Voice on Spectrum (2Max Romanoff)

Hi, All!
Sorry if this is off-topic.

Hello, Max!
>
AS> Listen, can you give me examples of how to build all this,
AS> because I really want the Spectrum to speak,
AS> even if it’s not in a very pleasant voice.
>
MR> so, you sample and that’s it.
Yes but how much memory will it eat up... how many sounds are there in Russian? And 2 kilos for each sound?
And then, from them it will be possible to obtain by transformation
only those voices that you get on a tape recorder when
accelerating/slowing down playback.
Well, will it be nice to listen to such garbage in toys?
Unfortunately, I only have a Talmud called
"Theory and application of digital processing signals"
L. Rabiner, B. Gould.
This old, 1978, book without examples with programs is just
theory :(. It devotes
to digital processing of speech signals
only about 60 pages of text out of 850.
There is also J. Keiter, “Computers - Speech Synthesizers” M. Mir, 1985.
This book more specifically, with algorithms and examples of implementations
in hardware, but also hopelessly old.
A talking chip (which could be bought separately for $12 with a dictionary
5,000 words 8-|, I came across only one, in Conrad,
it spoke German, was built into the measuring device ;)
pronounced the measurement results very coolly digital
multimeter ;) if only they had guessed and made an interface thoughIt would be
half-duplex and replace German with Russian-English ;)
There are also Chinese watches that speak Russian (with a terrible accent),
speaking calls, etc. But IMHO there are “digital tape recorders”, and
not speech synthesizers with frequency modulation.
When Caller IDs started talking (especially one of the versions that spoke
in a “computer” voice, I thought that finally the
Spectrums would start talking too. After all, someone threw in the source code for the speaker and a small
Polish speaker (not the source code).
But if only someone had explained how to synthesize speech with different
timbres. And in order not to consume all the resources of the z80. If this is
impossible, then at least what are the most used
operations required (and what should be the time for their execution),
that is, not just to say “impossible”, but to explain why it is impossible.
So you briefly threw out “samples”, so give an example of some kind of
synthesis (and, by the way, to make a library of sounds yourself). “manually”
you will be tortured, you need a program that pulls sounds out of the voice,
but this is automatically even more complicated than synthesis, just like
the conversion pcx -> txt is more complex than txt -> pcx) )
If you did this, explain what all this entails in terms of memory and
in terms of processor load.
Of course, all this is available for sound cards (and even I think frequency modulation),
but for some reason it’s not on the Spectrum :(
LanguagesWhat is it that no one wants to study? Yes, and music in demos with vocals would be more fun. And I’m silent about educational programs and toys.
And maybe someone has libraries of sounds and effects for frequency
modulation?
Or maybe someone can tell me how to rob these sounds from a pussy?
In a word, help the teapot, whoever can! I think these questions are interesting
not only to me. It’s not in the FAQ (this surprises me every time)
>
> haha, do you seriously think that in Aonah, with their 4Mhz Z80, they use
> some kind of voice packaging?! :)))
In the latest versions, the voices were recorded from a microphone - this is undoubtedly!
The more valuable is the version that spoke in the SYNTHETIZED voice of a robot
and was probably not “brought to mind”, i.e. to a normal human
voice (male, female, children's, senile, baritone, bass, mezo-soprano
etc.) :( We went, as always, “the path of least resistance.”
And then they made a digital answering machine and voice memory - it was clearly using the same technology: digitization and packaging, because without packaging you
not much in 64K write it down.
On ISKRA-226 (1983 April 1), when we were stupidly digitizing a 10 kHz ADC,
only the phrase “And your whole back is white,” said at a normal
tempo, got in.

With respect, Andrey Savichev/CGS