Free SP0256-AL2 Emulator Available for Download
Kenneth Lemieux
provides hex value output. Let me know what you think.
Thanks,
Ken
Chrisr
wrote:
Maybe someone could incorporate this into the vectrex emulators. I'd
love to know what the hell the invaders say in yasi.
Chris
Give a man a fire and he'll be warm for a day;
set him on fire and he'll be warm the rest of his life
Richard Hutchinson
Are you using the Naval Research Labs TTS algorithm (as found in the
CTS256) ?
I've used it (in CTS256), but found the pronunciation of many words to be
pretty bad (which, I suppose is expected, considering the limitations of the
chip and the impossibility of creating a perfect TTS algorithm).
In the end, I discovered that to get the best from the SPO256, was to
experiment manually with combinations of phonemes. And importantly, if
you're programming for the VecVoice, to use the actual hardware (as it
differs somewhat to a bank of wavs).
Although, now that the algorithm exists as a PC program, words could be
created and then the bad ones tweaked to sound better.
Richard H.
P.S Where did you get the phoneme wavs ? ;-)
Kenneth Lemieux
Yes, it's the Naval Research Labs algorithm but it's been tweaked to
work with the SP0256. Ya, the sound is pretty rough but I do find it
helps alot with getting words to sound right. You do have to
experiment with creative spelling but, at least for me, it's much
easier to use than looking up the codes and trying them.
I think we all agree that the SP0256-AL2/CTS256 is not a great way to
do text to speech but it has its appeal. :)
I would like to put VecVoice support in and I plan to once my VecVoice
arrives! Is it just a matter of sending 0 to 64 to the serial port?
And yes, I did "borrow" the waves. I really, really meant to send
e-mail to you and Alex Herbert before putting the program up last
night. (I did just send Alex a note) I guess I was excited once I had
it running last night and wanted to put it up. I thought the waves
were done by Steve Walz
http://groups.google.com/groups?q=ftp://ftp.armory.com/pub/user/rstevew/CTSPO256/allophones.zip+(246K)&hl=en&lr=&ie=UTF-8&oe=UTF-8&selm=38C23ECB.4831%40armory.com&rnum=1
because of the dates on the files so I didn't think it would be an
issue.
Of course, I checked that link today and found out that the waves
probably weren't the ones done by Steve. Do you know who captured
them?
Really sorry about not asking first. I hope it's alright. If it's a
problem, just drop me an e-mail and I'll correct it.
By any chance, does anyone have Steve's waves around? I'm curious as
to how they sound?
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3fd5f64c$0$25664$cc9e...@news.dial.pipex.com>...
Richard Hutchinson
http://www.wps.com/products/Story-Teller/technical/T2A/index.html
>And yes, I did "borrow" the waves. I really, really meant to send
>e-mail to you
I was only joking :)
I didn't make them either, just spent a few days sorting them from an
Odyssey Voice Emulator set. They weren't labelled, and I had to listen to
each one and compare it with the ones coming from the SPO256 and rename them
accordingly - it was a real pain !.
The length of most of them is wrong, and the low sample rate doesn't help
either.
>I would like to put VecVoice support in and I plan to once my VecVoice
>arrives!
Cool !
>Is it just a matter of sending 0 to 64 to the serial port?
Yes, just send 0 to 63 (with 0 (pause 1) at then end of the string to stop
the last one sustaining)
The serial needs to be inverted 9600 8n1
Richard H.
tim lindner
> I think we all agree that the SP0256-AL2/CTS256 is not a great way to
> do text to speech but it has its appeal. :)
You might be interested to know that Frank Palazzolo has extracted the
ROM from the CTS256.
So if you have a TMS70x0 disasembler and a whole lot of time, you can
figure out the exact algrothrims used.
--
The ears are too length.
--------------------------------------------------------------
tim lindner: First initial last name at ix dot netcom dot com
Kenneth Lemieux
I've been kicking around the idea of programming a PIC with the rules
to see what I can come up with. Maybe when I have more time. But
thanks for the information!
fa...@spam.com (tim lindner) wrote in message news:<1g5r8kt.95vbiz1g8950mN%fa...@spam.com>...
Kenneth Lemieux
That's the link where I got the source. It took me about 4 evenings
(about an hour each night) to get it converted to C++. I made a SP0256
object in C++ Builder and started adding features. I have just put in
the ability to save the audio to a WAV file and would like to add
support for other wave sets. I'd love get my hands on the -17 set if
you know where to find it.
