I'm very happy with my frontend. It talks to my backend well. Saying that outloud will make me laugh every time. Sounds really dirty, but it's the truth y'all.
Time to give FastSM another try again, thanks @Bri for making it you and others who work on this are my true heros.
Indeed, I personally do notice a change in performance. To really test, just hold the tab key down. Do you hear the startings of letters consistently and does speech stop after you release it on a control right away? If so, the new frontend is not introducing performance penalties, I'd say.
Aha, how fitting for NV Speech Player. eurpod.com/synths/nvSpeechPlay… is the first version to hit the 10th file milestone for my synthesizers. It's also the first time I'm using the new directory link. This copy fixes memory corruption problems in the frontend, Post-stop aspiration block was made non-brittle now, and the applyRules allocation performance fix. No new versions for awhile until we have tweaked more languages, so I'm done with new files at least until way later in my evening or tomorrow.
This entry was edited (14 hours ago)
one thing that I dislike about NVDA at the moment is that whenever addon updates are available, it pops up with a dialog in the middle of whatever I am doing to let me know, hey, you can update your addons! Can you just send me a toast notification? Please and thank you.
Sadly you can't change it to a tost notification. Like either you can disable thaht entire check I think and then you need to check yourself, or enable auto updates, which is, idk, is that really better?
Starting a btrfs convert. Wish me luck.
People asked for it, and I do think it's a great idea! All of my Speech synthesizers live at eurpod.com/synths - easy to remember URL, and if you want to check it for updates, easy to look at for a version difference. At most, we are only keeping the last 2 versions of a synthesizer, so they will not pile up here. Still uploading some of the larger files, but in the next hour this will all be updated.
The old URLs will work for now, a little bit. As time goes on I will remove them from their old locations and update the blog page to point to this folder too.
The old URLs will work for now, a little bit. As time goes on I will remove them from their old locations and update the blog page to point to this folder too.
This entry was edited (15 hours ago)
Zach Bennoui reshared this.
Jak zní tepelné čerpadlo v -15 ℃
Hehe, ty seš nadšenec, u nás bylo -17℃ a já byl tedy taky koukat, jak to běží, že to nepotřebuje použít ani elektrickou patronu, ale telefon jsem nechal doma a honem pospíchal za ním 
A upřímně, všímám si, že nejhlučnější je TČ v teplotách okolo nuly, asi jak je "hustší" vzduk plný vody a ten to protlačuje lamelami a i ventilátor běhá na vyšší otáčky. Jakmile je mráz a klesne vlhkost, tak je to paradoxně daleko tišší, protože ventilátor sníží otáčky. Sice se více namáhá kompresor, ale ten je u dnešních TČ tak kvalitně hlukově zaizolován, že není skoro vůbec slyšet a to ani při 17 pod nulou.
A upřímně, všímám si, že nejhlučnější je TČ v teplotách okolo nuly, asi jak je "hustší" vzduk plný vody a ten to protlačuje lamelami a i ventilátor běhá na vyšší otáčky. Jakmile je mráz a klesne vlhkost, tak je to paradoxně daleko tišší, protože ventilátor sníží otáčky. Sice se více namáhá kompresor, ale ten je u dnešních TČ tak kvalitně hlukově zaizolován, že není skoro vůbec slyšet a to ani při 17 pod nulou.
@electriccz to ráno jsem na střeše 15 minut ometal sníh z panelů, takže 15 sekund u čerpadla už byla maličkost. 🙂
Jinak mně nejhlučnější přijde, když jede na zpětný chod a chladí. To ho občas slyšíme i v koupelně, nad kterou je.
Jinak mně nejhlučnější přijde, když jede na zpětný chod a chladí. To ho občas slyšíme i v koupelně, nad kterou je.
terror reporting
We will be doing some performance tweaks in the C++ wrapper next update. I can already see people complaining that there's a slight hit. You want to know why? I made a mistake. Or maybe GPT. Well one of us did anyway. I wrote the base C++ outline for it and then it helped me. But uh, at the moment we are allocating memory for out and copying the entire text string for every single rule, even if the rule doesn't match. If we have 50 replacement rules (common in phonology), we are re-allocating and copying the full text string 50 times per function call. Ouch ouch. I'm fixing this.
