mastodon - Link to source

#NVDASR users.
I posted a few weeks ago about a physical NVDA remote.
A small battery powered device with a Perkins-style Braille input keyboard and speaker, which let you control your online computer running NVDA from anywhere (The remote protocol is very lightweight and works quite reasonably, even with a 3G data connection).
I was thinking SIM slot, USB-C for power, and a builtin speaker and 3.5mm headphone jack.

Would you buy one?

  • yes, and pay up to $100. (63%, 12 votes)
  • I'd pay more than $100 for this thing (36%, 7 votes)
  • Nope, you're crazy. (10%, 2 votes)
19 voters. Poll end: 1 month ago

#nvdasr

reshared this

in reply to Sean Randall

mastodon - Link to source

Tamas G

@matt problem could be that a lot of plans are going E-SIM at least here in the US - so getting something that can accept that is more challenging. Nonetheless, if both a Wi-fi only and one with a 3G/LTE module existed, that's not bad, and I would probably choose the Wi-FI one piggybacked to my iPad Mini's hotspot when on the go like that. Would be wonderful either way especially if it could have the translation layer for modifiers and such baked in.
@Matt Campbell
in reply to Sean Randall

mastodon - Link to source

Sean Randall

Thanks to everyone who's responded so far.

Now that Remote has become part of the NVDA core, I feel confident that shipping someone a physical product would have a bit of longevity about it. yes, you're reliant on NVDA being here in a few years, but no longer are you needing an addon to also be maintained.

The community threw about £12,000 at the NVDA remote project 10 years ago. Perhaps it's time for another hundred or so to own a gadget that let's you leverage that, even when you're not *at* your computer?

I've toyed with the idea of a physical remote for a long time.
The simplicity of the Apple TV and Amazon Fire stick remotes really appeals to me.
Seeing the potential with something like the size of the Hable One a couple years back, then having apple do so much with their Braille Screen Input command mode really got my brain working.

So fire questions at me and we'll see what sort of numbers we get.
We're expecting a baby toward the end of september, so although I'll be sleepless, I'll also be out of work for half a year.
This might be a good time for a bit of tinkering for me.

in reply to Jonathan

mastodon - Link to source

Sean Randall

@jonathan859 wifi would always be an option alongside a SIM card.
I have alt+escape mapped to space&t on my braille input devices. it's like alt+tab but cycles through all open windows. or I can braille backspace*a (to hold alt), then repeat braille+46 for tab. those are very customised to me, of course, but you get the idea. space+56 is my tab because I read the control I'm tabbing to with my left hand.
Connecting a keyboard would be ... an interesting idea. I'm not ruling it out, the same chipset would support a bluetooth keyboard, but then you can already do that with NVDA Remote on iOS or android. The whole idea of this project was something small, really taking the idea of the Perkins-style keyboard to keep the size down whilst offering flexibility in controlling your computer.
@Jonathan
in reply to Sean Randall

mastodon - Link to source

Tech Singer

Seriously, this is something I've wanted for years. I'm using a surface Go for this, but as you can imagine, it's huge for what it does! I would be overjoyed for a unit with USB C so I could plub in a QWERTY keyboard and charger. You're right about 3G working fine, I have had no issue with that and this surface go. The only issue is that the unit would have to produce its own speech.
in reply to Sean Randall

mastodon - Link to source

Tech Singer

That's fairly satisfactory, though I'm not sure whether I could plunge for more than $100 I would still have to carry around a phone to get eloquence through teamtalk/sonobus, but I would happily spend $100 just for the keyboard access. I would just mute the speech on the device and have the phone do the talking with the device doing the input. (yes, I can use virtualhere, but that does need another device and usually a bigger one than anything you could be thinking of.). The normal person, one without bad hearing, could probably just use Pico.
in reply to Martin

mastodon - Link to source

Tech Singer

@mcourcel If @cachondo gets the funding for this, I would probably use it with either Narwhal for the listening side or, as I said, sonobus or teamtalk for audio. Keeping in mind that both sonobus and teamtalk would allow you to get all audio data from the PC, though with a bit of delay. They would also allow it on the iPhone and Android. The pain point for this, on my end at least, is the keyboard and the delay. This wouldn't solve the delay if I still want Eloquence, but it would definitely solve the keyboard.
@Sean Randall @Martin
in reply to Tech Singer

mastodon - Link to source

Sean Randall

@techsinger well, for a test of viability I don't need much money at all. I can progress with silly pushbuttons on the board and won't need a battery, SIM card or case. I'll eventually need to turn my spaghetti into a Custom PCB Design and a 3d printed case, and will want a decent quality battery and buttons that feel good and take more than the fingertip. :)
@Tech Singer
in reply to Sean Randall

