Jump to content
  • entries
    143
  • comments
    451
  • views
    173,671

Speech Synthesizers


mos6507

2,123 views

I've actually had a lot of experience working with speech synthesizers. Back in the day I had S.A.M. That program completely captivated me. I'd say it was one of the best things I remember about the Atari 8-bit. It's also why I was sorely disappointed by the 1400XL prototype I picked up when I learned that its Votrax chip did not allow you to adjust all of the vocal attributes. I had always wished I had an Alien Voice Box because that thing supposedly could sing. I think that used the same chip as the 1400, but implemented properly.

 

My first dot com job I made Microsoft Agent characters. The API for these guys includes speech synth (and limited speech recognition). However, people mostly rememeber them for clippy, where they got their bad reputation.

 

clippy.png

 

Here is one of the guys I made using Lightwave 3D:

 

sizlik.gif

 

It's hard to describe what the business model was at that startup. They thought they could use these characters composited into analog television where they would stream in over the vertical blank. The ATI All in Wonder cards at the time had WebTV software from Microsoft that you could load up on your PC to create a kind of web sandbox with interactivity that would stream in this way. The video could then become a DirectX overlay. They were going to go around to local stations and try to sell them a special server that would constantly loop and stream data through a router which would do the VBI insertion. We ran one semi successful test with KWHY TV 22 in Los Angeles.

 

Anyway, we used to make these little presentations using the characters kind of like a puppet show. These were supposed to be like faux news broadcasts, so we made some "news anchor" characters like this one:

 

Cathie.gif

 

At the time I thought it was kind of neat that I was able to use some of my skills from S.A.M. in my career.

 

When the VecVox and AtariVox came out, it was clear to me that it was the direct successor of something like the Alien Voice Box. To date I've been disappointed with the quality of AtariVox speech in homebrew games. Even the powerup sound of the AtariVox seems off to me. It sounds like it says a-(as in cat)-tarivox instead of a-(as an cut)-tarivox. Maybe I'm ultra-sensitive to phonemes (well, SAM calls them phonemes, AtariVox calls them allophones). Maybe not everyone has the same enthusiasm for it, or the affinity. Instead of whining, though, I just volunteered to rescript the speech in Juno First for Chris Walton and so far he likes what I'm coming up with. Hopefully everyone else will as well. Maybe I'll carve a little niche for myself as the speechman for homebrews that use AtariVox. That would be nice.

 

avox-pic.jpg

 

 

UPDATE:

 

cd-w gave me a little 4K program so you can compose speech using a Cuttle Cart or Supercharger rather than relying on the USB interface. I also made an AtariVox header file for it that makes it easier to string allophones together. Let me know if this is useful.

 

 

helpme.zip

10 Comments


Recommended Comments

Maybe I'll carve a little niche for myself as the speechman for homebrews that use AtariVox. That would be nice.

It would indeed. That's one of the reasons I've been pushing for various means of making developing speech for the AtariVox more accessible to more people. We've had various people assist with music and graphics for 2600 homebrews, so why not speech?

Link to comment
To date I've been disappointed with the quality of AtariVox speech in homebrew games.

 

Have you tried Strat-O-Gems? There's a lot of speech in that one.

Link to comment
To date I've been disappointed with the quality of AtariVox speech in homebrew games.

 

Have you tried Strat-O-Gems? There's a lot of speech in that one.

 

Frequent speech. The phrases aren't that long. I'm not bitching. It's just that I'd like to see a game with more elaborate speech. I haven't tried Man Goes Down yet. Supposedly that has a lot.

Link to comment
Even the powerup sound of the AtariVox seems off to me.

That's my British pronunciation :cool:

 

BTW the start-up can be changed. I can instruct you how to do it if you want ?

 

 

I had always wished I had an Alien Voice Box because that thing supposedly could sing.

AVox can sing too.

 

 

 

Wouldn't it be cool if someone used the TTS algorithm in a game.

Link to comment
AVox can sing too.

 

Though it's odd that they call it a "three octave range" when the thing can't even sing up to middle C. I wonder what the limitation was? A range from two octaves below to one octave above would have been very nice. As it is, Strat-O-Gems wraps the pitch for "NINE" down an octave.

 

Wouldn't it be cool if someone used the TTS algorithm in a game.

 

How many Atari 2600 games have any meaningful amount of text that couldn't be stored just as well phonetically?

Link to comment
How many Atari 2600 games have any meaningful amount of text that couldn't be stored just as well phonetically?

 

Those that generate a lot of text dynamically.

Link to comment
Those that generate a lot of text dynamically.

 

Any examples?

 

In general, not thinking about the 2600 in particular...

 

- An expert system like Alice that is able to construct sentences on the fly.

 

- Redirecting network text chat to TTS. That was something we were working on with the Microsoft Agents, using them in a multiuser chatroom. The problem was the speech couldn't overlap and the speed of speaking was a lot slower than people can read so it would lag behind the text stream. But I digress.

 

Don't get me wrong. TTS will never match the quality of building speech phonetically and it should normally be avoided. It does have uses, though.

Link to comment
In general, not thinking about the 2600 in particular...

 

I can imagine text-to-speech perhaps being useful on platforms other than the 2600, though there aren't a whole lot of cases where a phonetic system wouldn't be better (in a multi-player game, rather than text-to-speeching each other's taunts, just use microphones). I can't think of any plausible 2600 game where text-to-speech would be useful, though.

Link to comment
Guest
Add a comment...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...