I am musing on the idea of making a new low-cost speech synthesizer for the CPC. One that people can afford (LambdaSpeak is still too expensive, due to the 39 $ click board).
It will be 45 $ device, consisting of
- an Atmega 328p
- the 25 $ SpeakJet chip
- a flipflop
- a bus driver
- an op-amp
(all through hole). It will be SSA-1 and DKtronics compatible, and have a SpeakJet mode as well.
Maybe it will also have AMDrum emulation.
It will be called "AmJet"
Would there be any interest in such a device?
yes [emoji846]
Good idea! However, I really hope that you still release the LambdaSpeak device. :)
Quote from: GUNHED on 15:01, 01 June 18
Good idea! However, I really hope that you still release the LambdaSpeak device. :)
In a sense, it is released - everybody what wants one can build it themselves:
https://github.com/lambdamikel/LambdaSpeak (https://github.com/lambdamikel/LambdaSpeak)
So what would be the difference with the Lambdaspeak ?
Different speech chip. Speakjet - cool video here (not mine!)
https://www.youtube.com/watch?v=IbpNK3B9eB8 (https://www.youtube.com/watch?v=IbpNK3B9eB8)
Sounds completely robotic. But cool in itself!
It is only 25 $.
Will it need a different RSX extension?
Quote from: GUNHED on 11:39, 13 June 18
Will it need a different RSX extension?
Not quite sure yet. For sure, there are major differences with DecTalk. Does it have different voices, spanish, english mode? Probably not.
These commands will just be ignored. DecTalk mode - not applicable.
Volume, speed? I will see. Haven't started yet looking at the chip in detail.
sounds interesting :)
Could be a perfect Dalek emulator! ;D
Quote from: GUNHED on 15:08, 14 June 18
Could be a perfect Dalek emulator! ;D
Yes, it's accent sounds quite British indeed - almost like a BBC speaker :laugh:
And, what's even better, you'll get one for free if you do the RSX port of course ;)
Haha! Yes indeed! The RSX Extension shouldn't be a problem at all. Since I already got the LS 1.5 I don't need the other one - even if I really like the sound of it. More thinking about getting a robot (see other Thread ZX as Dalek controller! :) ).
EDIT: Hmm.... maybe it's to early for that... However it could be an nice idea to do a poll to ask which Speech Synthy people would like to buy? Could make it more easy to order PCB's (Bryce I remember offered help to do a series of them?!?). I'm just looking forward! :)
Quote from: GUNHED on 23:47, 14 June 18
However it could be an nice idea to do a poll to ask which Speech Synthy people would like to buy? Could make it more easy to order PCB's (Bryce I remember offered help to do a series of them?!?). I'm just looking forward! :)
Right, I am leaving all potential future / follow-up action to Bryce if and when he finds the time. For me, LambdaSpeak is done. Maybe I will sell a couple of kits (PCB + parts) on Ebay with programmed chips, but not many. PCB assembly services are too expensive for small batches, the potential sales number too small, and I don't have enough time to solder them manually.
Spoiler alert:
https://youtu.be/uRDf5vs4q9s (https://youtu.be/uRDf5vs4q9s)
... Coming this Xmas... Stay tuned!
There will be another little surprise (or two, or three)... 8)
Not sure about the name yet. Obviously, the hardware specs have been changed slightly... XC9536 CPLD, ATmega 8535L 16 Mhz MCU, SpeakJet... and...
The card will also have PIN headers for SPI, I2C, and UART / Serial, and the firmware will be open source.
Project update - SSA1 / DKtronics emulation almost working, but still needs some allophone fine tuning:
https://www.youtube.com/watch?v=5Q-oyd2Qd9Y&feature=youtu.be
R2D2 phonemes? :D
Quote from: Gryzor on 11:49, 09 October 19
R2D2 phonemes? :D
Yes, some! But they do not come with meaning / translation unfortunately, so it cannot be used as a R2D2 babelfish device.
Update: Roland in Space and Tubaruba also working by now. Will post a video later this week.
Project update - I also have the serial interface working by now. Tested.