Testing things! Feel free to ignore.
@jonathan859 @jscholes Quite a bit opposite to most of the Windows clients though. Plans are for a 100% native desktop app, no scripting languages like Python or JavaScript, and for it to be closed-source shareware. If I'll actually get any users, I do not know.
@flyingpenguinMwauthzyx @jscholes Unpopular opinion these days, because not everything has to be open source. I don't know anyone who'd contribute to it and I don't see the requirement. Tweesecake also wasn't OSS and people loved it. I like OSS software yes, absolutely, but it's also work to respect kinda. I know that wasn't your intention just saying my thoughts. Of course Quin can and probably has a different reason.
@jonathan859 @jscholes that is true, and I can see that side of things as well. On the other hand, I can see why people want open source software as well. They get to modify it however they see phit if there is something that they dislike or if there are things that they want to change for instance.
Yeah I totaly agree on that point. Don't think any of the reasons for software to be OSS is invalid. I also prefer OSS where I can.
This entry was edited (14 hours ago)
@jonathan859 @jscholes like for instance, currently I am using a mastadon client. And yes, I know I probably spelled that wrong. Anyways, there are a few features that I want aded, so I might just vibe code it in. As unpopulare as that is, python is not my language of choice, so as a result, I am not good at it lol.
@jonathan859 @jscholes Okay, can't resist. So in the unregistered version, the app works perfectly except that every tenth time you post, an additional post is made that's an advert for your app, and at the end it says, "Please beg this user to register this program. Registration eliminates these posts."
The great thing is, people should not notice changes to existing sounds and tonality. What did change were expansions of rules we can define per language, but to everyone else, this update should not introduce changes in UK and US English, or how the voice sounds. That was not the goal. The goal was to decouple the IPA rule tokenizer into a nice, beautiful frontend.
This entry was edited (15 hours ago)
OK. Huge huge refactoring of NV Speech Player is now complete. Download: eurpod.com/nvSpeechPlayer-2025…
We now have per-language files in the packs folder. Check it out. It is currently defining 22 languages. I would like help improving these languages, now that everyone is able to modify rulesets for them. This has decoupled the rules from the IPA.py engine itself, allowing anyone to freely update and modify voice prosody, tonality, and inflection per letter and phoneme creating it. That is, huge.
Languages supported: English (US, UK), Bulgarian, Chinese (Mandarin), Czech, Danish, Dutch, Spanish (Spain, Mexico), Finnish, French, German, Hungarian, Italian, Polish, Portuguese (Brazilian and Portugal), Romanian, Slovak, Swedish. Speakers of these languages are invited to modify their language files and test to make it sound more native.
We now have per-language files in the packs folder. Check it out. It is currently defining 22 languages. I would like help improving these languages, now that everyone is able to modify rulesets for them. This has decoupled the rules from the IPA.py engine itself, allowing anyone to freely update and modify voice prosody, tonality, and inflection per letter and phoneme creating it. That is, huge.
Languages supported: English (US, UK), Bulgarian, Chinese (Mandarin), Czech, Danish, Dutch, Spanish (Spain, Mexico), Finnish, French, German, Hungarian, Italian, Polish, Portuguese (Brazilian and Portugal), Romanian, Slovak, Swedish. Speakers of these languages are invited to modify their language files and test to make it sound more native.
Pitermach reshared this.
@x0 @jaybird110127 yep that's coming next. So for secure sandboxing I'd have to include my own Espeak, and for just including a compiled copy of Espeak.exe, I think that's not breaking GPL V2/V3. What is would be is integrating it with the DLL more directly, or including Espeak's phoneme compiler so that the engine reads the phondata and phontab files Espeak uses directly + language dictionaries. That's the no-no line. So at first, going the included espeak.exe route isn't bad, just less direct. Haha I know US English will definitely get worse with 2007 Espeak, but maybe some of that was sound too, if we can get things to have better pronouncing then maybe the downgrade won't be so bad.
