What's new
NVDA 2024.1 compatibility
Changed the name to Sonata since we plan to support additional TTS models besides Piper in the future.
Support for fast variants for Piper voices. These fast v...
if there's a great way for me to train my existing voices into this new model, it would be great to know - will check out this new repo! Excellent work at the update, keep up the amazing talent you have.
@Tamasg If your existing voices are trained with Piper, then they'll work with this version. If they fail to work for any reason, you can copy the config from any working voice to your voice, and edit relevant values.
yeah, and I love the speed improvement it gives to some of those! Especially Mac Alex :) I did notice though that the faster models do a way better job at not cutting the starting fragments of words (older voices can still slightly do this.) I see that encode and decode are now two models so feels like splitting older voices into the new format is less straight forward without a full retrain, but I might be entirely wrong.
oooh this is so superb. So I could still continue to train with the existing notebook and eventually just convert the checkpoint CKPT file into the new RT voice, that's going to be amazing. At least it unblocks my work though with sourcing higher quality data and transcripts for the voices I've developed, so just already knowing that is huge help. I could point the variables to URLs in my drive folder for the checkpoint, then convert that way with the script you referenced.
thanks a lot for such precise steps on this! I noticed the tasks.py script was pulling the voice and doing these in one-go, which makes more sense to how the Fast voices got created for the existing ones, though I bet that in itself is quite a project to author on top of the new streaming model work.
ah interesting, the source command with that .sh file returns an error at this line: `this_dir="$( cd "$( dirname "$0" )" && pwd )"` (it says: `"dirname: invalid option -- 'b' Try 'dirname --help' for more information."` so something isn't passing properly to $0 there, interesting.
@Tamasg I didn't edit that script, it came from piper repo. Anyways it does not effect the installation. I encountered it myself when exporting voices.
ok. I'm not sure if this is good or bad, and I hope it doesn't insult your script code too much by gutting it, but I made a notebook file of your steps: https://eurpod.com/Export_piper_voice_RT.ipynb - hopefully this looks correct, it restructures it a bit to take a path to a file stored in drive after mounting. I'll try it in a little bit and see if it's a spectacular fail or not :D (minor correction to made for running the .sh file since we can't use virtual environments.)
Important notice!!! After installing this version, you will lose all of your installed voices. Please use the voice manager to re-install the voices again.
I some how got the audio themes add on to run on nvda 2024.1 just find but it's quite sluggish when using w SAPI also I would like to a sound which could be played when pressing enter
Pronunciations issues not withstanding, the speed increase here is legit impressive! And I know enough to know that training data is what's needed here. I hope, when someone comes up with the optimal dataset requirements, that we can find someone willing to submit a voice.
@nick A dataset designed specifically for screen reader usage, goes a long way toward creating a good quality voice. If guidelines are the issue, we can come up with a set of guidelines based on Microsoft/Google guidelines which are openly available.
I think that would be wise. I don't know if I have the time or voice discipline to make a voice myself, but I do think it might be sensible, at least, to have this available.
Wow, that's a huge speed increase. You also may wish to note that it only works if you use WASAPI, otherwise it throws: AttributeError: 'WinmmWavePlayer' object has no attribute 'setVolume'
@twynn I'm going to uninstall the voices I managed to download, uninstall the add-on, and reinstall. I'll get you the debug logs. BTW, the voices were those I downloaded through the voice manager.
Dear @Musharraf :verified: Can you please give me a hint how do I build the file sonata-grpc.exe is it a result of building https://github.com/mush42/sonata ? We are working on a slovak human sounding voice with friends and I am tweaking corresponding espeak-data along the way, so until I manage to get these pushed and merged to espeak-ng I imagine my best bet is rebuilding the addon with all the resources locally. Thanks for all the fantastic work you are putting into this.
@pvagner Here's how to build the sonata-grpc binary: git clone https://github.com/mush42/sonata cd ./sonata/sonata-grpc # With Rust installed cargo build --release
@pvagner If you just want to set the eSpeak-ng data directory, you don't need to re-build the binary. Just set the following environment variable before launching sonata-grpc: SONATA_ESPEAKNG_DATA_DIRECTORY=[your custom espeak-data directory parent]
Very impressed with the new Split #Braille Feature in #JAWS. Being able to read Teams chat with half my display whilst typing someplace else is tremendous. They seem to have well thought out the different modes and views. Shame about the cells wasted with the split and lack of customisability of the viewport widths. If I won the lottery I'd throw a few hundred k at #NVDASR for comparable #BrailleDisplay development
I have the same feeling for when I use Split Braille at work, it's become quite useful for splitting work between meeting chats / another document, sadly a new reason I've lately been switching to JFW during meetings.
