Sega SPEECH synthesis CSM Sounds (Page 2 of 2)

indeed

God damn it, that is TERRIFYING. I love it
Looking forward to play around with that ROM!

I did some encoding experiments that would fit this audio mode perfectly -- condensing a sample to its N most prominent sine components over a window of M microseconds. Then you just have to encode the partial number and its amplitude. For 4 components at 25 Hz, you'd probably be able to get it down to 800 bps. Here's an example of 8 partials at 25 Hz: https://dl.dropboxusercontent.com/u/501 … ariots.mp3. This has no amplitude quantization, though.

For speech synthesis, you could severely limit the spectrum with a pre-filter, but I'm not sure what would produce the best overall result for speech.

Great work !
a frequency range of 100Hz-5000Hz should be enough for speech analysis.
I would need something like that to make a nice tool for FMDrive
// Wav vocal sample >> FFT analysis >> Formant Freqs + amplitude >> MIDI
4 main formants in the speech>> FRQ to operators Fnumber >> midi pitch or midi notes
4 levels for the power of spectrum >> Db to operators TL >> midi volume or TL(cc)

These should output 4 midi files for OP4,3,2 and 1

can you make something like that ? that would be also cool as a wav sample to midi using each sine waves signals for each most prominent sine...

So here is some example of how to use the very unique CSM speech.
First a reminder :
Example of CSM speech in the MSX game Zeilard :
http://www.alyjameslab.com/tutorials/MSX_CSM.wav
Example of CSM random speech in FMDrive vsti
http://www.alyjameslab.com/tutorials/FMDrive_CSM.wav

Sine waves artifacts are noticeable in both of these examples because there is no decay on the envelope generator.

Now lets decay very quickly and the artifact are less noticeable at the end.
Example of an attempt at saying " HELLO" with fast decay
first MIDI data are played very quickly then slowed down.
http://www.alyjameslab.com/tutorials/hellocsm.wav

and for a visual explanation better than words..
Here is one instance of FMDrive in Cubase with automation lane and MIDI channels for the special mode..
Notice the similarity between the spectrogram and the automation data.

what you see is power of spectrum and variation of frequency in the time domain.
Sum of sine waves at different frequencies gives us what we call vocal formants.

Yep ^_^ let me hear those samples when done