I have a -60 sitting a drawer here and I know where to get a -80.
Love to make the wave set selectable.
>inverted 9600 8n1
I know what 9600 8n1 means but not sure what you mean by inverted?
Does that just mean 2 and 3 are crossed?
Thanks,
Ken
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3fd7148d$0$25669$cc9e...@news.dial.pipex.com>...
Richard Hutchinson
>Love to make the wave set selectable.
I had a -17 but it was damaged.
I have a -12 (Intellivoice) and a -72 (Clay sent me) which, on listening to
the samples, sounds like it came from some sort of medical equipment - most
likely a defibrillator.
>I know what 9600 8n1 means but not sure what you mean by inverted?
>Does that just mean 2 and 3 are crossed?
No, the actual serial data steam is inverted (i.e. 0's are turned to 1's,
1's turned to 0's).
It needed to be this way to work properly with the Vec. I use my VecVoice
through an hex inverter when using it with a PC -
www.vectrex.biz/VV-interface.gif
>I've been kicking around the idea of programming a PIC with the rules
>to see what I can come up with
The CTS256 was actually a pre-programmed PIC 7000
Richard H.
tim lindner
> The CTS256 was actually a pre-programmed PIC 7000
Just for the record: it is a PIC 7041. General Instruments repackaged a
Texas Instruments part: TMS 7041 that includes a special masked ROM.
<http://support.radioshack.com/support_supplies/doc17/17519.htm>
--
tim lindner
tlin...@ix.netcom.com Bright
Kenneth Lemieux
This is what I know for SP0256 related chips. Can you add to the list
or make corrections?
Chip Name Description Status
--------- ----------- ------
SP0256-AL2 Phoneme Set Available
SP0256-12 Intellivoice R Hutchingson
SP0256-17 & SPR016-117 ? K Lemieux
(currently on
loan)
SP0256-19 Odessey
Sp0256-60 ? K Lemieux
SP0256-72 Medical Equipment? R Hutchingson
SP0256-80 ? I have a link
for
this somewhere
Other Related Chips
-------------------
SP0250 ? early version of SP0256; required extensive control
logic from microcontroller
SP0264 ? These exist but I have not seen one in person.
Does anyone have a datasheet? Suspect it is a next
gen version of SP0256.
SP0256-NOW ? I run across these when I search for SP0256-AL2's
but have not seen one.
SP1000 Speech Recognition and generation chip. I have these
but
not enough supporting documentation to get one
running.
Steve C. at Circuit Cellar might have something.
?????? Companion interface chip to the SP1000. I have a
datasheet for this. Need to scan it.
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3fd858eb$0$7006$cc9e...@news.dial.pipex.com>...
Kenneth Lemieux
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3fd858eb$0$7006$cc9e...@news.dial.pipex.com>...
Clay Cowgill
news:74fa6435.03121...@posting.google.com...
> Other Related Chips
> -------------------
> SP0250 ? early version of SP0256; required extensive control
> logic from microcontroller
The SP0250 (Orator) is an LPC speech chip. The LPC stuff was a close
relative of the Speak and Spell chips and things like the TMS5220 (used in
all sorts of arcade games).
-Clay
Kenneth Lemieux
"Clay Cowgill" <cl...@yahoo.com> wrote in message news:<3l4Cb.502805$HS4.3881300@attbi_s01>...
Kenneth Lemieux
http://www.speechchips.com/downloads/chiptalksetup.exe
I added the ability to save the audio as a wave file.
ken...@maine.rr.com (Kenneth Lemieux) wrote in message news:<74fa6435.03120...@posting.google.com>...
Richard Hutchinson
>or make corrections?
No, that's more info than I've managed to acquire. I'd really like to see
the complete list of SPO's (someone must have it).
SPO256-80 - does that mean there's another 79 versions I wonder..
I do find it quite exciting discovering all the different phrases stored on
these chips. I purposely made the VecVoice able to access all 256 locations,
just in case a >64 sample chip surfaces.
Richard H.
BTW how did you come by so many chips - bankrupt stock buyout or something ?
Kenneth Lemieux
I do not beleive that there are 80 discrete versions. I think they
just did custom masks for different applications.
I am curious about the SP0264 as I see there are variations of that as
well.
SP0264-036
SP0264-044
SP0264/P057
As for my source, I do alot of research and call around alot (and I do
mean alot). The chips come from various supply houses in the US and
overseas. Most of these places don't sell to the general public and
require pretty large minimum orders.