Next steps - connect the "surprise chip"...and add support for I2C and SPI. There will be headers for all of these (UART, SPI, I2C).
https://youtu.be/zLsgOHT1fmA (https://youtu.be/zLsgOHT1fmA)
... for now, it is only a nano SwinSID... but nothing should prevent me from putting in a real SID.
The biggest obstacle is 12 V required for a "real" SID - but hey, aren't we lucky that the 664 and 6128 have a 12 V PSU? Not so sure what to do with CPC 464 users though. For now, maybe I stay with the SwinSID which is fine with 5 V only.
Also works with a real SID; but had to change the clock phase of the CPLD generated 1 MhZ clock (was inverted):
https://youtu.be/dJlccupSALY (https://youtu.be/dJlccupSALY)
You can clearly hear that the demo program showing OSC3 "LFO" addition resulting in vibrato from the C64 Programmer's Reference Guide
http://www.devili.iki.fi/Computers/Commodore/C64/Programmers_Reference/Chapter_4/page_203.html
works fine. With nano SwinSID, the sound is "flat" instead as OSC3 Readout is not supported.
Nothing better than a real SID!
Almost done...
Great job! ;)
I started a similar project some years ago with two SwinSID (6 channel, stereo) and a CPC version of a SID player (the code was based on the SAM Coupe's one).
Here you can see (and hear) the results and some documentation.
http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/new-amstrad-cpc-sound-board-(aka-sonique-sound-board)-sid-part-(wip)/ (http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/new-amstrad-cpc-sound-board-(aka-sonique-sound-board)-sid-part-(wip)/)
https://soundcloud.com/david-donaire-s-nchez/sets/amstrad-cpc-sonique-sound (https://soundcloud.com/david-donaire-s-nchez/sets/amstrad-cpc-sonique-sound)
I think the SID player was somewhere in my hard drive, hehehehe.
Congrats and continue with your hard work. :)
Cheers,
DaDMaN.
QuoteGreat job! (http://www.cpcwiki.eu/forum/Smileys/SoLoSMiLeYS1/wink.gif)
I started a similar project some years ago with two SwinSID (6 channel, stereo) and a CPC version of a SID player (the code was based on the SAM Coupe's one).
Thank you! Well, this one has only one SID - yes I looked at the SAM Coupe SIDPlayer code, and also there is another version based on that one that uses the AY for "SID Emulation".
https://cpcrulez.fr/applications_music-sid_player.htm (https://cpcrulez.fr/applications_music-sid_player.htm)
Both of them will required some work for me. If yours is readily available, I'd give that a try ;)
Some more pics - I have send the PCBs to OshPark for a first run.
Looks really nice, great job 8)
Can you allready estimate the price :D ??
*allreadyrunningtomycollectionlookingforsomesidchips*
Well, without the SID or SwinSID nano or ARM SID or whatever people prefer to plug into the socket:
- SpeakJet 25 $
- ATMega 8535 3 $
- Xilinx XC9536 XL 3 $
- other parts 8 $
- PCB 5 $
So it should be in the 50 $ range. Not sure that will include shipping or not yet (USPS postage is 14 $ for me to Europe).
The SID is optional.
The firmware sources will be open source. Notice the PIN headers - so far I have the GPIO and UART / Serial Interface working and supported by the current firmware, but especially I2C and SPI functionality I will probably just hand over to the community. If you want to extent the firmware, you will just need a standard AVR programmer and WinAVR. I will not delay publication of the board until I have generic I2C and SPI support fully implemented in the firmware; once the board has been tested I will upload all sources and Gerbers and such to GitHub. Probably in November.
Btw, the SpeakJet is not only a speech chip - it is also a 5 oscillator sound synthesizer
https://www.sparkfun.com/datasheets/Components/General/speakjet-usermanual.pdf (https://www.sparkfun.com/datasheets/Components/General/speakjet-usermanual.pdf)
Would be nice to see somebody write some music for it.