@jaybird110127 Ah, hell. Definitely gonna stay away from it for years then. Not just because it breaks my synths. It sounds like it'll break every damn add-on in the book, especially ones I rely on like translators which thrive on monkey patching.
@x0 @jaybird110127 unfortunately yeah. At least the re-write effort on the community will be a large one. I think they plan to give tools for sandboxing, I heard they were going to make a compatibility layer that lets you run any add-on with it or add-on devs can utilize it, but I have no idea how that's going or if it'll be ready by 2026.1, there's talk of it pushing to 2026.2. So, who knows RN LOL
@x0 @jaybird110127 No idea where you got your info, but no, the new add-ons runtime isn't coming in 2026.1. It also won't swap out and kill the traditional one, at least not as far as we know at this time.
@NikJov @x0 @jaybird110127 oh? This is why Eloquence was moved onto using it as well, I'm glad they delayed this then. Whew. That's for sure disaster averted, though I do think they could do it chunk by chunk, giving devs enough of an API change for something that big to prepare for a very specific future release very publically in each What's new so nobody has an excuse. Then I think it'll be acceptable :D
@x0 @jaybird110127 There's only one updated Eloquence release I'm aware of, which was just updated for compatibility with the new 64-bit NVDA releases by adding a separate process to communicate with the synth. I'm talking about github.com/fastfinge/eloquence…
It wasn't moved to any new add-ons API however, because it isn't even out yet.
It wasn't moved to any new add-ons API however, because it isn't even out yet.
GitHub - fastfinge/eloquence_64: Eloquence synthesizer NVDA add-on compatible with 64-bit versions of NVDA
Eloquence synthesizer NVDA add-on compatible with 64-bit versions of NVDA - fastfinge/eloquence_64GitHub
@NikJov @x0 @jaybird110127 ah yeah. So that's still changing then. Oof. Thee 64-bit thing is what's going to break add-ons, since many use 32-bit dlls.
@x0 @jaybird110127 Yep, a lot of synthesizers are affected. As far as I know the plan is to eventually have a compatibility layer, for example Mick is working on one for SAPI 4 and SAPI 5, see github.com/nvaccess/nvda/pull/…
Support for 32 bit sapi4 and sapi5 via a 32 bit synthDriver runtime by michaelDCurran · Pull Request #19432 · nvaccess/nvda
Link to issue number: Replaces pr #19412 Summary of the issue: With the switch to 64 bit, NVDA no longer supports sapi4, or 32 bit specific sapi5 synthesizers. NVDA should somehow continue to supp...GitHub
@NikJov @x0 @jaybird110127 I honestly wish they'd just keep building separate 32 and 64-bit builds, but that ain't gonna happen. Although, this does make things at least somewhat easier, since as long as the DLL itself is 64-bit it can still be loaded into NVDA, which at least means SpeechPlayer could be recompiled (other legacy synths, even like BestSpeech will still need a wrapper.) Definitely staying away until February / March at least, it's a terrifying mess :D
It’s amazing how these people talk about AGI as if it was actually possible (it’s not).
What’s being sold as “AI” currently is slightly random statistical token chain generation; it has zero to do with thinking, intelligence or creativity.
theguardian.com/technology/202…
‘We could hit a wall’: why trillions of dollars of risk is no guarantee of AI reward
Progress of artificial general intelligence could stall, which may lead to a financial crash, says Yoshua Bengio, one of the ‘godfathers’ of modern AIDan Milmo (The Guardian)
as you may have noticed: my temper and wiggle room for stupidity in security reports have shrunk to a minimum
“I’m sorry. You are totally right and I should have known better.
<Thinking 45s>
I went through the report again, and I think you misunderstand what I meant”
You are hereby banned from our program. We don't want your "help". Please never contact us again.
(just sent off to another "helpful" security researcher)
and then: hackerone.com/reports/3514263
curl disclosed on HackerOne: libcurl: Improper Authentication State...