That said, updating #JAWS is still a pain. It's ridiculous that the US has had such a cheap way of using JAWS at home for over half a decade now and the rest of the world pays stupid prices or has to buy at very specific times of the year which you only know about if you're quite involved in the community, And I'm still sad that workplace adjustments are still hugely JAWS driven, when NVDA Addon development should really have made this much more of an open thing. Hey ho, it's all about the choice I suppose and that's a good thing. Hopefully screen readers will keep enspiring each other to new things.
The "JAWS in the workplace" thing is also very country-specific. It seems to be more common in places where strong accessibility laws exist and employees can get funding for scripts. This is not the case here, for example, and almost nobody uses JAWS. I know people here who work in large companies, including banks and such, and NVDA is just fine.
Before I went back to work I was a full time nvda user. However, is it wrong to say that if I ever get a windows pc, I'd happely pay for an anual jaws subscription?
@cublanco pretty much in the same boat, although I get JAWS for free as part of the Hungarian blind citizen's program that gives it out there to qualifying people. It's a timed license that gets renewed for 2 years at a time by the contract though. Work has mostly made me get JAWS again for home just in case I need one of the same tools that can make my job a bit more productive that NVDA doesn't have.
@Tamasg @cublanco I spent a whole month writing research it rulesets I couldn't ever reuse, either. I wish there was a cross-compatible tool for snipping data out of web pages by rule like that. XPath was really clever at it.
@cublanco oh yeah, the webpage customization features are terrific. (I also haven't quite figured out on all my sites why certain regions are considered a "glance" area, but the method JAWS has for identifying those I would welcome in NVDA too.) It's those slight pamperings that do add up in time you spend doing it manually
@Tamasg The thing that gets me, is someone thought that I needed three hours of relearning how to use JFW, when it only took me at least, five minutes? Mind you, the last version I used was 12 at home.
Hot tip of the day, as it's one worth re-posting once in a blue moon: Use MMSYS.cpl after pressing Win+ R, to open the classic version of the sound control panel. From here, you can press the first letter of your sound device and press alt+S to set it as default, or press spacebar to open the property sheet for it. If muted: Press CTRL+tab, tab, spacebar to unmute it, or don't press tab after hitting CTRL+tab, but use arrows to adjust volume directly. Either way, it's sexy, it's old, it works.
@bermudianbrit yeah, to be honest I thought the writing was on the wall for it a long time ago when the "volume mixer" panel got added to the modern sound settings area, since that does give you one extra thing to change per-app audio device you can't do in the old one. But nope, still there in insider builds, too. Same with NCPA.CPL, which brings up, good old Network Connections from Windows XP days. Ah yes, 20+ year old code, yum yum.
@cublanco @bermudianbrit Windows. "When using the GUI takes longer than using the commandline!" :D Something coders never thought would become a reality but they always knew deep down even if command arguments get complex for most but boy do they get a lot done. xD
@cublanco @bermudianbrit The first Macintosh I used (they still called them Macintosh back then) was running System 7.1 in 1992. When I got my first modern Macbook Pro (2011), a lot of those old concepts still held true. So, there's that.
@FreakyFwoof @BorrisInABox @cublanco @bermudianbrit ooooh yeah that's a great one, especially since they quietly changed the windows +X menu to use the newer settings page for "apps and features" and not the one that only showed the old win32 programs, love that one.
This is my answer to what do I hate about modern interfaces. They all suck in some way and no longer allow this. Either it is no first letter navigation, up arrow does something unexpected, home/end keys don't move the value to the min/max value, focus jumps somewhere where it shouldn't, etc... Good old days where keyboard navigation was logical and predictable.
@jscholes ha. Yeah same same! Always trips me up when working in Unix shells. I only did it to have punctuation of various synthesizers spell it out for others to not have to review manually.
If you want to make a hotkey for this, go to c:\windows\system32 and find the file named mmsys.cpl. Context menu, send to, desktop as shortcut. From there, go to properties and make your hotkey. Same for ncpa.cpl.
does anyone have any recommendations for screen readers that run on Linux? I want to start testing webapps I code (and other people's sites) to make sure they are accessible to people using a screen reader but I'm not sure where to start.
Did you know if you press CTRL + ALT + Tab, you can release the keys and the ALT + Tab UI will stay open for you while you pick which window you want to switch to? 👀
How on this earth have I been using Windows for 30 years as a keyboard user and somehow missed this? I can't thank you enough for this tip. Please share more tips with the community as time allows. This is fabulous.
Fun fact: A trip down memory lane: When Hungarian as a voice was added to ESpeak, I was a 15-16 year old kid in high school, and reached out to Jonathan Duddington directly thinking I wouldn't get a response back. Not only did he respond, but he used my recordings of Hungarian text read out to create that voice. Since then other Hungarians have shaped the phonemes, but that story will live down in my mind as a good one. That and I got to talk with Jonathan Duddington, like how cool it was.
