Quote from: BSC on 22:59, 07 February 14
You seem to be making a lot of games :) Anywhere on the interwebs where I can see / download them?
And I hope you are not planning or hoping for sampled music like in the example videos you posted over there?
Like, with speech / singing? That would not be possible with the current version of AYAY. I was thinking about
using phonemes (a bit like in this oldskool speech rsx extension which sounded quite good) as wave forms
instead of the basic forms like triangle, sine etc. But that's a completely different story and I have currently no
clue whether that would be feasible at all - I might be dreaming away here ;)
Anyway: I think it should be possible to combine those kind of games with the AYAY engine. It could become a
quite painful developing process, though.. But we are developing on the CPC .. in 2014 .. we like pain, right? :D
Definitely no speech or singing!!! :D
With your SID program, integrating phonemes would be very interesting indeed.
If you go to 0:30...
Amstrad CPC Gems 2 - El Enigma de Aceps (http://www.youtube.com/watch?v=hPTSioCeqtE#)
El Enigma de Aceps was groundbreaking, the first text adventure in spanish with good graphics and synthesized voice. But the guy had a weird pronunciation :laugh:
[ot]El Enigma De Aceps made use of Superior Software's Speech. (Perhaps with spanish phonemes)
You can easily include it into games.
The program:
http://www.cpc-power.com/index.php?page=detail&num=4372 (http://www.cpc-power.com/index.php?page=detail&num=4372)[/ot]
Quote from: Devilmarkus on 10:48, 08 February 14
[ot]El Enigma De Aceps made use of Superior Software's Speech. (Perhaps with spanish phonemes)[/ot]
Could be, but the voice says "pilamide" instead of piramide.
Quote from: McKlain on 11:09, 08 February 14
Could be, but the voice says "pilamide" instead of piramide.
I wonder if by breaking down the words phonetically and fine tuning each phonetic part with the pitch and speech, you could get something sounding more human?
Amstrad CPC - Speech Synthesis (http://www.youtube.com/watch?v=Y9fK5foeDuo#)
Quote from: sigh on 12:37, 08 February 14
I wonder if by breaking down the words phonetically and fine tuning each phonetic part with the pitch and speech, you could get something sounding more human?
Amstrad CPC - Speech Synthesis (http://www.youtube.com/watch?v=Y9fK5foeDuo#)
Is it just me, or does this video don't have any speech audio? (I can hear the floppy drive noises from WinAPE but no AY output)
There's sound, forward a bit.
I watched the whole video... There's no audio for me, just crackling and, @ beginning the emulated floppy noises...
Edit: Hahahahaaaaaa Youtube doesnt playback on my left audio channel?!?!? WTF???? is there a balance slider somewhere?
Quote from: Devilmarkus on 20:33, 08 February 14
I watched the whole video... There's no audio for me, just crackling and, @ beginning the emulated floppy noises...
Edit: Hahahahaaaaaa Youtube doesnt playback on my left audio channel?!?!? WTF? ??? is there a balance slider somewhere?
Did not get any sound as well except (sort off) the emulated floppy either on windows 7 64bit or linux 64bit.
However, the sound is fine with the video saved locally and played with VLC. Flash player bug ?
Maybe, yeah... But we are going too off topic here, I think ;)
Lets get back to topic better
Quote from: Devilmarkus on 20:33, 08 February 14
I watched the whole video... There's no audio for me, just crackling and, @ beginning the emulated floppy noises...
Edit: Hahahahaaaaaa Youtube doesnt playback on my left audio channel?!?!? WTF? ??? is there a balance slider somewhere?
I couldn't hear it either. Just the loading noises.
Well, here's the DSK for Speech:
(perhaps useful for SID things, too?)
Quote from: sigh on 12:37, 08 February 14
I wonder if by breaking down the words phonetically and fine tuning each phonetic part with the pitch and speech, you could get something sounding more human?
At least that's what I was thinking about. I did some reading concerning phonemes (mostly vocals, though) and IIRC they basically consist of 2 predominant frequencies.
And I also remember the waveforms to be rather simple.. Anyway: I think that would be a cool project to try to figure out what audio data is needed to recreate a human voice (and reverse engineering this speech extension could be a good starting point) and if the wave forms turn out to be simple enough to fit with the AYAY player, it could be possible to make the CPC sing :D
Quote from: Devilmarkus on 13:03, 08 February 14
Is it just me, or does this video don't have any speech audio? (I can hear the floppy drive noises from WinAPE but no AY output)
No sound for me as well. But this video does have sound:
speech v1.1 for Amstrad CPC - YouTube (http://www.youtube.com/watch?feature=player_detailpage&v=BIn6mFaLjVY#t=27)
Quote from: BSC on 21:04, 09 February 14
At least that's what I was thinking about. I did some reading concerning phonemes (mostly vocals, though) and IIRC they basically consist of 2 predominant frequencies.
And I also remember the waveforms to be rather simple.. Anyway: I think that would be a cool project to try to figure out what audio data is needed to recreate a human voice (and reverse engineering this speech extension could be a good starting point) and if the wave forms turn out to be simple enough to fit with the AYAY player, it could be possible to make the CPC sing :D
That would probably be the ultimate 8 bit speech software!
What I find really impressive about this type of speech is that you can combine it with music.
Go to 0:42
SSA-1 Speech synthesizer used in CPC game "Roland in space" (http://www.youtube.com/watch?v=YcgUZG_jz20#)
I take it that the ram and cpu consumption must be very small using this type of speech?
