Last month, there was an organised walk around Gent Zuid, aimed at adding surveillance cameras to #OpenStreetMap with @MapComplete. This map of the area shows the high number of cameras in this area alone (this is zoom level 18!), but also, that they're now mapped, allowing anyone to check where there's cameras with a reasonable degree of accuracy 🥰
In 1976, Congress set fire to the country's libraries; in 1998, they did it again. Today, in 2024, the flames have died down, and out of the ashes a new public domain is growing. Happy Public Domain Day 2025 to all who celebrate!
--
If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
So now, my website has 3 speech synthesizer add-ons. Perhaps this weekend I'll make a simple page for it or extend the Piper voices page to include them: Brailab, developed in Hungary, Rom dump from original hardware box and dll obtained from Brailab source of the CD version. eurpod.com/brailab2024.nvda-ad… BeSTSpeech (Keynote gold): eurpod.com/BestSpeak.nvda-addo… and SoftVoice, with many unfixable bugs. eurpod.com/softvoice2024.nvda-…
Yeah, I noticed mellow mail is like a very childish choirboy! lol. But never been a fan of that synth, I'm more of a votrax, doubletalk and keynote guy
@rommix0 ooh, not anymore - it's long lost over multiple hard drive migrations. Perhaps @erion may know though. Would be good to have it again just to understand any speech parameters I've been missing.
@rommix0 @erion yeah, the dll is there :) And the add-on should give a fairly good rough idea of how the TTS_StartSay function works to send speech to, it doesn't require piping the text to a hidden window thread so a lot simpler to get going.
@rommix0 @erion think the only downside of that TTS is just the low-quality of the speech samples (think it's not even 11K but 8K) although the .bin files are probably direct 1:1 rom dumps of the hardware chip's that was producing it so probably not much to be done. What makes it unique is that it phonetically pronounces letters, rather than as their full-letter counterparts, so you can get it to beatbox and do other very cool sounds quickly :D ahaha
@rommix0 @erion I think I'm still astounded by the Resilience of some of these engines, the Brailab one we (Read: Erion) first made in like 2013 or 2014. I was still just barely out of college so in ten years time have just lost the source. Might be able to get in touch with the other contributor (áron) who also had it. Don't think we could ever release it since as I recall at the time BME-TMIT gave it to us for private-use only, though may be wrong on this.
@TomGrant91 @rommix0 @erion if someone could pull that off, I'd be super amazed. It would require a lot of tracks and mixing of them and normalizations, but now that we have access to them I guess there's never been a better time to create a speech synthesis choir with a bunch of parts.
@rommix0 @erion If only synthesisers could beat box! lol I know what you mean Rommix lol. I can just imagine Bestspeech beatboxing accompanied by DECtalk singing now!
@rommix0 @TomGrant91 @erion yeah, I remember you could run this one - Homelab preceded Brailab, and was a full speaking computer system prior to them just releasing the synthesizer component. Nice find though, I like that this has rom dumps from several generations so that may be something fun to play around with.
The Brailab DLL is not from Brailab CD, it's from the original hardware, just partially rewritten to compile and run on Windows. When I was working on translating NVDA and of course exposing Hungarian blind people to it, we made an agreement through a Hungarian organisation with the original developer to make it available for Hungarian NVDA users. We also needed some extra functionality for indexing, as well as playing wav files, which would have replaced the synth's spelling feature.
Implementing this was not quite feasible, but they did give us the source code, should we wish to do this ourselves. About a year later, Vocaliser became available and due to its more natural speech, Brailab was mostly forgotten. Some people who grew up using the hardware or later the CD variant still use the addon today. They have also developed an overlay for Android which uses Brailab as its synth.