I was told about that one. The Hungarian version of espeak just, sounds good, a bit too good, so I asked and I was told of part of its story. It was fun! :)
The more I do see the BTSpeak and use it though, the more I understand that it's not just a Raspberry Pi 4. A lot of critiques will say (myself included) that they aren't paying much to produce them, but it does look like the keypad is custom-made for the units with BRLTTY driving that keypad for the experience. Even that alone takes time and R&D to do, so having worked in corporate environments I get the cost and pricing in this regard more and more the deeper I do explore it.
a lot more goes into this than just a Pi-4. To those critiquing who say, "I could just buy my own Pi4 and do the same things!" you could, but that's if you're technical. The appeal for many folks this has is having that hardware and the R&D for that hardware in one package. Maybe producing them is a few hundred dollars of the cost, but also needing to do this in bulk, injection-mold cases, assemble carrier board +RPI units into the casing, verify hardware units work, pre-program info... A lot.
And while it's true that they are using an open-source board to drive their project and not contributing this back to the community (yet) is a bit of a sore spot for me, but it's also something that could in theory change once they have finished building the system in the product and then over-time open-sourced parts they've built in terms of scripts publicly. Who knows, let's not judge the food before the water for it has even boiled.
and really the more I read about the CM4, it's not unreasonable for them to use it in this industrial way. Nothing in utilizing a CM4 says that you then also need to make other aspects of the hardware open-source, like board/ modifications for modules you've added. As an example, they use the wm8960 with the CM4 (https://www.waveshare.com/wm8960-audio-hat.htm), a keypad which drives its own signaling for the 9-key input, battery, and more. It's a lot of small pieces put together into one product
I don't mind that they are using an RPi board, I just wish it was the 5! I know there will always be new versions but these things always feel obsolescent before they're even available.
At the same time, no one complained about the board of the BNS feeling "old," you just wanted it to do more stuff. So I recognize that this too is more of a hobbyist complaint than an actual negative for the device.
@objectinspace yeah, I wish the CM5 board were released already - sadly there's a delay between the foundation releasing the general-purpose board and then the 32 SKUs or more for the embedded, custom CM4. They could have gone with a slightly higher CM4 with 8 GB ram, but my guess is they expected people to not need that much for a good Linux desktop experience, which isn't wrong either. In theory you could send in units though and get them upgraded with a CM5 swap.
That's true they could swap out the boards, I hadn't thought of that. If they were willing to do that, that would increase the lifespan of the device quite a bit! They might want to prefer to make a new device though. But that's when I think they would lose people, I can get on the $200 yearly upgrade train but not the $4k-every-two-years-for-a-new-thing train. Hopefully they understand that.
@objectinspace Unless they need to make a significant change to the rest of the hardware, I would hope that these units could be upgraded to the new boards. And now that they're not selling the basic units, which were identical to the pro hardware as I understand, perhaps a combined PMA/SMA could give them a nice source of ongoing revenue.
@chrisn @ppatel @objectinspace yeah, that's not an invalid point. I think you could install a smaller passive heatsink that's low-profile and on the CPU, that could help it some but the case may not leave much room for it either right now, I haven't tried. The heat transfers through to the bottom but has not gotten uncomfortably warm for me, just quite toasty during heavy compile operations
it's not about the board for me, not at all. The problem is that they're using gpl software, and presumably slightly modifying it to fit their needs. They use mate in desktop mode, they use nano and maybe a gui editor soon, they use brltty, lots and lots of other foss components. Not open sourcing their work on those is a gpl violation if what I think is happening actually does happen, and even if not, it's kinda keeping the status quo and reinforcing the stereotype that assistive technology is expensive because governments and businesses have deep pockets, and it's not actually about the people, while not giving any of it back to the community, at least to the contributors on the backs of which they built their platform. I do like what I heard from the device, I like what's happening if taken in isolation, I just hope there's gonna be a gradual open sourcing of some of the components, especially those which touch gpl code. I'm mostly glad people are making stuff with linux, the OS everyone thought very hostile for blind and VI people, and that more people are discovering and may get some interest in it as time passes.
@esoteric_programmer +100 to this! I didn't deeply explore all the modifications, It makes sense that Dave Mielky leads a lot of the developments for it since he's the now-maintainer of BRLTTY as well, so a lot of what drives the underpinnings was done with collaboration of the developer directly. I also get that some of the applications are getting built by other developers, even if these are Python scripts with bash TUI underpinnings that are simpler, Blazie has rights to.