mastodon - Link to source

Tech Singer

Absolutely, I understand that Eloquence is going to be impossible. You'd basically need to duplicate what I have with the surface, and at that point, one might as well just get an actual surface or a gaming handheld and have done with it. All the issues of size/speed/battery life would be as bad as they are now. I would definitely spend the $150 to get the keyboard entry alone, as I said. It would be nice to have the thing able to mute itself and just use the keyboard, like we have with "mute remote" on the old add-on.
in reply to Tech Singer

mastodon - Link to source

Sean Randall

@techsinger I'm also looking at a "Typical battery life of up to 30 hours on Wi-Fi, up to 20 hours on cellular. Meaning with an active connection to a remote server. That's with the current setup, without the overheads of a full operating system. SO I don't think the tradeoffs for a wider variety of text-to-speech engines is worth the tradeoff in this situation.
@Tech Singer
in reply to Sean Randall

mastodon - Link to source

Tech Singer

@Jage I'm using Eloquence but again, this is a surface go. Battery life is an issue, heat is an issue, the LTE modem is only fairly good and loses power occasionally. The 100 meg is the amount of data being sent daily for reading books, typing... basically everything I do on the remote PC. I'm writing this with that now, for example. I do nothing on the surface as an actual computer, it's an NVDA remote thin client :)
@Jage
in reply to Jage

mastodon - Link to source

Sean Randall

@Jage I have no idea. If you want Android, you've already got cheap devices that do a lot more, absolutely.
But people do buy these blindy media players and other gadgets. I'm not expecting a wave of interest, but it'd be something I'd use, so why not ask, I guess?
Not sure I'd foot the bill for just me, but if i had a handful of commitments to buy so I could get everything designed I'd go for it.
@Jage
in reply to Dave Taylor

mastodon - Link to source

Sean Randall

@davetaylor2112 you're the 2nd person to mention qwerty. But I already have more bluetooth keyboards than I know what to do with, and my phone. The real plus of a braille input device for me is size, it's something about the size of my phone. I'm slower on braille than werty, but still clack along at about 70WPM. I grew up on a perkins and - outside of MS Office - have never had reliability issues with NVDA and braille input. I've mapped every key to a braille chord or combo and am just as happy keying in with a brailliant as a mantis, just to drop names in.
@Dave Taylor
in reply to Dave Taylor

mastodon - Link to source

Sean Randall

@davetaylor2112 I can't control my laptop from my phone via the hable. The need is small, I'll grant that. But so's the need for a button audio player, when phones can do audiobooks. It's a little device that would scratch my itch at least. it'll cost me about 80 quid in parts to see if it's doable, and hopefully keep my brain active with a new baby on the cards when I can't work full-time.
@Dave Taylor
in reply to Sean Randall

mastodon - Link to source

Sean Randall

So in designing the prototype for the physical remote for #NVDASR, I discovered that you can still buy the part of the #Doubletalk speech synthesizer that does the talking.
this makes me very happy and, alongside the NVDA project, I'm going to work on a lighter version of the old Book Port device when I take my work sabbatical. Broadly to meet my own needs first, of course.
This would mean a Telephone-style keypad for controls, an audio jack, at least 24 hours of playback on a single charge if I can manage it, USB-C for power and an SD card for storage. A belt/pocket clip is utterly essential for me.
Initial input would be text files only, progressing eventually to other text-based formats.
I have no interest in DAISY support or podcasts, although Internet radio is a distinct possibility if you're happy to halve your runtime.
RSS would be relatively straightforward too but, again, but decrease the battery life due to being online.
File transfer would be wireless, not USB.
FM Radio is on my wish list.
Voice memo style audio recording is also a potential, as I do use that with Rockbox at present.

Why, you ask?

Mostly because I haven't used Doubletalk in almost 2 decades and want to see if i can keep it going.

And largely for the cools. 'Oh yes, I'm just reading this book on a gadget I built myself.'

#nvdasr #doubletalk
in reply to Sean Randall

mastodon - Link to source

Jayson Smith

This is an interesting idea. I wouldn't personally buy it because Doubletalk's not my favorite, and especially not the newer style of Doubletalk that came out in the late 1990's, I'm assuming that's what he's still selling. I like the earlier style of Doubletalk as demonstrated in the attached recording.