I am legally bound by this agreement, so I am unfortunately unable to disclose the source code, sorry.
lol, especially with how vocalizer for the longest time struggled with Hungarian words that had double S's (like 'kissebb" for littleest) - would always say it as an sz. It's a bit of a shame really because I think having Profivox (which of course the blind community cracked out of JAWS anyway) is far better than Vocalizer as it's also home-made, but of course by that time JAWS had exclusivity on it so NVDA was stuck with the inferior TTS option. (There was also Multivox, but its voice is not one I particularly miss.) But yeah, figured we couldn't distribute that source. I had gone into maintaining the driver because nobody in Hungary cared about it. I even paid Tyler like $50 at one point to add Indexing for me.
I had more struggles, small short i or long í it's a pain! Especially if you're trying to study the language. Might I add, with 0 success, I still need a teacher. Hopeless me, people want ridiculous prices. Anyway, Vocalizer for Bulgarian also sucks balls. Absolutely awful, disgusting!
Espeak is better than Vocalizer in these two languages. Period. You can't argue with me. Grins evilly, but Bulgarian also has neural voices which suck when trying to read English. Speechlab from when I was born is the best. And because there aren't other blind Bulgarians anywhere I've seen on fedi, I can't get kicked in the replies for choosing the older speech. Also have signed NDA with the blind union here, and the voice contains my data, I initially had to activate it with, otherwise I could give you the thing to play with, although I have no source code, so enjoy disassembling it. :D
Anyway, me and understanding of tts isn't the best. But you all get a hug. Because I decided so. Also free, a word I can't spell in English free of illnesses, and virtual. Enjoy! And may all your tts endeavours be successful! I'm back to reading. And also, I beat Erion in books of a series read, so celebrate my achievement, or not. With a sip or 5 of water. :D
Ah, Multivox was the first Hungarian synth that everyone used with Jaws before it was translated, there was no eSpeak, only Brailab, WinTalker Voice and Multivox. Then Profivox came, which was licensed exclusively for Jaws, and the underrated SpeakBoard hich was sold even in supermarkets on CD, then Vocalizer, which is everywhere now.
It has loads of issues, Nuance clearly did not test it, nor listened when lots or developers (including myself) complained. They have added horrible pronunciation rules that cause common words to be mispronounced and there's only a partial fix. But I digress.
There was no manual that I know of, we just got a zip file with the source dump and a few lines on how to compile it. It had 7 various parameters to my knowledge: rate, volume, pitch, emphasis, spelling, weird mode (this just modified formants I believe), and another one I am unable to recall at the moment. Most of these were implemented in the NVDA addon, the useful ones anyway.
I might be able to find the source, so please let me know if you need me to look up anything.
Let's. Just stick to English. And I also don't want to correct every Bulgarian sentence. Laughs. I'm pretty honest, but that's like giving fire to my OCD. The more incorrect something of a language I know and I'm fluent is, the worse I am. But I digress from our TTS subject. So, we have neural engine synths for Bulgarian, but you can get them only if you're a member of the blind union here, and the Gergana you mentioned, that's also Speechlab, is super natural with the neural engine thing.
I have somewhere an installer of this neural thing. Despite it being available from the blind union, somebody did an illegal share *smiles*. Btw, i am member of the Bezmishka mailing list.
If you have an ID number, a full name and an activation code or managed to crack it, good for you. If you have two voices, I think you might have the latest version. If someone managed to make this crap neural tts faster, that would be great. But there's this thing, Bulgarians are assholes, they don't want to give out source codes.
I wish we could get Profivox neural somehow, I heard that's becoming a thing but beyond some web demos haven't seen much. Ah well. Maybe some day, for now I still don't mind the Hungarian Wintalker voice which, miracles, runs on Win11 :D
oh yeah! So far the RHVoice Katalin is actually fairly good, except when it says the letter F her voice cracks horribly, but I mean in general I like that one over the other 2 Piper-based ones right now up on rhvoice.hu xD
I'd say it has more problems with the letter S. But that's understandable, all these problems are because of the database itself. You should have heard how the original one sounds like. It's horrible.
This cracking will be fixed when i fix the pboblematic sentences in dataset. I need to grab somewhere the exception words in Hungarian. Maybe you can help me? thinks like cukroszacskó and stuff.
