongst other text things.
2. A second service pulls tweets from the database and puts them together into a speech-ready form of mark-up (essentially HTML-for-speech-synthesis), then feeds this to an open source speech synthesis engine called Festival
3. A specially configured web server that serves up processed sound files

[Steven Hasty]

Our twitter radio "host" could also solicit calls-in via http://www.tweetcall.com/
and read aloud the ones in @reply.

On May 1, 11:36 am, "Matt L. Hackett" <mlhack...@gmail.com> wrote:
> I'm definitely a supporter of open source, but at this point it's not enough
> of a well-polished piece of software that I'd feel comfortable having the
> source out there with my name on it. It may get there as I finish up, we'll
> have to see.
>
> I'll be writing up both a technical and a prose blurb about the project for
> the site, the former explaining how it works. The very simple version is
> that there are three components (one of which is not yet done):
> 1. A service that collects tweets, filters, scrubs, reformats them, and
> saves to a database. Google's Language APIs are used to pick out only
> coherent English tweets, amongst other text things.
> 2. A second service pulls tweets from the database and puts them together
> into a speech-ready form of mark-up (essentially HTML-for-speech-synthesis),
> then feeds this to an open source speech synthesis engine called Festival
> 3. A specially configured web server that serves up processed sound files
>

> On Fri, May 1, 2009 at 5:45 AM, Bradley Hope <bradleyh...@gmail.com> wrote:
>
> > This is very cool. I think we should make it open source --- how
> > exactly does it work, anyway?
>

> > On Thu, Apr 30, 2009 at 11:59 PM, Zack Sultan <zcsul...@gmail.com> wrote:
> > > THIS IS SOOOO FUCKING COOL.
>

> > > On Thu, Apr 30, 2009 at 10:38 AM, Matt L. Hackett <mlhack...@gmail.com>

> > > wrote:
>
> > >> HI all-
>
> > >> As the new Moon approaches, I wanted to give you all another taste of
> > >> TweetRadio, now more audible, more in English, and with lots of
> > >> complicated nerdy behind-the-scenes improvements that I won't get
> > >> into.
>

> > >>http://mhackett.net/tweetradio/twradio_beta_slt_sable2.mp3[links to an MP3]

[Matt L. Hackett]

Though this example doesn't show it, the application actually does
take @replies to @tweet_radio and render them in the stream -- lots of
interesting possibilities here. TweetCall makes that even better.

[diana kolsky]

http://www.youtube.com/watch?v=HrlS9_n8GX4