Welcome, Guest. Please login or register.


Author Topic: Speakeasy, HDIGI, digitalization, allophones and read-aloud. (Read 17393 times)

Offline gflorez

  • EP addict
  • *
  • Posts: 3607
  • Country: es
    • Támogató Támogató
Re: Speakeasy, HDIGI, digitalization, allophones and read-aloud.
« Reply #60 on: 2016.June.26. 22:34:43 »
I can see some differences on the two allophones lists, Speakeasy and Amsoft Speech.

Almost all of them match, even the representation of the sound, of course there are sounds that are not represented on the two list. Speech has more sounds, but Speakeasy has more variations of the same sound.

Speakeasy chip sb0256: http://courses.cit.cornell.edu/ee476/Speech/SPO256-AL2.pdf

Amsoft Speech list:

"%"  Pause 2
"1" to "9" stress or tilde on vowels and "/H". "9" lower note, "1" higher note, contrary to what it may seem.
You can put a number just after every vowel. I thing Speech can almost sing.

"." End of line
"?" Interrogation

"AI"                      "ER"
"A0"                      "EE"
"AH"                      "EH"
"AY"
"AW"                      "OO"
"AE"                      "OW"
"AA"                      "OY"
                            "OH"
"UU"
"UH"                      "/H"
"UW"
"UX"
------------------------------------
"D"                       "CT"
"DH"                      "CH"
"DR"
"DU"                      "Z"
                          "ZH"
"T"
"TH"                      "S"
"TR"                      "SH"

"N"                       "B"
"NX"
                          "R"
"L"
                          "M"
"V"
                          "K"
"P"
                          "W"
"J"
                          "Y"
"F"
                          "G"

Who has peeked the disassembly code could have realised that the sounds are formed of playing sample chunks of 63 bytes(but almost all of them measure 64) a number of cycles from 1 to 9.

There are complex sounds composed of two or three different samples.

But the most strange thing is that there are two chunks that are double than normal. The extra sample code is not used by the playing routines, so  I think they are discarded sounds, not used on the commercial release.
« Last Edit: 2016.June.27. 11:56:39 by gflorez »

Offline gflorez

  • EP addict
  • *
  • Posts: 3607
  • Country: es
    • Támogató Támogató
Re: Speakeasy, HDIGI, digitalization, allophones and read-aloud.
« Reply #61 on: 2016.June.28. 02:09:57 »
Injecting the Speech sample numbers on a Wav header file(6Kherz, 8bit) I have been able to ear some of the voices with Wavelab, looping the chunks.

I've put at least 16x7Fh between chunks to mark the start and the end of the chunks.



Edit: 127(7f) is the intermediate point, not zero, that is the lower limit of the wave.
« Last Edit: 2016.June.29. 11:30:12 by gflorez »

Offline gflorez

  • EP addict
  • *
  • Posts: 3607
  • Country: es
    • Támogató Támogató
Re: Speakeasy, HDIGI, digitalization, allophones and read-aloud.
« Reply #62 on: 2016.July.01. 16:28:29 »
I was totally wrong(but not as wrong....).

The samples on Amsoft's Speech are packed on 4 bits. The playing routine first takes the lower nibble and shifts it left two bits, and secondly the high nibble is shifted right another two bits. Then the range of the played samples is from 0 to 64 in multipliers of 4. There are 126 samples on every chunk, not 63.

To hear them on a PC I have unpacked the nibbles, shifted them correctly and added 95 to get 127 as the medium point.

Curiously the resulting wave and sound is very similar.

The upper wave is the wrong one.
« Last Edit: 2016.July.01. 20:24:37 by gflorez »

Offline gflorez

  • EP addict
  • *
  • Posts: 3607
  • Country: es
    • Támogató Támogató
Re: Speakeasy, HDIGI, digitalization, allophones and read-aloud.
« Reply #63 on: 2016.July.04. 08:02:47 »
Unfortunately similar can not be done, or can, but most of the musics would not be so good.
In these routines we should store the values, and play them back in 50 hz interrupt by ay simulation routine, or here store the values, and play it back with ay emulation routine. About 100 additional bytes are needed.

I've disassembled(with unknown memory addressing) the APPSYAMP.EXE SymbOS application, looking for the CPC AY output routines.

I think that partial EP sound can be achieved if Prodatron releases the player's source.

Code: [Select]
1a3a 11f482    ld      de,82f4h
1a3d 43        ld      b,e
1a3e ed49      out     (c),c
1a40 01c0f6    ld      bc,0f6c0h
1a43 ed49      out     (c),c
1a45 0e00      ld      c,00h
1a47 ed49      out     (c),c
1a49 04        inc     b
1a4a 3e92      ld      a,92h
1a4c ed79      out     (c),a
1a4e cbf1      set     6,c
1a50 06f6      ld      b,0f6h
1a52 ed49      out     (c),c
1a54 43        ld      b,e
1a55 ed78      in      a,(c)
1a57 0100f7    ld      bc,0f700h
1a5a ed51      out     (c),d
1a5c 05        dec     b
1a5d ed49      out     (c),c
1a5f c9        ret