And yeah, Profivox is the best thing we've ever had, withthe exception of not supporting other languages, such as russian for example. I would need both hungarian and russian, that's why I wanted RHVoice to be created at the first place.
Hmm, so would that need recordings or is it phonetic adjustments you're more after on them? For sure interesting! I think we do have the best chance with RHVoice as far as a solid cross-platform one, at least one which would well rival the competition like Vocalizer and be better than it.
~17% of registered voters in the riding of Cloverdale-Langley City have elected the same despicable pro-conversion therapy CPC candidate and MP it has in the past...
Nový OP obsahuje aj nový vizuál. Na oboch stranách sa nachádzajú slovenské motívy. Zmena sa týka i rozmerov fotografií, farby preukazu, doplnené boli aj
do teraz som ju nepočul a sú to kkti nie fašisti, s prepáčením. V škole dnes v triede na stene ako sedí učiteľ som si mal možnosť prečítať prvé dve state hymny a niečo z ústavy, neviem čo si mám o tom myslieť.
This is weird, I'm not getting the weird artifacting with Bestspeak all the time, but I was somehow able to almost consistantly reproduce it when going between the desktop and the Tweesecake window, I wonder what you guys think @Tamasg and @rommix0
that's an odd one! It might be a conflict with another add-on (sounds like you have the earcons one, haven't tested it alongside it admitedly) - it may be that the interruption gets locked mid-chunk while the audio plays, or at the very least it's causing some delay to bstShutup and invalidating the handle. I think it's interesting because for 90% of people it's working fine, now there's just that last 10% of random glitches and quirks that some are still getting here and there but I'm having a much harder time to repro in my environment. (would be interesting if you can get it back and if so, whether disabling that add-on and restarting would cause it.)
Yeah, I use Navigation sounds, so it could be that, maybe try installing that addon or something idk, its very strange, as it was also throwing a fit when in the live chat of a friend's YouTube stream, clickable then random sounds, and at one point button id icon beep *NVDA reboots itself*
So you can cut a whole extra door into the side for easier maintenance access to the toilets but it is impossible to arrange the windows so they fit the interior?
Wow, a nice surprise! We just heart that @codelutin@mastodon.libre-entreprise.com are to donate 4K EUR to our developments -- see their post (which unfortunately doesn't currently federate correctly it seems)
(FWIW we have a money-contribution page at delta.chat/en/donate and are happy about donations to mitigate a potential financial crunch early next year)
Donate Money Help Delta Chat develop and grow as an independent project serving the people who use it: IBAN DE86100777770428658900, Account owner (please always specify): Björn Petersen, BIC NORSDE...
UPDATE 17 December 2024 1:45pm (GMT/UTC): We are back up and running. Happy Mapping. 🗺 ❤ 🚀 OpenStreetMap.org and a number of related services are currently offline. 15 December 2024 starting approximately 4:00AM (GMT/UTC).
This is a rewrite of the lead section, as I feel the previous one was overly long-winded, yet did not meaningfully summarize the most important points of the article in a way that enables readers t...
We've edited and uploaded another 24 videos from the recent #LibreOffice Conference 2024! So the playlist is now 51 videos – enjoy learning about the technology and community behind our free and private office suite: youtube.com/watch?v=lsMoOtAxqh… #foss #OpenSource
A quick look back at the LibreOffice and Open Source Conference 2024, which took place in Luxembourg from 10 - 12 October.LibreOffice is a free and private o...
Here at TetraLogical we put people first. One of our founding principles is ensuring everything we do is ethical. We make decisions for the good of our customers and their customers, our team, and for the world at large.
The @w3c has just released their Ethical Web Principles in the belief that the web should empower an equitable, informed, and interconnected society.
The web should be a platform that helps people
and provides a positive social benefit.
As we continue to evolve the web platform,
we must therefore consider the consequences of our work.
The Ray-Bans are getting live AI too according to the following mac rumours article: https://www.macrumors.com/2024/12/16/meta-smart-glasses-live-ai/ Trouble is, only for those on the early access program and only in North America.
