Have any blind Udio users figured out how to use the new Inpainting feature? An Inpainting section of the page appears, there are unlabeled buttons for what I think are the fragments that can be selected, but clicking any of them just starts/stops playback of the entire track. In YouTube videos, people click on fragments, and they hear the fragment, making it more clear what part they’re about to regenerate or delete.
It's hardly a hot take, but I still marvel, from time to time, that Windows is still a thing. As operating systems go, it is insanely old. Even back in 2000, working at MS in a product group, we thought it was old, having been around over 15 years. I don't think any of us would have honestly believed it would still be a thing in 25 years. At that time, the Mac OS was seen as way too old, and it had been around 17 years by then. Windows is this shitty curse that will seemingly last forever.
@bryansmart How is performance, particularly with VO on the web? I've heard some people still getting the dreaded busy busy messages even on m3 machines. Why does that happen, in the first place? I've heard conflicting things, like that VoiceOver only uses one core of a processor.
AirChat, the new Tiktok style chat app that's only audio, is somewhat accessible on iOS with VoiceOver. There's some unlabeled buttons on the main screen, but the audio quality is really good.
Youtube music's podcast experience has annoyed me. For one thing, 2X is too slow for me! But also because the episodes play in reverse chronological order, which is just insane. So I guess I need a new #podcast app! What do folks recommend for #android?
Since the BTSpeak will be somewhat emulating behavior of the BNS, I wonder if that also means it will include BNS BASIC? Of course, Python is a better noob language for 2024, but, still. Lot of old BNS BASIC software out there that would be a little interesting to run.
So no Eloquence on the Pixel 8 because 64bit only, which means, funnily enough, a synthesizer you would not expect to be 64bit ready (though then again maybe so because AOSP) Pico, in the form of SmartVoice, can be installed on 64bit only devices if you so choose.
Mobile Dectalk is mostly safely doable on Android from a legal ambiguity standpoint, and now that Android has made sideloading even easier than it already is, I don't think installing an APK would be too much to ask.
@jackf723 I heard a demo live. It is using some form of older dectalk, because it doesn't sound like the newer versions. The closest it sounds to is 4.3
I know this might be a long shot, but does anyone have a steinberg plug-in that was around in the early 2000s called Voice Machine? Boosts appreciated.
@BorrisInABox Are you the one who wrote Where's My Keyboard? I think I remember finding it in Multisynth many ages ago. It's stuck in my head and I would love to find a render of it without having to set up the oldschool multisynth nonsense.
@simon@BorrisInABox@x0@cordova5029 I have Hypersonic 2 working on win 11. If you need orriginal Hypersonic, it won't work on anything newer than XP 32bit.
Right so have a track I wrote last year, when I was feeling pretty contemplative. I called it fountain of life. i"ll also provide a link to the actual flac file, because the don will transoce it like a jerk. Enjoy, and I"m curious what you people think of this side of me. I should share some of my other stuff for those who haven't heard it. Anyway, enough woffle. Description: this track is primarily in c sharp major. A rather ambient piece, including several pads, and a fountain/stream sound. All sounds are curtasy of my Roland fA06 keyboard. And here's the file link if you want to hear it less transcoded. https://www.dropbox.com/s/so1jbt6cjay5f5l/fountain%20of%20life.flac?dl=1
@bryansmart@cordova5029 Christian music is so bland these days. Just slap four chords on it, repeat the chorus like eight times, add a guitar (with no solos,) insert the word Jesus in as many times as you can, and you have your worship song. Hell, you could take a Christian song instrumental, get a non Christian acapella of a popular song, and it'll still sound ten times better than the orriginal Christian song. Appoligies if I offended anyone.
So, I'm not sure if I've said this here, but with Emacspeak, a lot of stuff that's hard for a blind person to do in, say, a terminal or manpages or info documents, is really simple, because it's a keyboard-driven interface. Want to move up to the previous prompt to review output? Control + C, then Control + P. Or it may just be C-c p. Want to move to the next heading in a manpage? Just hit N or P. Like, it's a bit like Powershell, where normally terminals are just a stream of text, but Emacs is objects. And objects, like elements on the web, can be navigated or used as landmarks, or searched through if nothing else works.
Now, I'm not saying Emacspeak is perfect. It still suffers from being a programmer's tool, so definitely not for everyone. But what it does do well, especially not messing with my tags in HTML unlike VS Code's nonsense where it'll just jumble everything up for you, and even stuff like reading through processes in ProcEd-mode, Emacspeak makes that really visual information a lot simpler for audio.
Made this with the Piper Training notebook using Google Colab. Using NVDA. Now, you to can have Blaire Stainton read you anything you want. This is still a work in progress, so some inflections may be off.
I'm finding that if you use the latest fine-tuned Demucs model (htdemucs_ft) on a true lossless song, the result is very, very close to perfect separation. Sometimes I actually can't tell the track was separated using AI. The jump from MDX to this new model seems as big as the jump from Spleeter to Demucs. The problem is that finding lossless songs is hard if you don't know what you're doing, and when you try with a lossy song, you quickly become aware of just how much data the compression removes. Even some of the lossless tracks I get from Deezer are not truly lossless, which is annoying. I don't have an accessible way of verifying it, so I could easily have hundreds of GB of fake FLAC songs. But meanwhile, the results I get are mind-blowing.
@simon FYI, MV Sep dot com has a model for the Demucs Fine tune. Mind you it's slow, depending on how many other files are in the cue for processing, but there is that option for people who don't want to deal with command line and anaconda environments. Mind you, you are sending your audio to a server to be processed, so there's that as well for the potential downsides of it as well.
I'm sorry everyone but we need to take back whatever we all said about Apple not doing machine learning right.
iOS 17's Photos app can now identify and explain laundry symbols in pictures and this is amazing. No more need for apps that do this; just swipe up on a photo.
My first ever LG phone (the 4600, no letter identifier) had this startup sound. LG made a ton of phones with variations on the sound; I'll link to a directory of 14 or 15 variations. My mother lost that phone in early 2007 and I never heard the startup sound again, until I found a YouTube video with all LG phone startup/shutdown sounds in 2023. I recognized it instantly. I'm 100% positive that out of all of the variants, this is the one that was on my phone. Sound memory is weird. Or maybe it's just mine. Anyway, I've always liked the sound, for some reason.
@simon I remember that startup sound. LOL, I remember the TTS for reading contact names sounded terrible, like a software version of a doubletalk, but with like no inflection. and some weird pronunciations. and the menu readout was pretty bad as well.
So we have mastercards with tactile distinctions, which is a great start don't get me wrong. But let's talk about the fact that this still does not address the bigger issue - the number not being independently readable. The tech is already out there. Why aren't any of these card providers partnering with Wayaround? I.E. waytag containing an activation code, you go to the card issuer's website, enter the activation code, last four digits of your ssn, validate, and then your card info is revealed?
@bryansmart@talon@FreakyFwoof that would be awesome if they did that. At first I thought that the "big new accessible music production" thing was Ableton finally beeing made accessible. My hopes were so high LOL. And then I find out that its just some stupid native instruments overlay thing that you have to use a Maschine controler to use. That's the thing that urks me. Make the damn software accessible without using a stupid overlay to the controler. Rant over.
@bryansmart@FreakyFwoof@pitermach@talon Thanks for clarifying. My appoligies for flying off the handle. I have a Komplete Kontrol keyboard. I just wish we could create our own sounds from scratch rather than relying on presets, but I guess that is the way of the future.