in reply to Jayson Smith

mastodon - Link to source

Sean Randall

@jaybird110127 @Tamasg the monologue95 one does very much sound like the newer doubletalks to me.
I'm kinda excited about the opportunity to play with the hardware and make a gadget, but also frustrated to see the software is doing the same thing and locked away behind a brain-twisting set of machine code I can't disentangle.
It gives me hope that one day when all the hardware has been used up, there might be options, though.
@Jayson Smith @Tamas G
in reply to Jayson Smith

mastodon - Link to source

Sean Randall

@jaybird110127 @matt My only problem with RIM is one of the benefits of the current protocol is it's lightweight, which means it works over a data connection. SO although it solves the issue of needing on-device tts, it adds the complexity of a system taking up more bandwidth unless you're on a solid connection.
My own use case for the device was largely when not at home.
SO I'm not against it as an idea, but as the design of a device I'd want to use, I'd prefer the on-device speech with a less latent system when travelling that would work over GSM.
@Matt Campbell @Jayson Smith
in reply to Sean Randall

mastodon - Link to source

Sean Randall

@jaybird110127 @matt that said, there's absolutely nothing in the hardware that would limit me, or anyone else, from just sticking different firmware onto the thing.
Want to use it to control NVDA via NVDA remote? flash that.
Want to have it sending text to and receiving audio from RIM? Flash that firmware.
Want it to be a standalone UEB or nemith braille input calculator?
Flash that thing.
Prefer a Braille in, voice out notepad?
Flash the baby.
Want it to act as a bluetooth remote for your media centre? You've got buttons and bluetooth. code it. Flash it. Change channels.

I can't afford to build any of this stuff without community input, so the result's just going to be utterly reusable hardware.

@Matt Campbell @Jayson Smith
in reply to Sean Randall

mastodon - Link to source

Sean Randall

@matt @jaybird110127 I don't imagine either product, as in the NVDA remote controller or the doubletalk-based text file reader, will be big enough to justify the design changes of several hundred dollars initial outlay to get a second type of remote controller wired with hardware speech output.
The remote controller will be perkins-style keypad and software speech output, and the file reader telephone-style keypad and doubletalk output.
@Matt Campbell @Jayson Smith
in reply to Sean Randall

mastodon - Link to source

Sean Randall

Woow. Great design sesh following up on this book reader with the makerspace.

the belt clip I'd planned has now turned into a variety of different connectors. A little desk stand and a suction mount are pretty cool possibilities.
Most exciting of all for me is the ease with which I can integrate a Qi wireless charging coil into the case. I'll literally be able to drop this thing onto a charging pad at home, whereupon it'll enable its wifi to let me add books over the network. Prototyping suggests I'll be able to drop text files in from my Windows or Mac computers, or the files app on my phone, which is awesome. I'm not building a high-end device, but I don't know of any other book readers for the blind that support wireless charging, although admitedly I've been very dismissive of them all since they all use software text to speech engines, none of which I like.

USB connectivity for copy and paste file operations will still work of course, unless you want to use the thing as a hardware speech synthesizer, in which case you'll have to change the mode in the settings. I can't run both USB classes at a time, but I imagine the serial port thing is only going to be a rarity.

The guys at the hackspace all think I'm mad for wanting the Doubletalk, they say it's ridiculously fast or stupidly robotic.
But They're great for championing the work and I know I can call on them when I inevitably blow up speakers or connect the wrong battery terminals and melt a board or something.

in reply to Sean Randall

mastodon - Link to source

Sean Randall

battery life has now settled at between 21 and 26 hours of play time. I'm hoping for the higher end, there's something stupidly cool about a machine that can just talk solidly for 24 hours without dying.

Standby time is weeks, the SBC's have craaazy low standby power draws, even whilst keeping time. The one I left this time last week has dropped by about .002%. Mad. in theory, therefore, an alarm is also doable.

Never understood why the book courier had an alarm but no speaker, did they think I'd sleep in my earphones?
I'm not holy convinced as to the precision of a real time clock on these things yet though, so honestly an alarm clock is not high on the trello
Was a bit worried my initial design didn't allow for USB file transfer, that would really annoy people. But looks as though that's a nonissue with the better board too.
Wow.
This might actually go somewhere.

in reply to Sean Randall

mastodon - Link to source

Sean Randall

One of my maker pals has said before I even think about kickstarter, write the user's manual I'd want to read.
This will give my backers an insight into the features I'm planning for, and me a bunch of things that are requirements for the code.

It sounds like good advice.

I hate writing documentation.
I've hated writing it from when I released my very first computer program offered to the public at my local library in 1998 and the intervening 2 and a half decades or so have done nothing to disabuse me of the notion.

It's an owner's manual, anyway, not a user's. If this thing gets off the ground, it'll be because the people who paid for it bought its parts and its code and bought into the desire for wanting the brilliant, ancient voice.

in reply to Sean Randall

mastodon - Link to source

Tim Ward ⭐🇪🇺🔶 #FBPE

I have no idea whether this applies to you, but one common reason I've found for people not liking writing documentation is that they've never been taught to touch-type and not surprisingly find hunt-and-peck two-finger-typing to be a right ****ing pain.