@jackc Občas při startu se objeví závada SOS volání, ale je to jednou za čas asistent co hlídá odstup. Jednou jsem měl, že nefunguje asistent co hlídá odstup. Na mě jde tam té elektroniky až zbytečně moc. @s173nc3r
btw po tejto aktualizácii pozorujem vysokú spotrebu batérie, tuším mám pocit že pri vypnutom displeji.. argument že po aktualizácii sa to môže stať neobstojí lebo sa to nestáva a už je to druhý deň
- kartu nepoužívám na účte mám mínus - porno sledujem len také "obyčajné" - som čašník, nie politik kurva - kde bývam a sa pohybujem zistí aj amatér z fotiek - rusi si môžu vysrať oko tak či tak - trestnú činnosť nepácham
mno, píšem si otázky čomu vôbec nerozumiem a tým by som chcel aj začať, medzitým profesora "spracovať" do tej pozície že poviem iba to čo viem... děj se vůle má 😬
Apple has deprecated Secure-Transport capabilities since iOS 13.0 (which is several years by now) with a recommendation of porting to. use Network.framework instead. Therefore building libCURL with...
For validating certs using the platform trust store while not using the platform TLS library, we have an example you can use: github.com/apple/swift-nio-ssl…. Swift code, but I’m sure a dedicated contributor could translate it to C.
Storm Dragon pushed changes to the testing branch of the audiogame-manager project Hopefully hide some messages that were making the launch menu ... interesting.
also, a lot of games fail to download, could you eventually either remove them or host them in a more permanent place? when I try to download the super mario bros audio rendition and a lot of other games, I get error downloading <file_name>, installation can not continue and stuff like that. Is it only me? O yeah, and also stuff like danger at the wheel is giving me ShellExecuteX failed, what's that even? agm doesn't give me any more logs besides that
We've been asked about it for years, and we are pleased to confirm that we have finally setup an official NV Access mailing list. You can subscribe to receive:
- Our fortnightly In-Process blog - NV Access organisational updates - NVDA Release announcements (Stable, and / or beta) - Good news stories - Job listings
@TwoThousandStu Yes, some of the contact info we had was from awhile ago, so we apologise if it's not a good address for you - there is a link in the email you should be able to update your details (or unsubscribe).
I’m one of just three passengers in the “Business” silent zone carriage on this SBB train. 1st class. And one of the other two is playing music out of his phone
This is the step free access for 55cm and 76cm platforms. It’s both ingenious and peculiar, as the floor needs this sort of ramp that has metal strips across it for grip
This SBB Giruno, built by Stadler… well I have mixed views
It’s a low floor high speed train, but all except 2 doors have steps still at Swiss platforms. The ramps inside over the wheels result in many odd spots and corners. Door vestibules are very cramped
It’s like the ICNG I took in Netherlands on Sunday - turning a regional train design into an InterCity results in some strange design aspects
A Child’s Christmas in Wales by Dylan Thomas, Read by Christine Malec I had occasion to record this last Holiday season. It lingers in my memory as sweet, fun, funny, intentionally nostalgic, and successfully emotive. patreon.com/posts/childs-chris…
@cachondo Oh lovely! You make me even more humble to have dared that reading in my Canadian accent, which now sounds flat to me. Oh how his words flow like better than poetry. I too always noticed well-spoken people, and I think I unconsciously vowed as a child, to grow up to be one if I could. The, "Dark is Rising," series by Susan Cooper was set partly in Wales, and the reader who's name I've forgotten did a beautiful job with the Welsh accent, in my opinion anyway
@cachondo I was enjoying your recitation of the prose, you have a nicely measured cadance, and of course your accent is an accent to me as a Canadian, so more interesting for that.
That's kind. I was enjoying reading it! But Teams popped up and I am sure my colleagues wouldn't appreciate it. It's still on my Braille display, though. Maybe I'll record the whole thing over the holidays!