Following the recent advisory for **CVE-2025-14524**, I conducted an investigation into how libcurl manages OAuth2 credentials during complex redirect chains. I have confirmed that while the...HackerOne
This could be a theatre play:
“Drop the AI”
“I sincerely apologize, you are absolutely right”
youtube.com/shorts/6eA_o9qZBuU…
“Say Potato!” Stream on Apple and Spotify!! #funnysong #tinder #dating #music
Užívejte si videa a hudbu, kterou máte rádi, nahrávejte originální obsah a sdílejte vše s přáteli, rodinou i celým světem na YouTube.YouTube
I wonder if at this point the training data of LLMs itself contains slop reports for curl
But Wait!
Anyway, BIND 9 now also has Bug Bounty program via #YesWeHack (fosstodon.org/@iscdotorg/11576…) and we got exactly one genuine issue out of 15 (and that's issue that has been previously independently reported). The rest was:
⁃ Cryptographic Weakness in BIND 9.20.15 PRNG Enabling DNS Cache Poisoning (Bullshit AI Slop; it just proved lack of randomness in provided PoC 
:)
⁃ Multiple EC/TLS Private Keys Committed to Public Bind9 Repository (yeah, in system tests) (1/2)
When the European Commission approached us about funding a bug bounty for BIND 9, we were impressed with the proposal. We have a policy against bug bounties (because we were frustrated with people wasting our time), but under this proposal, the YesWeHack team would do initial triage, and use their expertise to minimize the 'slop' reports. This is a game-changer for a small development team.The bounty program is active, and we are looking for our first valid report.
yeswehack.com/programs/bind-bu…
BIND 9 Bug Bounty Program bug bounty program - YesWeHack
BIND 9 Bug Bounty Program bug bounty program detailsYesWeHack #1 Bug Bounty Platform in Europe
we are about announce that we are ending out bug-bounty program in an attempt to slow down the tsunami of terror reporting
“ I sincerely apologize. You are absolutely right.”
lol. Don’t need any AI detection software here 🤪😂
Seinfeld soup nazi vibes, absolutely love it. The use of direct copy/paste AI output in these reports is so incredibly disrespectful to you and anyone else who works on your program. And shame on H1 for allowing this nonsense.
Games I love like Warsim, The Wastes, and usurper inspired me to think about creating my own console-based game. Then I wrote 750 lines of code just to make a reusable system for console menus. And realized the save system is going to be another 500 lines, probably. And then the settings system. And after thinking about 2000 lines of code before I can get to anything even resembling the simplest game mechanic, I'm not inspired, anymore. Why is 99 percent of programming doing the least interesting part of any project?
johann reshared this.
Early in my programming career I wrote a whole series of functions that did the mundane things in a business program. Then I wrote a program that allowed me to create input and output screens WYSIWYG. Those screens called my functions. In a few weeks of work, which was kinda fun, I eliminated many days of tedium for each application I was to write over the next decade.
Keep that in mind when you build the foundation. You do it once.
@RegGuy Hmmm. I only wish I found that kind of thing fun. But it's just a lot of thinking about menu lists and settings objects and what gets passed where and how and what's global and blah.
the thing is there are enough people who find this fun that there are multiple text-based game engines out there.
@MostlyBlindGamer@RegGuy Right, but it would take me just as long to understand someone else's engine as it would to right my own. And none of them do quite what I want, anyway. IE a terminal window with options you select by entering a single number or character and pressing enter. They're all command-based. But if I want sighted players I have to think about output colors. And so that means a settings object. And some people want ASCII boarders ("---------" etc) and some don't. So now we have a bunch of case statements and need to check colour and ascii art settings. And if we're doing that we might as well do sound, too, to have an earcon when prompting for an option and a click when an option is selected. But now that means an audio library. And utilities to think about file paths and including assets. And now we're at 2000 lines just for menus.
remember KISS for keep it simple (and) stupid? The new thing is YAGNI for you’re not gonna need it.
@MostlyBlindGamer@RegGuy And we haven't even talked about localization. Because if I want that it has to be done at the start, or else it becomes a huge impossible thing later.
so you don’t want to make a game engine, but the only thing that matters is making a engine?
What about making a proof of concept for the game?
What about making a proof of concept for the game?
@MostlyBlindGamer But I can't really make a proof of concept that works the way I want it to until all the plumbing exists.