If so, it would be interesting to hear one of these voices over the richness of the SID tracker.
Didn't the SSA1 mix it's own sound output with the AY output using it's internal stereo mixer? My cousin had one and it came with two little speakers.
This hardware speech is worst than software speech done on some games like Robocop.
Robocop uses samples, as far as I know.
Quote from: McKlain on 12:28, 10 February 14
Didn't the SSA1 mix it's own sound output with the AY output using it's internal stereo mixer? My cousin had one and it came with two little speakers.
Yes, both the SSA-1 and the DKtronics Speech Synth have their own amplifier that mixes the 3 AY channels with the voice output, so both can play at once. Both also use the SPO-256 speech chip which is phonetic based speech, so it will always be worse than sampled speech. It is possible to make the speech much clearer by fine tuning the words, but it doesn't sound like the Roland programmers spent much time trying to do this.
Bryce.
The Amstrad SSA-1 just collects few bytes sent to it's port and plays the phonemes for itself.
So no real CPU power needed here.
The voice is already generated in the SPO-256 chip and just played back by this series of bytes.
DK'Tronics speech works in the same way.
Both don't need much CPU power.
As you can see in my video for "Roland In Space", the CPC is talking and also simulteanously loading the game from disk...
In my video I extra turned the volume for AY output down to concentrate on voice output.
You can try the game also here:
(With emulated SSA-1 speech)
(http://cpc.devilmarkus.de/games/-R-/rolandinspace.png) (http://cpc.devilmarkus.de/game.php?id=-R-&title=rolandinspace)
If you get trouble in allowing java applets after the latest java update, just ask me... There's a solution for it....
Quote from: McKlain on 13:16, 10 February 14
Robocop uses samples, as far as I know.
And Morteville Manor ?
The Techni-musique speech syntesizer sound definitively better. (MEA8000)
Ordinateur Amstrad cpc 6128 synthetiseur vocal (http://www.youtube.com/watch?v=Qwnc8j6vDAs#ws)
The main reason why I don't like hardware speech synthesizers very much is their limitation in one languaged phonemes only...
Sure, the Techni-Musique sounds pretty nice, but it's limited to French language...
Edit: This of course also included the limited phonemes inside the SPO speech chip.
Quote from: TotO on 13:53, 10 February 14
And Morteville Manor ?
The Techni-musique speech syntesizer sound definitively better. (MEA8000)
Ordinateur Amstrad cpc 6128 synthetiseur vocal (http://www.youtube.com/watch?v=Qwnc8j6vDAs#ws)
The techni-musique plays on another league compared to the SSA1.
Yep, I have to agree on this.
The SPO256 barely sounds like english, while the Techni-musique can sing...
Quote from: Bryce on 13:44, 10 February 14
It is possible to make the speech much clearer by fine tuning the words, but it doesn't sound like the Roland programmers spent much time trying to do this.
Bryce.
Is the Techni-musique the same in ram/cpu consumption as the SPO?
What interests me is having this used
while playing an actual game. Software speech is too ram hungry for this to happen.
The C64 with it's SID chip was able to handle speech for games like Way of the exploding Fist while playing.
It makes me wonder if that could of been possible on the CPC by using the SP0 instead of samples.
Man....so much unexplored areas with this machine :D
I think that everything on the c64 took less cpu time and less ram than on the cpc.
Quote from: sigh on 16:56, 10 February 14
What interests me is having this used while playing an actual game. Software speech is too ram hungry for this to happen.
The C64 with it's SID chip was able to handle speech for games like Way of the exploding Fist while playing.
It makes me wonder if that could of been possible on the CPC by using the SP0 instead of samples.
I don't know what you mean by SP0 but I think that trying to use phonemes (or formants) is a really interesting approach.
It might not be (very) usable in a game, but I think that depends.
I wonder if anyone here happens to have done anything related to speech synthesis using formants. A couple of samples
would be a great start, I think. I googled around a bit yesterday and one thing I wound was this:
AUE: Formant values for 14 Fontana vowels (http://alt-usage-english.org/Fontana_formant_numbers.html)
So it's basically 3 frequencies per vowel. Another interesting site I found:
Lecture 017 Instrument IV (http://www.feilding.net/sfuad/musi3012-01/html/lectures/017_instruments_IV.htm)
Quote from: Devilmarkus on 13:45, 10 February 14
As you can see in my video for "Roland In Space", the CPC is talking and also simulteanously loading the game from disk...
Yes, and this is wrong :D .
On a real CPC "Roland In Space" will start loading only at the end of the speech (More precisely, only once the last phoneme has been send)
While is possible to speak during disk activity (like it's done with the AY), Roland In Space does not do it !
old discussion on SPO256 / SSA1 : Amstrad SSA-1 Speech Synthesizer ROM needed (http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/amstrad-ssa-1-speech-synthesizer-rom-needed/)
Sorry - I meant the phonemes:)
With the Roland example and El Enigma de Aceps, even creating voices like grunts and shouts would be an incredible step.
I'm guessing the more complicated the the speech pattern, then the more difficult to use.
.
I wonder if anyone has every tried to generate or steal phonemes or formants samples and play them on the CPC.
Apart from the great speech RSX extension I mean..
(*dream mode on*)
It would be sooo cool if someone out there had done this already and could provide me with a
set of samples which represent all the formants needed for e.g. english language (others like samoan, middle-old north-western welsh
or kölsch could always be done later) which I could incorporate in AYAY to play around with.
(*dream mode off*)