Day 17 - #adventOfIOSAccessibility. Check the traversal order of elements in your app. Sometimes the default top-left to bottom-right order might not be the most logical one. Sometimes you may consciously want to tweak the order. Other times, grouping is the answer.
co třeba? Se pochlub ať vím o co přicházím 
Tom Grant the tincel
in reply to Tamas G • • •Tamas G
Unknown parent • • •Tamas G
Unknown parent • • •Tamas G
Unknown parent • • •Tamas G
Unknown parent • • •Tamas G
Unknown parent • • •Tom Grant the tincel
Unknown parent • • •Tamas G
Unknown parent • • •Guillem Leon
in reply to Tamas G • • •Erion
in reply to Tamas G • • •The Brailab DLL is not from Brailab CD, it's from the original hardware, just partially rewritten to compile and run on Windows. When I was working on translating NVDA and of course exposing Hungarian blind people to it, we made an agreement through a Hungarian organisation with the original developer to make it available for Hungarian NVDA users. We also needed some extra functionality for indexing, as well as playing wav files, which would have replaced the synth's spelling feature.
Implementing this was not quite feasible, but they did give us the source code, should we wish to do this ourselves. About a year later, Vocaliser became available and due to its more natural speech, Brailab was mostly forgotten. Some people who grew up using the hardware or later the CD variant still use the addon today. They have also developed an overlay for Android which uses Brailab as its synth.
I am legally bound by this agreement, so I am unfortunately unable to disclose the source code, sorry.
Winter blue tardis
in reply to Erion • • •Erion
in reply to Winter blue tardis • • •Winter blue tardis
in reply to Erion • • •Tamas G
in reply to Winter blue tardis • • •But yeah, figured we couldn't distribute that source. I had gone into maintaining the driver because nobody in Hungary cared about it. I even paid Tyler like $50 at one point to add Indexing for me.
Winter blue tardis
in reply to Tamas G • • •Sean | Ginsenshi The blindwolf
in reply to Tamas G • • •Winter blue tardis
in reply to Winter blue tardis • • •Winter blue tardis
in reply to Tamas G • • •Erion
in reply to Tamas G • • •Ah, Multivox was the first Hungarian synth that everyone used with Jaws before it was translated, there was no eSpeak, only Brailab, WinTalker Voice and Multivox. Then Profivox came, which was licensed exclusively for Jaws, and the underrated SpeakBoard hich was sold even in supermarkets on CD, then Vocalizer, which is everywhere now.
It has loads of issues, Nuance clearly did not test it, nor listened when lots or developers (including myself) complained. They have added horrible pronunciation rules that cause common words to be mispronounced and there's only a partial fix. But I digress.
Winter blue tardis
in reply to Erion • • •Erion
Unknown parent • • •There was no manual that I know of, we just got a zip file with the source dump and a few lines on how to compile it. It had 7 various parameters to my knowledge: rate, volume, pitch, emphasis, spelling, weird mode (this just modified formants I believe), and another one I am unable to recall at the moment. Most of these were implemented in the NVDA addon, the useful ones anyway.
I might be able to find the source, so please let me know if you need me to look up anything.
Zvonimir Stanecic
in reply to Winter blue tardis • • •Winter blue tardis
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Winter blue tardis • • •Winter blue tardis
in reply to Zvonimir Stanecic • • •Winter blue tardis
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Winter blue tardis • • •Winter blue tardis
in reply to Zvonimir Stanecic • • •Winter blue tardis
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Winter blue tardis • • •Tamas G
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Tamas G • • •Tamas G
in reply to Zvonimir Stanecic • • •Winter blue tardis
in reply to Tamas G • • •Nikolett Till
in reply to Tamas G • • •Zvonimir Stanecic
in reply to Tamas G • • •thinks like cukroszacskó and stuff.
Nikolett Till
in reply to Tamas G • • •Nikolett Till
in reply to Tamas G • • •Nikolett Till
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Nikolett Till • • •Zvonimir Stanecic
in reply to Nikolett Till • • •Nikolett Till
in reply to Zvonimir Stanecic • • •Tamas G
in reply to Zvonimir Stanecic • • •Zvonimir Stanecic
in reply to Tamas G • • •