@esoteric_programmer ah yeah, just to them in general, since those pieces of code were contractually developed for the product and constitute its user interface. I'd also be curious as to what extent they had to modify some of those other open-source tools and whether those changes could have potential to be used outside of their product environment (which could add a wrinkle to contributing them back if they are there to accommodate more than just BRLTTY)
yeah, no one says anything about their own components developed for them, they can open source that...or not, what matters is they should at least follow the gpl and contribute back whatever they modified that's covered by the gpl, doesn't matter the kind of patch and the reason for it. Offtopic for a sec, but does the unit still heat up if you make it go to sleep? if so, I'm assuming the command that's being ran for making it do so is systemctl suspend, isn't it? and when you shut it down with the button, it does a systemctl hybernate, right? if not, in the first case it won't suspend properly, and in the second, without hybernation, it's an actual cold boot every time, which is gonna make the unit turn on as slow as my computer, which is not acceptable for a note taker imo. There are ways to do it without the systemctl commands, but those are the most recommended, universally supported and portable across architectures.
@esoteric_programmer yeah, technically it looks like the Power button itself sends a command and executes it through BRLTTY to also lock the rest of the keypad buttons. It's not even S3-sleep though or anything right now, just a keypad lock with background operations continuing or idling at 600 mHZ, whereas holding it down just does a full system shutdown, so your cold boot theory isn't wrong - it's awake in 25 seconds or so in that instance, which is OK but not great.
@Jage This hardware has been available for years and years. Anyone could have made one, at any time, but didn't. I've been in multiple serious discussions about making a Pi-based notetaker with bonafide people with a history of developing hardware and software AT products. Talks have always ended with "It's a lot of work, and blind consumers will be cheap entitled idiots." Respect to the Blazy folks, but it's going about as we expected. Sad.
@Jage That's right. These are people who are outraged a sandwich is sold for 3 times the cost of it's ingredients. But a sandwich is more than ingredients. The restaurant is paying rent, utilities, kitchen and wait staff, insurance, taxes, credit card fees, and trying to recover costs for tables, chairs, kitchen equipment, plus costs for spoiled food, on and on. They can charge 3 times the cost of a sandwich's ingredients, and still barely make any money.
Today I have found out when enabling subtitles on #youtube the area where the subtitle track is playing is no longer configured as so called live region so youtube subtitles are no longer read automatically with a #screenReader if it's running. I've also looked at #invidious which has aria-live set to off. This is unfortunate as I was expecting this to just work. It can be tested on this video if you like https://youtu.be/watch?v=3i7gkN-1sAIhttps://yewtu.be/watch?v=3i7gkN-1sAI
The desktop Linux ecosystem has changed dramatically since accessibility was first implemented in GNOME over 20 years ago, and even since the accessibility stack was migrated to D-Bus roughly 15 years ago.
From the Lighthouse for the Blind, Inc.: The world at your fingertips: Essential Tech for Travelers Who Are Blind or Visually Impaired; Wednesday, April 24, 7:00 PM Eastern Time https://groups.io/g/tech-vi/message/6769
Microsoft announced their latest round of FOSS fund recipients. We're thrilled to share that @NVAccess are among this quarter's recipients. From: https://github.com/microsoft/foss-fund
"A project of the Microsoft Open Source Programs Office, the FOSS Fund provides up to $10,000 USD in sponsorships to open source projects as selected by Microsoft employees."
Congratulations also to The GNU Compiler Collection, Urllib3, CLAP & MSW.
The Microsoft FOSS Fund provides a direct way for Microsoft engineers to participate in the nomination and selection process to help communities and projects they are passionate about. The FOSS Fun...
Congrats, both for being selected this year and for possibly being the only project to get funding twice. It also shows there aren't just a tiny scattering of people who are onboard with accessibility at Microsoft, which is a useful reminder sometimes.
@simon MS seems like the only big company who really takes enterprise accessibility seriously. If you've never seen Azure for example, you'll be astonished how fricking good it is. Sure, it's not perfect, but the vast majority of things are labeled, modals are actual aria modals, notifications come through live regions, the cloud shell is actually accessible, controls have proper roles, landmarks are there etc. I even wrote about this on Twitter once and the feedback was tremendous, I'm reliably informed that people "up top" in the Azure team saw and discussed my tweet in some kind of executive meeting. I've also played briefly with other enterprise tools of theirs, including Azure Data Studio, PowerBI and whatever the SQL Server thing was called, and they were all great accessibility-wise. Same with VS Code, Visual Studio, the now discontinued Visual Studio for Mac, it just all works. Compare this to something like App Store Connect, which I fortunately never had the misfortune to use, but which barely works with a screen reader if you can say even that.
@lizhare @simon Um, that's not Microsoft I don't think, that's Salesforce. They're another company that cares, but I don't exactly consider them to be "big tech".
