mastodon - Link to source

Love this track, just found it today. Have known Charley's music for a while, love her song Timebomb as well. youtu.be/XeJwH3iuN0U?si=45dBpv…

#music #newmusic #pop


Try (vs. Charley) - Fly By Midnight (Official Lyric Video)

Official lyric video for original single "Try" (vs. Over October) by Fly By Midnight. Stream: https://stem.ffm.to/thefastesttimeofourlives | Subscribe: http:...
YouTube
#music #pop #newmusic

mastodon - Link to source
Okay, I'm having too much fun with this. Here's a stupid voice of my own creation (no idea why the pitch isn't right) reading a stupid story I wrote about a hyper-intelligent hedge trimmer. I'm sure @BorrisInABox and @cordova5029 will get a chuckle out of this.

@Xantastic @Borris

mastodon - Link to source

Because much commentary is now being posted about the killing of a man in Minneapolis today by armed and masked federal agents, and because my previous thread about this today has grown long, I'm starting a separate thread about this now.

"There is some dispute now whether that was ICE or Customs and Border Patrol, wrestling an individual to the ground. He appeared to be motionless."

~ Jennifer Rubin

#Minneapolis #Trump #ICE #DHS #MaskedThugs
/1

contrarian.substack.com/p/exec…


Execution in Minneapolis- Jen Rubin and Mary McCord go live

A recording with Jen Rubin and Mary McCord
Jennifer Rubin (The Contrarian)
#Trump #dhs #ice #minneapolis #maskedthugs
in reply to William Lindsey

mastodon - Link to source

William Lindsey

"Another day, another dead Minnesotan. Shot in the street while getting pinned down by masked federal thugs. And Donald Trump and Kristi Noem and Greg Bovino wasted no time putting out narratives that don’t at all square with the videos we’re seeing. But who you gonna believe? Them, or your lyin’ eyes?"

Matt Labash

#Minneapolis #Trump #ICE #DHS #MaskedThugs #AlexPretti #murder
/37

mattlabash.substack.com/p/a-pr…


A Prayer For The Unjust

Let the hard rain of justice rain down upon them
Matt Labash (Slack Tide by Matt Labash)
#Trump #dhs #murder #ice #minneapolis #maskedthugs #alexpretti
in reply to William Lindsey

mastodon - Link to source

William Lindsey

"RIP, Alex Jeffrey Pretti, an intensive-care nurse by trade. Intensive-care nurses, not typically being he most war-like people. And who the goon squad is already slandering as a 'domestic terrorist.' Yeah, I’m for stopping domestic terrorism, too. Which is why someone should arrest Stephen Miller, who takes demented pleasure in waging war on his own people. While prompting his corrupt boss to do the same."

#Minneapolis #Trump #ICE #DHS #MaskedThugs #AlexPretti #murder
/38

#Trump #dhs #murder #ice #minneapolis #maskedthugs #alexpretti
This entry was edited (16 hours ago)

mastodon - Link to source
This is the second time that I have had OmniDescriber just totally lose my configuration. This time, I did literally nothing as far as I can tell. Windows Defender says my computer isn't infected with anything, but I'm still managing to encounter this issue. I have no idea what I'm doing wrong, but I'm not sure I can rely on this program anymore. If anyone has any thoughts on this feel free to let me know. Thanks.
in reply to Brandon Tyson

mastodon - Link to source

Jonathan

The fact it's no longer a local tool I can put my API keys in, it's a cloud solution where I have no idea what models they use internally and don't pay for my API usage, rather I pay for the pricing they decide to put up. I'm happy for the developer he got this, chance, and yes accessibility wise the web version is ok and does what it's supposed to, but well. I preferred the local native app, I had control, it couldn't go down, etc etc.

mastodon - Link to source
I was using Debian earlier today--the GUI, not over SSH. I think I see why Linux isn't more popular. I had a .txt file, and I pressed enter on it. Any computer user would expect the file to open. At worst, they would expect to be prompted to choose an application. Nope. Debian popped up a dialog saying the file was executable, and asking if I wanted to run it. Come on, Linux. It's a text file! What are you doing? This is (part of) why nobody likes you.

