Talker/80 - New Voice Synthesizer in the Making

[LambdaMikel]

A first batch has been produced - so far I have a 100% success rate.
Seems also the Seeed PCBs are of good quality (green ones).

 

[LambdaMikel]

"Talking Eliza" with the Talker/80:

https://www.youtube.com/watch?v=Vv27CSFUAi0&feature=youtu.be

 

[LambdaMikel]

A GitHub repository is being populated as well currently - the sources and Gerbers are still missing, but will show up there eventually:

https://github.com/lambdamikel/Talker-80

 

[LambdaMikel]

This is the final Talker/80 demo video. It shows some of the BASIC demo programs that can be found on the talker80.jv3 disk image in the Github repo.

• Spanish Speaking
• Advanced Voice Control
• DECtalk Singing from BASIC
• Welcome Program
• Jeff Shrager's "Creative Computing" Eliza with natural sounding DECtalk speech

All of this runs directly from Level 2 (Misosys) disk BASIC, no driver and no phonemes required. Adding speech to a BASIC program can be very simple. For the Eliza program, for example, I only needed to add some "out 11, asc(c$)" to an already present "character by character" printing routine. 10 Minutes of work.

https://youtu.be/ctfGo_UebRY?list=PLvdXKcHrGqhcJzzogLMI-J4Or-ap6UVSY

That's it for now, folks! Thanks for the interest.

I'll open another thread when I am going for the Model 3 / Model 4, but probably not before next year.

 

[MikeModifed]

https://www.instagram.com/p/B_8ljvpFHCz/?igshid=1wl3wcnb6mvlx :clap2:

 

[LambdaMikel]

Oh, nice! I didn't post it there - maybe they got it from Hackaday

 

[LambdaMikel]

Update - I just learned from another thread that Talker/80 can also be connected directly to the "Screenprinter Expansion Port" of the Expansion Interface. I was wrong in assuming that WR is not being passed through by the EI. I got missinformation from the book "Hardware Interfacing with the TRS-80" from Uffenbeck. Or there is something else I don't understand. This book states, for example, in Experiment 6 "Memory-Mapped I/O using the 8255" (and in all experiments that use Memory-Based / Mapped IO):

"This experiment cannot be done on a Model III.... The experiment also cannot be done using the expansion interface and a Model 1 computer. This is because the RD and WR control signals are not enabled on the expansion interface bus (you may want to disconnect the expansion interface temporarily for this experiment)".

This is wrong. Talker/80 TRS Voice Synthesizer Emulation Mode works perfectly fine if connected directly to the EI Expansion Port (the port in the fron left corner which connects to the FreHD usually).

So, in that light, the Talker 80's expansion port edge connector passthrough would not have been necessary, but it is still convenient for people that don't have an EI and may want to connect another expansion (e.g., the Quinterface or similar). Then the Model 1 port is not blocked.

But feel free to connect it to the EI Expansion Port interface instead!

 

[Tarkie101]

Mike,

Very interested in putting one of these together, as I've an original TRS-80 Voice Syth which needs bringing back from the dead, and having a working one to drive me forwards may help

Any idea on the availability of the gerbers?

 

[LambdaMikel]

Mike,

Very interested in putting one of these together, as I've an original TRS-80 Voice Syth which needs bringing back from the dead, and having a working one to drive me forwards may help

Any idea on the availability of the gerbers?

Hi Tarkie,

the plan is to hear from the first three customers first and check if it worked nicely with their setups. I sold 3 on Ebay last week (within one day, actually!), and they should arrive soon. If there are no unexpected major problems, I am going to put the Gerbers and the Firmware (including sources) online. I will also continue selling 2 or 3 every month on Ebay, but I cannot really produce a lot, and "development cost recovery" is unlikely anyhow, given that the profit per unit is in the 10 $ range for me. So selling them on Ebay is mostly "community service" for folks that can't or don't want to assemble it themselves (or don't have a programmer that can do the rather difficult GAL22V10B's... )

So, please stay tuned, hopefully by the End of May everything should be online.

Best,
Michael

 

[Tarkie101]

So, please stay tuned, hopefully by the End of May everything should be online.

Bouncing with excitement.

Steve

 

[pski]

So selling them on Ebay is mostly "community service" for folks that can't or don't want to assemble it themselves (or don't have a programmer that can do the rather difficult GAL22V10B's... )

This is greatly appreciated. I know there is no profit in selling completed units at a reasonable price. But, a significant number of the community members either don’t have the time or skills to build these.

 

[LambdaMikel]

This is greatly appreciated. I know there is no profit in selling completed units at a reasonable price. But, a significant number of the community members either don’t have the time or skills to build these.

Thanks Pete - I'll do what I can!