At this point, the only 2 sources I know of in the world for SP0256 in
any quantity are myself and the source in the UK. If anyone knows of
any other place to buy them, please feel free to mention where.
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3fd9e24c$0$28699$cc9e...@news.dial.pipex.com>...
Rob Mitchell
Ken:
The VecVoice that I build for you will have the inverter chip
installed :)
Rob Mitchell, Atlanta, GA
Kenneth Lemieux
I was working on comm port support last night. I'll put in an invert
option in case some one needs it down the road.
Ken
Rob Mitchell <solde...@my-deja.com> wrote in message news:<3FDABD57...@my-deja.com>...
Frank Palazzolo
I have a vectrex and love it, and I've actually done a lot of research on
these chips, as well as worked on MAME and MESS. Tim Lindner pointed me
over to this thread.
The SP0256 is actually a simple microprocessor (microsequencer actually),
with an internal ROM,
along with an LPC speech generation portion which is virtually identical to
the SP0250. Here are the details of the instruction set:
http://spatula-city.org/~im14u2c/intv/tech/sp0256_instr_set.html
Joe Zbiciak and I figured out how all the internals work in order to emulate
it. I have successfully dumped the binaries from several of these, using
the test mode of the chip. There are also auxiliary serial speech roms
available for this which expand the internal ROM. See these data sheets:
http://www.dsplib.com/chips/sp025x.html
Status:
SPO256-AL2 - dumped (are there any other language versions of this?)
SPO256-012 - dumped (some other speech code in intellivision cartridges as
well)
SPO256-019 - dumped, 16K more serial data in serial voice ROMs in the O2
Voice module (8 SPR016s)
Also, I believe there are some serial speech ROMs in a few O2 carts too.
(Sid the Spellbinder has an SPR128-004, for example)
SPO256-017 & SPR016-117 - acquired from ebay, planning to dump over the
holidays
(BTW - reports of damaged 017's may be just that the code tries to jump
out to code in the companion SPR016 which is not connected, we'll know soon
enough...)
060, 072, 080, etc. - would like to
acquire/borrow/dump/analyze...(please..please? :) )
I'm thinking of putting up a binary archive of these on the web very soon -
at least until the copyright police arrive.
A couple other notes. There is also a U.S. patent (4,296,279) with a
detailed description of the predecessor to the SPO250.
The SPO250 is recently emulated in MAME, and the SPO256 is emulated in Joe
Z.'s jzintv emulator too. Joe's code has the ability to make wav samples
directly from the ROM data. Ideally, the SPO256 could be added to the MESS
codebase, using the SPO250 to build on. This would enable support for the
SPO256 in all the consoles emulated by MESS, including the Vectrex.
I haven't heard of the SP0264 before, or seen any other references to it.
I'm interested if any more info shows up.
The CTS256-AL2 and the CoCo Speech/Sound PIC7040's are another story -
TMS7040's in disguise. As Tim said, I dumped these as well.
Thanks,
Frank
Kenneth Lemieux
Glad you could join us. I realise how many people where interested in
the o' SP0256 line.
You must be really dideciated to go to all the trouble of dumping the
roms for all these. I can loan you my -60. Send me your address and I
will mail it right out.
I will see what I can do to track down a -80. I have already sent
e-mail to try and get an SP0264 datasheet. I'll let you know what I
hear.
When you say dump the ROM's, do you mean the code or the actual sound
from each? I would love to add samples from all the sets to my
program.
Also, let me know if you need web space to host anything. I would love
to have an SP0256 archive on my site. I've been thinking about such a
thing for awhile but didn't feel that I had significant content.
Also, when you emulated the SP0256, I assume you mean emulated the
sound production and not just by stringing waves together. Is that
right?
I have never seen anything to suggest that there were other language
versions of the SP0256.
I wouldn't be too worried about the copyright police just yet. I have
tried numerous times to find out who owns the masks for the SP0256 but
with no luck. General Instruments was bought by Microchip so, in
theory, they should own the mask. I have called Microchip directly
and they insist they have never heard of the part number. I did find
the Rochester Electronics http://www.rocelec.com/ owns the masks and
has a significant quantity of the GI SP1000. They also know nothing
about the SP0256. I did find it interesting that they want $45 per
SP1000 but could not provide a datasheet or any application notes.
Thanks for the great links too!
Ken
"Frank Palazzolo" <pala...@NOMORESPAMPLEEZ.comcast.net> wrote in message news:<8rqdnVaQz_N...@comcast.com>...