Quote from: LambdaMikel
Thank you! Well, this one has only one SID - yes I looked at the SAM Coupe SIDPlayer code, and also there is another version based on that one that uses the AY for "SID Emulation".
https://cpcrulez.fr/applications_music-sid_player.htm (https://cpcrulez.fr/applications_music-sid_player.htm)
Both of them will required some work for me. If yours is readily available, I'd give that a try ;)
Some more pics - I have send the PCBs to OshPark for a first run.
Searching on my harddrive all I found is a DSK file with the working code (RUN"SWINSID") and one piece of ASM code (WinApe32 friendly) with the player generator adapted by SyX.
Here you have :)
Very very cool, thanks a lot! That should be easy to adapt for Speak&SID...
Just lost 2 Z80's and my M4 due to the "SID CHIP OF DEATH" - hence a warning to not repeat my mistake ::)
Thought I could save a couple of $$$ on Ebay by buying an "untested" 8580 R5 - I should have known better. Supplied the SID with its 9 V, turned on the CPC - and that was it. Of course, I had to try an other CPC ::) - same result. :doh: :picard2:
Luckily and to my surprise, the CPC's did survive the data / address bus 9 V overvoltage without any further damage - both RAM and Gate Arrays are fine, but the Z80's as well as the M4 got fried :-[ The CPC's are already up and running again (I have a couple of spare Z80's). The M4 won't boot anymore and locks the CPC now :-[ :(
So, either there was a short in the SID, or it was a fake / relabeled one....
Well, you're using c64 chips. They are evil! They are satanic! They are killers! :o Don't use that dark stuff anylonger and come back to the purity of Z80 world. Add any other sound chip if desired, but not that commodore crap! I'm really sorry to hear that the M4 is still not working. :(
Quote from: LambdaMikel on 08:56, 24 October 19
Just lost 2 Z80's and my M4 due to the "SID CHIP OF DEATH" - hence a warning to not repeat my mistake ::)
Thought I could save a couple of $$$ on Ebay by buying an "untested" 8580 R5 - I should have known better. Supplied the SID with its 9 V, turned on the CPC - and that was it. Of course, I had to try an other CPC ::) - same result. :doh: :picard2:
Luckily and to my surprise, the CPC's did survive the data / address bus 9 V overvoltage without any further damage - both RAM and Gate Arrays are fine, but the Z80's as well as the M4 got fried :-[ The CPC's are already up and running again (I have a couple of spare Z80's). The M4 won't boot anymore and locks the CPC now :-[ :(
So, either there was a short in the SID, or it was a fake / relabeled one....
And that's the reason I never used original SID ICs from canivalized C64 computers. Only "new" and 5v TTL signal compatible ICs (and 3.3v with TTL signal tolerance, of course) will be connected to my beloved CPC when designing new hardware projects, hehehehe. ;)
Quote from: GUNHED on 12:26, 24 October 19
Well, you're using c64 chips. They are evil! They are satanic! They are killers! :o Don't use that dark stuff anylonger and come back to the purity of Z80 world. Add any other sound chip if desired, but not that commodore crap! I'm really sorry to hear that the M4 is still not working. :(
"I AM YOUR ... SID" :o :laugh: SID = 666 = (|IS (SID rotated by 180 degrees!)
But hey, it IS a Halloween Project :)
I will continue prototyping with SwinSID and ARM SID should come soon too :)
Playing with 9 V / 12 V for the different original SIDs in combination with the CPC is of course a bit like playing with fire. There isn't much you can do to protect the CPC busses from that if things go haywire... welll you could probably check in a C64 first, that way the damage is not so severe if the s**** hits the fan. But if you have a known working "good" SID from the Light Side of Force, then it should be fine.
It's a pity, you put so much time, work and dedication into this (and other) project(s), and then this crap happens. Hope Duke can help you with the M4 in case in needs to travel overseas. :-\
Quote from: GUNHED on 15:32, 24 October 19
It's a pity, you put so much time, work and dedication into this (and other) project(s), and then this crap happens. Hope Duke can help you with the M4 in case in needs to travel overseas. :-\
Yes, @duke really rocks, he immediately offered to help me out :) Thanks a lot!!!
My ARMSID arrived - it works great and sounds a bit more authentic (and warmer / more analog) than the SwinSID.