In the meantime, here is some more testing - this time with Quinnterface + FreHD.

https://youtu.be/EcCoxA2afwY
https://youtu.be/4Sgb11Q7sNk

I know, I should Talker/80 do the talking... but I am afraid that this time it required some explanations. Hope you can get through the accent

 

[dittman]

I've had my Talker/80 for a couple of days.

My first impression is this is very nice board and the fact that it emulates multiple voice synthesizers for the Model I is amazing.

My first test was to connect it to my Model I and Expansion Interface and use the HFE image provided in the Github page. I ran all the included tests successfully. I was very impressed.

My next test was to connect the Talker/80 to the Model I directly and then connect a Quinnterface/FreHD to the Talker/80. For this the power on sequence is critical. You must power on the Talker/80 first, then the Q/F combo. I didn't wait long enough and the Talker/80 would hang. The power on sequence is mentioned in the documentation and I found that if I waited for the Talker/80 to finish with the startup message before powering on the Q/F I was able to boot. It may not be necessary to wait that long but I didn't test because it's only a matter of a couple of seconds.

My latest test was to try the Talker/80 with the MIRE and MISE. First I connected the MIRE to the Model I and then the Talker/80. I had this all on a power switch and they powered on at the same time and there wasn't an issue. I had the Talking Eliza program loaded on my MIRE emulated floppy from testing a Radio Shack Voice Synthesizer so I set the dip switches on the Talker/80 to start up in RS VS mode and reset. I then ran Talking Eliza and it worked perfectly. I then powered everything off and connected the MISE to the expansion connector on the Talker/80 and powered everything up again. Again everything came up without any issues. The Talking Eliza program still worked.

Did I mention that I am very impressed?

 

[LambdaMikel]

Thank you very much for the kind words - I am super relieved that it mostly worked out ok for you.
Especially the MIRE and MISE... I don't own these expansions, but know that they are very popular and frequent in the field, so it is important that Talker/80 works with these.

Did you encounter anything that might hinder or potentially warrant a delay of the publication of the sources, or further distribution of more Talker/80s?

Thanks again for supporting the project and the great testing effort! Really appreciate it.

 

[LambdaMikel]

I have another small batch (3 assembled) available on eBay.
And the Gerbers and sources are available on the Github by now.

 

[Tarkie101]

I have another small batch (3 assembled) available on eBay.

Thanks for the heads up, snaffled one of the assembled ones.

Steve

 

[LambdaMikel]

Thanks for the heads up, snaffled one of the assembled ones.

Steve

Thanks Steve. It'll be in the mail on Tuesday morning (Monday is a holiday here).

Overall, six Talker/80 have been sold by now.

Currently, I have a "kit version" online for DIY - it requires some decent soldering skills though (you need to change one SMD resistor, but the rest is through-hole). I will assemble some more at some point.

It will take some time to assemble another batch (I still have material for another batch of 3).

 

[Joe Zwerko]

A word of advice: buy more components than you think you're going to need. Purchased in volume, a hundred resistors or capacitors don't cost much.
There's clearly a strong demand for Talker/80, and you'll be surprised how quickly you go through parts and PCBs.
MISE and M3SE started out slowly too, but there are now hundreds of each "in the wild". It wasn't planned that way.

 

[LambdaMikel]

A word of advice: buy more components than you think you're going to need. Purchased in volume, a hundred resistors or capacitors don't cost much.
There's clearly a strong demand for Talker/80, and you'll be surprised how quickly you go through parts and PCBs.
MISE and M3SE started out slowly too, but there are now hundreds of each "in the wild". It wasn't planned that way.

Thanks for the advice! Yes, I am surprised how quickly they sell (3 per day, basically).

I have some previous experience in selling a very similar device to the Amstrad CPC Community. And it took more than 2 years to sell more than 20 there... This is a very different experience here. I expect the demand to level out after, say, 20 devices as well though. We will see!

I cannot really afford buying the Click text to speech boards in larger quantities though; with my last order I was happy to get a discount that usually applies if you order more than 10 click boards. Resistors etc. are of course not a big cost factor. But the click! boards are (i.e., their normal price tag is 39 $ per piece, plus shipping...)

I am not sure how many devices I can produce in that way. It seems that my capacity is at ~ 3 boards every 2 weeks. This is why it is all open source by now - I welcome other folks to assemble them as well. I don't really have the capacity for "mass production" (and not convinced yet that this will be required )

We will see!

 

[Joe Zwerko]

More advice: get a cheaper PCB supplier. My largest board, the MIRE, costs less than $6 including delivery from PCBway, with immersion gold. It's about 7" x 6.5".

Edit
----
Just realized you meant the daughter board. Yeah those can be expensive. The Ethernet on MISE/M3SE runs $20 and the FPGA module around $15.