Frank Palazzolo
> You must be really dideciated to go to all the trouble of dumping the
> roms for all these. I can loan you my -60. Send me your address and I
> will mail it right out.
Great, I did that just now! :) I've got a -72 coming too so more news on
this later.
> I will see what I can do to track down a -80. I have already sent
> e-mail to try and get an SP0264 datasheet. I'll let you know what I
> hear.
Also great!
> When you say dump the ROM's, do you mean the code or the actual sound
> from each? I would love to add samples from all the sets to my
> program.
It is both. The code and data are all part of the ROM image. You can make
noise-free samples from that image using Joe. Z's emulator. I haven't done
that before but maybe I can help figure it out, or maybe we can get Joe to
do it :)
> Also, let me know if you need web space to host anything. I would love
> to have an SP0256 archive on my site. I've been thinking about such a
> thing for awhile but didn't feel that I had significant content.
Ok, I will.
> Also, when you emulated the SP0256, I assume you mean emulated the
> sound production and not just by stringing waves together. Is that
> right?
Exactly - except Joe did it, not me. I just helped figure it out.
-Frank
Joseph Zbiciak
: "Kenneth Lemieux" <ken...@maine.rr.com> wrote in message
The SP0250 and TMS5220 are similar in some ways (both are LPC), but
significantly different in others, at least from what I understand.
Specifically, I believe the TMS5220 implements a lattice filter
structure for its 10-pole LPC implementation at 8kHz. The SP0250
(and SP0256) implement cascaded 2-pole sections to provide a 12-pole
LPC at 10kHz.
FWIW, Frank Palazzolo and I reverse engineered the SP0256, and I have
documentation here:
http://spatula-city.org/~im14u2c/intv/tech/sp0256_instr_set.html
Unfortunately, my webserver is down right at the moment, so you can
use the Google cache here:
Our reverse engineering isn't 100% accurate. Another member of the
Intellivision community found some engineering documents related to the
SP0256 which shed complete light on the part. The one aspect we got
wrong in our reverse engineering efforts is the behavior of the
microsequencer when an opcode is prefixed with a repeat count of 0. So
far, I have seen no data in the wild which uses this feature, which
leads me to believe it may not actually work. (I haven't yet tested
that hypothesis.)
Regards,
--Joe
--
------------------------------------------------------------------------------
Joseph Zbiciak http://www.primenet.com/~im14u2c/ Not your average "Joe"
R$+@$=W <-- sendmail.cf {$/{{.+ <-- modem noise
!@#!@@! <-- Mr. Dithers swearing Zbiciak <-- Joe's last name
---------------------- Member of the Intellivisionaries ----------------------
Joseph Zbiciak
:> from each? I would love to add samples from all the sets to my
:> program.
Frank Palazzolo <pala...@nomorespampleez.comcast.net> writes:
: It is both. The code and data are all part of the ROM image. You can make
: noise-free samples from that image using Joe. Z's emulator. I haven't done
: that before but maybe I can help figure it out, or maybe we can get Joe to
: do it :)
Yes, my emulation can make "perfect" 8-bit 10kHz waveform files from the
sound data. It emulates the ROMs directly. I put perfect in quotes since
I'm never 100% sure that I've emulated all the timing quirks of the
microsequencer. But so far, it sounds pretty good.
It was really cool hearing all the Sid Spell samples coming out of
jzIntv. (Yes, Sid Spell from the O2.) I used the dumped ROMs, and
mapped them into the Intellivoice's address space. I could play all
the samples within the emulator, triggering them from Intellivision
assembly code. It was a unique experience. :-)
Kenneth Lemieux
http://www.speechchips.com/downloads/SP0264.pdf
I looks like it has some nice features:
"Frank Palazzolo" <pala...@NOMORESPAMPLEEZ.comcast.net> wrote in message news:<6tqdneXQJLo...@comcast.com>...
Frank Palazzolo
At first glance this appears to be equivalent to an SPO256 with 8K of
internal ROM instead of 2K.
-Frank
"Kenneth Lemieux" <ken...@maine.rr.com> wrote in message
news:74fa6435.03121...@posting.google.com...
Kenneth Lemieux
64KROM and supports L.P.C Synthesis, Format Synthesis and allophone
Synthesis.
Of course, I'm not sure if these actually exist. Note the stamp in the
upper left corner... preliminary information. No idea if this
actually went into production or not.
"Frank Palazzolo" <pala...@NOMORESPAMPLEEZ.comcast.net> wrote in message news:<qpCdncWwgL9...@comcast.com>...