It also does the OSC3 readout correctly - it is a bit more expensive, but the developers did a great job I must say. Impressed.
Allright, it is done :)
The first design worked out of the box - once again, breadboard prototyping paid off.
Software and SID player are coming soon. And then the Open Source Github repo with everything.
I hope that some people will have fun with it. A generic UART / Serial Port and GPIO CPC interface is already supported by the firmware. I still need to add a generic CPC interface for I2C and SPI functions. But since the firmware will be open source on Github, everbody can also easily add support for their I2C / SPI device on their own. And communication with the CPC (in both directions) is of course already implemented.
EDIT: I will have the SID oriented the other way in the next PCB revision...
Quote from: DaDMaN on 12:18, 23 October 19
Searching on my harddrive all I found is a DSK file with the working code (RUN"SWINSID") and one piece of ASM code (WinApe32 friendly) with the player generator adapted by SyX.
Here you have :)
Thanks again! Wasn't a lot of work to get it working with Speak&SID - after studying the ASM for 2 hours ;)
So, here is your SID Player with Speak&SID playing the original SID 6581.
https://youtu.be/xVo5ycUuM5Q
@LambdaMikel (http://www.cpcwiki.eu/forum/index.php?action=profile;u=2172) : have you already tried the FPGASid (http://fpgasid.de/)?
Quote from: HAL 6128 on 07:57, 04 November 19
@LambdaMikel (http://www.cpcwiki.eu/forum/index.php?action=profile;u=2172) : have you already tried the FPGASid (http://fpgasid.de/)? (http://fpgasid.de/)?)
Not yet, and it seems it is a bit too large to fit in the socket? I could make another PCB revision with more space though, if that's important to people... Also, it is a bit too pricey IMHO (74,95 €....)
EDIT: No, sorry. The price of FPGASid is prohibitive, I am not going to support it. There are three perfect SID options already (real SIDs, ARMSID, SwinSID). More is unnecessary.
Github will be populated here
https://github.com/lambdamikel/Speak-SID
New video - line out recording and LEDs
https://youtu.be/FXDS3pdf-w8 (https://youtu.be/FXDS3pdf-w8)
In the meantime, the 2nd PCB revision arrived. I can make about 5 to 6 of these for now.
Please PM if you like to get one (50 $ + 10 $ Shipping). Without (Swin / ARM / real) SID of course.
I have uploaded the firmware C source to the GitHub by now.
The firmware still misses some generic functionality for I2C and SPI (UART / Serial Interface and GPIO is there), but I will add some CPC IO functions for these soon.
I'll also port my MIDI CPC demo, and then probably do the Real Time Clock for I2C and maybe SD Card demo for SPI.
Quote from: LambdaMikel on 19:38, 11 November 19
I'll also port my MIDI CPC demo, and then probably do the Real Time Clock for I2C and maybe SD Card demo for SPI.
Will the RTC be compatible to the RTC of the LS3?
Quote from: GUNHED on 20:43, 11 November 19
Will the RTC be compatible to the RTC of the LS3?
The answer is Yes and No:
Yes: the Speak&SID firmware is open source, so if you wish to customize it such that it provides, for I2C, the same clock functions that LS3 does, then you can do that with not so much effort. This, however, results in a I2C / SPI-device specific API, which is not generic. Hence, the answer should really be:
No: the idea for the SPeak&SID card really is to provide "generic" API functions for I2C and SPI to the CPC in the firmware. That means, rather than having I2C or SPI-device specific API functions such as "set hours", "get time", etc., let the CPC construct the right I2C / SPI message instead, and also do the decoding of the corresponding responses received from these devices. The Speak&SID firmware will only provide generic low level I2C / SPI functionality - buffering of input and output message, and send and receive.
So, "Yes" will require that you put the high level SPI/I2C-device specific functions such as "set hours / get hours" into the AVR firmware yourself, and you can make it any way you want.
And "No" will imply that you will have to wait for me to provide these generic I2C / SPI API functions, and then the device-specific code (set hours, get hours, ...) will be dealt with by the CPC.
