Has anyone ever reverse engineered narrator.device?

[Dunny]

Originally Posted by Thomas Richter

Use the source, Luke!



http://aminet.net/package/util/libs/translator42


http://aminet.net/package/dev/src/trans42src


Essentially, it is a big list of rules, along with an even bigger list of exceptions, by word. I remember that I played a lot with it attempting to create a somewhat better version of Geman translator rules - without really succeeding.



Unfortunately, a lot of the melody of speech are built into the narrator.device, and thus not accessible on the level of the translator. This is why every language sounds as if an american attempts to speak a foreign language (or, shortly, "not convincing").

Oh, thanks for that - it's a good start. Just need to figure out the actual speech synth now. I understand that translator.library basically converts english words to phonemes, hopefully I can figure out how to generate them as waves

[Thorham]

Originally Posted by Thomas Richter

This *IS* a toy - it was used (as on the Mac) as a tech-demo for the Amiga.

A tech demo of what appears to be a serious product is hardly a toy

[Bruce Abbott]

Originally Posted by Matt_H

As I recall, Cloanto negotiated a new agreement with SoftVoice some years ago so that they could include narrator.device, etc. with Amiga Forever.

Thanks for that. So that confirms it - Cloanto do have a license to distribute narrator device with Amiga OS.

Originally Posted by Dunny

Oh, thanks for that - it's a good start. Just need to figure out the actual speech synth now. I understand that translator.library basically converts english words to phonemes, hopefully I can figure out how to generate them as waves

Here's something that might help,

SPO256-AL2 Sample Pack

This is a set of waveforms recorded from the General Instruments SP0256 speech synthesizer chip. They might not match up very well with SoftVoice's samples in the narrator device, but have the advantage of being freely distributable for non-commercial purposes.

Another chip that might be worth looking at is the Votrax SC-01 (I had one of these in 1981). Here's an emulator for it with samples:-

votrax-speak

[Thomas Richter]

Originally Posted by Bruce Abbott

Here's something that might help,

SPO256-AL2 Sample Pack

This is a set of waveforms recorded from the General Instruments SP0256 speech synthesizer chip. They might not match up very well with SoftVoice's samples in the narrator device, but have the advantage of being freely distributable for non-commercial purposes.

The narrator.device is not sample-based, neither this chip. The samples really help little - the narrator.device does not work by sequence sample by sample, but also computes the transition from sample to sample, and modulates the pitch of the voice to create the right voice melody.

[Dunny]

Originally Posted by Bruce Abbott

Thanks for that. So that confirms it - Cloanto do have a license to distribute narrator device with Amiga OS.



Here's something that might help,

SPO256-AL2 Sample Pack

This is a set of waveforms recorded from the General Instruments SP0256 speech synthesizer chip. They might not match up very well with SoftVoice's samples in the narrator device, but have the advantage of being freely distributable for non-commercial purposes.

Another chip that might be worth looking at is the Votrax SC-01 (I had one of these in 1981). Here's an emulator for it with samples:-

votrax-speak

And again many thanks. I've emulated the SPO256 previously for my currah uSpeech emulation, but as Thomas points out it's sample based and very poor compared to Narrator (which is a pretty low bar to clear, it has to be said but there it is!).

I may go a sample-based route for my speech synth, but I'd prefer not to if possible.

[Bruce Abbott]

Interesting podcast interview with Mark Burton (creator of SAM) in 2020.

"Mark Burton: ...we were starting to do a German version, then when they started putting text to speech systems into Windows, as a standard thing for free - that was bad... we were done".

"ANTIC: The SoftVoice text to speech website is still up, but it looks like it hasn't been updated for a long time, and ... nobody responds to those emails.

Mark Burton: No, nobody does... I believe he kept that open - and he kept the SoftVoice corporation alive - for medical insurance purposes."

"ANTIC: A few years ago... somebody reverse-engineered SAM, figured out it worked, and then ported it to C. And there's a couple of versions of that online... exactly like SAM.

Mark Burton: Yeah, it's pretty exact - I've heard that."

"Mark Burton: We asked Steve Jobs, do you want an exclusive - and he said no, [he] didn't care if we licensed it to others. So that was great because right then the Amiga came into existence, and they wanted the same thing. So Macintalk on the Mac is narrator on the Amiga - it's pretty much the exact same software.... and we actually collected a royalty per Amiga sold - so that was a nice ongoing revenue stream for a while..."

discordier / sam

This is a vanilla Javascript port of the Text-To-Speech (TTS) software SAM (Software Automatic Mouth) for the Commodore C64 published in the year 1982 by Don't Ask Software (now SoftVoice, Inc.).

It is based on the adaption to C by Stefan Macke and the refactorings by Vidar Hokstad and 8BitPimp

It includes a Text-To-Phoneme converter called reciter and a Phoneme-To-Speech routine for the final output.

It aims for low memory impact and file size...

vidarh / SAM

Sam is a very small Text-To-Speech (TTS) program written in C, that runs on most popular platforms. It is an adaption to C of the speech software SAM (Software Automatic Mouth) for the Commodore C64 published in the year 1982 by Don't Ask Software (now SoftVoice, Inc.). It includes a Text-To-Phoneme converter called reciter and a Phoneme-To-Speech routine for the final output. It is so small that it will work also on embedded computers. On my computer it takes less than 39KB (much smaller on embedded devices as the executable-overhead is not necessary) of disk space and is a fully stand alone program. For immediate output it uses the SDL-library, otherwise it can save .wav files.

Adaption To C

This program was converted semi-automatic into C by converting each assembler opcode...

License

The software is a reverse-engineered version of a commercial software published more than 30 years ago. The current copyright holder is SoftVoice, Inc. (www.text2speech.com)

Any attempt to contact the company failed. The website was last updated in the year 2009. The status of the original software can therefore best described as Abandonware.

Drawing speech: Software Automatic Mouth

[Don_Adan]

If my memory not failed, I remember also french speech (text to phenomene). It was adventure game, perhaps from Lankhor.

[saimon69]

Originally Posted by Don_Adan

If my memory not failed, I remember also french speech (text to phenomene). It was adventure game, perhaps from Lankhor.

Maupiti Island is the title you are looking for

[Don_Adan]

Yes, but Delphine Software also uses text to speech option.

[ Show youtube player ]

I played only in english version of Operation Stealth, then dont know if also french, german, italian and spanish version has this option.

[Thomas Richter]

No, they didn't. That was a third party hack that never really worked reliably.