Good time for @podcast to interview these folks about the future of screen reading technology and other matters of agreements and disagreements. Would be fun cc @doubletap
Now we just need to find a way to run .bas files on the BT Speak. I'm sure something like this might be easy to implement: https://robhagemans.github.io/pcbasic/
Supporting tinkering and general "power user" shenanigans in your software is a genuine accessibility thing tbh. I see some people complaining about Linux users who use a non-systemd init system, X11 instead of wayland, or just generally some "non-standard" configuration. The attitude seems to be that since these users are "smart enough" to stray off the sacred defaults to suit their needs better, there's no need to consider them at all because "they can figure out how to make it work".
Obviously it's not possible to satisfy every accessibility need and use case simultaneously, but it's still harmful to categorically dismiss and actively remove support for a certain group of users just because you perceive them as "gifted nerds". Someone who extensively modifies their system to make it fit their needs may be able to solve the problems you've introduced to their usage of your software on their own, but every single application that needs this extra wrangling adds work and friction to the pile. It's kind of like replacing a wheelchair ramp with stairs because "they have it better than armless amputees, they can just bring a plank of wood with them"—yeah, TECHNICALLY they can do that, but such an arrangement is only desirable if your baseline reference is "being pelted with stones".
Also, while we're on that subject, "just use the defaults" is kind of like suggesting that wheelchair users just get someone to carry them around like a sack of potatoes all day. I don't think it's acceptable to force people to give up more autonomy/control over their lives than they absolutely have to in order to get support, especially in circumstances where a better, more equitable solution could easily be found. Perhaps I'm being somewhat extreme with my comparisons, but it's more about the similar mentalities and attitudes at play, not equating the severity of the access needs.
My overall point is that the split between "software for nerds" and "software for normal people" is due to a difference in use cases and access needs, not just a superiority complex in the minds of either group. You can build software that accommodates both groups, but you have to actually listen to them and not unilaterally dismiss one as egotistical and obstinate. And you should treat choosing not to accommodate either group similarly to choosing to ignore any other access need: necessary in order for you to better serve a specific demographic, perhaps, but certainly not a decision with zero negative consequences.
@matt Thanks Matt! I should also stress that this is a piece you kinda need to read all the way through. A few people have come away from reading the headline with the idea that I just don't like git commit signing specifically and they should turn it off as opposed to turning it on. For some people, in some situations, it makes perfect sense and it's a fine security primitive. To quote one of the last paragraphs of the blog:
> Git commit signing itself is not particularly consequential
As Snikket is still not available on the #Google Play Store, we've published a longer blog post about the situation, how to work around it using #FDroid, and the long path of stupid Google review responses that led us here.
very minor clarification: Google had accepted the version that reinstated the contact permission. This was live for a couple of days/weeks until they suddenly changed their mind. They didn't just outright rejected the update. https://gultsch.social/@daniel/112084175241312889
Interesting that you still had contact list integration soft disabled. That confirms that Google believes contact permission == uploads data. Considering how Google makes money that kinda makes sense I guess...
Google had accepted a version of #Conversations_im with address book integration that explicitly asks the user for consent to process the contact list locally on their device before requesting contacts permission. Our privacy policy included explicit wording w.r.t. local processing of the contact list.
However it’s unthinkable for Google that someone would request contacts permission and then not upload them. A few days later they changed their mind and threatened to remove the app again. #XMPP
What if Mastodon had a way to save the ALT text you entered each time you used a new image. So if you used the same image later, it would some how know what you used the last time and give you the option to automatically use that instead of having to type something each and every time.
Seems to me there would be some way to be able to do this easily and to save all of us time from having to type the same thing for ALT text on a regular basis.
An EXIF metadata field would address this and more. It wouldn’t be a perfect replacement for alt-text in an HTML attribute or ARIA because it wouldn’t describe the image in context, but it’d still be a welcome addition.
Ok, so the word "pistol" has a fascinating etymology.
In English it comes from French "pistole", which comes from German "pistole," which originates from Czech "pišt'ala" meaning "whistle" (like a tin whistle or flute).
But in Czech the word for "pistol" is not "pišt'ala." It is, in fact, "pistole." This is a borrow word from either French of Italian "pistole," which comes from German, which comes from Czech "pišt'ala." In effect, the Czech word for "pistol" is a borrow-word-twice-removed from Czech.
Dnes si dovolím odkázať na veľmi vydarený český hlas pre hlasový výstup #TTS #RHVoice. Sám autor o tom píše tu: https://groups.io/g/Blind-android/message/1743 . O použití pre #Windows #Linux a #Android sa dočítate aj na jednoduchej komunitnej stránke https://hlas.ondrosik.sk/ . Gro používateľov sú ťažko zrakovo postihnutí používatelia, ktorí si bez kvalitného hlasového výstupu a čítača obrazovky nedokážu svoj digitálny život predstaviť, možno ale aj vám by sa mohol hodiť takýto hlas pre váš počítač či smartfón nezávislí od obrovských korporácií. Mohli by ste ho napr. použiť na čítanie kníh alebo počas navigácii cez GPS.