It should be noted that the latter approach will only work for "simple enough" devices, but not for very complex ones (i.e., it is impossible to set up the Epson speech chip like this for example).
The documentation is almost done by now.
https://github.com/lambdamikel/Speak-SID (https://github.com/lambdamikel/Speak-SID)
Thanks to the first buyers / customers for their support - getting ready to ship the first batch. :)
And the new (final) PCB revision is also done.That's it from me on this project.
(The backside has ground copper btw)
Actually, there is one more thing that I am going to add to the firmware:
the ability to read MIDI data from the UART, pass it to the CPC, and the CPC sending sound data to the SID.
I just realized that the hardware already supports this. I can hence turn this into an inexpensive SID + AY MIDI Synthesizer for the CPC. What does not work is to use the UART / MIDI and SpeakJet at the same time though. But that's ok.
Something like this might be nice for the CPC:
https://www.youtube.com/watch?v=womHSih6dQU (https://www.youtube.com/watch?v=womHSih6dQU)
The dude is using the C64 Cynthcart and MIDI port expansion - Speak&SID already has the UART and only needs a MIDI breakout header.
https://www.amazon.com/ubld-it-MIDI-Breakout-Board/dp/B00YDLVLVO/ref=sr_1_1?keywords=midi+breakout+header&qid=1575390754&sr=8-1 (https://www.amazon.com/ubld-it-MIDI-Breakout-Board/dp/B00YDLVLVO/ref=sr_1_1?keywords=midi+breakout+header&qid=1575390754&sr=8-1)
Quote from: LambdaMikel on 17:22, 03 December 19
Something like this might be nice for the CPC:
[amazing youtube video]
The dude is using the C64 Cynthcart and MIDI port expansion - Speak&SID already has the UART and only needs a MIDI breakout header.
That's some incredible stuff, thanks for sharing. Also, that bloke is hilarious and oozes talent.
Quote from: BSC on 21:19, 04 December 19
That's some incredible stuff, thanks for sharing. Also, that bloke is hilarious and oozes talent.
Check out his Furby Organ.. he is just entirely crazy :D In the most positive sense of the word of course! :)
[size=78%] [/size]
Yep, he's completely crazy! ;D
But, I love the energy and the will to try thing no one else tried before.
You can see the guy in action in this video:
https://www.youtube.com/watch?v=ZB27L6OuxHE
Quote from: LambdaMikel on 19:38, 11 November 19
I'll also port my MIDI CPC demo, and then probably do the Real Time Clock for I2C and maybe SD Card demo for SPI.
Here is a first cut at a MIDI IN SID+AY CPC Synthesizer:
https://youtu.be/4esnHlCg138
It is weird that the SID and AY are slightly detuned from each other!SID is currrently only 1 channel... will add polyphony soon, and then I need controls for filter resonance and stuff as well of course.
5 more shipping out on Monday - if you had ordered one in the last 2 weeks, yours will be in this batch!
I am running out of SpeakJet chips... I can make only 2 more before Christmas.
If you want yours before Christmas, please order it no later than coming Tuesday.
Great Work!!!
Do you think I would play SID with AMSDAP+Playsoniq(MSX SID) with my CPC and your SID player or Owen/DaDMaN?
Playsoniq SID (and CMB-MOS SID) use &H00 to &H1C address
With AMSDAP when MSX see #xx , CPC see #FFxx
It is a mad idea??? I will fry anything??? :-X
(I asked too in AMSDAP thread sorry the Off Topic)(Please respond there if you consider)
Would it be possible to achieve something like this in the CPC... ?
https://youtu.be/W3LoVzl4JQk
ASAIK with LambdaSpeak III - yes. :)
Quote from: GUNHED on 09:19, 12 February 20
ASAIK with LambdaSpeak III - yes. :)
No extra hardware is required to "play" MIDI from a 8-bit computer, but a proper cable.
"I am playing a track that I have not yet finished and demonstrating the MIDI capabilities of the Sinclair ZX Spectrum 128k Plus 2, in 128k basic mode.
