Joi Ito's Web

Joi Ito's conversation with the living web.

Dan Gillmor came over to my house yesterday and we were going to talk about the book he was writing, "Making the News". He asked me whether I had read the outline... no... Oops. Sorry! We ended up making another appointment and spent the rest of the time geeking out. Anyway, that's not the point I want to make.

I felt really guilty, and came up with a great idea. I opened my mac in my car and got it to start reading the outline to me using text to speech. This is the first time I really tried it and it worked well. (Although the woman's voice reading Dan's words was a bit weird.) So then I thought about this some more. What I really want is a text to speech to mp3 converter that took my RSS feed and dumped it into an mp3 file that I could listen to my iPod on the way to work.

The other thing I could do is make my blog available in mp3 format. Has someone already done this? Is there a text to mp3 tool somehwere? Is this a stupid idea?

Update: Dan just blogged about yesterday. He writes about our geeking out. What we did was get ssh and port-forwarding running on his Mac. I totally don't understand why more people don't use ssh port-forwarding for mail and sftp and scp for file transfers. Welcome to the society of ssh lovers Dan. ;-)


You're not the first one to have this idea, I saw this exact idea on LazyWeb, March 3rd.

joi, what you can do is get a tts translator to convert your file into an audio file (typically .wav), and save it in a temp file. Then use one of the scripts out there that will command line convert it into an mp3, and timestamp it for you. We've pieced together a lot of stuff like this to launch a cool mobile product where a person can sing a song (karaoke) with audio backing into their phone, or record a personal voice message mixing some funny fx like a helicopter, farting, gun shots, etc. in the bkgd, which is then converted into an mp3, which is then incorporated into a multimedia flash file (music artist graphics, lyrics, a cartoon, or some marketing image), which is then automatically sent to a friend's email or MMS as a personal greeting.

here's one

and another

and another...

but these are standalones, don't know if you can strip away the GUI and just use the kernel. i prefer stuff that's command line controlled in order to run off our mobile platform.

good luck!

This is not a stupid idea at all.

The more I think about the idea the more cool ideas I think of doing with it (I can already see tags that let me pick different voices for each post (see in my mind that is) ).

Here are three quick thoughts I have:

1) I see this as another way to create an audioblog. Actually, a way of creating an audioblogging post that makes for easy editing.

2) Many of the present bloggers feel more confortable writing than speaking into a microphone (I say present because I believe there are many future audiobloggers that would prefer to speak over writing). If the audio output of the process you describe is exceptable, it becomes an efficient, productive way to produce audio from an already written post without having the blogger doing extra work. Plus they don't have to hear their own voice which some just can't get used to.

3) Possible method that would assist in enabling mobloggers and basebloggers that speak different languages into the blogging conversation.

I would say maybe the idea is just a little ahead of its time like audioblogging technology in general.

The day is not far off when mobloggers will need to part of the blogging conversation. I believe when moblogging two-way converations become real, audio produced with methods as you described will play a role.

I use iSpeakIT didn't see it mentioned at LazyWeb.

Voice Box has some example AppleScripts for fetching RSS feeds and speaking it. I think it does AIFF only, but the iPod plays AIFF.

Anyway, should be trivial to script "import into iTunes and sync iPod" afterwards


I let my Mac read my inbox aloud, and record it to MiniDisc, while I get ready to leave. I then listen to it on the way to work, with my MiniDisc hooked up to the car stereo. The MiniDisc recorder is auto start, so I just pause it for record, and when the Mac starts reading aloud, it starts recording...when the Mac stops, so does the recording.

You could use Audio Hijack Pro to grab the sound, and then drop it into your iTunes music folder. Sync to your iPod and go.

Audio Hijack Pro had a piece on creating a VoiceXML application from an RSS feed a little while back. This is quite easy to do with the Perl XPath module and an "//item" expression. I recently wrote something similar, and was able to make a male voice TTS by using Bevocal's platform and setting voice="male" inside the prompt tag.

Voice technology is steadily coming along. I often thought it would be nice to listen to the NY Times in my car on the way to work. Maybe soon I can even pick the voice talent I want it read in. Al

One thing that might come in handy is a style sheet for the audio output. Something that emphasizes headlines in one way, bold text in another and hyperlinks in yet another way. Yunno, different style sheets for different browsers.

Based on this posting, and in need of a distraction, I added this functionality to Any blogs on will now have an autogenerated audioblog in mp3 format of the latest entries. Go read about it at this posting. It's kinda fun...

I think that this is a terrific idea.

Lots of GREAT feeback. Thanks everyone. I ended up VoiceBox and AppleScript as my solution. I've got Dan's outline on my iPod and I'm a happy camper.

Jamie, that was fast! Nice feature.

Yes, audio is still in its infancy. That's why it's so fun to explore. ;-)

ROBOT RADIO has all the CLI tools you need for Mac OSX. I just don't know how to put the pieces together.


Great idea! I thought about something in a similar vein a couple years back when I was working for a language technology company in Tokyo, but the first Rio was just starting to take off, MP3's weren't a de facto standard yet, and the text-to-speech technology was still rudimentary. I thought it would be a great application to be able to review information in an audio format daily on the long Chuo-sen ride home - whether that information came from emails, news, webpages, or documents.

Bottom line on this one: great to see this exploding into feasibility, especially with weblogs! I'll have to try it out.

Of course it would be a lot cooler if every source had its own voice (specified in the XML file... nice). Only 2-3 Macintalk voices are bearable for listening to longer texts.

VoiceBox supports pitch and rate settings per voice, also voices can be switched, so there is some margin to fiddle.


Great minds think alike :)

We've got this working - it is pretty cool. Look for a release of the script this Tuesday.

Read more about our RSS to iPod AppleScript on my blog.

Good evening, just launched this week. Talkr monitors RSS feeds and converts them to podcasts.

We have a program called Talkr Partners ( to provide fully hosted podcasts. We would be very happy to host a podcast of your blog -- if you're interested drop me a note at cbrooks at talkr daught com


Here is another product that will work on your RSS feed - Text Speaker text to speech software for your pc. It has customizable pronunciation, reads anything on your screen, and it even has talking reminders. The bundled voices are well priced and sound very human. Voices are available in English, French, Italian, Spanish, German, and more. Easily converts blogs, email, e-books, and more to MP3 or for listening instantly.

1 TrackBacks

Listed below are links to blogs that reference this entry: Blogaud - I want to listen to my RSS feed on my iPod.

TrackBack URL for this entry:

Weblog 之丰富体验 生活会从每个Blogger 的角度反映出来其本来的面貌,或更丰富它,这会随着Blogger 群体的扩大而更加鲜明。在早期我们讨论Blog的时候,曾经非常欣赏哪个城市(例如NewYork)的B... Read More