@snikket_im just started their new federated instant messaging hosting service (based on #XMPP) for everyone.
This can really be a gamechanger for instant messaging.
You can bring your own domain name if you want to.
This means: for about 5.50€ (6$) a month (according to the website) you get ten instant messaging accounts for you and your family/friends using your personal domain name (not included).
And you are able to talk to every other XMPP user out there.
I've been working on a new app to use your #Linux phones and tablets as secondary wireless "side displays" (not mirrors). This is based on #GStreamer, x264, #Mutter APIs, and low-latency UDP streaming, and will be designed to work on both ARM and Intel, regardless of hardware acceleration support. (Testing on a #librem5)
This is still at an early stage, and will take some time to become usable. Thoughts and feedback?
I discovered a new GPS app called WaveOut recently that I am extremely impressed with. It uses directional sound similar to Soundscape, but instead of having the sound come from the direction of your destination, it creates a turn by turn route, and the sound shows you where the next point on your route is. In addition to GPS, it uses your phone's camera to better determine where you are. I tested it on a route where the directions from Google Maps are inaccurate and hard to follow, since there is a confusing roundabout crossing, places where there's only a sidewalk or grass with no street along side, and some very weird intersections, and I've been using Voice Vista, an improved version of Soundscape, and trying to choose the best path that leads me towards the sound. I was very skeptical that WaveOut would give me better directions, but it was pretty much perfect! When there was an upcoming turn, or when I got slightly off route, it played a sound leading me directly to where I should go, and otherwise it was silent. It led me a slightly different way than I'm use to, but it never gave me inaccurate or misleading directions or made me guess where to go, and it was definitely easier than trying to figure out where to go based on the beacon in Voice Vista and Google Maps directions. The sound also kept up with the direction I was facing better than Voice Vista, and it also seemed to track my location more accurately than other apps. https://www.dreamwaves.io
We’re seeking input from #FOSS maintainers as we design a fellowship program pilot. We want to test a support mechanism that addresses structural issues in the FOSS ecosystem, and support maintainers who work on open digital infrastructure in the public interest.
We're excited to announce the release of #GNOME46! This release brings many updates and improvements thanks to the hard work of #GNOME contributors. 🎉 https://youtu.be/r_QyRJf3rtQ
The GNOME Project is proud to announce the release of GNOME 46, ‘Kathmandu’. This release brings remote login with RDP, variable refresh rates (VRR), many accessibility improvements, enhanced notifications and improved Settings.
Great Release! All those constant refinements really enhance the usability even further! 😀 And thanks, especially, for moving the current file actions to the sidebar. Gnome 45s version was not visible enough, so I missed, something was still going on, several times. I absolutely love Gnome and don't regret engraving the logo on the super key of my @tuxedocomputers laptop ❤️
Vielen Dank und ich freue mich auf diese Fedora Version wo #GNOME 46 der Standard ist. Aber seien wir ehrlich. Welche der mitgelieferten Apps sind brauchbar und was fehlt? Was ist mit Card/Cal DAV Support? .
Wohooo! Congratulations, and well done to everyone involved!
Thank you devs, thank you designers, thank you translators, thank you bug reporters, thank you QA people, thank you feature suggesters, thank you money donators, thank you evangelists, thank you documenters, thank you people working tirelessly in the background with keeping up the websites and communication channels and schedules and meetings and brewing coffee and everything else that's needed for #Gnome to be all that it is!
VRR finally landed! Can't wait to not have to keep updating mutter-vrr from the AUR to enjoy games on GNOME! GNOME's VRR implementation has been way better than KDE's in my experience using the patched mutter.
..still no mouse driven, simple add a desktop App icon? till you learn this .. remain banished in the paleolithic, no deployment, till you you can demonstrate basic userland concepts . . . (without jammin' your mandatory 'efficiency' #BS work flow agenda down my neck, how the f___ do you know what I need?? You did not ask! ) .. so, its #KDE #XFCE #LXDE for me and my users.. who sometimes just ~ want~ it on the desktop.
Thank you to everyone involved for all your hard work! There's no OS that comes close to GNOME + Linux for helping me concentrate and get things done. You've created something magical that gets better and better each update.
If you had code on GitHub at any point it looks like it might be included in a large dataset called “The Stack” — If you want your code removed from this massive “ai” training data go here:
I found two of my old Github repos in there. Both were deleted last year and both were private. This is a serious breach of trust by Github and @huggingface.