reshared this

mastodon - Link to source

Friendly reminder that Signal—while maybe not perfect—is a much better choice for communication than RCS, especially if you are communicating anything sensitive. Doubly true if you are on iOS or communicating with anyone who is, as Apple still does not even use end-to-end encryption for RCS like Google does on Android.

signal.org/download/


Download Signal

Download Signal for Android, iOS, Linux, macOS, and Windows.
Signal Messenger
This entry was edited (17 hours ago)
Unknown parent

momostr - Link to source

Vyram Kraven

I'm very relaxed. Giving feedback isn't crying mr.mellowdramatic it's feedback & you don't like it okay cool. I'm not gunna stop saying it though because thats my review on the app. You don't want to change it back okay don't that's just more review sauce for me.
The way back is by not updating & if you kill older versions well then i'm just -1 to using the app & i'll try something like Cherogram which offers webxdc as well & then advertise it till there's a change that makes it worse then better.
Unknown parent

momostr - Link to source

Vyram Kraven

Yep & when something is put into this world the experience made is the one kept. If by drowning in a glass of water you mean being public with review & experience then I will happily do so. If you didn't care you wouldn't be taking the time to anticdoting feedback as crying or drowning in a glass of water. If things get worse maybe I will look for an alternative one minus isn't going to make me go jump off. Theres so many factors to what makes a person stay or not.
in reply to daniel:// stenberg://

mastodon - Link to source

Roy -- the dull one

Thanks for your toot!

The urge to connect every thing to the Internet astonishes me. I enjoy other people's observations and complaints about the Bidness World's belief that we are all idIoTs.

This thread reminded me of a story from a galaxy far, far away - timewise. Before the Internet. Even before FRED, the first PC. The story is "The Robot Millennia", in "Galaxies Like Grains of Sand" by Brian Aldiss, New American LIbrary (1960). It was prescient, and fun. But no IoT.

mastodon - Link to source

Will you be at #FOSDEM or in Brussels next week and want to meet some of the team? 👀

We have a few more spots open to register for the event at Little Delirium on Jan 31! If you haven't registered yet, do so before space runs out 👇🏼

#HomeAssistant #SmartHome #HomeAutomation #OpenSource
luma.com/ptshwcqn


Meet the Home Assistant team - FOSDEM Brussels

We'd love for you to join us 😎 Join the Open Home Foundation and key members of the Home Assistant team for a meetup in Brussels during FOSDEM 2026! 🤩 We're hosting a casual evening for the open-source community to connect.
luma.com
#opensource #homeassistant #fosdem #smarthome #homeautomation

mastodon - Link to source

#AndroidAppRain at apt.izzysoft.de/fdroid/?radd=1… today brings you 21 updated and 1 added apps:

* Be nice: lets you launch apps in split-screen mode 🛡️

Enjoy your #free #Android #apps with the #IzzyOnDroid repository :awesome:


IzzyOnDroid F-Droid Repository

This is a repository of apps to be used with your F-Droid client. Applications in this repository are official binaries built by the original application developers, taken from their resp. repositories (mostly Github, GitLab, Codeberg).
IzzyOnDroid App Repo
#android #free #apps #AndroidAppRain #izzyondroid

mastodon - Link to source
Hi @Tamasg I'm trying NVSpeechPlayer 2026 v10, I'm not sure if this is the latest version, but the voice sounds a bit different, even when using Adam, than the original SpeechPlayer. Were there some changes that could change how the overall voice quality sounds? It's hard to pinpoint it, but the voice itself sounds less harsh and consonents like P sound very short now. I see several settings in the voice dialog but I have no idea what any of them do, but they did not seem to make much of a difference. Thank you very much for your time and I look forward to any input you may have. Thanks.
@Tamas G
in reply to Brandon Tyson

