Anyone tried to MIDIfy any vintage speech synth chips using a MIDIBox core? I’m thinking of the old SP0256 and Votrax chips that used to be available as addons to personal computers in the '80s, and were apparently also used in some arcade games.
a|x
Hi.
I remember playing around with the SP0256 20 (odd) years ago! It was find but pretty basic.
There is a MIDIbox project that has similar (but much improved) functionality based around the SpeakJet IC.Take a look at http://ucapps.de/mbh…c_speakjet.html for more information and http://www.midibox.o…idibox_speakjet
Cheers
Phil
Hi ,
actually seems that midification of speech syntesisers with MBHP is only for Speakjet.
I used years ago SPO256AL2 to give voice to VIC20 and ZXSpectrum and more recent on LPT port
of my old 486 and used to sample it on my old S1100
Time ago I thought the same project as this chip are very easy to manage ( at least 6 bit word and two
control lines ) but I am a BASIC programmer
and actually I havent enough skills for design such an application.
I think it could be interesting project…another way to generate chip voice 
Anyway I can give hardware effort
Regards
Antix
Hi Phil, antix, thanks for the rapid replies.
I know about the SpeakJet project. I’m about to build one based on TKs new version of audiocommander’s original code, tweaked to run on the LPCXPRESSO board. I was just wondering if anyone had done something similar for any older chips.
I have the datasheet for the SP0256-AL2, and it looks like getting it to output basic allophones wouldn’t be too challenging, for someone who knew what they were doing (unlike me). It doesn’t seem to have any information about pitching the oscillators though.
The Votrax also sounds intriguing. There’s apparently a patent application from Kraftwerk (no less) for a circuit to add MIDI control for this IC. I have a suspicion it’s what they used on their remake of Autobahn from their The Mix album.
a|x
The MBHP_CORE_LPC17 would be powerful enough to handle speech synthesis by itself - in background (because it wouldn’t really time consuming work @ 120 MHz…)
It even supports a I2S output for an audio DAC at 48 kHz (stereo)
Problem: I don’t know free (open source) code for such a synthesis, and developing it wouldn’t be a task of one evening…
Best Regards, Thorsten.
Problem: I don’t know free (open source) code for such a synthesis, and developing it wouldn’t be a task of one evening…
That could be interesting, have you seen espeak? http://espeak.sourceforge.net/
Cheers
Phil
The MBHP_CORE_LPC17 would be powerful enough to handle speech synthesis by itself - in background (because it wouldn’t really time consuming work @ 120 MHz…)
It even supports a I2S output for an audio DAC at 48 kHz (stereo)
Problem: I don’t know free (open source) code for such a synthesis, and developing it wouldn’t be a task of one evening…
Best Regards, Thorsten.
Interesting. I was more interested in the sound of those vintage chips, but under MIDI control. Having said that, an entirely software-based solution, would also be quite cool. Would you envisage that being possible within a MIOS framework, TK?
Incidentally, did you get my pvt message re. the the SpeakJet circuit?
a|x
That could be interesting, have you seen espeak? http://espeak.sourceforge.net/
Cheers
Phil
That does sound cool. I guess if you were going strip it down to MIDI-controlled phoneme synthesis, you could probably strip out the majority of the quoted 1.4MB of program code, too, leaving just the phoneme tables and formant engine.
Never done any hardware programming in C though, so I may be talking out of my hat.
a|x
Nice solution - it seems to be portable to embedded systems.
But still not a “weekend project” for myself - maybe for somebody else?
Best Regards, Thorsten.
Nice solution - it seems to be portable to embedded systems.
But still not a “weekend project” for myself - maybe for somebody else?
Best Regards, Thorsten.
Definitely not a ‘weekend project’ for me. More a 6-month project, probably 
a|x
I have found this which looks interesting http://www.circuitcellar.com/archives/viewable/Birsan207/index.html
If I get a spare few hours (days?) I may have a play with this…
Cheers
Phil
uhm…
Probably it need less time find a SPO256AL2 and write an application on a pic core to move some bits on MIDI event
But as I say I did in BASIC some years ago and I will try when I will have more MIOS knowledge.
Anyway my SPO256AL2 works fine yet
Regards
I have found this which looks interesting http://www.circuitcellar.com/archives/viewable/Birsan207/index.html
wow, great walkthrough!
Best Regards, Thorsten.
Would these be of any help?
Cantarino - work in progress Arduino speech synthesiser
Cantarino was reimplemented in some form in the Shruthi-1
The LPC17 is running *much* faster than an Arduino (although the LPCXPRESSO is cheaper), so that more complex algorithms can be used which should lead to higher audio quality. 
Best Regards, Thorsten.
Just been listening to some of the Cantarino demos. That sounds great! Not very understandable, but lovely and crunchy. Would be cool to hear that running from a MIDIbox, under MIDI control.
I wonder if it was written in C.
a|x
The LPC17 is running *much* faster than an Arduino (although the LPCXPRESSO is cheaper), so that more complex algorithms can be used which should lead to higher audio quality.
You say ‘higher audio quality’, Thorsten, but I quite like the crunchiness of low-quality speech synthesis (hence my interest in vintage speech ICs)
Just out of interest, is it possible/advisable to output audio directly from the LCPXPRESSO, or is some kind of external DAC required?
a|x
You say ‘higher audio quality’, Thorsten, but I quite like the crunchiness of low-quality speech synthesis (hence my interest in vintage speech ICs)
Adding some distortion would be simple as well
Just out of interest, is it possible/advisable to output audio directly from the LCPXPRESSO, or is some kind of external DAC required?
Yes, there are multiple possibilities:
Maybe the first option isn’t that bad, especially since it would allow to generate different sounds in parallel on multiple outputs to simplify the Fx chain
Best Regards, Thorsten.
Interesting.
Going back to the hardware approach, I’ve discovered another speech synth chip that seems to have a nice voice, and is quite well documented. It’s called the Soundgin, and it seems to be quite like the SpeakJet in some ways.
a|x
The soundgen is very close to the speakjet, but is geared more towards general synth as well was speech. I’ve got a demo board around here, (somewhere). The synth design is modeled after a SID, I think, but of course does’t sound as good. Had thought about modding the speckjet project for a soundgen, but my project ‘to do’ list is already very long.
Yogi