Also the repos i found of mine i’m sure were private, but even if they were public at some point, for a brief time, in the past that isn’t my consent to use them for purposes beyond their intent.
--- Edit 2 -- I see this made it to HN, which is a level of attention I do not want nor appreciate....
For all those wondering about the private repo issue -- No, I am not 100% sure that these ancient repos weren't at some point public for a split second before I changed it. I do know that they were never meant for this and that one of them didn't even contain any code.
If my accidentally making a repo public for a moment just so happened to overlap with this scraping, then I guess that's possible. But it in no way invalidates the issues, and the anger that i feel about it.
Hello. I made a recording of the BT Speak. Do you wish to listen? It's here. @btspeak @dblazie It goes into upsides, downsides, the power button problem, more. Update: Here's the edited recording file, which includes info on the Micro SD slot now too, and normalizes volume a bit across it: https://eurpod.com/BTSpeak_demo.mp3
@technocounselor oh thanks! It was definitely on the whim, and what was crazy? I posted about the power button on the BTSpeak mailing list. After I made that recording Bryan Blazie called me to help me fix it even though he was in between flights at an airport and only had 5 minutes. Just wow. Immediate customer support. When else do you get that? Give me one company.
@technocounselor I remember the old Blazie Engineering company had virtually 24/7 tech support. Like even on weekends, if you had something wrong, you called them up and they'd either tell you how to fix or get you to send it in and find ways you could ship it to them. Absolutely zero has changed about that spirit today and that probably excited me more than it should have too about it.
@technocounselor If this thing could somehow help me learn linux, I'd be up for it. Don't know if there aare any text/audio tutorials out there etc. For example, I installed a text adventure/interactive fiction game interpreter on the Icon Braille plus through its command line. Don't remember what the name of it was or exactly how I did it but it ran great, spoke all text automatically and everything. Wonder if the pro version of the BT can do that?
@jpellis2008 @technocounselor I don't see any reason why the BT Speak couldn't assist in learning Linux. Bear in mind that Blazie Technologies doesn't itself offer Linux tutorials but there are tons of them available online. Maybe some enterprising user(s) will craft a series of "Learn Linux With Your BT Speak" tutorials.
just a bit of a worning for anyone who has the new BT speak, do not allow the linux kernel to be upgraded. my unit is now bricked and no longer boots after the system upgraded the installed packages with sudo apt update then sudo apt upgrade. I'll have to look into this one as the system had to be hard powered off and now I get no feedback after pressing the power button.
Because apparently Google can't be bothered implementing decent keyboard #accessibility for Google Calendar, I'm currently trying to throw together a Greasemonkey script which changes right arrow so that it moves to the next day when you hit the end of the current view instead of just doing nothing. You can currently press n to go to the next day, but then, it does this absurdly useless thing where it... focuses the first day in the scrolled view instead of the last, which is normally at least a week before the day that just had focus. 🤦♂️ It'd be nice if these trillion dollar companies could do things properly - this stuff just isn't that hard - but as usual, I find myself doing their jobs for them.
Would you perhaps consider making the script available for others? I’ve been dealing with this issue for a very long time and a fix would be lovely. understand if not, though.
Greasemonkey scripts to improve the accessibility of various websites (e.g. Pandora) and to solve common web accessibility problems (e.g. windowless Flash). - jcsteh/axSGrease
All of the cell technologies I've encountered at the low-cost end of the market have usability downsides, so I end up buying the more expensive Piezoelectric braille displays. The most interesting technologies I'm aware of currently are those by Dot Inc., and by Tactile Engineering, both capable of supporting multi-line Braille/graphics-capable displays. A patent search will show numerous other proposals developed over the past several decades, few of which have ever become publicly available products.
The more I learn about Glidance, a system that purports to replace the cane, the more it seems to have taken all the criticisms of "smart canes" into account. It might be the first genuine high-tech navigation aid that could be beneficial, replacing the cane in most circumstances. I'm usually highly skeptical about these things because people just havent thought through the implications including ergonomics, weather conditions, etc
A new era of independence for the visually impaired Glide is the first self-guided mobility aid designed to help individuals with sight loss navigate the
@vick21 @Tamasg From all I've seen, the team is taking a very thoughtful approach to the design and implementation of various features. I'd like to see them succeed. I know many blind people who will never be good with guide dogs but can use them. For variety of reasons, they will also not be good cane travelers. This product is going to be positioned in that middle spot for those people.
I'll probably by this. A guide dog is something that's unreasonable for me to have given my current environment and job, and being able to combine my GPS navigation, that I rely on heavily anyway, with my primary mobility ade would be incredibly useful. Due to my traveling places without any or very little notice, route learning is not something I can do much of.