Lots of people struggle to prioritise fun over flexibility. My suggestion is to organically build those boring boilerplate systems bit by bit as they're needed. No menu system might mean manually editing stuff in a file between runs during initial development, but is probably better than no game.
@jscholes If I do it that way, it just means refactoring later, though. And that's even harder and less fun because now you're rebuilding things that already exist.
Refactoring is inevitable.
If you haven't built a menu system to begin with, there'll be nothing to rebuild. The refactoring will mostly be limited to slotting it in once you start.
If you've built one early that doesn't actually solve the problems you end up needing it to, or solves them in a way that ultimately makes it hard to integrate, that's when you'll end up rewriting stuff.
@jscholes Maybe. I've played enough of this style of game that I have pretty strong opinions about how it should all work, though.
That said, thinking about object graphs, what gets passed where, and other architectural stuff is legitimately my favourite part of software dev. But I can't really do it without at least a semi-useful or interesting project to apply it to.
@jscholes Hah. So this all started when I decided I wanted a class where I could create a new menu, add the selection key, a name, and a callback function for each item, then call the menu to print itself and do all the input and error checking, and call the callback for whatever item was chosen. And then it just kept growing. I'm not even at the object graph stage. I just hate most of the text games I mentioned, where every menu looks and acts different, because print statements and case statements are just sprinkled all over the code.
@jscholes If I ever finish the basics, the thing I'm aiming for, probably in 70 years or so, is a singleplayer trade wars/elite style space game. Planetary exploration, trading, combat, procedurally generated galaxy, etc. All the stuff we have in that style is either multiplayer PVP, or only semi-accessible like smugglers5. Interface via console menus, because that feels both simple and retro. But with stuff happening in realtime like textspaced used to be (it's now discontinued), and high quality sound. The idea is you just let it run on your taskbar and check in every 20-30 minutes when you hear something that needs doing. So kind of casual all-day play while multitasking. I've dreamed of something like that for like 20 years, but nobody has done anything even close. Textspaced was the nearest, but it never quite got there, and it's gone, now.
From the standpoint of accessibility-first, need recommendations for office suite software that is not Microsoft Office? It doesn't have to do everything that MS Office does. Documents and spreadsheets are all I really need. I'm also rather hesitant to go the Google route for this stuff. Libre Office? OpenOffice? Any ideas?
Ontario opposition NDP Leader Marit Stiles said it best. “It’s the same Doug Ford playbook; do nothing, then complain to grab headlines," Stiles said. "Saskatchewan Premier Scott Moe flew to China with the Prime Minister to advocate for his province’s canola farmers.
"Meanwhile, Ford was sitting at home, leaving Ontario workers with no one to fight for them. If you’re not at the table, you’re not fit for Premier." cbc.ca/news/canada/toronto/dou…
"Meanwhile, Ford was sitting at home, leaving Ontario workers with no one to fight for them. If you’re not at the table, you’re not fit for Premier." cbc.ca/news/canada/toronto/dou…
London PR firm rewrites Wikipedia for governments and billionaires
Founded by Keir Starmer’s comms chief, Portland helps rich clients ‘protect their reputation’ – with a shady, off-the-books service
thebureauinvestigates.com/stor…
Founded by Keir Starmer’s comms chief, Portland helps rich clients ‘protect their reputation’ – with a shady, off-the-books service
thebureauinvestigates.com/stor…
Hello #Fediverse,
as #FOSDEM 2026 is approaching I was thinking about bridging the livestreams to #PeerTube as well.
Streams are under #CreativeCommons license that allows sharing videos, so there is no big deal in it.
As there are plenty of rooms (and we dibs Social Web for #VHSky :) ), it would be fun to coordinate streaming accross multiple instances to stream as much rooms as possible without overloading single instance.
What do you think, anybody interested?
Also, anybody has idea how to actually bridge m3u to PeerTube? @j4n3z maybe?
reshared this
I’ve found an issue with syllable start detection compared to Python. In Python, the detection starts at a consonant if it's followed by a vowel, which influences pitch path calculation. This could explain some differences. But there’s also an exception for stressed vowels. Interesting!