In my generation it was "only girls get typing lessons at school". But my mother was a typing teacher so she taught me at home.

in reply to Sean Randall

mastodon - Link to source

Sean Randall

Still waiting on everything else, but I wrote a very short entry in the development blog about the first parts of this project.
[Thoughts on the first parts | Randall Reader](seanrandall.github.io/Randallr…)

Thoughts on the first parts

I can’t actually believe I’m going ahead with this project. I didn’t quite think I would until, on Thursday, I was holding a little keypad in one hand and a microcontroller in the other.
Randall Reader
in reply to Sean Randall

mastodon - Link to source

Jayson Smith

Even though Doubletalk isn't my thing, and really hasn't been since I got a DECtalk PC for Christmas in 1993, I'm still strangely excited about this project. I guess it's just the thought of an individual, rather than a company, trying to make something like this. At some point I would love to see one, but given the speech used, I wouldn't actually use one near enough to justify buying one, for which I'm extremely sorry.
in reply to David Dunphy

mastodon - Link to source

Sean Randall

@startrek2025 i'm not abandoning the NVDA remote idea at all, but I've shifted to this project first for 2 reasons.

Firstly, for my first hardware effort, it requires no custom parts. the Keypad for this device will be an off-the-shelf component, whereas a Braille keypad will be a custom job in itself. So seeing how the development chain goes will help with that.

second, the availability of the doubletalk chips is a precarious point. I know I can buy a couple dozen of them today. But RC Systems is a small outfit with the guy retired. Can I have more than the initial batch now? Maybe. Will I be able to get them in a year or 5? Who knows.

@David Dunphy
in reply to Andre Louis

mastodon - Link to source

Jayson Smith

@FreakyFwoof That's actually one of the things I was wondering about. Is the design going to be coded exclusively for the included Doubletalk, or will there be flexibility in the source code to throw in another set of speech commands for changing parameters, etc? I thought of this the other night when I imagined a scenario where the legal situation with DECtalk is untangled (Ha ha, yeah I imagine pigs will be piloting passenger aircraft first), and someone makes a DECtalk on a chip project.
@Andre Louis
in reply to Jayson Smith

mastodon - Link to source

Sean Randall

@jaybird110127 both The source code and hardware will be entangled with the Doubletalk to quite an extent:
The chip provides its own audio amplification circuit and specific output pins to both headphones and speakers, as well as an input to mix other audio (which is what makes the possibility of a built-in FM radio possible for v2 of the board). SO you'd need a pretty impressive replacement to drop DECtalk in as a hardware module.
The software is just sending strings of text to the synthesizer though.
\8Sthis is a test\r
Tells doubletalk to say 'this is a test' at speed 8.
[:rate xx] this is a test
is the same for DECtalk etc.
So yeah, if I write the module properly, you'll just use a different class to communicate with the synthesizer.

On the other hand, given how much DECtalk source is out there, it's quite reasonable to expect someone to give a crack at porting it to run natively on the platform. You'd then just need to change the software to do everything rather than send text to the connected chip and replace the doubletalk with an amplifier.

@Jayson Smith
in reply to Sean Randall

mastodon - Link to source

Jayson Smith

I've known some rubber keys to wear down over time and become much harder to hit than normal. When we were moving away from our home in Louisville, we had a Yamaha keyboard purchased in 1992, and we didn't bring it with us. One contributing factor, aside from where's it gonna go and why do we even need this, is that the last time I'd played with it, many of the rubber buttons were extremely hard to make actually work.
in reply to Jamie Teh

mastodon - Link to source

Sean Randall

@jcsteh so far, I'm literally connecting pre-soldered components. The keypad and Microcontroller are complete components in and of themselves, so I'm literally just running cables between them.
Things will get dicier when I move to testing on battery power because the colours will matter then, and when I come to get the FM tuner chip into the circuit, which I'll have to solder in myself.
But baby steps.
@Jamie Teh
in reply to Sean Randall

mastodon - Link to source

Sean Randall

Monday.
Little V-Stamp speech synthesizer has made the big trip across the water. Seems a bit weird it can move 4,900 miles so quickly and then be sat there for a whole weekend not doing the last 70 miles or so, though.

Some life admin before work in terms of paying bills with my first coffee of the day. I have to sort the breakfast dishes and put the bins out before work, so that's the next 20 minutes swallowed up, before returning to the corporate slog for another 40 hours.

a UPS tracking page showing a shipment starting in Everett, WA on 31st July to multiple locations, ending in Castle Donington, United Kingdom by August 1st.
⇧