@fireborn It's a perfect solution for people like us. If I know that I'm going somewhere, I plan my route with Soundscape. On Occasion, I'd like to just not worry about it too much.
Did you know that a lot of #Dune 's Fremen language is just Arabic? Some examples:
Muad'Dib: Teacher Lisan al-Gaib: Teller of things yet to come. Shai-Hulud: Eternal thing, or Eternal Shaikh/Sheik (old man). Mahdi: The guided one Arrakis: the dancer, the name of a star (Mu Draconis A)
The accuracy of Mozilla Location Service (MLS) has steadily declined. With no plans to restart the stumbler program or increase investments to MLS we have made the decision to retire the service.
Suggestion I would like to specify an alternative text for all media content uploaded to Matrix including casual uploads, user avatar and room avatar, and I think this should be supported to comply...
Suggestion I would like to specify an alternative text for all media content uploaded to Matrix including casual uploads, user avatar and room avatar, and I think this should be supported to comply...
Musharraf :verified:
in reply to Musharraf :verified: • • •Tamas G
in reply to Musharraf :verified: • • •Musharraf :verified:
in reply to Tamas G • • •If your existing voices are trained with Piper, then they'll work with this version.
If they fail to work for any reason, you can copy the config from any working voice to your voice, and edit relevant values.
Tamas G
in reply to Musharraf :verified: • • •Devin Prater :blind:
in reply to Tamas G • • •Musharraf :verified:
in reply to Tamas G • • •@Tamasg
If you have the original checkpoint, you can convert it to the new format.
Take a look at this script, which I used to export Piper's checkpoints:
https://github.com/mush42/piper-rt-maker/blob/main/tasks.py
I need to update the docs, and add a section on training voices.
piper-rt-maker/tasks.py at main · mush42/piper-rt-maker
GitHubTamas G reshared this.
Tamas G
in reply to Musharraf :verified: • • •Musharraf :verified:
in reply to Tamas G • • •Here're the steps to convert the checkpoint to fast format:
# Clone piper fork containing export code
git clone https://github.com/mush42/piper
cd ./piper
# Checkout streaming branch
git checkout streaming
cd ./piper/src/python
pip3 install -r requirements.txt
# Upgrade torch
pip3 install --upgrade torch pytorch-lightning onnx
source ./build_monotonic_align.sh
# Export. Edit paths
python3 -m piper_train.export_onnx_streaming --debug [checkpoint path] [export directory]
GitHub - mush42/piper: A fast, local neural text to speech system
GitHubTamas G
in reply to Musharraf :verified: • • •Tamas G
in reply to Musharraf :verified: • • •Musharraf :verified:
in reply to Tamas G • • •I didn't edit that script, it came from piper repo.
Anyways it does not effect the installation. I encountered it myself when exporting voices.
Tamas G
in reply to Musharraf :verified: • • •Musharraf :verified:
in reply to Musharraf :verified: • • •After installing this version, you will lose all of your installed voices. Please use the voice manager to re-install the voices again.
Aryan
in reply to Musharraf :verified: • • •Nick Giannak III
in reply to Musharraf :verified: • • •Musharraf :verified:
in reply to Nick Giannak III • • •A dataset designed specifically for screen reader usage, goes a long way toward creating a good quality voice.
If guidelines are the issue, we can come up with a set of guidelines based on Microsoft/Google guidelines which are openly available.
Nick Giannak III
in reply to Musharraf :verified: • • •Timothy Wynn
in reply to Musharraf :verified: • • •AttributeError: 'WinmmWavePlayer' object has no attribute 'setVolume'
Pratik Patel
in reply to Timothy Wynn • • •Musharraf :verified:
in reply to Pratik Patel • • •Which voices?
Custom voices or the ones downloaded from the voice manager?
I'd appreciate it if you can provide NVDA logs.
Pratik Patel
in reply to Musharraf :verified: • • •Peter Vágner
in reply to Musharraf :verified: • •Thanks for all the fantastic work you are putting into this.
Musharraf :verified:
in reply to Peter Vágner • • •Here's how to build the sonata-grpc binary:
git clone https://github.com/mush42/sonata
cd ./sonata/sonata-grpc
# With Rust installed
cargo build --release
GitHub - mush42/sonata: A cross-platform engine for neural TTS models.
GitHubPeter Vágner likes this.
Musharraf :verified:
in reply to Peter Vágner • • •If you just want to set the eSpeak-ng data directory, you don't need to re-build the binary.
Just set the following environment variable before launching sonata-grpc:
SONATA_ESPEAKNG_DATA_DIRECTORY=[your custom espeak-data directory parent]
Peter Vágner likes this.
Peter Vágner
in reply to Musharraf :verified: • •