Let’s be honest, Ring was already some technocratic, dystopian BS, but if you needed a reason to finally, finally kill it with fire, here’s your reason: Ring is partnering with Flock to help ICE spy on you and your neighbors for the government 👀
#Technology #InfoSec
techcrunch.com/2025/10/16/amaz…
Amazon's Ring to partner with Flock, a network of AI cameras used by ICE, feds, and police | TechCrunch
Agencies that use Flock can request that Ring doorbell users share footage to help with "evidence collection and investigative work."Amanda Silberling (TechCrunch)
Tampereella hurja trilleri! Olli Rehn älähtää
The amount of times Google Gemini gave me a solid code correction and answer while Chat GPT sits there thinking and thinking is staggering, honestly it's making me disavow it because it's not like this is better in "pro" mode. I noticed Gemini "thinks" in smaller chunks but it could be just that they update the process more often. Regardless, GPT, you're beginning to really really suck. Gemini can give me the same quality of C++ and Python code recommendations, it's still shit when pasting back full code for you so better when patching, which is about the only thing holding me back from it as opposed to GPT. Gemini can read attachments but it doesn't quite get the same containerization treatment - GPT mounts your files from /mnt/data, Gemini clearly stores it somewhere and can read but when writing it is restricted from spitting back a file for you and envoking commands in a container. Huge downside to Google, ugh.
Something that using a local Gemini Cli instance could solve though, honestly.
Something that using a local Gemini Cli instance could solve though, honestly.
This entry was edited (17 hours ago)
Jonathan reshared this.
You wanna talk about Gemini? They fixed the uploading multiple files thing in AI Studio, now the response formatting is all fucked up so each file isn't a single code block anymore and the download/copy buttons are gone. Fix one thing, enshitify another.
I bloggered a post.
It's about shortcomings of FLOSS and a possible next thing.
My Next Project Won't be FLOSS:
pointless.one/my-next-project-…
#FLOSS #FOSS #FreeSoftware #OpenSource #GNU #GPL #OSI #MIT #BSD #BTPL #PolyForm
My Next Project Won't be FLOSS @ PointlessOne
Where FLOSS fails and what to do about it. With a little bit of history.PointlessOne
Some of the variables you can define in the yaml now:
# Base defaults for NV Speech Player frontend.
#
# If a setting is not listed here, the C++ defaults are used.
# Packs are merged in this order:
# default.yaml -> <lang>.yaml -> <lang-region>.yaml -> <lang-region-variant>.yaml
#
# Keep this file small. Put language-specific tweaks into their own files.
settings:
# Timing / stress shaping
primaryStressDiv: 1.4
secondaryStressDiv: 1.1
# Stop closure insertion ("click" before stops/affricates)
stopClosureMode: vowel-and-cluster
stopClosureClusterGapsEnabled: true
# Length mark (ː) handling
lengthenedScale: 1.05
lengthenedScaleHu: 1.3
applyLengthenedScaleToVowelsOnly: true
# Output gain defaults (same as ipa.py)
defaultPreFormantGain: 1.0
defaultOutputGain: 1.5
# Normalization cleanup
stripAllophoneDigits: true
stripHyphen: true
# Tonal languages
tonal: false
toneDigitsEnabled: true
toneContoursMode: absolute
This entry was edited (18 hours ago)
»Ein #Barrierefreiheit-Feature, dass nur Menschen mit einer #Behinderung hilft, ist wahrscheinlich schlecht.«
Das sind so die Dinge die ich eigentlich nur im Fediverse lese.
Pro-Tipp: Ihr müsst gar nicht alle Gedanken die ihr habt mit der Welt teilen.
Prüfbericht zur #Barrierefreiheit mit Testimonials von KI-generierten behinderten Menschen.
How about nein einfach nur nein.
Why are Google's higher quality tts voices gated behind Google talkback? What kind of fucking insanity is this. The testing versions didn't work this way, so what gives? Get all the way to fuck.
Sigh. We could pass the cast pointer from the callback since it points to valid C++ memory, but it could be riskier in some cases, especially with temporary DLL memory. The safer approach would be to create a new speechPlayer.Frame, set its fields, apply modifications, and then send it to queueFrame. We also need to handle cases where the frame is NULL, meaning silence. So much work.
hey man, just wanted to say thanks so much for working on this, this adon is the only one I know of where you can update the formants of the adon using a UI. Keep up the good work and you're doing good!