I use an old lightgun lead for a MIDI lead and instead of a lightgun on the end it has a 5 PIN MIDI Din plug on it. Which gives me 8 MIDI channels in basic."
Quote from: TotO on 10:25, 12 February 20
No extra hardware is required to "play" MIDI from a 8-bit computer, but a proper cable.
That is a bit missleading... the 128 K Speccy has RS232 support (and MIDI) built-in. So making a cable and accessing it from BASIC is easy, since already supported by the OS / Firmware / BASIC.
http://oldmachinery.blogspot.com/2012/11/spectrum-128-and-midi.html (http://oldmachinery.blogspot.com/2012/11/spectrum-128-and-midi.html)
True, it is a software serial interface. "Bit banging" is used. And that can be used to implement RS232 on most old computers - PROVIDED they are fast enough to bit bang MIDI baud rate of 31250. That would be a stretch, say, for a ZX81. And CPU load is very high with that method.
Virtual Net 96 for the CPC also uses "bit banging", resulting in high CPU load:
http://www.cpcwiki.eu/index.php/Virtual_Net_96 (http://www.cpcwiki.eu/index.php/Virtual_Net_96)
In addition, MIDI OUT is easy to implement, even from BASIC once you have fast serial interface API in ASM (be it software serial interface via Bit Banging or real UART hardware).
MIDI IN is much more difficult. I don't think we will see MIDI IN being implemented on the Speccy in that way.
For that, real RS232 / MIDI hardware and ASM to process the real time MIDI messages is required.
So, saying that "only a proper cable is required" to play (= MIDI OUT) on "8-bit home computer" is a bit of an oversimplification. Certainly, as always - it depends. Is the the CPU fast enough to do bit banging at 31250 bauds? Do I have a fast ASM implementation of a serial interface (certainly, you cannot bit bang from BASIC!), etc.
Quote from: VintageAdvantage on 18:37, 13 February 20
That is a bit missleading... the 128 K Speccy has RS232 support (and MIDI) built-in. So making a cable and accessing it from BASIC is easy, since already supported by the OS / Firmware / BASIC. [...] So, saying that "only a proper cable is required" to play (= MIDI OUT) on "8-bit home computer" is a bit of an oversimplification. Certainly, as always - it depends. Is the the CPU fast enough to do bit banging at 31250 bauds? Do I have a fast ASM implementation of a serial interface (certainly, you cannot bit bang from BASIC!), etc.
Think about that using the Tape port on CPC 6128.
Quote from: TotO on 18:49, 13 February 20
Think about that using the Tape port on CPC 6128.
Give it a try - would be curious to learn what you find! I probably think that a digital output is better suited for bit banging out serial data, but who knows!
For a real MIDI interface with both IN and OUT, I would always spend the extra buck to get a real hardware UART (on a microcontroller) to do the job, with proper buffering and such, so CPU can focus on MIDI message processing only.
Quote from: VintageAdvantage on 18:56, 13 February 20For a real MIDI interface with both IN and OUT, I would always spend the extra buck to get a real hardware UART (on a microcontroller) to do the job, with proper buffering and such, so CPU can focus on MIDI message processing only.
Sure.
Tried to order Speak&SID from LambaMikel on Ebay. Bad news: he run out of SpeakJet chips. He had sold more than 30 Speak&SID by now, and depleted the SpeakJet stock in the US! :o Guess I have to buy USIFAC for now for serial interface and midi.
The chip is back-ordered though, so he thinks he can make some more end of the year. SID ROCKS!! 8)
Holy lord! People buy that SID expanson like crazy, but LambdaSpeak III is sooooooooooooo much better. :) :) :) Get the LS3 - it can do everything! :) :) :)
IMHO LambdaPeak is too damn expensive, and it ain't got no SID man! 8)
Quote from: VintageAdvantage on 16:07, 18 February 20
IMHO LambdaPeak is too damn expensive, and it ain't got no SID man! 8)
That's the great thing about it: no SID. 8) 8) 8)
Seriously, it's not too expensive. Only postage from USA is expansive. But if someone would do a batch inside the EU it would be cheap enough. :)