mastodon - Link to source

Tamas G

yeah, there were like 3 missing phoneme fields: preFormantGain. , GlottalOpenQuotient, ETC. these got added in the later pack and are now tunable per vowel, in the past they would just be set to the engine default and everything got the same treatment. So this allows more flexibility but some things could have shifted with the change. Every letter now needs to define these new 3, they are very important for shaping and leaving them out in the past until now has been a disaster.
in reply to Tamas G

mastodon - Link to source

Brandon Tyson

Awesome, I'll take a look. I also notice weird inconsistencies with phonemes in general with US English, not just in speech player, but in eSpeak in general. Are we stuck with eSpeak's limitations? Or can I make other changes in the phoneme editor? For instance, certain words with "ong" in them sound more like "onng" instead (with Eloquence, "onng" will produce more of an "ah" sound instead. It's like eSpeak has specific exceptions where "ong" sounds correct. Can I overwrite this for speech player? Or are we stuck with this weird inconsistency?
in reply to Brandon Tyson

mastodon - Link to source

Tamas G

This is a really good observation, and you’re not imagining it.
What you’re hearing isn’t specific to SpeechPlayer, it comes from how eSpeak phonemizes certain vowel + nasal combinations in English, especially around “ng”. Depending on the word, eSpeak may output slightly different vowels before /ŋ/, and that can make some “ong” words sound more like “onng” or overly nasal.
Eloquence handled this differently: it didn’t rely on a generic phonemizer, but used hand-tuned rules that deliberately adjusted vowel quality before nasals. That’s why those cases often sound more natural there.
The good news is that we’re not stuck with eSpeak’s output as-is. For SpeechPlayer, we can override or normalize specific IPA patterns in the language pack (for example, adjusting vowel quality before /ŋ/ or mapping those cases to dedicated internal vowel variants). We’ve already done similar overrides for other tricky cases like “fear/hear” and UK vs US differences.
So while eSpeak itself has limitations, SpeechPlayer isn’t limited to reproducing them exactly — it just means the fixes live in the language/phoneme rules rather than in the core engine. You can check out our other normalization rules to get an idea for this.
in reply to Brandon Tyson

mastodon - Link to source

Tamas G

ahaha darn it. Now I hear it too. On words like "beyond" - yeah, it might be really an ESpeak thing. We're not limited to it as our phonemizer, actively exploring others, but bundling any other one could drastically increase the size of SpeechPlayer as the add-on. Right now we kind of highjack Espeak from within NVDA itself, and that gives us a neat little advantage but does lock us in.
in reply to Tamas G

mastodon - Link to source

Brandon Tyson

Oh and we can't forget about the "hazz bean" bug that's in some earlier NVDA and eSpeak versions... I'm still on NVDA 2024 and I heard that with Speech Player and eSpeak. I need to see if it's fixed in 2025.3, I think it might be. I'll keep using it, but also the O phoneme in words like "box" sounds like it's way overdone. Book still sounds like it's in UK english too. Support has a wrong "or" sound, this one's again on eSpeak, eSpeak keeps breaking US english and making it way worse. This used to be handled correctly, but I think they broke it in 2024 and it still sounds like that today.

mastodon - Link to source
NVSpeech Player 2026 v10: Removes all pops/clicks from stream, 100% for sure now.
Spanish can now say "yo" properly. First attempt at "hear" and "dear" but it's not perfect yet. Other tweaks to voice quality.
Sample rate added. You can now choose from 4 of them if you don't like the default.
eurpod.com/synths/nvSpeechPlay…
This entry was edited (16 hours ago)
in reply to Cleverson

mastodon - Link to source

Tamas G

@clv1 You’re right, the schwa rule was present, but it didn’t trigger before "rã" in pt-br because we normalize ã into a Brazil-specific nasal vowel symbol Ã, and our PT_VOWELS class didn’t include Ã. I’ve added à to PT_VOWELS (and PT_NASAL_VOWELS) in pt.yaml, so consonant + ɾ + à now correctly becomes consonant + ᵊɾ + Ã. That should fix the missing schwa before rã / ram / ran-style sequences.
@Cleverson
⇧