@GoemonIshikawa next one will include a /packs folder for everyone. the packs folder will have phonemes.yaml and inside lang, per-language rulesets. this will make it way more of a community effort. It's coming.
Do you know if there's a way where you could make a small UI within the adon for making phonemes and language rules so we could make our own languages for you to include? Banger ideas your having there, I'll for sure help with making things for this adon, sadly though sliders and entering via UI is about the only thing I can do as although I wanted to do the same thing for espeakNG, I'm not good with working with just a notepad and a list of formants. But hopefully if you want to go the UI direction as well I can for sure make languages and accents with it along with you! I've been meaning to make Barbados English with synths, as I come from a part of the world that includes Barbados people.
@GoemonIshikawa oh yeah, that is not out of the question whatsoever. But I first want to lock down linguistic rules, even if that means future-proofing now for things like tonal languages and other dialectical characteristics. That's what I'm trying to smoothen out next before improving or working on other languages as it will really serve as a foundation.
For sure totally! Don't you think though that you should have stuck to English and just worked on that only before putting other languages in though? I do have a little understanding as I hang around friends who speak much languages, though sadly I don't know how you would present that in code.
@GoemonIshikawa Haha not a chance with Hungarian being my first :) I think I had two simultaneous goals: Do really good English, but also make a good Hungarian speech synthesizer, because there is not a good one right now. So by nature of supporting Hungarian and US English (which is like its own language), it expanded things so much that by that point I was like, "wellp, let's add all the languages we can!" xD
That does sound very good, but from the looks of it just as an outsider looking in it kinda looks like some languages when updated are having the rules mixed, as words like Ahmed are becoming ashMed. But yes take a brake my brother, and resume when you can! I myself have always wanted to see if I can someday make a version of English US that between sentences or in certain punctuations we can put a filter on some of the consonant sounds to make like sudo breath in and out, to really push formant synthesis!
@GoemonIshikawa is it only me, or does the synth not show up in the synthesizer selection dialog? Am running nvda 2025.3
@flyingpenguinMwauthzyx @GoemonIshikawa for sure odd - I tried the latest driver directly and installed over my working copy just to verify, but it shows up here from that link's download.
@flyingpenguinMwauthzyx @GoemonIshikawa try the link at eurpod.com/nvSpeechPlayer-2025… - that one can't be cached as it is, but the base eurpod.com/nvSpeechPlayer-2025… is also always the latest version.
@flyingpenguinMwauthzyx Hey man I think you might have messed with something, as some words like level now come out like levva when reading.
@GoemonIshikawa @flyingpenguinMwauthzyx yeah noticing this too! I'm in the process of making US and UK better but these are the types of feedback that I'd really like. It's very close and I tried to make sure we calculate aspiration, pitch, all that the same way so the rules are kept, hopefully I can be in a place where that can be a change in one of the rules or phonemes and not the frontend code itself by now.
@GoemonIshikawa @flyingpenguinMwauthzyx Got some really good feedback from @TomGrant91 about the word soundings so I'm refining based on what he said too since I think he's right. At least it's not missing vowels and phonemes like it was in that first build, couldn't even say the word manager or priveleges :D ahaha. Coming together, but slowly.
@flyingpenguinMwauthzyx @TomGrant91 If you want I can try and help with some languages like Japanese, for starters do you know how the synth determins that kind of lettering? Like does it have a letter list of sorts such that it can say the letter?
So I wasn't going to post about this but I just got off the phone with Tesco customer service and I'm shocked.
Last night I was having some soup, Tesco brand Tuscan inspired Bean soup, to be exact, bit down on a chunk of glass that was in there. Taste of blood in my mouth, dent in my tongue, possibly swallowed a small bit... I was worried about it. Also, fucking hell like.
Today I call em up, had already sent an email, they are uninterested, tell me I have to complain in store.
What the fuck.
@fireborn I think it might have put me off canned soup for good, I'll just have to make my own soup from now on.