Frank Palazzolo
64K bits = 8K bytes, hence marketing used the bigger number. GI always does
that on their speech chip data sheets. Also, those different synthesis
types are the exact same marketing blurb as the SP0256B data sheet:
http://www.dsplib.com/chips/sp0256b.zip
By "supports" they meant "if you write new firmware" you can do any of that
stuff. :)
-Frank
Richard Hutchinson
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=2584122707&category=4663
Is that the matching speech ROM with it ?
Richard H.
Richard Hutchinson
>Includes SPR016 Serial ROM chip
Er..that would be a yes then ;-)
Richard H.
Richard Hutchinson
Kenneth Lemieux
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message news:<3ff439cd$0$9387$cc9e...@news-text.dial.pipex.com>...
Cecil Casey
Hi Cecil,
I believe when Microchip acquired the GI (General Instruments) division
back in the early 90's, many of the those devices and their documentation
were incorporated into our product portfolio. Over the years, most if not
all were no longer being produced or supported. My understanding is that
all of those speech devices have not been produced for a long time and the
datasheet for the SPO264 is not available. We in the technical support
group have most, if not all, of the older datasheets for GI devices. So if
we don't have it, then it is not possible to get it. Unfortunately, since
this device was originally designed by GI, we don't have anyone here that
has working knowledge of this part.
My best recommendation would be to try to locate information on the
internet.
Regards,
Dennis Lehman
Corporate Applications Engineer
Microchip Technology, Inc.
Technical Support
Telephone (480) 792-7627
tech.s...@microchip.com
"Cecil Casey" <crcas...@earthlink.net>
12/23/2003 11:30 PM
Please respond to crcasey934
To: <tech.s...@microchip.com>
cc:
Subject: RE: Request for older data sheet.
As a bit of background I am part of a group that uses your older speech
chips in conjunction with classic video game hardware. I am trying to
research the development of the SPO series hardware. I have studied several
versions of the SPO256-x chips but I understand you were working on a more
advanced part. The part I was looking for was the SPO264. I am not sure it
ever made it into production; the only data sheet I have seen on it was a
Microchip document, but it was marked as a 'work in progress'. Maybe you
could clarify if this part ever made it to production. If the SPO264 was
produced could you give me production dates for that part please? If not I
am looking for more background, would you have names of people that were
working on the development team.
Thank you for any insight you may be able to provide.
Cecil Casey
-----Original Message-----
From: tech.s...@microchip.com [mailto:tech.s...@microchip.com]
Sent: Monday, December 22, 2003 10:32 AM
To: Cecil Casey
Subject: Request for older data sheet.
Hello Cecil,
The only remaining related documentation that we have covers the following
devices:
SP0256
Regards,
Dennis Lehman
Corporate Applications Engineer
Microchip Technology, Inc.
Technical Support
Telephone (480) 792-7627
tech.s...@microchip.com
__________________
Please follow up.
eh
----- Forwarded by Eric Hawthorne/Marcom/US/Microchip on 12/22/2003 10:12
AM -----
"Cecil Casey" <crcas...@earthlink.net>
12/20/2003 08:27 PM
Please respond to crcasey934
To: <lit_i...@microchip.com>
cc:
Subject: Request for older data sheet.
I am looking for a data sheet for a product that you produced I believe in
the mid to late 80's. I have searched your web site and you do not seem to
keep any older data sheets available online.
The data sheet I am looking for is on the SPO-264 narrator speech processor.
Sources show it to be a 12 page document, but no amount of searching has
turned this document up. I was hoping you had a scanned copy in your
archives. If not I am willing to pay a reproduction fee for this document.
Thank you for any help you can provide.
------------------------
-Cecil Casey
crcas...@yahoo.com
"Richard Hutchinson" <richard.h...@dsl.pipex.com> wrote in message
news:3ff433c7$0$9390$cc9e...@news-text.dial.pipex.com...
jacobtim...@gmail.com
I also got them working with my arduino. Most of the rest of the addresses are filled with the same sort of static noise, and every here and there there's something that could be a sound effect/tone.
Here's the eBay listing for them (Ended):
http://www.ebay.com/itm/59-vocal-tract-Speech-synthesizer-GI-SP0264-021-like-SP0256-Arduino-circuit-bent-/191811552237?hash=item2ca8dc03ed:g:KiUAAOSwezVWyuQK
The only datasheet I've found is this one, for the SPO264 in general (12 pages): http://www.datasheetarchive.com/dl/Scans-061/DSA2IH0094616.pdf