@Tutanota Is it planned to work on the agenda view while working with email? It would be a very practical improvement.
Thank you!
Thank you!
Thnaks for the feedback, our teams are working to improve the calendar and we will share this with them. :)
"AI Features" is just another buzzword for "Information Harvesting" - thanks, but no thanks!
AI exists to siphon as much data from the user as possible. For AI to benefit the people, it must be implemented with care akin to how Ente handles facial and object recognition, Proton Lumo is end-to-end encrypted including metadata, and Confer which aims to have a trustless, verifiable AI with no way to access the data.
Bluesky Corporation have just started hosting and verified the US agency ICE:
bsky.app/profile/icegov.bsky.s…
In other words, Bluesky are collaborating with racist violent thugs and murderers.
(Bluesky Corporation were already hosting the White House (bsky.app/profile/whitehouse-47…), the Department of War (bsky.app/profile/deptofwar.bsk…), JD Vance (bsky.app/profile/jd-vance-1.bs…), Homeland Security (bsky.app/profile/homelandgov.b…) etc. This is a long-term collaboration with Trump admin.)
This entry was edited (1 day ago)
reshared this
@gajim@fosstodon.org
@gajim@community.xmpp.net
With Gajim version 2.4.2 windows-key + arrow keys still don't work.
Back to 1.9.5 again.
This is a GTK issue, see dev.gajim.org/gajim/gajim/-/is…
Windows: No window snapping / no snap layout (#12348) · Issues · gajim / gajim · GitLab
Please search for similar issues first. Versions OS: Windows 10 Gajim: 2.3.2 (Installed through Windows...GitLab
Thank you.
It's a shame snapping isn't working. Because of this, I keep having to go back to v1.9.5. Snapping does work with this older version. Hopefully, this will be fixed in the future.
This entry was edited (19 hours ago)
It feels so good to know I'm localy archiving/saving podcasts with audiobookshelf now. Noone ever taking old episodes away just because. Really wish I had done that way sooner. Guess some surely nostalgic stuff will never see the light of the world again.
Seit gestern steht eine Cherry Stream auf meinem Schreibtisch, die die gute alte G83 abgelöst hat. Wenn man jahrzehntelang auf so einer Tastatur geschrieben hat, muss man sich erst einmal bemühen, die neue nicht gleich wieder durch zu festes Tippen kaputtzumachen. 😁 Aber es ist ein absoluter Traum auf der Stream zu tippen, und sie kostet keine 30 Euro. Mal sehen wie lange sie durchhält.
In wie fern Traum? Hier kommt der Blindie in mir raus aber gibts mal ne Audio demo? Bin grad auch noch am Überlegen was ich mir fürs Büro hole weil puh, die Dinger da sind irgendwie meh.
@jonathan859 Sie hat halbhohe Tasten, also nicht ganz so flach wie ein Laptop, aber tippt sich trotzdem fast genauso leicht. Man kommt auf ganz andere Geschwindigkeiten als mit der vergleichsweise schwergängigen G83. Habe dadurch bisher auch weniger Buchstabendreher und sonstige Tippfehler, vor allem in Verbindung mit Großbuchstaben am Wortanfang. Bei der G83 musste ich oft nachkorrigieren. Ach und sie ist weitestgehend standardkonform, findet man heute auch eher selten.
@jonathan859 Ich schicke nachher mal ein Audio. Sie ist nicht besonders flüsterleise, im Vergleich zu anderen Tastaturen, aber darauf kam es mir auch nicht an. Leiser als die G83 definitiv.
Back to nicer topics, finally getting my free weekend for the Open Beatz festival 2026, absolutely looking forward to this. My 1st real festival.
Sadly they seem to have a kinda strict recording policy, so we'll see wether I can take the Zoom with me. But well it's still a half year so plenty of time to overthink jk.
Sadly they seem to have a kinda strict recording policy, so we'll see wether I can take the Zoom with me. But well it's still a half year so plenty of time to overthink jk.
And my dad got his ticket to come with me. Shit can't wait this will be fire lol, definitely something great to look forward to in 2026, my 1st festival for real.
Alex Chapman
in reply to Tamas G • • •